Upload
others
View
7
Download
0
Embed Size (px)
Citation preview
PHYSICOCHEMICAL AND SENSORY PROPERTIES OF
SOYMILK FROM FIVE SOYBEAN LINES
_______________________________________
A Thesis
presented to
the Faculty of the Graduate School
at the University of Missouri-Columbia
_______________________________________________________
In Partial Fulfillment
of the Requirements for the Degree
Master of Science
_____________________________________________________
by
JING TANG
Dr. Fu-hung Hsieh, Thesis Supervisor
DECMEBER 2013
The undersigned, appointed by the dean of the Graduate School, have examined the
thesis entitled
PHYSICOCHEMICAL AND SENSORY PROPERTIES OF
SOYMILK FROM FIVE SOYBEAN LINES
presented by Jing Tang,
a candidate for the degree of or Master of Science,
and hereby certify that, in their opinion, it is worthy of acceptance.
Dr. Fu-hung Hsieh, Food Science
Dr. Ingolf Gruen, Food Science
Dr. Mengshi Lin, Food Science
Dr. Mark Ellersieck, Statistics
ii
ACKNOWLEDGEMENTS
I would like to give my sincere thank to Dr. Fu-hung Hsieh for being my advisor.
Especially for his patience guiding me through my research problems, he has always been
a good mentor. He always opens my mind for any kind of research conversation, and
always being a good listener. Thank you Dr. Hsieh. Without your warm heart and support,
I don’t think I can make it here.
I am also grateful to my committee members, Dr. Ingolf Gruen, Dr. Mengshi Lin,
and Dr. Mark Ellersieck, for their guidance on many research problems and giving me
very valuable comments.
I would like to express my thanks to Dr. Kristin Bilyeu and her lab members for
their kindness to provide soybeans and details of soybeans to support my research.
I want to say thank you to Harold Huff, Lakdas Fernando, and Rick Linhardt for
their patience and kindness to help me working in the laboratory and solving the
problems of making soy ice creams.
I would like to express my thanks to Linda Little and JoAnn Lewis. Thank you for
your patience and always giving me correct answers for all kinds of questions.
I am very thankful to the valuable friendships with all my friends. Thank you for
sharing your life with me and encouraging me when I had problems.
iii
Last but not least, I would like to express my thanks to my parents and my
grandparents. I could not finish my degree without your support and love. Your loves
always make me have a faith in me. I love you.
iv
TABLE OF CONTENT
ACKNOWLEDGEMENTS ................................................................................................ ii
LIST OF TABLES ............................................................................................................ vii
LIST OF FIGURES ......................................................................................................... viii
ABSTRACT ....................................................................................................................... ix
CHAPTER
1. INTRODUCTION .......................................................................................................... 1
1.1 Background .................................................................................................................. 1
1.2 Objectives .................................................................................................................... 3
2. LITERATURE REVIEW ............................................................................................... 4
2.1 Soy Products ................................................................................................................. 4
2.2 Soy Foods and Health Promotion ................................................................................. 5
2.2.1 Soy Foods and Coronary Heart Disease .............................................................. 6
2.2.2 Soy Foods and Cancer ......................................................................................... 7
2.2.3 Soy Foods and Women’s Health ......................................................................... 9
2.2.3.1 Soy Foods and Menopausal Symptoms ..................................................... 9
2.2.3.2 Soy Foods and Calcium Binding Properties ............................................ 10
2.2.4 Soy Foods and Chronic Kidney Disease ........................................................... 11
2.2.5 Soy Foods and Cognitive Function ................................................................... 12
2.3 Disadvantages of Soy Foods ....................................................................................... 14
2.3.1 Soy Foods Allergy ............................................................................................. 14
2.3.2 Soy Foods Intolerance ....................................................................................... 15
2.3.3 Beany Flavor ..................................................................................................... 15
2.4 New Soybean Lines .................................................................................................... 16
v
2.5 Sensory Evaluation ..................................................................................................... 19
2.5.1 Descriptive Sensory Evaluation ........................................................................ 19
2.5.2 Discrimination Test ........................................................................................... 23
2.5.3 Consumer Sensory Evaluation .......................................................................... 23
3. MATERIAL AND METHODS .................................................................................... 25
3.1 Sample Preparation ..................................................................................................... 25
3.1.1 Soy Milk Samples Preparation .......................................................................... 25
3.1.2 Soy Ice Cream Samples Preparation ................................................................. 26
3.2 Analytical Methods ..................................................................................................... 27
3.2.1 pH Value ........................................................................................................... 27
3.2.2 Color .................................................................................................................. 28
3.2.3 Viscosity ............................................................................................................ 28
3.2.3.1 Viscosity for Soymilk .............................................................................. 28
3.2.3.2 Viscosity for Soy Ice Cream Mix ............................................................ 28
3.2.4 Overrun .............................................................................................................. 29
3.2.5 Melting Rate ...................................................................................................... 29
3.2.6 Hardness ............................................................................................................ 29
3.2.7 Total Solids Content .......................................................................................... 30
3.2.8 Fat Content ........................................................................................................ 30
3.2.9 Protein Content .................................................................................................. 31
3.3 Descriptive Sensory Evaluation for Soymilk .............................................................. 33
3.3.1 Training of Panelists .......................................................................................... 33
3.3.2 ABX Test ........................................................................................................... 33
3.3.3 Descriptive Sensory Evaluation ........................................................................ 34
vi
3.3.4 References Preparation ...................................................................................... 34
3.4 Consumer Acceptance Test for Soymilk .................................................................... 37
3.5 Statistical Analysis ...................................................................................................... 37
4. RESULTS AND DISCUSSION ................................................................................... 43
4.1 Soymilk Fat, Protein, and Total Solids Contents ........................................................ 43
4.2 Soymilk Color ............................................................................................................. 45
4.3 Soymilk Viscosity ....................................................................................................... 46
4.4 Descriptive Sensory Evaluation for Soymilk .............................................................. 48
4.5 Consumer Acceptance Test for Soymilk .................................................................... 55
4.6 Soy Ice Cream Color ................................................................................................... 56
4.7 Soy Ice Cream Texture Properties .............................................................................. 57
5. CONCLUSION ............................................................................................................. 61
APPENDICES .................................................................................................................. 63
BIBLIOGRAOHY ............................................................................................................ 83
vii
LIST OF TABLES
Table Page
2.1 Sensory studies of different degree of lipoxygenase for beany flavor in soymilk.16
3.1 Initial list of attributes of soy milk samples……………...…..…………………..40
3.2 Final list of attributes of soy milk samples…………..…………………………..42
4.1 Composition of soymilk samples………..……………………………………...44
4.2 Soluble carbohydrate content of soybean lines…………….………….………....45
4.3 pH value and color of soymilk samples……….……………………….………...46
4.4 Viscosity of soymilk samples………….…………………………….…………..47
4.5 Sensory data with significance values from ANOVA and jack-knife estimates...51
4.6 Variance explained by the seven principal directions…………………….……..52
4.7 Consumer acceptance of soymilk samples…………...…………….……………56
4.8 Color of soy ice cream mixes……………………………………………….……57
4.9 Viscosity of soy ice cream mixes……………………………………………..….58
4.10 Overrun value and hardness of soy ice cream samples………………..………..59
viii
LIST OF FIGURES
Figure Page
4.1 Scores on principal direction 1 and 2 from PCA on mean centered data……..…53
4.2 Loadings on principal direction 1 and 2 from PCA on mean centered data……..54
4.3 Percentage of melting down of five soy ice cream samples….…..….........……..60
ix
PHYSICOCHEMICAL AND SENSORY PROPERTIES OF
SOYMILK FROM FIVE SOYBEAN LINES
Jing Tang
Dr. Fu-hung Hsieh, Thesis Supervisor
ABSTRACT
Soymilk may provide health benefits due to its soy isoflavone compounds,
polyunsaturated fatty acid compositions and fibers. The main objective of this study was
to investigate the effects of soybean varieties on descriptive sensory evaluation, consumer
acceptance test, and texture properties of soy ice creams. Plain flavor soymilks made
from five soybean lines were made in the laboratory using a soymilk maker for physical,
chemical, descriptive sensory analysis, and making soy ice creams. Soymilks were placed
at room temperature for 2 h first prior to the sensory evaluations. For the consumer
acceptance test, soymilks were flavored with sugar and vanilla.
Soymilk fat content, protein content, and redness did not show significant
differences among samples. However, high oleic acid soymilk sample and high sucrose
content soymilk sample had significantly higher results on total solids content and
yellowness. Moreover, the high sucrose content soymilk sample was significantly lighter
than others due to its clear hilum skin. For the viscosity, high oleic acid soymilk sample
was significantly more viscous than other samples because of the oleic acid composition
and the molecular conformation of the triacylglyceride molecules. The viscosity also
contributed to the overrun value of soy ice cream. The more viscous the mix, the lower
the overrun of the soy ice cream became. Thus, the low overrun was associated with
harder texture and higher melting rate.
x
The results from descriptive sensory evaluation showed that lipoxygenase free
soymilk was the least beany flavor and most rice-like flavor, and high sucrose content
soymilk was significantly sweeter than others, but all of them did not significantly change
the texture of soymilks. On the other hand, high oleic acid content soymilk was the
thickest one, but was not associated with flavor aspects. Furthermore, consumer
acceptance test showed that soymilk made from sample KB10-5#1137 was the best.
1
CHAPTER 1
INTRODUCTION
1.1 Background
Soy foods are traditional foods from soybeans in Asia, and now become popular
in Western Countries. In 2011, total U. S. soy foods sales exceeded $5 billion, lead by
soy milk beverages, which reached 994 million dollars according to the Soyfoods
Association of North America (SANA) (2011). Soy foods have a high plant protein
content and contain polyphenol components, such as isoflavones. Thus, soy foods are
classified as a functional food. In addition, some papers indicated that soy foods may
decrease the risk of coronary heart disease (Malik and others 2004), have anti-cancer and
anti-inflammation properties (Yang and others 2009; Peng and others 2009), improve
menopausal symptoms and increase the Calcium absorption for women (Charoenphun
and others 2013; Bao and others 2008), provide positive effects for Type 1 or Type 2
diabetes (Zimmermann and others 2012), and maintain or even relieve dementia
symptoms for patients who suffer from Alzheimer’s disease (Duffy and others 2003).
Soy foods, especially soymilk, are considered a good substitution for dairy
products for individuals who have milk intolerance. Milk intolerance including cow’s
milk protein allergy (CMPA), cow’s milk protein intolerance (CMPI), and lactose
intolerance is prevalent in the world, especially in children. CMPA can affect from 2 to 6%
of children. Studies have shown that 80% to 90% of these children will resolve CMPA
within their fifth year (Host and others 2002; Wood 2003; Gopalan 2011). The important
and widely used method used to treat milk intolerance is the exclusion of all forms of
2
animal milk and milk products from the diet. Because of the high plant protein content
and similar texture as milk, soymilk is a healthy protein source for those milk intolerant
individuals. In addition, cow’s milk contains saturated fats that might increase the risk of
cardiovascular diseases (Astrup and others 2011; American Heart Association Nutrition
Committee. 2006; Siri-Tarino and others 2010). Soymilk, which contains much lower
amounts of saturated fat than cow’s milk, could be a good choice for individuals who are
concerned about heart disease.
Although soymilk and soy foods are nutritious and their consumption has many
health benefits, soy milk has some limitations such as beany flavor, possible allergic
reactions and undesirable microbial fermentation after intake by some consumers.
Consequently, scientists have tried to resolve these problems by creating some new lines
of soybeans, for instance, low and ultra-low raffinose family of oligosaccharides
soybeans, lipoxygenase free soybeans, high oleic acid soybeans, low P34 allergen
soybeans, and high sucrose soybeans. It would be of great interest to know if soy milks
made from these new lines of soybeans could address these problems. This has not been
reported in the literature.
Traditional dairy ice creams are popular and sales continue to increase, 4.1%
increase in 2011 reaching 10.7 billion dollars and 4% growth in 2012 (International
Dairy Foods Association 2013). For those people who cannot or do not want to consume
any dairy products, soymilk ice creams might be an acceptable replacement. Therefore,
physical properties of soy ice creams made from new types of soymilks were tested
because texture and mouth feel are the most important characteristics for ice cream.
3
1.2 Objectives
The objectives of this study were
1) To produce soymilks and soy ice creams from five lines of soybeans;
2) To determine the effect of soybean lines on physical and chemical characteristics
of the soymilks, including color, viscosity, pH value, total solids content, fat
content, and protein content;
3) To design a descriptive sensory test and a consumer acceptance test of the
soymilk samples and to study the effect of soybean lines on sensory evaluation;
4) To investigate the effect of soybean lines on texture properties of the soy ice
cream samples, including overrun value, viscosity, hardness, and melting
properties.
4
CHAPTER 2
LITERATURE REVIEW
2.1 Soy Products
Soy based foods were developed in Asian countries and accepted in Western
countries within decades. In 1917, soymilk was produced commercially in the United
States, introduced by Li Yu-ying in 1913 (SoyInfo Center 2001). Soy foods become
popular due to two reasons: the first is that soy foods contain many nutrients, and the
second is that a high percentage of populations in Western countries have allergic
reactions to dairy. The American College of Allergy, Asthma and Immunology (ACAAI)
reported that around 2.5% of children younger than three years old are allergic to milk in
the USA and approximately 80% of these children with milk allergy will not outgrow the
allergic reaction until the age of sixteen (2010). Moreover, thirty to fifty million
Americans, including adults and children, have lactose intolerance reported by the Ohio
State University Wexner Medical Center (2007).
Traditional soy foods are classified into two categories: non-fermented and
fermented. Typical non-fermented soybean products are soy milk, tofu, tofu skin, soy
meat alternatives, etc., while fermented products include soy sauce, fermented bean paste,
natto, sufu, miso, onchom, tempeh, etc. In the United States, the commercial soy based
products include soy protein products (meat analogs), soy milk, soy cheese, soy nut butter,
soy yogurt, soy frozen dessert, and soy based infant formulas (United Soybean Board
2012). In 2011, the soy based food industry sales exceeded $5 billion, the most
5
significant product being soy milk beverages, of which sales reached 994.3 million
dollars, according to the Soyfoods Association of North America (SANA) (2011).
2.2 Soy Foods and Health Promotion
Soybeans are one of the highest sources of plant protein, containing around 40%
protein, in addition to being rich in essential amino acids (Xiao 2008). They also contain
35% carbohydrates, including soluble di- and oligosaccharides, 18% to 22% fat,
including polyunsaturated fatty acids, tocopherols, phytosterols, phospholipids, minerals,
B vitamins, and fiber (Liu 1997). Soy foods are classified as functional foods because
they contain polyphenol components, including isoflavones, glycosides and malonate
conjugates, which contribute to their antioxidant activity (Xiao 2008; Mackinnon and
others 2011). The amount of isoflavones in soy products varies with the type of soybean,
the condition and area of cultivation, and processing. Green soybeans and tempeh are
good sources of isoflavones with more than 48.95 mg isoflavones per 100 g product.
Tofu and soy milk contain approximately 30 to 40 mg isoflavones per gram of protein
(USDA 2008).
However, soybeans contain potentially anti-nutritive compounds, such as lectins,
phytic acid, protease inhibitors, and phenolic acids (Xiao 2008). They may adversely
affect iron and zinc balances in some individuals who do not eat a well-balanced diet.
Despite the potential anti-nutritive effects, a large amount of research papers reported
health benefits of soy foods in the area of disease prevention.
Non-fermented soy foods contain a high amount of soybean isoflavones, which
contribute to lower rates of osteoporosis (Taku and others 2011; Wei and others 2012).
Previous studies showed that non-fermented soy foods may be associated with reducing
6
gastric cancer risk (Wu and others 2000; Messina and others 1994; Kim and others 2011)
and breast cancer risk (Wu and others 1996). On the other hand, fermented soy food may
improve the bioavailability of dietary zinc, iron, and vitamins, since it has much less
phytic acid than raw beans, which may obstruct the absorption of minerals (Hirabayashi
and others 1998; Kasaoka and others 1997; Denter and others 1998; Jiang and others
2011). In addition, after fermented soy foods are digested and adsorbed in the human
body, oligopeptides can be generated, which have a more rapid absorption and higher
biological activity than free amino acids in the small intestine (Gibbs and others 2004;
Maebuchi and others 2007).
2.2.1 Soy Foods and Coronary Heart Disease
Coronary Heart Disease (CHD) is a serious heart disease and a major cause of
death in most developed countries. The high fat and sugar content of the diet particularly
in the United States and many other variable factors, such as smoking and drinking
alcoholic beverages, contribute to an individual’s risk for CHD. Low-density lipoproteins
(LDL) play key roles in the atherosclerotic process. LDL is oxidized by free radicals
along the walls of blood vessels and forms a glue-like material which builds up and
eventually may block the vessel. High-density lipoproteins (HDL) have the protective
effect against LDL oxidation and eliminate the cholesterol along the blood vessel wall
(Malik and others 2004). Soy foods contain a high quantity of soy isoflavone
components that have effects on reducing the risk for CHD because of their antioxidant
(Anderson and others 1998; Tikkanen and others 1998) and anti-inflammatory effects
(Sadowska-Krowicka and others 1998; Gooderham and others 1996). A study with rats
reported that soy isoflavones appear to be incorporated with LDL and protect it from
7
oxidation (Anderson and others 1998). In addition, soy proteins could be another
contributor for reducing risk of CHD; in a clinical study of 38 humans, soy protein intake
significantly decreased serum concentrations of LDL-cholesterol, while increasing the
HDL concentration in 34 of the 38 individuals (Anderson and others 2001).
Furthermore, during the atherosclerotic process, blood vessels could be damaged
by the abnormal spasm or “clamping down” which occurs when the atherosclerotic
plaque is exposed to stress. Soy proteins and soy isoflavones may reduce the blood clot
formation (Anderson and others 2001) and restore damaged blood vessels to protect
blood vessels because of their anti-inflammatory effects (Honore and others 1997). Since
soy proteins and their isoflavones may correct the balance of blood lipid levels, protect
blood vessels, and have antioxidant and anti-inflammatory effects, in 1999 the U.S. Food
and Drug Administration (FDA) approved a health claim that eating soy protein as part of
a low saturated fat and cholesterol diet may decrease an individual’s risk for CHD (FDA
1999). That same year, the U.S. FDA published that 25 g of soy protein per day provided
an appropriate intake to reduce risk for CHD (FDA 1999).
2.2.2 Soy Foods and Cancer
Currently, more than 13 million people in the United States have had some type
of cancer (American Cancer Society 2012). The main types of cancer are lung cancer,
colon cancer, breast cancer, prostate cancer, and skin cancer. The high rate of cancer is
often because of unhealthy lifestyles, such as smoking, high-calorie diets, drinking
alcohol, lack of physical activities, etc.
Results from animal and in vitro studies have suggested that soy foods have a
positive effect against cancer initiation due to its high isoflavone content. The one
8
possible mechanism of soy’s effects against cancer is that the structure of isoflavones is
similar to estrogen’s structure, thus having a weak estrogenic effect. Consequently, the
association of soy foods with cancer risk was first researched for breast cancer and
prostate cancer prevention. Several recently published papers indicated that soy foods or
isoflavones may decrease the risk of breast and prostate cancers. Yan and Spitznagel
(2005) examined the relationship between soy intake and prostate cancer risk by meta-
analysis and demonstrated that consumption of soy food could be related to a lower risk
of prostate cancer in men. In China, 73,223 Chinese women participated in a Shanghai
Women’s Health Study to investigate the association of soy food intake with breast
cancer risk. The result of this prospective cohort study showed that soy food intake
significantly reduced the risk of premenopausal breast cancer (Lee and others 2009).
However, Nechuta and others (2012) demonstrated that soy food intake did not result in a
significant reduction in the risk of breast cancer, yet it had a statistically significant
reductive effect on the risk of cancer recurrence.
The other possible mechanism of soy foods to prevent cancer is similar to the
mechanism of blood vessel protection. Formation of new blood vessels is the key step for
the growth of human cancers and facilitation of tumor cell metastases. Soy isoflavone is
the major protective components, since it can restore the structure of blood vessels
(Honore and others 1997) and inhibit cell transformation and tumor cell proliferation
(Anderson and others 2001). A prospective cohort study provided strong evidence that
soy food consumption may protect against colorectal cancer in women (Yang and others
2009). In addition, in a large case control study, researchers demonstrated that soy food
consumption may reduce the risk of colorectal cancer in men (Budhathoki and others
9
2011). Soy food consumption has a positive effect on lung cancer as well. A meta-
analysis demonstrated that soy isoflavones were associated with a 27% risk reduction in
lung cancer (Yang and others 2011). A recent epidemiologic study which involved 444
nonsmoking women supported this claim with similar conclusions (Yang and others
2013).
2.2.3 Soy Foods and Women’s Health
2.2.3.1 Soy Foods and Menopausal Symptoms
The main menopausal symptoms women experience are vasomotor symptoms,
hot flashes, and night sweats. The prevalence of those symptoms is much more common
among North American and European women than those in China and Japan.
Approximately 20% of Chinese women and less than 25% of Japanese women complain
of hot flashes and night sweats, compared to around 85% of North American and
European women (Anderson and others 2001; Reed and others 2013; Freeman and others
2007). McNagny and others (1997) indicated the undesirable symptoms could be
eliminated or treated with estrogen substitutions, called hormone replacement therapy
(HRT), which is chosen by 47% of postmenopausal women’s physicians. However,
commercial HRT drugs may have some side effects such as heart attack, breast cancer,
and blood clots in the lungs and legs (Björn and others 1999). Furthermore, the
differences in prevalence of menopausal symptoms in various countries are hypothesized
to be partly related to the content of dietary soy consumption. The average soy intake per
day of Chinese women is up to nine times greater than that of North American women
(Reed and others 2013). The high concentration of soy’s estrogen-like isoflavones could
10
make it a good hormone replacement option for postmenopausal women. Consequently,
soy foods may be a healthy alternative to HRT pills for women.
2.2.3.2 Soy Foods and Calcium Binding Properties
Calcium is known to make up about 1.5-2.2% of bodyweight. Bones and teeth
contain approximately 99% of the body’s calcium. The other 1% is dispersed in
intracellular and extracellular fluids (Kitts 2006). Intestinal calcium absorption regresses
in ovarian hormone deficiency and may partially lead to bone loss of women during
menopause (Arjmandi and others 2002). Common sources of calcium-binding peptides
are casein (Kitts 2006), shrimp, carp muscle (Charoenphun and others 2013), and soy
protein (Charoenphun and others 2013; Bao and others 2007).
One possible mechanism of regressive calcium absorption is that estrogen
deficiency leads to a down-regulation of key calcium transport molecules in intestinal
cells (Aloia and others 2013). Previous studies have shown that since soy isoflavones
have a similar structure to estrogen receptors, which could enhances calcium transport in
intestinal cells, so that soy isoflavones may exert partly estrogen activity in tissues and
alter intestinal calcium absorption just like estrogen receptors do (Arjmandi and others
2002). In addition, soybean protein contains a number of asparagine and glutamine,
because the free carboxyl group content of the hydrolysates significantly increased by
treatment with asparaginase or glutaminase (Bao and others 2007). Bao and colleagues
(2008) found that calcium binding increased linearly with the increment of carboxyl
group of acidic amino acids in soy protein hydrolysate.
Though the calcium-binding capacity of soy protein is lower than that of casein
(Pathomrungsiyounggul and others 2010), soy isoflavones may alter intestinal calcium
11
absorption, particularly in women, because of its structural similarities with estrogen
receptors (Arjmandi and others 2002) and soy protein hydrolysates may increase calcium
absorption as well (Bao and others 2008 and Bao and others 2007).
2.2.4 Soy Foods and Chronic Kidney Disease
Dietary protein plays a critical role in the prevention and treatment of chronic
kidney disease. Compared with meat protein, vegetable protein has a temperate effect on
renal glomerular filtration (Kitazato and others 2002). Soy food is one of the most
frequently used plant protein sources in our diet. One hundred fifty years ago, researchers
studied the effects of soy protein on kidney diseases (Anderson and others 2001).
Recently, soy protein and their isoflavones have been demonstrated to provide some
positive effects in individuals with Type 1 or Type 2 diabetes (Anderson and others 2001;
Zimmermann and others 2012; Boué and others 2012), end-stage renal disease (ESRD)
(Ogborn and others 2010), and polycystic kidney disease (Peng and others 2009). The
mechanisms of how soy protein prevents kidney diseases have not been fully explained;
however, from previous studies, researchers know that soy isoflavones, amino acid
composition, and antioxidant and anti-inflammatory properties could be the major
contributors.
Type 2 diabetes mellitus results from insulin resistance, which causes an increase
in hepatic glucose production. The ultimate goal of diabetes treatment is to maintain or
decrease the level of hyperglycemia (Moller 2001). Several animal and human studies
indicated that soy food can play a role in the improvement of glycemic control.
Consumption of soy foods is associated with lower fasting insulin levels, reduced risk of
type 2 diabetes, and glycemic levels (Jang and others 2010; Davis and others 2007; Lee
12
2006 and Zimmermann and others 2012). For instance, consumption of cheonggukjang, a
fermented soybean paste, has been shown to reduce glycemic level and maintain
pancreatic β-cell in diabetic animal models (Kwon and others 2007).
Furthermore, nearly one-third of those with diabetic nephropathy have
exacerbation renal disease, such as ESRD and polycystic kidney disease (Anderson and
others 2001). Renal prostanoids and maintenance of normal renal function are sensitive to
physiological challenges. In pathological conditions, renal prostanoids may rapidly
change in inflammatory and proliferative processes and develop into renal injury (Peng
and others 2009). Consequently, when pressure increases between glomerular capillaries,
capillary dilation will stimulate some intracellular pathways and alter level of prostanoids.
Soy isoflavones may block some of those pathways and thereby inhibit kinase activity
(Hao and others 2008 and Anderson and others 2001). In addition, because of their anti-
inflammatory effects, soy isoflavones may decrease the inflammatory processes which
contribute to the pathogenesis of renal failure.
2.2.5 Soy Foods and Cognitive Function
Alzheimer’s is a chronic and progressive disease, which is the most common form
of dementia. The symptom of Alzheimer’s disease (AD) in the early stages, is memory
loss, and in the late stage, is losing the ability to carry on a conversation and respond to
their environment. AD is the fifth leading cause of death among aged 65 and older in the
United States, and also is the only cause of death among the top 10 without a way to cure
it. In 2013, 5.2 million Americans suffer from AD, including 200,000 individuals
younger than age 65 who have younger-onset Alzheimer’s (Alzheimer’s Association
2013). Plaques and tangles are observed in older individuals and far more in AD patients.
13
These two abnormal structures are prime suspects in damaging and killing nerve cells.
However, researchers and scientists are still working on what role plaques and tangles
play in AD. Recently, soy isoflavones, as partial estrogen agonists, have been proved to
modulate cognitive function, particularly in postmenopausal estrogen-deficient women.
Genistein and daidzein are structurally similar to estrogen and have the ability to
bind to estrogen receptors, hence exerting both agonistic and antagonistic estrogenic
effects (Miksicek 1995). Two types of estrogen receptor have been found in the nucleus:
ERα and ERβ. ERβ is the major subtype of receptor in the brain so that it is hypothesized
that soy isoflavones may exert potentially biological activities in the brain (Lee 2005).
Some human studies have shown an association between soy isoflavones intake
and cognitive function. Duffy and others (2003) indicated a beneficial contribution of soy
isoflavones in the prevention of Alzheimer’s and enhanced cognitive function. They
examined the short term, 12 weeks, effect of soy extract containing isoflavones of thirty
three postmenopausal women (50-65 years) on cognitive tests and regulatory behaviors.
The group receiving isoflavones supplement showed significantly improvements in recall
of pictures and in a sustained attention task. Kritz-Silverstein and others (2002)
demonstrated that category fluency was higher in women receiving soy isoflavones
supplement than in controls. In addition, a case-control study found that soy isoflavones
supplementation improved cognitive function in men (File 2001). However, based on a
double-bind, randomized, placebo-controlled trial of 202 healthy postmenopausal women,
Kreijkamp-Kaspers and colleagues (2004) reported that there were no differences
between the control group receiving milk protein and the experimental group receiving
14
25.6 g of soy protein containing 99 mg of isoflavones in cognitive function of these
women.
2.3 Disadvantages of Soy Foods
2.3.1 Soy Foods Allergy
Soy is a popular vegetarian alternative to meats, and many processed foods
contain soy fiber or soy protein to get a higher nutritional value. In addition, soy food is
also an important source of nutrition for infants with milk allergy (Sicherer and others
2006). However, approximately 0.4% of children are allergic to soy based products,
making soy allergy about half as common as peanut allergy (Savage and others 2010).
Therefore, it is urgent to identify the protein components which elicit allergic
manifestations of soy products. Immunoblotting and enzyme-linked immunosorbent
assay (ELISA) have been widely used to detect and determine the major allergens by
using allergen-specific antibodies (Ogawa 2006). According to immunoblotting patterns,
Herian and others (1990) demonstrated that allergens in soybean could be classified into
three categories. Furthermore, the protein components binding with immunoglobulin E
(IgE) antibodies in the sera of soybean sensitive patients indicated three major allergenic
proteins in soybean, called Gly m Bd 30K, Gly m Bd 28K and Gly m Bd 60K (Ogawa
2006; Ogawa and others 1991). Gly m Bd 30K, which is identified as P34, is a common
allergenic protein in soybean, because two thirds of soybean-sensitive patients examined
have shown IgE antibody binding to Gly m Bd 30K (Ogawa 2006). However, it is hard
to define how many specific IgEs in soybean will cause clinical soy allergy. In a
15
retrospective review, Komata and others (2009) demonstrated that soy-specific IgE levels
of between 20 and 30 kU/L predicted a 50% chance of leading a soy allergy.
There are several methods that could reduce or remove these three major allergens.
They are molecular breeding, genetic modification, physicochemical reduction,
enzymatic digestion, chemical modification, extrusion cooking (Ogawa 2006) and
transgenic suppression (Bilyeu and others 2009).
2.3.2 Soy Foods Intolerance
After consumption of soy foods, some sensitive individuals may generate
undesirable microbial fermentation in the intestines, leading to digestive distress. Some
sensitive individuals cannot fully digest α-galactosides with α-galactosidase completely
in their small intestine. So in their large intestine, non-digestible oligosaccharides are
digested by microorganisms, liberating huge amounts of gas (Slominski 1994). These
nondigestible oligosaccharides, including the stachyose and raffinose families of
oligosaccharides (RFOs), cannot be eliminated by usual processing (Bilyeu and others
2009).
2.3.3 Beany Flavor
Normal mature soybeans contain three lipoxygenase (LOX) isozymes, LOX-1,
LOX-2, and LOX-3 (Torres-Penaranda and others 1998). In oxidation reactions, LOX
isozymes require substrates containing polyunsaturated fatty acids, including linoleic acid
and linolenic acid (King and others 2001). These polyunsaturated fatty acids contribute to
beany or grassy flavor in soy foods. Table 2.1 shows that the differences in LOX may
lead to different results in sensory studies of soymilk.
16
Table 2.1 Sensory studies of different degree of lipoxygenase for beany flavor in
soymilks
No Lipoxygenase Variety Beany Flavor (Comperd with
Normal Soymilk)
References
1 LOX-free Significant difference on beany
flavor
Torres-Penaranda
and others 1998
2 LOX-free Significant difference on beany
flavor
Torres-Penaranda
and others 2001
3 LOX-free Significant difference on beany
flavor
Kobayashi and
others 1995
4 LOX-1 No significant difference on beany
flavor
Kobayashi and
others 1995
5 LOX-free No significant difference on beany
flavor
King and others
2001
6 LOX-free No significant difference on beany
flavor
Suratman and others
2004
2.4 New Soybean Lines
Soy foods have high amounts of nutrients so that consuming soy foods has a
number of health benefits for humans. However, eating soy foods has been limited by
allergic reactions, soy intolerance of sensitive individuals, undesirable beany flavor and
other limitations. Consequently, scientists developed several varieties of soybeans either
to lower the risk of undesired sensitive reactions, such as anaphylaxis and soy intolerance,
or alleviating unwanted beany flavor. Nowadays, several soybean lines, such as the high
oleic acid soybean lines, the low raffinose soybean lines, lipoxygenase free soybean lines,
17
low allergen content soybean lines, and some other soybean lines, were developed by
scientists to fix these limitations and may potentially improve the health benefits.
Commercial soybeans typically contain an average of 20% oil (dry weight).
Soybean oil is widely used for heat intensive using such as cooking and frying. However,
heat stability and oxidative stability of soybean oil has been terminated due to the
unwanted production of trans fats from hydrogenation of soybean oil. The alteration of
fatty acid compositions in soybean to improve soybean oil quality has been a long-time
goal of soybean researchers. High oleic acid soybean oil can target the recent demand of
trans-fat free soybean oil. Oleic acid is considered the most functional for food uses due
to its oxidative stability. High oleic acid soybeans typically contain at least 75% oleic
acid content and 6% polyunsaturated fatty acids (linoleic acid and linolenic acid), while
commercial soybeans have approximately 20-25% oleic acid content and 60% of
polyunsaturated fatty acids (Pham and others 2012). The high concentration of
polyunsaturated fatty acids contribute to the low oxidative and frying stability of soybean
oil, leading to rancidity, and shortened storage time of food products (Warner and Fehr
2008).
The raffinose family of oligosaccharides (RFOs) included in raffinose and
stachyose is made up of different β-galactosyl derivatives, which can be found in soybean
seeds (Hernandez-Hernandez and others 2011; Martíenz-Villaluenga and others 2007).
Raffinose and stachyose cannot be digested by humans or other monogastric animals
because of the lack of α-galactosidase activity within the gut. As a result, the RFOs are
considered as anti-nutritional factors. The low concentration level of raffinose and
stachyose in soybeans is a result of raffinose and stachyose synthase enzyme activities,
18
which initiate a galactosyl transfer from galatinol to sucrose (Dierking and others 2008).
Generally speaking, removing raffinose and stachyose from soybeans may increase
metabolism energy of a person on a soybean-supplemented diet and may reduce flatulent
productions (Coon and others 1990; Dierking and others 2008; Parsons and others 2000).
The soybean lipoxygenases are capable of producing unsaturated fatty acid
hydroperoxides by oxidizing the polyunsaturated fatty acids, such as linoleic and
linolenic acids. The polyunsaturated fatty acids are associated with the undesirable beany
or grassy flavors for soymilk (Torres-Penaranda and others 1998). In addition, the total
contents of hexanal in soymilk have been significantly associated with the lipoxygenases
contents in soymilk beverages (Wang and others 1998). Hexanal is the critical chemical
compound which provides the beany flavor in soymilk. Yuan and Chang (2007) indicated
that lower beany flavor and beany aroma (Kobayashi and others 1995) were present in
the soymilk produced by lipoxygenases-deficient soybean varieties.
The major allergenic soy protein, P34, has been classified as an immunodominant
allergen (Herman and others 2003). Herman and others (2003) and Helm and co-workers
(2000) reported up to 65% of soy sensitive participants in their studies had an allergic
reaction to this protein. The one mechanism of protein allergic reaction may be the
homology in the amino acid sequences among various allergenic proteins so that the
allergic reactions may occur due to the cross reactivity between similar allergens (Wilson
and others 2005). For instance, the allergy caused by peas and peanuts may be due to the
cross reactivity relationship between IgE antibodies against pea vicilin and peanut vicilin
(Wensing and others 2003). Since P34 contains approximately 70% sequence homology
with peanut’s main allergen (Ara h 1) and more than 50% with the cow’s milk allergen
19
(α-S1-casein), it will cause an allergic reaction when it meets the allergen, IgE antibodies,
just like the peanut vicilin. So reducing the P34 allergen content of soybeans may
decrease the possibility of an allergic reaction to soymilk beverages (Bilyeu and others
2009).
However, not many research studies investigated the sensory properties of these
new beans. Although soybeans have many benefits, it is useless if customers will not
accept them. Therefore, sensory evaluation of the new varieties of soybeans should be
done so as to successfully obtain and understand each single aspect of flavor in soybeans.
Since directly eating soybeans may be unacceptable to the panelists, plain flavor soymilk
could be used as a research project to study these soybeans.
2.5 Sensory Evaluation
Sensory analysis is widely used in food companies to investigate the influences on
ingredients, processing variables, and storage conditions of their products. It provides
developers with a perceptive of product quality, profiles of competing products and
consumer expectations (Stone and others 2004).
Sensory analysis methods consist of three main techniques, Descriptive Sensory
Evaluation, Discrimination Testing and Consumer Sensory Test.
2.5.1 Descriptive Sensory Evaluation
Descriptive Sensory Evaluation has been regularly used to distinguish among
aroma, flavor, appearance, after-taste and texture characteristics and quantify the
perceived intensities of the sensory attributes of a food product (Lawless and others
2010). According to these aspects of the perceived food product, sensory judges’ job is
20
discussing a list of attributes as comprehensive as possible, and then quantifying the
intensities of all the attributes. In addition, descriptive sensory evaluation can provide the
basis for mapping the similarities and differences of products so that the effect of specific
ingredient or process variables can be determined and explained. As a result, the
descriptive sensory analysis can describe both the qualitative and quantitative sensory
components of a food product by well-trained sensory panelists (Stone and others 2004).
Furthermore, descriptive analyses can also distinguish the differences among several
products on specific attributes. It doesn’t like other sensory analyses which provide
overall results (Lawless and others 2010). Consequently, descriptive sensory analysis is
frequently used in product development in food industry to determine how competitive
the new product is and how close this new product target consumer’s demands. However,
the panelists of descriptive analyses should never be consumers, since the sensory judges
should be trained, at least should be reproducible.
Generally speaking, a descriptive sensory analysis has between 8 to 12 panelists
which have been well trained. A brief statement or presentation of the objective before
training is necessary so that panelists won’t misunderstand the purpose of the sensory
study. Many sensory analyses failed because of misconstruing of the objective of the
study (Larmond 1987). During training sessions, reference standards are provided by the
panel leader to help panelists to understand and agree on the meaning of the attributes
used. After they reach agreements on all of the attributes, a quantitative scale for intensity
which allows the data to be statistically analyzed, should be provided by leader. Panelists
won’t be asked for their hedonic responses to the products, though it may bring few
21
influences on the results. However, this few influences can be eliminated by sensory
approaches (Lawless and others 2010 and Larmond 1987).
Descriptive Sensory Analysis including Flavor Profile® method, Texture Profile
®
Method, Quantitative Descriptive Analysis® (QDA), Spectrum
® Method, Quantitative
Flavor Profiling, Free-choice Profiling and Diagnostic descriptive analysis. These first
three methods are frequently used (Caul 1957; Brandt and others 1963 and Stone and
others 1974). The Flavor Profile® method focuses on describing the intensity of flavors
and aromas of food products, including the aftertaste and the overall impression. While,
the Texture Profile® method is the description of textural characteristics and mouth feel
of food products. QDA involves an unstructured scale, which associates with the specific
descriptive attributes about both flavor and textural attributes and indicates the intensity
(Stone and others 1974).
QDA is a developed method from the Flavor Profile® method and Texture
Profile® method. Compared with these two methods, QDA is not a consensus technique.
The data may not reach agreement by all of the sensory judges, though after a consent
discussion, because of the sensitivity of each judge. In addition, panel leader is not
participating in during the discussing session in QDA method (Meilgaard and others 2006;
Murray and others 2001; Stone and others 1980).
During the QDA training sessions, 10-12 panelists are asked to list all the possible
properties of the products and generate a series of terms to describe differences among
the product samples on their own, as completely as possible. After the individual work,
they discuss to develop a standardized definition and terminology for each sensory
22
attribute and to decide the standard references which are appropriately to be used to
anchor the descriptive terms (Murray and others 2001; Drake and others 2005; Drake and
others 2007). When the list of attributes is generated, judges are trained by creating a
consensus vocabulary. The leader’s responsibilities are only directing discussion,
providing materials, standard references and samples which are required by the panelists,
and offering individual training sessions based on the statistical analysis of the panelist’s
performance. As mentioned before that QDA is not a consensus method, during training,
one of the challenges is how to help judges determine the intensity of each characteristic.
Most sensory scientist assume that each penal will use different parts of the intensity
scale to make their decisions so that the isolate scale values are not the essential factor
(Civille and others 1986). Consequently, the QDA data must be regarded as relative
values instead of absolute values. Some statistical analysis methods, such as t-tests and
ANOVA can eliminate the effect of using different parts of the scale (Lawless and others
2010).
Therefore, QDA method has been widely used due to several reasons, including:
(1) It can evaluate all the sensory attributes and properties of product samples, even a
completely new one;
(2) It only need a limited number of sensory judges, usually 10-12 judges, to evaluate
the samples;
(3) It creates a sensory language only by panelists, and has no influence from the
panel leader;
(4) It has capability to estimate multiple products samples in one test;
(5) It can be analyzed by statistical method easily.
23
2.5.2 Discrimination Test
Discrimination Evaluation only allows the sensory panelists to answer questions
that whether the two samples are different or not. Panel leaders cannot determine that two
product samples differ from one another on a specific characteristic. In some cases,
sensory panelists may not distinguish the differences between samples, which can be
used in food industry as well (Meilgaard and others 2006; Stone and others 2004). For
instance, an ice cream food company may want to use the cheaper vanilla extract to
replace the expensive one in their vanilla ice cream. Meanwhile, they may not want the
consumers to perceive a difference in the product as well. In this situation, the
discrimination test can be used. If the result shows that judges cannot tell the difference
between two products, making this substitution will decrease the cost. However, if the
judges can perceive the variance, making this substitution will not be an appropriate way
to reduce the cost (Lawless and others 2010). Paired Comparison Tests, Triangle Tests,
Duo-Trio Test, ABX Discrimination Task, and A-Not-A Test, belong to the
discrimination analysis.
2.5.3 Consumer Sensory Evaluation
Consumer Sensory Testing is usually the last step before a product goes into
market. The panelists participating in this test are usually un-trained sensory judges, most
of them are consumers. There are two main consumer sensory tests for food products, the
preference testing and the acceptance testing. Both of them are called hedonic or affective
tests. In the preference testing, panelists have to make a choice between two products or
among several products. If product samples are only two, it is called a paired preference
test, which is the simplest and most popular sensory test to consumers, because paired
24
preference test mimics the processing of shopping, choosing among alternatives (Lawless
and others 2010). However, it is possible that a product wins in the paired preference test,
but loses in the real market, may be because of the insufficient of sample capacity, or
other reasons. Acceptance testing with a hedonic scale is designed to fix this problem
(Coetzee and others 1996).
In the acceptance or liking testing, panelists are required to mark the acceptability
of products on a hedonic scale without any comparison. The most common scale is 9-
point hedonic scale, which including Like Extremely, Like Very Much, Like Moderately,
Like Slightly, Neither Like or Dislike, Dislike Slightly, Dislike Moderately, Dislike Very
Much, and Dislike Extremely. “9” means Like Extremely, and “1” means Dislike
Extremely (Lawless and others 2010). Sensory scientists supply the samples with
randomly number to panelists one at a time, and then ask panels to indicate their hedonic
response on the 9-point hedonic scale so that the problem mentioned can be fixed. For
instance, sample A got a 5 scale in a liking test, while sample B got a 3 scale in the liking
test, though sample A had a higher score which can be regard as wining in the paired
preference test, but it didn’t indicate sample A will be the first choice for consumers. In
addition, a previous study demonstrated that whether the scale was printed horizontally or
vertically, or whether the scales was started from “Dislike Extremely” or “Like
Extremely”, did not show a significantly difference (Jones and others 1955).
25
CHAPTER 3
MATERIAL AND METHODS
3.1 Sample Preparation
3.1.1 Soy Milk Samples Preparation
Soy milk samples from five soybean lines were used in this study. The
preparation of each soy milk sample was in a food-grade lab. The following describes the
primary contents of interest for each soy milk product:
1) The first line of soybean had an ultra-low concentration of raffinose. The seeds
had a stachyose of content 0.3%, with no detectable raffinose (KB10-23#1681).
2) The second line of soybean had a high level of oleic acid (KB10-5#1137).
3) The third line of soybean not only contained a low level of raffinose, but was also
lipoxygenase free (KB10-25#1235).
4) The fourth line of soybeans had a low concentration of raffinose, a low
concentration of phytic acid, and also a low content of Gly m Bd 30K (P34)
allergen (KB09-5(2) #15).
5) The last line of soybean contained a high level of sucrose and clear hilum
(534545).
All these soybeans were provided by Dr. Bilyeu who is an adjunct professor at the
University of Missouri-Columbia. Moreover, all these beans were not genetically
modified.
26
Soybeans from each treatment were soaked in distilled water for 18 h at room
temperature, and then they were drained and washed with distilled water before making
the soymilk samples. The samples were made with soybeans and distilled water (the ratio
of soybeans and distilled water was 1:9 w/w). The beans and distilled water were added
into a soy milk maker (G3 soy milk maker, Soya Joy, Dandridge, TN). The blender in the
soy milk maker ground the beans with water and the temperature reached 100°C for 5
min to pasteurize these ingredients. Soymilk samples were removed from the soy milk
maker, filtered twice with a 7-inch stainless steel strainer and transferred into a clean
container. After that the samples were stored in a refrigerator at 4°C for 24 h. Before
conducting the sensory evaluation, all the samples were placed at room temperature for 2
h.
Plain soymilk samples were prepared for descriptive sensory evaluation and
physical and chemical properties measurement. Flavored soymilk samples were provided
for consumer acceptance sensory evaluation.
Flavored soymilk samples were prepared at the same ratio as the samples in the
descriptive sensory evaluation. In addition, 5% (w/w) sugar and 0.03% (w/w) vanilla
extract were added into the plain soymilk. The procedure for the flavored soymilk
samples was the same as the plain soymilk samples.
3.1.2 Soy Ice Cream Samples Preparation
Raw materials of soy ice cream were sugar (local grocery store), lecithin powder
(ADM, Decatur, IL), xanthan (Kelco, Chicago, IL), guar gum (Bob’s Red Mill,
Milwaukie, OR), pure vanilla extract (local grocery store), soybean oil (local grocery
27
store) and plain soy milks made from five varieties of beans described in the last
paragraph.
Soy ice cream samples in this study did not contain any dairy ingredient. They
were made by the five soy milk samples separately. All the solid ingredients, including
12% (w/w) sugar, 5% (w/w) lecithin, 0.15% (w/w) xanthan and 0.15% (w/w) guar gum
were mixed together and added into soymilk (74.4% w/w), then followed by adding 0.3%
(w/w) extract vanilla flavor into the system. After these ingredients were mixed well, 8%
(w/w) soybean oil were added and the mix was homogenized by using a lab homogenizer
(T25 digital ULTRA TURRAX®, IKA, Wilmington, NC) for 15 min. The speed of
homogenizer was set at 21,000 rpm. Next step was pasteurization. Each mix was
transferred into a 1.5 quart sauce pan, and heated to ca. 75°C for 20 s. The mixes were
cooled to room temperature, and then stored at 4°C overnight for ageing. Each aged mix
was poured into a homemade ice cream maker (a small freezer, 4 Quart Plastic Bucket
Ice Cream Maker, WESTBEND, West Bend, WI). The soy ice creams were packed in
500 mL plastic containers with lids and stored at -18°C for hardening.
3.2 Analytical Methods
3.2.1 pH Value
The pH value of five soymilk samples and five soy ice cream mixes was
determined with a pH meter (S20 SevenEasy TM
pH, Mettler Toledo, Columbus, OH).
Soymilk samples were placed in room temperature for 2 h before measured. Soy ice
cream mixes were stored in -4°C overnight, and measured at room temperature.
28
3.2.2 Color
The color of five soymilk samples and five soy ice cream mixes was measured
with a colorimeter (Konica Minolta CR-410, Mahwah, NJ) at room temperature. L*, a*,
b* represent lightness, redness and yellowness, respectively. L* ranges from 0 to 100.
Positive readings of “a*” indicate redness, while negative readings indicate greenness.
Positive readings of “b*” mean yellowness of the sample, and negative readings mean
blueness. Three readings were taken from each sample. The second and third readings
were taken after rotating the sample 90 degrees.
3.2.3 Viscosity
3.2.3.1 Viscosity for Soymilk
Physical viscosity of the soymilk samples was measured in triplicate at room
temperature. A HAAKE RS100 RheoStress Viscometer (HAAKE Buchler Instruments,
Paramus, NJ) was used and the rotation speed was 100 s-1
. In addition, the gap between
the two parallel plates was 0.5 mm. The diameter of top plate was 35 mm, and the
diameter of bottom plate was 35 mm.
3.2.3.2 Viscosity for Soy Ice Cream Mix
The viscosity of the aged mixes was measured using a concentric cylinder
viscometer (HAAKE, VT550 ThermoMC, Brookfield, Stoughton, MA), with a
cylindrical cone (60 mm of diameter) and a cylinder (41 mm of diameter). A thermostat
(TC-602, Brookfield, Stoughton, MA) adjusted the temperature of the sample.
Approximately 15 mL of sample were placed at 4 ± 0.3°C for 5 min, then subjected to a
program that increased the shear rate from 2.52 to 19.96 s-1
over 4 min. The power law
29
model was used to determine the flow behavior index and the consistency coefficient.
The power law was calculated as follow.
[τ is the shear stress (Pa), K the consistency index (Pasn), γ the shear rate (s
-1), and n is
the flow behavior index.]
3.2.4 Overrun
An overrun measurement was taken per sample by comparing the weight of ice
cream mix and ice cream in a 100 mL beaker. Overrun (%) was calculated as follows:
3.2.5 Melting Rate
A sample of ice cream was cut from the quarter gallon container, and weighed
100 to 120 g. The sample of ice cream (initially at -18°C) was placed on a wire screen
(around 8 holes per cm) on top of a funnel that was attached to a 50 mL beaker. The ice
cream was placed in a room temperature environment at 25°C. Every 5 min, for up to 90
min, the dripped weight was recorded (Muse and others 2004).
3.2.6 Hardness
Hardness of ice cream after 3 days at -18°C was measured by using a TA-HDi
Texture Profile Analyzer (Texture Tech. Corp., NY). The probe dimensions were 0.3 cm
in diameter and 15 cm in height. Before the tests, the samples were tempered at -14°C
overnight. The conditions for the analysis were as follows: 50 mm penetration distance
30
and probe speeds were 2 mm/s pre-penetration, 3 mm/s during penetration, and 5 mm/s
post-penetration (Pereira and others 2011). The hardness was measured as the peak
compression force (N) during the penetration of the sample. Three measurements were
taken from each sample and the results averaged.
3.2.7 Total Solids Content
The total solids content of soymilk samples was determined by drying 5 g of
samples in an air oven at 105ºC for 12 h. Total solids content was calculated as follows:
3.2.8 Fat Content
Fat content determination was according to the Modified Mojonnier Ether
Extraction Method (AOAC Official Method 989.05. 2007).
Dry weighing dishes were numbered and the weights were recorded.
Approximately 10 g soymilk sample was pipeted into a flask. One and half mL NH4OH
(Fisher Scientific, Fair Lawn, NJ) was added into the flask and the solution was mixed
well to neutralize any acid present. Three drops of phenolphthalein indicator
(phenolphthalein solution with alcoholic 1.0% volumetric indicator, Fisher Scientific,
Fair Lawn, NJ) were added into the solution to help sharpen the visual appearance of
interface between ether and aqueous layers during extraction. Ten mL 95% alcohol
(Fisher Scientific, Fair Lawn, NJ) were added into the flask and the flask was shaken
about 15 s. For the first extraction, 25 mL ethyl ether (Fisher Scientific, Fair Lawn, NJ)
were added and the flask was shaken about 1 min, then 25 mL petroleum ether (Fisher
31
Scientific, Fair Lawn, NJ) were added into the solution, and the flask was centrifuged at
600 rpm for 30 s (Models 5804R, Eppendorf 15 Amp Centrifuge, Hauppauge, NY) to
obtain clean separation of the aqueous and ether phases. Ether solution was decanted into
a weighing dish.
For the second extraction, 5 mL 95% alcohol were added into the aqueous, the
flask was shaken about 15 s. Next, 15 mL ethyl ether were added into the solution, the
flask was shaken about 1 min. Fifteen mL petroleum ether were placed in the solution,
and then the flask was centrifuged at 600 rpm for 30 s. For the third extraction, the steps
of the second extraction were repeated. Next, weighing dishes were placed in a hood
overnight to evaporate solvents. The weights of each weighing dish plus fat were
recorded. Fat content (%) was calculated as follows:
3.2.9 Protein Content
Protein content determination was according to the Mirochemical Determination
of Nitro-Micro Kjeldahl Method (AOAC Official Method 960.52. 2007).
Approximately 100 mg of soymilk sample were weighed and placed into a
digestion flask with about 40 mg CuSO4·10H2O (Fisher Scientific, Fair Lawn, NJ), about
1.9 g K2SO4 (Fisher Scientific, Fair Lawn, NJ), and 4 to 5 boiling chips (No.10 Sieve).
Ten mL 95%-98% H2SO4 (Sigma-Aldrich, Chemical Company, Milwaukee, WI) were
added to rinse any sample on the neck of the flask into the bulb. The digestion flask was
32
placed on a heating device for 2 h, until digestion completely finished (color was shown
as clear with light blue-green color). After the whole system was cooled down to room
temperature, the system was diluted into a 50 mL volumetric flask with distilled water up
to the 50 mL. Next, ten mL of the solution in the volumetric flask were taken out and
prepared for the distillation.
Five mL H3BO3 (2% (w/v) aqueous solution with mixed indicator, HMIS,
Arlington, TX) were placed in a 125 mL Erlenmeyer flask under a condenser with the tip
extending below the surface of the H3BO3 solution. The digestion solution and 10 mL 50%
NaOH (Fisher, Scientific, Fair Lawn, NJ) solution were added to the still, about 20 mL
distillate (indicator color changed from blue to green) were collected by the distillation
apparatus (Labcon North America, Petaluma, CA). The distillate was titrated by 0.05 M
HCl (Fisher, Scientific, Fair Lawn, NJ) to the end point (color changed from green to
purple). The volume of used HCl solution during titration was recorded. Blank was
determined. The test was duplicated for each treatment. Nitrogen content (%) was
calculated as follows:
Nitrogen constitutes 16% of most proteins and, hence, the total nitrogen content
of a sample was multiplied by 6.25 to arrive at the value of the crude protein.
33
3.3 Descriptive Sensory Evaluation for Soymilk
3.3.1 Training of Panelists
Twelve judges (nine female and three male) from the University of Missouri-
Columbia, all with a food science background, were selected for this descriptive sensory
evaluation. In the first training session, the leader gave a brief introduction about the
methodology and the procedure for making the soy milk samples. Meanwhile, five
soymilk samples were given to each of the judges. After the introduction, judges were
asked to taste the samples and then describe the sensory characteristics of these products
using their own language. After completion their individual task, judges shared their first
feeling about the soymilk samples, and decided collectively on an original list of
attributes describing these samples (the initial list of attributes is shown in Table 3.1).
Some references were prepared based on previous studies (Torres-Penaranda and others
1998 and Torres-Penaranda and others 2001 and Chamber and others 2006). From the
second to the fifth training session, the sensory leader provided all the references of each
attribute to help panelists correct the definition and terminology for all the attributes. In
the sixth training session, judges selected a final list of attributes (shown as Table 3.2)
and did a pre-test. According to the results of the pre-test, three panelists needed more
individual training sessions focusing on practicing the ABX test which belongs to the
discrimination test.
3.3.2 ABX Test
Panelists who need to do the ABX test finished the test individually. During the
ABX test, three samples (A, B and X) were given to each panelist. The sensory leader
34
told the panelists that the first two samples (A and B) were different, and asked each
panelist to identify which one was the same as the third one (X). This test was used to
train them to discriminate the specific attribute which they were not familiar with. The
test was replicated three times.
3.3.3 Descriptive Sensory Evaluation
Forty mL of the samples and references were presented in 3-oz white plastic cups
coded with three-digit numbers with lids on at room temperature (KB10-23#1681 was
represent by 177, KB10-5#1137 was represent by 742, KB10-25#1235 was represent by
882, KB09-5(2) #15 was represent by 216, and 534545 was represent by 529). Judges
were requested to record their responses on a 15 cm intensity line scale. They were
required to taste the reference first to perceive the intensity of the reference, and then to
taste and swallow the samples, and put the vertical marker on the 15 cm scale for one
attribute. After finishing one attribute, they were asked to rinse mouth by water (and
crackers, if necessary). After tasting each sample, they were instructed to rinse mouth by
water (and crackers, if necessary) and to take a break for at least 5 min. The order of
samples for each panelist was random. Panelists were asked to do three replicates of each
soymilk sample.
3.3.4 References Preparation
1% sucrose solution: 4 g sugar (local grocery store) was weighed and dissolved in
a 500 mL beaker with 400 mL distilled water.
2% sucrose solution: 8 g sugar was weighed and dissolved in a 500 mL beaker
with 400 mL distilled water.
35
0.02% caffeine solution: 0.08 g caffeine (Sigma, St. Louis, MO) was weighed and
dissolved in a 500 mL beaker with 400 mL distilled water.
0.04% caffeine solution: 0.16 g caffeine was weighed and dissolved in a 500 mL
beaker with 400 mL distilled water.
0.05% alum solution: 0.1 g alum (McCormick, Sparks, MD) was weighed and
dissolved in a 500 mL beaker with 400 mL distilled water, and then slightly
warmed with a stirring hot plate.
0.1% alum solution: 0.4 g alum was weighed and dissolved in a 500 mL beaker
with 400 mL distilled water, and then slightly warmed with a stirring hot plate.
0.2% alum solution: 0.8 g alum was weighed and dissolved in a 500 mL beaker
with 400 mL distilled water, and then slightly warmed with a stirring hot plate.
hexanal in oil (50 ppm): 12.5 mg of hexanal (Sigma-Aldrich Chemical Company,
Milwaukee, WI) was weighed and dissolved in a 20 mL beaker with soybean oil
(local grocery store). The solution was the transferred to a 250 mL volumetric
flask and rinsed 3 times and brought to the volume with soybean oil.
5% rice solution: 50 g of rice (local grocery store) was weighed and placed in a
blender (6640-022, 10 Speed Oster blender, Boca Raton, FL) with 500 mL
distilled water for 5 min. The rotate speed of blender was set at high speed-Grind.
After blending, the solution was transferred to a 1000 mL beaker and rinsed 3
times with 500 mL distilled water and stirred well. Supernatant was used as
reference during sensory evaluation.
10% rice solution: 100 g of rice (local grocery store) was weighed and placed in a
blender with 500 mL distilled water for 5 min. The rotate speed of blender was set
36
at high speed-Grind. After blending, the solution was transferred to a 1000 mL
beaker and rinsed 3 times with 500 mL distilled water and stirred well.
Supernatant was used as reference during sensory evaluation.
1.67% starch solution (unheated): 6.68 g starch (local grocery store) was weighed
and dissolved in a 500 mL beaker with 400 mL distilled water, and then stirred
the solution on a stirring hot plate for 10 min, without heating.
1.67% starch solution (heated): 6.68 g starch was weighed and dissolved in a 500
mL beaker with 400 mL distilled water, and then stirred and warmed with a
stirring hot plate to 90 °C for 10 min.
1% mushroom solution: 5 g dry mushroom (local grocery store) was weighed and
placed in a blender with 500 mL distilled water for 2 min, and the speed of the
blender was set at high speed-Blend. After blending, placed the mushroom
solution in a 600 mL beaker for 2 h. Supernatant was used as reference during
sensory evaluation.
3% mushroom solution: 15 g dry mushroom was weighed and placed in a blender
with 500 mL distilled water for 2 min, and the speed of the blender was set at high
speed-Blend. After blending, placed the mushroom solution in a 600 mL beaker
for 2 h. Supernatant was used as reference during sensory evaluation.
0.3% guar gum: 1.2 g guar gum was weighed and dissolved in a 500 mL beaker
with 400 mL distilled water, and then slightly warmed with a stirring hot plate.
0.4% guar gum: 1.6 g guar gum was weighed and dissolved in a 500 mL beaker
with 400 mL distilled water, and then slightly warmed with a stirring hot plate.
37
0.03% citric acid: 0.12 g citric acid (Sigma-Aldrich Chemical Company,
Milwaukee, WI) was weighed and dissolved in a 500 mL beaker with 400 mL
distilled water.
0.04% citric acid: 0.16 g citric acid was weighed and dissolved in a 500 mL
beaker with 400 mL distilled water.
References were prepared and completed in a food-grade lab, and were stored in a
refrigerator at 4°C for 24 h. Before doing the sensory evaluation, all the references were
placed at room temperature for 2 h.
3.4 Consumer Acceptance Test for Soymilk
For the consumer acceptance test, the flavored soymilk samples were provided.
Seventy five panelists from the University of Missouri-Columbia including students and
staff participated in this consumer acceptance evaluation. They were instructed in a brief
orientation session about the evaluation and characteristics of the samples. They were
asked to indicate their hedonic responses on unstructured 9 points line scales, anchored
by extremely like and extremely dislike on each end. Thirty mL of soymilk were
provided in 3-oz white plastic cups coded with 3-digit numbers at room temperature. The
order of samples for each panelist was random.
3.5 Statistical Analysis
The data was analyzed by Analysis of Variance (ANOVA) and the GLM
procedure in SAS 9.3 software (SAS Institute Inc., NC) to determine significant
parameters factors for the physical and chemical properties.
38
For sensory evaluation, ANOVA method was used to test the significant
difference among the soymilk samples for each attribute. The assessment scores of each
attribute were regressed on soymilks (fixed effect), judges (fixed effect), the interaction
of soymilks by judges (fixed effect), and replicates by judges (random effect). Least
square estimates and p-values were extracted from SAS output.
Furthermore, PCA method was applied to the mean centered data to estimate
loadings of each attribute at each principle direction. In descriptive sensory studies,
panelists are usually asked to evaluate several attributes of the samples. It is very possible
that not only one attribute has significantly influence on the samples, but also several
descriptors describe the same characteristic of the samples. For example, when sensory
panelists evaluate an ice cream sample on both aroma and flavor aspect. The milky aroma
and dairy flavor of the sample might be redundant and affect the same attribute of the ice
cream. In this situation, when some dependent variables are correlated with others, PCA
is the suitable technique to use (Heymann and others 1989). The redundancies, which
have the linear relationships, will be transferred into the new space to form a new set of
dependent variables, called principal components (PCs). Products will have new values
on these new variables (PCs) just as the original attributes which evaluated by judges.
The first principal component treats at the first eigenvalue and explains the largest
possible of the total variability in the data set. The following PCs demonstrate smaller
amounts of the total variance in the data set and have no interrelationships with prior PCs,
which mean that each PC is vertical to others (Lawless and others 2010; Heymann and
others 1989). If all the PCs are reserved, the PCA will act as a method that data are 100%
transferred. However, in fact, the major goal of PCA is amounting for as much of the
39
variability of the original data as possible by using as few of PCs as possible
(Borgognone and others 2001). Consequently, sensory scientists only analyze the first
few PCs to explain the main variance in the data set. In general, when the amount of
variance reaches 80% can be accepted (Kaiser 1960).
PAC is extensively used on sensory evaluation. A few examples are Owusu and
others (2011), Heenan and others (2010), Ghosh and Chattopadhyay (2012), Resano and
others (2010), He and others (2009), and Duart and others (2010). In addition, PCA can
be used to create an external preference mapping, which provides a revelation of the
relationships between the product samples and the individual characteristics (Lovely and
others 2009; Schmidt and others 2010).
40
Table 3.1 Initial list of attributes of soy milk samples
Attributes Reference and intensity
Flavor Sweetness 1% sucrose solution = 4.0
2% sucrose solution = 7.5
Bitter 0.02% caffeine solution = 4.0
0.04% caffeine solution = 7.5
Astringency 0.05% alum solution = 4.0
0.1% alum solution =7.5
Beany fresh bean sprout =7.5
hexanal in oil (50 ppm) = 5.0 (smell)
Rice-like 5% rice solution = 4.0
10% rice solution = 7.5
Oily Wesson vegetable oil =5.0 (smell)
Starchy 1.67% starch solution (heated) = 5.0
1.67% starch solution (unheated) = 7.5
Earthy 5 g ground mushroom soak in 500 mL
distilled water for 2 h = 4.0
15 g ground mushroom soak in 500 mL
distilled water for 2 h = 10.0
Mouth-
feel
Mouth Coating whole milk = 7.5
Teeth-etch 0.2% alum solution =7.5
Thickness 0.3% guar gum = 4.0
0.4% guar gum = 7.5
Silky Betty Crocker whipped butter cream
frosting = 10.0
41
Aftertaste Sourness 0.03% citric acid = 4.0
0.04% citric acid = 7.5
Bitter 0.02% caffeine solution = 7.5
42
Table 3.2 Final list of attributes of soy milk samples
Attributes Reference and intensity
Flavor Sweetness 1% sucrose solution = 4.0
2% sucrose solution = 7.5
Bitter 0.02% caffeine solution = 4.0
0.04% caffeine solution = 7.5
Astringency 0.05% alum solution = 4.0
0.1% alum solution =7.5
Beany fresh bean sprout =7.5
Rice-like 5% rice solution = 4.0
10% rice solution = 7.5
Mouth-feel Thickness whole milk = 7.5
Graininess 0.05% guar gum = 4.0
0.1% guar gum = 7.5
43
CHAPTER 4
RESULTS AND DISCUSSION
4.1 Soymilk Fat, Protein, and Total Solids Contents
Fat content and protein content of each sample did not show significant difference
(Table 4.1). For fat content, the range was from 2.09 to 2.14%. Sample KB10-5#1137
was slightly higher than others because of its high oleic acid amount. The range of
protein content was from 3.55 to 3.96%. Sample 534545 contained the highest amount of
protein compared to others. However, there were significant differences among samples
for total solids content. Sample KB10-5#1137 and sample 534545 contained significantly
more total solids than others. Because sample KB10-23#1681 contained ultra-low
amount of RFOs, which belong to oligosaccharides; sample KB10-25#1235 was made by
low RFOs and lipoxygenase free soybean; sample KB09-5(2)#15 not only contained low
amount of RFOs, but also contained low amount of P34 allergen and phytate. In addition,
Table 4.2 shows that the soluble carbohydrate content of each soybean was significantly
different (data provided by Dr. Bilyeu). The sucrose content and stachyose content of
sample 534545 was significantly higher than others that may contribute to the higher total
solids contents. Since the sucrose content determinations were finished by two batches,
the sample KB10-5#1137 was lower than others and raffinose and stachyose content were
not available. pH value of soymilk samples were shown in Table 4.3. The lowest pH
value was 6.54 from sample KB10-23#1681, and the highest pH value was 6.69 from
sample KB09-5(2)#15. However, there was no significant difference among samples.
44
Table 4.1 Composition of Soymilk Samples
Name of
soymilk
samples
RFOs Lipoxyge
-nase
P34
Allergen
Fatty
acids
Fat
(%)
Protein
(%)
Total
Solids
(%)
KB10-23#1681
Ultra-
low
RFOs
2.11a 3.91
a 4.72
b
KB10-5#1137
High
Oleic
Acid
2.14a 3.61
a 4.80
a
KB10-25#1235 Low
RFOs
l×1 l×2 l×3
null 2.13
a 3.55
a 4.76
ab
KB09-5(2)#15 Low
RFOs
Low
P34
Allergen
2.09a 3.63
a 4.54
b
534545 (Clean
Hilum) 2.11
a 3.96
a 4.92
a
*Means sharing the same letter within the same column do not differ at p < 0.05
45
Table 4.2 Soluble carbohydrate content of soybean lines (% of seed weight)
Soybean Lines Sucrose Raffinose Stachyose
Williams 82 (Commercial
soybeans) 6.7
c 1.0
a 5.6
a
KB10-23#1681 9.5a 0
d 0.3
c
KB10-5#1137 5.3 Not
Available
Not
Available
KB10-25#1235 8.3b 0.1
c 1.8
b
KB09-5(2)#15 8.3b 0.1
c 1.8
b
534545 10.0a 0.9
b 5.5
a
*Means sharing the same letter within the same column do not differ at p < 0.05
4.2 Soymilk Color
Color is an important property of food products. It relates to the quality of
products and could affect consumer acceptance. Generally speaking, consumers of
soymilk expect the product to have a pale yellow color. Therefore, commercial
manufacturers tend to maintain this natural color of soybean without adding color agent.
In this research, natural soybeans were used and no color agent added. Average of L, a, b
value were shown in Table 4.3.
Lightness, represented as L* value of soymilk samples varied between 86.90 and
96.76. The highest was from the sample 534545, which was significantly different from
others. In sample 534545, there was no hilum on the skin of soybeans, which usually is
black color, so that contributed to lightness of soymilk sample 534545. There was no
significant difference in redness (a* value) among each sample. Yellowness is
represented as positive value of b*. The range of yellowness was from 17.58 to 18.18.
46
There was significant difference between low RFOs group (sample KB10-23#1681,
sample KB10-25#1235, and sample KB09-5(2)#15) and regular RFOs content group
(sample KB10-5#1137 and sample 534545). Since the content of raffinose was low in
these soybeans, the content of raffinose polyester produced by raffinose and long chain
fatty acid may decrease. Raffinose polyester contributes to the yellowness of soymilk and
soybean oil (Akoh and Swanson 1987). As a result, these three soymilks had less yellow
color.
Table 4.3 pH value and color of soymilk samples
Samples pH Value Color
L* a* b*
KB10-23#1681 6.54a 86.90
b -3.27
a 17.58
b
KB10-5#1137 6.59a 87.92
b -3.23
a 18.18
a
KB10-25#1235 6.61a 91.65
b -3.34
a 17.63
b
KB09-5(2)#15 6.69a 87.36
b -3.24
a 17.66
b
534545 6.64a 96.76
a -3.29
a 17.99
a
*Means sharing the same letter within the same column do not differ at p < 0.05.
4.3 Soymilk Viscosity
Viscosity is an essential attribute for beverages. In commercial soymilk,
emulsifiers and stabilizers, such as guar gum, κ-carrageenan, and xanthan gum, may be
added to improve the quality of soymilk by promoting its thickness. The viscosity of
commercial soymilk is usually around 4.7 cP (20°C, 60 rpm). Some studies even
indicated increasing soymilk’s viscosity by 5 cP might significantly lower the astringency
flavor in soymilk (Courregelongue and others 1999). The mean values of viscosity of
47
soymilk samples were shown in Table 4.4. The range of viscosity was from 3.2 to 5.1 cP.
Sample KB10-5#1137 was significantly thicker than others, and sample KB09-5(2)#15
was the least viscous soymilk. Sample KB09-5(2)#15 contains low concentration of
raffinose, phytic acid and P34 allergen, and it also had least total solids content.
Therefore, the low amount of total solids content may contribute to low viscosity. The
thickest texture of sample KB10-5#1137 may be due to the oleic acid composition and
the molecular conformation of the triacylglyceride molecules. A cis double bond, present
in a long chain fatty acyl molecule, may lead to a bend or kink of the chain. Therefore,
the longer the chain of the fatty acyl groups, the more viscous the soymilk. The same
result was proved by Wang and Briggs (2002) in different soybean oils with modified
fatty acid compositions.
Table4.4 Viscosity of soymilk samples
Samples Apparent Viscosity (cP)
KB10-23#1681 4.1b
KB10-5#1137 4.9a
KB10-25#1235 4.3b
KB09-5(2)#15 3.2c
534545 4.0b
*Means sharing the same letter do not differ at p < 0.05
48
4.4 Descriptive Sensory Evaluation for Soymilk
When developing existing food products with new ingredients or new compounds,
it is critical to improve or at least maintain the sensory quality of the new one. If a new
compound or formulation fails to appeal to consumers, the new product will be
considered as an unsuccessful development. Therefore, using sensory evaluation to test or
monitor the potential changes of new products is very important. Descriptive sensory
evaluation, ABX test (for training step), and consumer acceptance test were used in this
study to determine changes in sensory attributes in the soymilk samples.
The ANOVA method helps to identify noticeable difference among soymilks for
each attribute. The ANOVA method yielded least square estimates for each combination
of soymilk and attribute, and they are listed in Table 3.2. The p-values for testing
significance among different soymilks are summarized in row 7 in Table 4.5. P=0.1 was
used as the cutoff for type I errors since it is not critical error to accept an attribute as
significant at this level. ANOVA method identified attributes “Beany” (p<0.0001),
“Rice-like” (p=0.054), “Sweetness” (p=0.078), and “Thickness” (p=0.082) to be
significant. Specifically, soymilk made by soybean KB10-25#1235 is least beany flavor;
soymilk sample KB10-25#1235 and KB10-5#1137 have most rice-like flavor among all;
soymilk samples 534545, KB10-23#1681, and KB09-5(2)#15 are sweeter than soymilk
samples KB10-5#1137 and KB10-25#1235; soymilk samples KB10-5#1137 is thicker
than others, and KB09-5(2)#15 is least thick among all. For other attributes, “Bitter”,
“Astringency”, and “Graininess”, there were no significant differences.
The possible reason for least beany flavor in KB10-12#1235 sample might be that
this sample was lipoxygenase free. The beany flavor components are mainly the
49
oxidation products of unsaturated lipids catalyzed by lipoxygenases (Yuan and others
2007). When no lipoxygenase exist in soybeans, polyunsaturated lipids will not be
catalyzed to produce the undesirable flavor. The result was proved in some other’s
studies (Torres-Penaranda and others 1998; Torres-Penaranda and others 2001;
Kobayashi and others 1995). During the cooking step, temperature reached 100 °C, the
high amount of oleic acid in sample KB10-5#1137 may degrade into octanal, heptanal,
nonanal, and decanal, which may contribute to rice-like flavor (Yang and others 2008).
The soymilk sample KB10-25#1235 has the highest result of rice-like flavor may due to
its least beany flavor. Since no obvious beany flavor cover rice-like flavor, panelists may
easily distinguish rice-like flavor. For the sweetness attribute, soymilk samples 534545
and KB10-23#1681 were the sweetest two samples, which got the same results on
sucrose content (Table 4.2, provided by Dr. Bilyeu). In sample 534545 and sample
KB10-23#1681, the sucrose contents were significantly higher than others including
William82, a commercial soybean lines. Although the sucrose content of KB09-5(2) was
significantly lower than sample 534545 and sample KB10-23#1681, the sensory result on
sweetness did not show significant difference between sample KB09-5(2)#15 and sample
534545 and no significant difference between sample KB09-5(2)#15 and sample KB10-
23#1681 either. This result may be due to sensitivity of panelists. The 1.7% difference of
sucrose content can be determined by apparatus and equipment, but it may not be easily
distinguished by human. Sample KB10-5#1137 is the thickest sample among others. This
may be due to its high oleic acid content. The high molecular weight of oleic acid (C18:0)
may cause the viscosity increase leading to the increase in the thickness of the soymilk
sample. The other possible reason is oleic acid could be considered as an emulsifier.
50
Soymilk is an emulsion system, contains carbohydrate, protein, lipid, and sugar
(Lakshmanan and others 2006). Increasing long-chain fatty acid molecules (oleic acid)
content may have same effect of adding emulsifiers (Matsumiya and others 2011).
The PCA method studies the relationships among attributes to analyze the
descriptive sensory data. The results show that the first four principal directions
explained the majority of the variation (80%) in the observations, PC1 explained 35.72%
of the variance in the data set, PC2 explained 18.40%, PC3 explained 15.84%, and PC4
explained 10.76% (Table 4.6). This suggests that the first four principal directions
explain the majority relationships of the seven attributes in the study.
51
Table 4.5 Sensory data with significance values from ANOVA and jack-knife estimates
Soymilk Sweetness Bitter Astringency Beany Rice-
like Thickness Graininess
KB10-
23#1681 2.813
a 3.434
ab 3.630
a 5.538
a 3.400
b 4.796
a 7.105
a
KB10-5#1137 1.792b 3.700
ab 3.841
a 5.607
a 4.155
ab 4.857
a 6.978
a
KB10-
25#1235 2.173
ab 3.752
a 3.671
a 3.828
b 5.175
a 4.814
a 7.486
a
KB09-5(2)#15 2.807a 2.534
b 2.888
a 5.670
a 4.159
ab 3.377
b 7.160
a
534545 2.830a 3.247
ab 3.449
a 5.594
a 3.762
ab 4.114
ab 6.963
a
S.D. 0.48 0.49 0.37 0.79 0.66 0.65 0.21
p(ANOVA) 0.078 0.138 0.107 <0.0001 0.054 0.082 0.958
p(PC1) 0.048 0.013 0.259 0.021 0.000 0.006 0.070
p(PC2) 0.340 0.067 0.058 0.287 0.405 0.491 0.272
p(PC3) 0.176 0.415 0.455 0.269 0.159 0.136 0.098
p(PC4) 0.281 0.013 0.312 0.077 0.007 0.443 0.407
* Pairwise comparisons use Scheffe’s multiple test adjustment at p-value=0.1.
52
Table 4.6 Variance explained by the seven principal directions
Variance Explained (%) Cumulative Variance (%)
PC1 35.72 35.72
PC2 18.40 54.12
PC3 15.84 69.96
PC4 10.76 80.71
PC5
PC6
PC7
7.44
6.55
5.30
88.15
94.70
100.00
Figure 4.1 is the scatter plot of scores for principal directions 1 and 2. The
principal direction 1 explains 36% of the variation compared to the principal direction 2
that explains 18% of the variation. The square symbol with a cross sign inside represents
scores for soymilk KB10-25#1235, and the plus sign denotes scores for other soymilks.
The arrow denotes the direction of “beany” attribute projecting on the 2-D plane. The
scores from KB10-25#1235 soymilk are at the opposite space compared to the “beany”
direction, which suggests that KB10-25#1235 is less beany than other soymilks. This
result is consistent with the ANOVA test.
53
Figure 4.1 Scores on principal direction 1 and 2 from PCA on mean centered data
Figure 4.2 is the scatter plot of loadings of all seven attributes for the principal
components 1 and 2, i.e. the projection of the directions of the seven attributes to the 2-D
plane spanned by the principal directions 1 and 2. The principal direction 1 was
dominated by “Beany”, “Sweetness”, “Thickness”, and “Rice-like”. The principal
direction 2 was controlled by “Astringency”, “Bitter”, and “Graininess”. By applying the
jack-knife estimation method, the significance of the loadings for each attribute for the
first four principal directions was tested. This is because the first four principal directions
explain the majority of the variation in observations. The p-values from tests are
summarized in Table 4.5. Specifically, the variation at the direction of principal
-4 -2 0 2 4 6
-4-2
02
46
Scores, Prin1 36%
Sco
res,
Prin2
18%
Beany
KB10-25
others
KB10-25
others
KB10-25
others
54
component 1 is explained by (at the significance level 10%) attributes “Sweetness”,
“Bitter”, “Beany”, “Rice-like”, and “Thickness”. Variation for principal direction 2 is
explained by attributes “Bitter” and “Astringency”. Similarly, principal direction 3
identifies attribute “Graininess”, and principal direction 4 identifies attributes “Bitter”,
“Beany”, and “Rice-like” as significant. Although “Beany” and “Rice-like” significantly
influence the same principal components, they are of opposite directions, which mean the
seven attributes in the study each represents a unique taste attribute of soymilk.
Figure 4.2 Loadings on principal direction 1 and 2 from PCA on mean centered data
-0.4 -0.2 0.0 0.2 0.4 0.6 0.8
-0.4
-0.2
0.0
0.2
0.4
0.6
0.8
Prin1 36%
Prin
2 1
8%
Sweetness
Bitter
Astringency
Beany
RicelikeThickness
Graininess
55
4.5 Consumer Acceptance Test for Soymilk
Soymilks made from five lines of soybeans were evaluated by 75 consumers.
Flavored soymilk samples with 5% (w/w) sugar and 0.03% (w/w) vanilla extract were
provided. The sensory properties of soymilks are shown in Table 4.7. There were no
significant differences of appearance and flavor among soymilk samples. In this study,
the sugar and vanilla extract may cover the beany flavor, rice-like flavor, and other
flavors so that consumers may not be able to distinguish the flavor property among the
soymilk samples. In addition, the light in each sensory booth may be different, so the
appearance characteristic may be influenced. However, there were significant differences
of overall and mouth-feel. Sample KB10-5#1137 had the highest score on mouth-feel
aspect. Combining with the result from descriptive sensory evaluation that sample KB10-
5#1137 was the thickest sample which may be preferred by consumers. As a result, the
overall score of sample KB10-5#1137 was the highest.
56
Table 4.7 Consumer acceptance of soymilk samples (Score as 9 = extremely like, 1 =
extremely dislike).
Samples Overall Appearance Flavor Mouth-Feel
KB10-23#1681 5.60b 6.57
a 5.51
a 5.54
b
KB10-5#1137 6.12a 6.44
a 5.28
a 6.03
a
KB10-25#1235 5.43b 6.53
a 5.32
a 5.46
b
KB09-5(2)#15 5.74b 6.48
a 5.43
a 5.41
b
534545 5.75b 6.81
a 5.28
a 5.60
b
*Means sharing the same letter within the same column do not differ at p < 0.05
4.6 Soy Ice Cream Color
The soy ice cream made from sample 534545 was significantly higher than others
in lightness, because there was no hilum on the skin of soybeans, which usually is black
color, so that contributed to lightness of the soy ice cream. However, there was no
significant difference in redness and yellowness between each sample. Although there
was significant difference of yellowness among the soymilk samples, adding lecithin,
which is yellow powder, might weaken the influence coming from oleic acid.
57
Table 4.8 Color of soy ice cream mixes
Samples Color
L* a* b*
KB10-23#1681 84.14b -2.50
a 27.96
b
KB10-5#1137 83.61b -2.87
a 28.68
a
KB10-25#1235 82.79b -2.63
a 27.26
b
KB09-5(2)#15 83.27b -2.38
a 27.05
b
534545 88.99a -1.88
a 29.59
a
*Means sharing the same letter within the same column do not differ at p < 0.05
4.7 Soy Ice Cream Texture Properties
Texture properties relate to the mouth feel of ice cream. The texture
characteristics of soy ice cream samples were determined by measuring the viscosity of
the mixes (Table 4.9), overrun, hardness (Table 4.10), and melting rate (Figure 4.3) of
soy ice creams. Sample KB10-5#1137 was significantly viscous than others. Since all of
the soy ice creams were added 12% (w/w) sugar, 5% (w/w) lecithin, 0.15% (w/w)
xanthan and 0.15% (w/w) guar gum, the only variance was the variety of soybeans, the
main reason that affect the viscosity of soy ice cream mixes may be as same as the
viscosity of soymilks. Though xanthan and guar gum are worked as emulsifier and
stabilizer to increase the viscosity of the mixes, sample KB10-5#1137 contained high
amount of oleic acid so that the double bond from oleic acid, present in a long chain fatty
acyl molecule, may cause a chain bending or chain kinking. Therefore, the viscosity of
soy ice cream mix from sample KB10-5#1137 was higher than other mixes.
58
Table 4.9 Viscosity of soy ice cream mixes
Samples Apparent Viscosity (cP)
KB10-23#1681 894.87b
KB10-5#1137 994.54a
KB10-25#1235 932.78b
KB09-5(2)#15 886.24b
534545 922.52b
*Means sharing the same letter do not differ at p < 0.05
Air is a critical component in ice cream affecting the physical and sensory
properties as well as the storage stability. Overrun which represents air bubble content,
affects to the creamy properties and the physical properties of melt down and hardness of
ice cream. Overrun value equals 100% means that the air makes up 50% of the ice cream
volume. The higher overrun value, the creamier texture the ice cream. Previous studies
showed that high apparent viscosity provided low overrun after whipping, which means
that if the mix is viscous, it might decrease air incorporation into ice creams (Kim and
others 2013). In this study, although milk was replaced by soymilk to make ice creams,
the same result was obtained. The mix of sample KB10-5#1137 was significantly thicker
than others and the overrun of this sample was significant lower than that of other soy ice
creams.
Hardness and melting down properties are influenced by the overrun, size of the
ice crystals, ice phase volume, and fat globule destabilization (Sofjan and Hartel 2004).
The range of hardness was from 7500.40 to 10985.27 N. Sample KB10-5#1137 was
significantly harder than other samples, and sample 534545 was significantly softer than
59
others. The lower overrun led to the harder soy ice cream, while the higher overrun led to
the softer soy ice cream. Furthermore, there was no soy ice cream samples melted before
30 min. Sample KB10-5#1137 started melting down at the 35 min and melted 78.3% of
sample during the 90 min period, while sample 534545 started melting down at the 50
min and melted only 32.3% of sample during the 90 min period. Samples were shown
significantly different among each other from the 35 min, and at the 90 min sample
KB10-5#1137 significantly melted more than others (data is shown in Appendix 8). The
lower overrun led the ice creams to melt more quickly. Muse and Hartel (2004) obtained
the same result in regular dairy ice cream. One possible reason is that high overrun could
contribute to a reduced rate of heat transfer due to a larger volume of air. Consequently,
for soy ice creams in this study, the high overrun related to soft texture and low melting
rate; the low overrun related to hard texture and high melting rate.
Table 4.10 Overrun value and hardness of soy ice cream samples
Samples Overrun (%) Hardness (N)
KB10-23#1681 26.18bc
9301.82b
KB10-5#1137 15.58c 10985.27
a
KB10-25#1235 33.94ab
8463.03b
KB09-5(2)#15 32.17b 8717.87
b
534545 38.30a 7500.40
b
*Means sharing the same letter within the same column do not differ at p < 0.05
60
Figure 4.3 Percentage of melting down of five soy ice cream samples
61
CHAPTER 5
CONCLUSION
The varieties did not significantly affect the fat content and protein content of the
soymilks, while sample 534545 was significantly higher than the other samples on total
solids content, because of its high sucrose content. In addition, sample KB10-5#1137 was
significantly higher than the other three samples on total solids content, but the reason is
not fully understood. Because of the high amount of fatty acid, sample KB10-5#1137 was
significantly more viscous than other soymilks. Since the soybeans that made sample
534545 did not have hilum on the seed skin, sample 534545 had the lightest color of the
soymilks. Moreover, samples that contained low RFOs concentration had a significant
lower yellow color than the samples that contained regular RFOs content. However,
there were no significant differences among samples on redness.
The results of descriptive sensory evaluation showed that lipoxygenase free
soymilk (sample KB10-25#1235) had the least beany flavor and most rice-like flavor.
High oleic acid content soymilk (sample KB10-5#1137) was the thickest one, due to its
fatty acid composition. Sample 534545 was the sweetest one. Consumer acceptance test
showed that sample KB10-5#1137 was the most accepted one. Though in descriptive
sensory analysis sample KB10-5#1137 was the least sweet one, the flavored soymilks
that were provided to consumers did not show any significant difference with other
samples on flavor and appearance aspects. In conclusion, even though sample KB10-
5#1137 showed some differences in the beany flavor attribute from other samples in the
descriptive analyses, there was significant difference in consumers’ preference. Therefore,
62
high oleic acid soymilk sample KB10-5#1137 can be considered an ideal formulation for
new soymilks.
The color of the soy ice creams showed the same result as soymilk color in that
the clear hilum sample was the lightest one, and low RFOs content samples had less
yellow color. High fatty acid content soy ice creams had higher viscosity and lower
overrun, due to a weak ability of air incorporation. Lower overrun led to hard and
unstable texture. Thus, sample KB10-5#1137 was the hardest soy ice cream and melted
more quickly than other samples. Although the high oleic acid soymilk got the best score
in the overall category from the consumer acceptance test due to its thicker mouth feel, it
did not show a good texture when making a soy ice cream.
Future studies could include:
1) To compare the five soymilk samples made from new soybean lines with the
commercial soymilk and cow milk;
2) To design a descriptive sensory evaluation of the soy ice cream samples;
3) To produce some other applications, such as tofu, tempeh, soy yogurt, etc and
investigate the effect of soybean lines on the physical, chemical and sensory
properties.
63
APPENDICES
APPENDIX 1. INFORMED CONSENT FOR CONSUMER
ACCEPTANCE TEST
I, (Date_________________) consent to participate in this research project and
understand the following:
PROJECT BACKGROUND: This project involves gathering human sensory data on
soy milk made from several varieties of soybeans. The data will be collected for analysis
and may be published. You must be at least 18 years of age to participate.
PURPOSE: The purpose of this sensory test is to study the beany flavor intensity and
consumer acceptance soy milks.
VOLUNTARY: This sensory test is entirely voluntary. You may refuse to answer any
question or choose to withdraw from participation at any time without any penalty or loss
of benefits to which you are otherwise entitled.
WHAT DO YOU DO? All participants of the sensory panel will attend a 5-min
evaluation session in which they taste several soy milks.
BENEFITS: Soy milk is a good source of protein and is lactose-free. For people who
are lactose intolerant or allergic to casein, soy milk is an excellent substitute of cow’s
milk. Drinking soy milk is often an acquired taste. To improve the acceptance of soy milk,
new varieties of soy beans are being developed to reduce the intensity of beany flavor in
soy milk. Your participation in this sensory test will facilitate the development of new
varieties of soy beans for soy milk application in the future.
RISKS: The expected risks are none other than those encountered in normal daily food
consumption. All products have either been prepared under sanitary conditions or are
commercially available products bought in a grocery store. If you know that you are
allergic to soy products (e.g. soybeans or soy proteins), you may NOT participate in this
study.
CONFIDENTIALITY: Your confidentiality will be maintained in that a participant’s
name will not appear on the ballot or in the published study itself. The data will only be
reported in aggregate form. Score sheets/data will be stored for a period of seven years
and then destroyed.
Thank you for your assistance in developing these soy milks. Although great
strides have been made in the instrumental analysis of foods, the development of new
foods still requires the human sensory response and feedback. Your efforts are greatly
appreciated. If you have any questions regarding the study, please contact Dr. Fu-hung
64
Hsieh at (573) 882-2444. If you have questions regarding your rights as a participant in
research, please feel free to contact the Campus Institutional Review Board at (573) 882-
9585.
Please keep one copy of this consent form for your records!
65
APPENDIX 2. CONSUMER ACCEPTANCE TEST
Date: .
This sensory test is about soy milk. Please swallow some of the soy milk at one time for
each sample, give the score of the sample, and then rinse your mouth with water to
continue to the next sample. Remember do not re-taste the sample during the first part of
this test.
First Part: Please choose one of the score in the forms below for each sample,
according to your experience.
Sample: #
Extremely
dislike
1
2 3 4
Neither
like or
dislike
5
6 7 8
Extremely
like
9
Appearance
(Color)
Flavor
Mouth feel
Overall
Sample: #
Extremely
dislike
1
2 3 4
Neither
like or
dislike
5
6 7 8
Extremely
like
9
Appearance
(Color)
Flavor
Mouth feel
Overall
66
Sample: #
Extremely
dislike
1
2 3 4
Neither
like or
dislike
5
6 7 8
Extremely
like
9
Appearance
(Color)
Flavor
Mouth feel
Overall
Sample: #
Extremely
dislike
1
2 3 4
Neither
like or
dislike
5
6 7 8
Extremely
like
9
Appearance
(Color)
Flavor
Mouth feel
Overall
67
Sample: #
Extremely
dislike
1
2 3 4
Neither
like or
dislike
5
6 7 8
Extremely
like
9
Appearance
(Color)
Flavor
Mouth feel
Overall
Second Part: Please rank the sample in an order from best sample to worst sample:
Sample # > Sample # > Sample # > Sample # >Sample #
68
APPENDIX 3. INDIVIDUAL TRAINING SESSION FOR
DECRIPTIVE SENSORY EVALUATION
ABX Test
Judge # Sample # Session # Date .
Directions: 1. Taste the sample based on the given order;
2. The left two samples (Reference A and Reference B) are different, please
determine which of the two samples match the last one and indicate by cycling the
code of the sample;
** If no similarity is apparent between the unknown sample and references, you must guess.
Reference A Reference B Unknown sample X
Code xxx xxx xxx
69
APPENDIX 4. DESCRIPTIVE SENSORY EVALUATION OF SOY
MILK
Judge # Sample # Session # Date:
Directions:
1. Please follow the order, from left to right, to test the sample. Write down your name
and the Sample Code;
2. Taste the reference for the first attributes, expectorate or swallow, to perceive the
intensity of reference;
3. Clean and rinse your mouth with water and crackers if needed;
4. Taste adequate sample to evaluate the intensity. Expectorate or swallow the sample;
5. Place a vertical mark on the horizontal line at the position that best describes the
perceived intensity of the given attribute;
6. Rinse your mouth by water (or crackers if needed) before you start the next attribute;
7. Start the next attribute by following Step 2 – 6;
8. After you finish one sample, please rinse your mouth by adequate water and crackers.
ATTENTION:
1. If you eat crackers, please make sure that you get rid of all crackers in your mouth,
since the crackers may have some negative influences on your results.
2. If you find a new attribute, which is not listed, please use the “Others” line on the
bottom of the last page to judge its intensity, so that this unknown attribute will not
affect your evaluation of the listed attributes. (You DO NOT need to write down the
attribute’s name of the “others”).
3. You can re-taste the sample among all the attributes, but please DO NOT go back to
samples you have already completed.
70
Flavor:
Sweetness:
None Strong
Bitter:
None Strong
Astringency:
None Strong
Beany:
None Strong
Rice-like
None Strong
71
Mouth-feel:
Thickness:
Thin Thick
Graininess:
Silky Grainy
Others:
None Strong
PLEASE:
1. Rinse your mouth by water and crackers;
2. Have a 5 min break, at least;
3. Make sure that all the residue of crackers are cleaned in
your mouth;
4. Start the next sample.
72
APPENDIX 5. CODE SERVING CONTAINERS FOR CONSUMER
ACCEPTANCE TEST
Subject No. Order in Tray
1 2 3 4 5
1 742 529 216 882 177
2 882 216 742 529 177
3 882 177 742 216 529
4 177 216 742 882 529
5 177 742 216 882 529
6 742 177 216 882 529
7 882 177 216 529 742
8 742 529 177 216 882
9 529 177 216 742 882
10 216 882 742 177 529
11 177 742 882 216 529
12 216 529 882 742 177
13 529 177 882 742 216
14 882 216 529 177 742
15 742 882 177 529 216
16 216 529 742 177 882
17 177 216 742 529 882
18 529 742 216 177 882
19 529 177 882 216 742
20 177 742 529 882 216
21 177 882 529 216 742
22 529 216 742 177 882
23 529 882 742 216 177
24 177 216 742 882 529
25 177 529 216 742 882
73
26 529 742 177 216 882
27 177 882 529 216 742
28 882 216 742 177 529
29 742 882 216 177 529
30 529 216 742 177 882
31 529 882 177 742 216
32 177 742 216 882 529
33 177 216 742 529 882
34 882 529 177 216 742
35 742 177 529 882 216
36 742 216 882 177 529
37 882 177 216 742 529
38 529 177 742 882 216
39 177 216 529 742 882
40 529 177 882 216 742
41 177 882 216 529 742
42 882 529 742 216 177
43 742 529 177 216 882
44 742 882 529 216 177
45 177 529 742 216 882
46 742 529 177 216 882
47 882 216 177 529 742
48 177 882 742 216 529
49 742 882 216 529 177
50 177 529 742 216 882
51 742 529 177 216 882
52 177 529 882 742 216
53 742 177 882 529 216
54 216 177 882 529 742
74
55 529 742 882 216 177
56 177 742 882 529 216
57 882 177 216 529 742
58 216 177 882 742 529
59 742 529 882 216 177
60 742 216 882 177 529
61 742 177 529 216 882
62 216 742 529 177 882
63 529 216 742 177 882
64 177 216 882 742 529
65 742 882 177 216 529
66 177 742 529 216 882
67 216 177 882 529 742
68 216 882 177 529 742
69 882 177 216 529 742
70 882 742 216 529 177
71 177 216 529 882 742
72 177 742 529 882 216
73 177 529 882 742 216
74 882 529 177 742 216
75 177 882 742 529 216
76 177 742 216 529 882
77 216 529 742 177 882
78 742 529 177 216 882
79 177 529 742 216 882
80 742 177 216 882 529
75
APPENDIX 6. CODE SERVING CONTAINERS FOR INDIVIDUAL
TRAINING SESSION (ABX TEST FOR SOME ATTRIBUTES)
Attributes Order in Tray
Reference A Reference B Unknown X
Astringency 177 529 X (same as 529)
177 529 X (same as 177)
216 177 X (same as 216)
216 177 X (same as 216)
Bitter 529 742 X (same as 529)
529 742 X (same as 742)
177 882 X (same as 882)
177 882 X (same as 177)
Beany 742 177 X (same as 177)
742 177 X (same as 742)
882 216 X (same as 882)
882 216 X (same as 882)
Thickness 882 529 X (same as 529)
882 529 X (same as 882)
216 529 X (same as 216)
216 529 X (same as 529)
Graininess 882 742 X (same as 742)
882 742 X (same as 882)
742 216 X (same as 742)
742 216 X (same as 216)
76
APPENDIX 7. CODE SERVING FOR DESCRIPTIVE SENSORY
EVALUATION
Session No. Judge
No.
Order in Tray
1 2 3 4 5
Session One 1 742 529 216 882 177
2 882 216 742 529 177
3 882 177 742 216 529
4 177 216 742 882 529
5 177 742 216 882 529
6 742 177 216 882 529
7 882 177 216 529 742
8 742 529 177 216 882
9 529 177 216 742 882
10 216 882 742 177 529
11 177 742 882 216 529
12 216 529 882 742 177
Session Two 1 141 183 418 417 304
2 141 304 183 418 417
3 304 418 141 183 417
4 141 417 304 183 418
5 417 183 418 141 304
6 418 417 183 141 304
7 304 183 418 141 417
8 304 417 141 418 183
9 141 418 183 417 304
77
10 141 183 418 304 417
11 417 304 141 183 418
12 418 141 304 417 183
Session
Three 1 556 132 868 359 849
2 868 132 556 849 359
3 849 359 556 868 132
4 849 868 556 132 359
5 849 132 359 868 556
6 868 849 359 132 556
7 359 868 849 132 556
8 132 868 556 849 359
9 849 556 132 868 359
10 132 849 359 868 556
11 868 132 556 359 849
12 556 132 868 359 849
*Sample 177, 141,132 are the same sample; Sample 742, 418,849 are the same sample;
Sample 882, 417, 556 are the same sample; Sample 216, 183,868 are the same sample;
Sample 529, 304, 359 are the same sample.
78
APPENDIX8. STATISCAL RESULTS OF MELTING DOWN
Time (min)
Melted Sample (%)
KB10-
23#1681
KB10-
5#1137
KB10-
25#1231
KB09-
5(2)#15 534545
10 0 0 0 0 0
15 0 0 0 0 0
20 0 0 0 0 0
25 0 0 0 0 0
30 0 0 0 0 0
35 0b 1.8
a 2.0
a 2.0
a 0
b
40 0.9b 2.7
a 3.0
a 3.1
a 0
c
45 1.9b 3.6
a 4.1
a 4.0
a 0
c
50 3.0b 4.7
a 5.0
a 5.0
a 0.8
c
55 5.1a 5.7
a 6.1
a 6.1
a 1.9
b
60 8.1a 8.8
a 9.1
a 8.0
a 3.9
b
65 11.0a 12.3
a 12.0
a 11.1
a 8.0
b
70 13.0b 16.0
a 16.0
a 15.0
a 11.0
c
75 16.9c 26.1
a 23.1
b 22.9
b 16.1
c
80 22.9c 40.2
a 35.9
a 31.0
b 20.3
c
85 30.1c 60.9
a 54.9
a 43.0
b 24.8
d
90 44.8c 78.3
a 71.2
a 58.1
b 32.3
d
*Means sharing the same letter within the same row do not differ at p < 0.05
79
APPENDIX 9. CODE FROM STATISTICAL ANALYSIS FOR
SENSORY EVALUATION
Average 3 datasets.
d1 <- read.csv("/Users/Tieming/consulting/Jing/data/data1.csv", header=TRUE)
d1 <- as.matrix(d1)
d2 <- read.csv("/Users/Tieming/consulting/Jing/data/data2.csv", header=TRUE)
d2 <- d2[,1:10]
d2 <- as.matrix(d2)
d3 <- read.csv("/Users/Tieming/consulting/Jing/data/data3.csv", header=TRUE)
d3 <- as.matrix(d3)
## dataset d: dim (60, 9, 3), each third dimension is a time of survey.
## First x Second dimension:
## obs x (7 attributes + addition evaluation + judge index).
d <- array(c(as.vector(d1[,-1]), as.vector(d2[,-1]), as.vector(d3[,-1])), dim=c(60, 9, 3))
mean.d <- apply(d, c(1,2), mean)[,1:7]
colnames(mean.d) <- colnames(d1)[2:8]
var.d <- apply(d, c(1,2), var)[,1:7]
colnames(var.d) <- colnames(d1)[2:8]
median.d <- apply(d, c(1,2), median)[,1:7]
colnames(median.d) <- colnames(d1)[2:8]
head(median.d)
## remove judge effect for each attribute.
soy <- as.factor(rep(1:5, times=12))
judge <- as.factor(rep(1:12, each=5))
data <- NULL
for(att.index in 1:7){
o <- lm(median.d[,att.index] ~ judge + soy)
summary(o)
o$coefficients
tmp <- median.d[,att.index] - rep(c(0, o$coefficients[2:12]), each=5)
data <- cbind(data, tmp)
}
colnames(data) <- colnames(median.d)
rownames(data) <- NULL
# Prepare mean-centered data
median.mean <- apply(data, 2, mean)
center.d <- data - array(rep(median.mean, each=60), dim=c(60, 7))
head(center.d)
Prin 1 vs. Prin 2
par(mar=c(5,5,1,1))
plot(o$scores[,1], o$scores[,2], xlab="Scores, Prin1 36%", ylab="Scores, Prin2 18%",
cex.lab=1.5, cex.axis=1.5, pch=19, xlim=range(o$scores[,c(1,2)]), ylim=range(o$scores[,c(1,2)]),
cex=1.3, type="n")
index.882 <- seq(3, 60, 5)
points(o$scores[index.882,1], o$scores[index.882,2], pch=7, cex=1.3)
points(o$scores[!(1:55)%in%index.882,1], o$scores[!(1:55)%in%index.882,2], pch=3, cex=1.3)
80
legend("topright", c("KB10-25","others"), cex=1.5,pch=c(7,3))
#install.packages("plotrix")
library(plotrix)
abline(h=0, v=0)
# draw 'beany' vector.
arrows(x0=0, y0=0, x1=o$loadings[4,1]*o$sdev[1]*4, y1=o$loadings[4,2]*o$sdev[2]*4,
length=0.1, lwd=2)
text("Beany", x=o$loadings[4,1]*o$sdev[1]*4, y=o$loadings[4,2]*o$sdev[2]*4+0.5, cex=1.5)
index.177 <- seq(1, 60, 5)
points(o$scores[index.177,1], o$scores[index.177,2], pch=19, col="blue", cex=1.5)
points(o$scores[seq(2,60,5),1], o$scores[seq(2,60,5),2], pch=19, col="purple", cex=1.5)
points(o$scores[seq(4,60,5),1], o$scores[seq(4,60,5),2], pch=19, col="yellow", cex=1.5)
points(o$scores[seq(5,60,5),1], o$scores[seq(5,60,5),2], pch=19, col="green", cex=1.5)
# Table for test
sweetness.est <- tapply(median.d[,1], soy, mean)
sd(sweetness.est)
bitter.est <- tapply(median.d[,2], soy, mean)
sd(bitter.est)
astringency.est <- tapply(median.d[,3], soy, mean)
sd(astringency.est)
beany.est <- tapply(median.d[,4], soy, mean)
sd(beany.est)
ricelike.est <- tapply(median.d[,5], soy, mean)
sd(ricelike.est)
thickness.est <- tapply(median.d[,6], soy, mean)
sd(thickness.est)
graininess.est <- tapply(median.d[,7], soy, mean)
sd(graininess.est)
# ANOVA p values
median.d <- as.data.frame(median.d)
o <- lm(Sweetness ~ soy + judge, data=median.d)
summary(o)
# PCA p values
dim(center.d)
head(center.d)
o <- princomp(center.d)
P <- o$loadings
tmp.pc1.loadings <- NULL
tmp.pc2.loadings <- NULL
tmp.pc3.loadings <- NULL
tmp.pc4.loadings <- NULL
for(soy.index in 1:5){
tmp.d <- center.d[soy!=soy.index,]
# center tmp.d
tmp.center.d <- apply(tmp.d, 2, mean)
tmp.center.d <- tmp.d - array(rep(tmp.center.d,each=48), dim=c(48, 7))
tmp.o <- princomp(tmp.center.d)
81
tmp.P <- tmp.o$loadings
#tmp.o$loadings
squared.sum.1 <- sum(((P-tmp.P)^2)[,1])
squared.sum.2 <- sum(((P+tmp.P)^2)[,1])
if(squared.sum.2 < squared.sum.1){tmp.P <- -tmp.P}
tmp.pc1.loadings <- cbind(tmp.pc1.loadings, tmp.P[,1])
squared.sum.1 <- sum(((P-tmp.P)^2)[,2])
squared.sum.2 <- sum(((P+tmp.P)^2)[,2])
if(squared.sum.2 < squared.sum.1){tmp.P <- -tmp.P}
tmp.pc2.loadings <- cbind(tmp.pc2.loadings, tmp.P[,2])
squared.sum.1 <- sum(((P-tmp.P)^2)[,3])
squared.sum.2 <- sum(((P+tmp.P)^2)[,3])
if(squared.sum.2 < squared.sum.1){tmp.P <- -tmp.P}
tmp.pc3.loadings <- cbind(tmp.pc3.loadings, tmp.P[,3])
squared.sum.1 <- sum(((P-tmp.P)^2)[,4])
squared.sum.2 <- sum(((P+tmp.P)^2)[,4])
if(squared.sum.2 < squared.sum.1){tmp.P <- -tmp.P}
tmp.pc4.loadings <- cbind(tmp.pc4.loadings, tmp.P[,4])
}
pc1.p <- NULL
for(i in 1:7){
sd <- sqrt(sum((P[i,1]-tmp.pc1.loadings[i,])^2)*4/5)
tmp.p <- pt(abs(P[i,1]/sd), 5, lower.tail=FALSE)
pc1.p <- c(pc1.p, tmp.p)
}
round(pc1.p, digits=3)
pc2.p <- NULL
for(i in 1:7){
sd <- sqrt(sum((P[i,2]-tmp.pc2.loadings[i,])^2)*4/5)
tmp.p <- pt(abs(P[i,2]/sd), 5, lower.tail=FALSE)
pc2.p <- c(pc2.p, tmp.p)
}
round(pc2.p, digits=3)
pc3.p <- NULL
for(i in 1:7){
sd <- sqrt(sum((P[i,3]-tmp.pc3.loadings[i,])^2)*4/5)
tmp.p <- pt(abs(P[i,3]/sd), 5, lower.tail=FALSE)
pc3.p <- c(pc3.p, tmp.p)
}
round(pc3.p, digits=3)
pc4.p <- NULL
for(i in 1:7){
sd <- sqrt(sum((P[i,4]-tmp.pc4.loadings[i,])^2)*4/5)
tmp.p <- pt(abs(P[i,4]/sd), 5, lower.tail=FALSE)
82
pc4.p <- c(pc4.p, tmp.p)
}
round(pc4.p, digits=3)
apply(tmp.pc2.loadings, 1, sd)
83
BIBLIOGRAOHY
Alzheimer’s Association. 2013. Alzheimer’s Facts and Figures. Available from:
http://www.alz.org/alzheimers_disease_facts_and_figures.asp#prevalence. Accessed
2013 August 17.
Aloia J, Shieh A. 2013. Intestinal Calcium Absorption Efficiency in Women and the
Influence of Menopause. In: Martin CJH, Watson RR and Preedy VR. Nutrition and Diet
in Menopause. Human Press. P67-78.
American Cancer Society. 2012. Cancer Basics. Available from:
http://www.cancer.org/cancer/cancerbasics/questions-people-ask-about-cancer. Accessed
2013 August 15.
American Heart Association Nutrition Committee, Lichtenstein AH, Appel LJ, Brands M,
Carnethon M. 2006. Diet and lifestyle recommendations revision 2006: a scientific
statement from the American Heart Association Nutrition Committee. Circulation 114:
82-96 Erratum in: Circulation (2006) 114(23):e629; Circulation (2006) 114(1):e27.
Anderson JW, Diwadkar VA, Bridges SR. 1998. Selective effects of different
antioxidants on oxidation of lipoproteins from rats. Proc Soc Exp Biol and Med 218:376-
81.
Anderson JW, Smith BM, Moore KA, Janna TJ. 2001. Soy Foods and Health Promotion.
In: Watson RR. Vegetables, Fruits, and Herbs in Health Promotion. Boca Raton: CRC
Press. p117-29.
84
Andani Z, Jaeger SR, Wakeling I, Macfie HJH. 2001. Mealiness in apples: towards a
multilingual Consumer vocabulary. J Food Sci 66(6):872-9.
AOAC Official Method 989.05. Fat Extraction. 2007
AOAC Official Method 960.52. Micro Kjeldahl Method. 2007
Akoh CC, Swanson BG. 1987. One-stage synthesis of raffinose fatty acid polyesters. J
Food Sci 52(6):1570-6.
Arjmandi BH, Kjalil DA, Hollis BW. 2002. Soy protein: Its effects on intestinal calcium
transport, serum Vitamin D and insulin-like growth factor-I in Ovariectomized Rats.
Calcif Tissue Int 70:483-7.
American College of Allergy, Asthma and Immunology. 2010. Milk Allergy. Available
from: http://www.acaai.org/allergist/allergies/Types/food-allergies/types/Pages/milk-
allergy.aspx. Accessed 2013 Auguet 17.
Astrup A, Dyerberg J, Elwood P, Hermansen K, Hu KB. 2011. The role of reducing
intakes of saturated fat in the prevention of cardiovascular disease: Where does the
evidence stand in 2010? Am J Clin Nutrit 93:684-8.
Bao XL, Lv Y, Yang BC, Ren CG, Guo ST. 2008. A study of the soluble complexes
formed during calcium binding by soybean protein hydrolysates. Food Chem 73:117-21.
Bao XL, Song M, Zhang J, Chen Y, Guo ST. 2007. Calcium-binding ability of soy
protein hydrolysates. Chin Chem Lett 18: 1115-8.
85
Bilyeu K, Ren CW, Nguyen HT, Sleper DA. 2009. Association of a four-basepair
insertion in the P34 gene with the low-gllergen trait in soybean. The Plant Genome 2:
141-8.
Björn I, Backström T 1999. Drug related negative side-effects is a common reason for
poor compliance in hormone replacement therapy. Maturitas 32:77-86.
Borgognone MG, Bussi J, Hough G. 2001. Principal component analysis in sensory
analysis: covariance or correlation matrix? Food Qual Prefer 12:323-6.
Boué SM, Isakova IA, Burow ME, Cao HP, Bhatnagar D. 2012. Glyceollins, soy
isoflavone phytoalexins, improve oral glucose disposal by stimulating glucose uptake. J
Agr Food Chem 60(25):6376-82.
Brandt MA, Skinner EB, Coleman JA. 1963. Texture profile method. J Food Sci
28(4):404-9.
Budhathoki S, Joshi AM, Ohnaka K, Yin G, Toyomura K. 2011. Soy food and isoflavone
intake and colorectal cancer risk: The Fukuoka Colorectal Cancer Study. Scand J
Gastroentero 46:165-72.
Caul JF. 1957. The profile method of flavor analysis. Adv Food Res 7:1-40.
Charoenphun N, Cheirsilp B, Sirinupong N, Youratong W. 2013. Calcium-binding
peptides derived for tilapia (Oreochromis niloticus) protein hydrolysate. Eur Food Res
Technol 236:57-63.
Chambers E, Jenkins A, Mcguire BH. 2006. Flavor properties of plain soymilk. J Sens
Stud 21: 165-79.
86
Civille GV, Lawless HT. 1986. The importance of language in describing perceptions. J
Sens Stud 1: 217-36.
Coetzee H, Taylor JRN. 1996. The use and adaptation of the paired-comparison method
in the sensory evaluation of Hamburger-type patties by illiterate/semi-literate consumers.
Food Qual Prefer 7: 81-5.
Coon CN, Leske KL, Akavanichan O, Cheng TK. 1990. Effect of oligosaccharide-free
soybean meal on true metabolizable energy and fiber digestion in adult roosters. Poult Sci
69:787-93.
Courregelongue S, Schlich P, Noble AC. 1999. Using repeated ingestion to determine the
effect of sweetness, viscosity and oiliness on temporal perception of soymilk astringency.
Food Qual Prefer 10:273-9.
Davis J, Higginbotham A, O’Connor T, Moustaid-Moussa N, Tebbe A. 2007. Soy protein
and isoflavones influence adiposity and development of metabolic syndrome in the obese
male ZDF rat. Ann Nutr Metabol 51:42-52.
Denter J, Rehm HJ, Bisping B. 1998. Changes in the contents of fat-soluble vitamins and
pro-vitamins during tempeh fermentation. Int J Food Microbiol 45:129-34.
Dierking EC, Bilyeu KD. 2008. Association of a soybean raffinose synthase gene with
low raffinose and stachysose seed phenotype. The Plant Genome 1(2):135-45.
Dijksterhuis GB, Punter P. 1990. Interpreting generalized procrustes analysis “analysis
variance” tables. Food Qual Prefer 2:255-65.
87
Drake MA, Yates MD, Gerard PD, Delahunty CM, Sheehan EM. 2005. Comparison of
differences between lexicons for descriptive analysis of Cheddar cheese flavor in Ireland,
New Zealand, and the United States of America. Int Dairy J 15: 473-83.
Drake MA, Jones VS, Russell T, Harding R, Gerard PD. 2007. Comparison of lexicons
for descriptive analysis of whey and soy protein in New Zealand and the U.S.A. J Sens
Stud 22:433-52.
Duart WF, Dias DR, Oliveira JM, Vilanova M, Teixeira JA. 2010. Raspberry (Rubus
idaeus L.) wine: Yeast selection, sensory evaluation and instrumental analysis of volatile
and other compounds. Food Res Int 43(9):2303-14.
Duffy R, Wiseman H, File SE. 2003. Improved cognitive function in postmenopausal
women after 12 weeks of consumption of a soya extract containing isoflavones. Plants
and the Central Nervous System 75(3):721-9.
Etaio I, Elortondo FJP, Albisu M, Gaston E, Ojeda M. 2008. Effect of wine-making
process and addition of white grapes on the sensory and physicochemical characteristics
of young red wines. Aust J Grape Wine Res 14:211-22.
File SE, Jarrett N, Fluck E, Duffy R, Casey K. 2001. Eating soya improves human
memory. Psychopharmacology 157: 430-6.
Food and Drug Administration. Department of Health and Human Services. Food
Labeling: Soy Protein and Coronary Heart Disease. 21 CFR Part 101. Docket No.98P-
0683. Federal Register 64:57700-2.
88
Freeman EW, Sherif K. 2007. Prevalence of hot flushes and night sweats around the
world: a systematic review. Climacteric 10:197-214.
Ghosh D, Chattopadhyay Parimal. 2012. Application of principal component analysis
(PCA) as a sensory assessment tool for fermented food products. J Food Sci Technol
49(3):328-34.
Gibbs BF, Zougman A, Masse R, Mulligan C. 2004. Production and characterization of
bioactive peptides from soy hydrolysate and soy-fermented food. Food Res Int 37:123-31.
Gooderham JM, Adlercreutz H, Ojala S, Wahala K, Holub BJ. 1996. A soy protein
isolate rich in genistein and daidzein and its effects on plasma isoflavone concentrations,
platelet aggregation, blood lipids and fatty acid composition of plasma phospholipid in
normal men. J Nutr 126:2000-6.
Gopalan S. 2011. Cow’s milk protein allergy and intolerance – practical issues in
diagnosis. Apollo Medicine 8(4):305-6.
Hao CM, Breyer MD. 2008. Physiological regulation of prostaglandins in the kidney.
Annu Rev Physiol 70:357-77.
He W, Hu XS, Zhao L, Liao XJ, Zhang Y. 2009. Evaluation of Chinese tea by the
electronic tongue: Correlation with sensory properties and classification according to
geographical origin and grade level. Food Res Int 42(10):1462-7.
Heenan SP, Dufour JP, Hamid N, Harvey W, Delahunty CM. 2010. The influence of
ingredients and time form baking on sensory quality and consumer freshness perceptions
in a baked model cake system. LWT - Food Sci Technol 43(7):1032-41.
89
Helm RM, Cockrell G, Connaughton C. 2000. Mutational analysis of the IgE-binding
epitopes of P34/Gly m Bd 30K. J Allergy Clin Immunol. 105(2): 378-84.
Herian AM, Taylor SL, Robert KB. 1990. Identification of soybean allergens by
immunoblotting with sera from soy-allergic adults. Int Arch Allergy Imm 92:193-8.
Herman EM, Helm RM, Jung R, Kinney AJ. 2003. Genetic modification removes an
immunodominant allergen from soybean. Plant Physiol 132:36-43.
Hernandez-Hernandez O, Côté GL, Kolida S, Rastall RA, Sanz ML. 2011. In vitro
fermentation of alternansucrase raffinose-derived oligosaccharides by human gut bacteria.
J Agric Food Chem 59:10901-6.
Heymann H. 1994a. A comparison of free choice profiling and multidimensional scaling
of vanilla samples. J Sens Stud 9:445-53.
Hirabayashi M, Matsui T, Yano H. 1998. Fermentation of soybean flour with Aspergillus
usamii improves Availabilities of zinc and iron in rats. J Nutr Sci Vitaminol 44:877-86.
Honore EK, Williams JK, Anthony MS, Clarkson TB. 1997. Soy isoflavones enhance
coronary vascular reactivity in atherosclerotic female macaques. Fertil Steril 67(1):148-
54.
Host A, Halken S, Jacobsen HP, Christensen AE, Herskind AM. 2002. Clinical course of
cow’s milk protein allergy/intolerance and atopic diseases in childhood. Pediatr Allergy
Immunol 13(15):23-8.
90
International Dairy Foods Association. 2013. We all scream for ice cream – Ice cream
and frozen novelty sales on the rise. Available from: http://www.idfa.org/resource-
center/industry-facts/ice-cream. Accessed 2013 August 19.
Jang EH, Ko JH, Ahn CW, Lee HH, Shin JK. 2010. In vivo and in vitro application of
black soybean peptides in the amelioration of endoplasmic reticulum stress and
improvement of insulin resistance. Life Sci 86:267-74.
Jiang B, Mu WM, Obiro W, Shi J. 2011. Traditional Chinese Functional Foods. In: Shi J,
Ho CT and Shahidi F. Functional Foods of the East. Boca Raton: CRC Press. p 13-20.
Jones LV, Peryam DR, Thurstone LL. 1955. Development of a scale for measuring
soldiers’ food preferences. Food Res 20:512-20.
Kaiser HF. 1960. The application of electronic computers in factor analysis. Educ
Psychol Meas 20:141-51.
Kasaoka S, Astuti M, Uehara M, Suzuki K, Goto S. 1997. Effect of Indonesian fermented
soybean tempeh on iron bioavailability and lipid peroxidation in anemic rats. J Agr Food
Chem 45:195-8.
Kim HJ, Bot A, Vries ICM, Golding M, Pelan EG. 2013. Effects of emulsifiers on
vegetable-fat based aerated emulsions with interfacial rheological contributions. Food
Res Int 53:342-51.
Kim J, Kang M, Lee JS, Inoue M, Sasazuki S. 2011. Fermented and non-fermented soy
food consumption and gastric cancer in Japanese and Korean populations: A meta-
analysis of observational studies. Cancer Sci 102:231-44.
91
King JM, Chin SM, Svendsen LK, Reitmeier CA, Johnson LA. 2001. Processing of
Lipoxygenase-Free soybeans and evaluation in foods. J Am Oil Chem Soc 87:353-60.
Kitazato H, Fujita H, Shimotomai T, Kafaya E, Narita T. 2002. Effects of chronic intake
of vegetable protein added to animal or fish protein on renal hemodynamics. Nephron
90:31-6.
Kitts DD. 2006. Calcium Binding Peptides. In: Mine Y and Shahidi F. Nutraceutical
Proteins and Peptides in Health and Disease. CRP press. New York. P 11-22.
Kobayashi A, Tsuda Y, Hirata N, Kutota K, Kitamura K. 1995. Aroma constituents of
soybean [Glycine max (L.) Merril] milk lacking lipoxygenase isozymes. J Agr Food
Chem 43:2449-52.
Komata T, Soderstrom L, Borres M, Tachimoto H, Ebisawa M. 2009. Usefulness of
wheat and soybean specific IgE antibody titers for the diagnosis of food allergy. Allergol
Int 58:599-603.
Kreijkamp-Kaspers S, Kok L, Grobbee DE, de Haan Eh, Aleman A. 2004. Effect of soy
protein containing isoflavones on cognitive function, bone mineral density, and plasma
lipid in postmenopausal women: a randomized controlled trial. J Amer Med Assoc
292:65-74.
Kritz-Silverstein D, Goodman-Gruen DL. 2002. Usual dietary isoflavone intake, bone
mineral density, and bone metabolism in postmenopausal women. J Altern Complement
Med 11:69-78.
92
Kwon DY, Jang JS, Hong SM, Lee JE, Sung SR.2007. Long-term consumption of
fermented soybean-derived Chungkookjang enhances insulinotropic action unlike
soybeans in 90% pancreatectomized diabetic rats. Eur J Nutr 46:44-52.
Lakshmanan R, Lamballerie MD, Jung S. 2006. Effetc of Soybean-to-water ratio and pH
on pressurized soymilk properties. J Food Sci 71:E384-91.
Larmond E. 1987. Sensory Evaluation can be Objective. In: Kapsalis JG. Objective
Methods in Food Quality Assessment. CRC. Boca Raton. p1-14.
Lawless HT, Heymann H. 2010. Sensory Evaluation of Food Principles and practices. 2nd
Edition. Springer. New York. 6p.
Lawless HT, Heymann H. 2010. Sensory Evaluation of Food Principles and practices. 2nd
Edition. Springer. New York. 227p.
Lawless HT, Heymann H. 2010. Sensory Evaluation of Food Principles and practices. 2nd
Edition. Springer. New York. 231-237p.
Lawless HT, Heymann H. 2010. Sensory Evaluation of Food Principles and practices. 2nd
Edition. Springer. New York. 79p.
Lawless HT, Heymann H. 2010. Sensory Evaluation of Food Principles and practices. 2nd
Edition. Springer. New York. 305p.
Lawless HT, Heymann H. 2010. Sensory Evaluation of Food Principles and practices. 2nd
Edition. Springer. New York. 326p.
93
Lawless HT, Heymann H. 2010. Sensory Evaluation of Food Principles and practices. 2nd
Edition. Springer. New York. 433p.
Lee JS. 2006. Effects of soy protein and genistein on blood glucose, antioxidant enzyme
activities, and lipid profile in streptozotocin induced diabetic rats. Life Sci 79:1578-84.
Lee SA, Shu XO, Li HL, Yang G, Cai H et al. 2009. Adolescent and adult soy food
intake and breast cancer risk: results from the Shanghai Women's Health Study. Am J
Clin Nutrit 89(6):1920-6.
Lee YB, Lee HJ, Sohn HS. 2005. Soy isoflavones and cognitive function. J Nutr
Biochem 16: 641-9.
Liu KS. 1997. Chemistry and nutritional value of soybean components. In: Soybeans:
Chemistry, Technology and Utilization. Gaithersburg: Aspen Publication, Inc. p 25-113.
Lovely C, Meullenet JF. 2009. Comparison of preference mapping techniques for the
optimization of strawberry yogurt. J Sens Stud 24:457-78.
Luan F, Liu HT, Wen YY, Zhang XY. 2008. Classification of the fragrance properties of
chemical compounds based on support vector machine and linear discriminant analysis.
Flavour Frag J 23:232-8.
Mackinnon ES, Rao LG. 2011. Traditional Chinese Functional Foods. In: Shi J, Ho CT
and Shahidi F. Functional Foods of the East. Boca Raton: CRC Press. p 277-96.
Maebuchi M, Samoto M, Kouno M, Ito R, Koikeda T. 2007. Improvement in the
intestinal absorption of soy protein by enzymatic digestion to oligopeptide in healthy
adult men. Food Sci Technol Res 13(1):45-53.
94
Malik S, Wong ND, Franklin SS, Kamath TV, L’Italien GJ, Pio JR, Williams GR. 2004.
Impact of the metabolic syndrome on mortality from coronary heart disease,
cardiovascular disease, and all causes in United States adults. Circulation 110:1245-50.
Martin N, MolimardP, Spinnler E, Schlich P. 2000. Comparison of odor profiles
performed by two independent trained panels following the same disruptive analysis
procedures. Food Qual Prefer 11: 487-95.
Martíenz-Villaluenga C, Gómez R. 2007. Characterization of bifidobacteria as starters in
fermented milk containing raffinose family of oligosaccharides from lupin as prebiotic.
Int Dairy J 17:116-22.
Matsumiya K, Nakanishi K, Matsumura Y. 2011. Destabilization of protein-based
emulsions by diglycerol esters of fatty acids – The importance of chain length similarity
between dispersed oil molecules and fatty acid residues of the emulsifier. Food
Hydrocolloid 25(4):773-80.
Meilgaard M, Civille CV, Carr BT. 2006. Sensory Evaluation Techniques. 4th
Edition.
CRC Boca Raton. 284p.
Messina MJ, Persky V, Setchell KD, Barnes S. 1994. Soy intake and cancer risk: a
review of the in vitro and in vivo data. Nutr Cancer 21:113-31.
McNagny SE, WengerNK, Frank E. 1997. Personal use of postmenopausal hormone
replacement therapy by women physicians in the United States. Ann Intern Med
127:1093-6.
95
Miksicek RJ. 1995. Estrogenic flavonoids: Structural requirements for biological activity.
Medicine. Res Exp 208:44-50.
Moller DE. 2001. New drug targets for type 2 diabetes and the metabolic syndrome.
Nature 414:821-7.
Murray JM, Delahunty CM, Baxter IA.2001. Descriptive analysis: past, present and
future. Food Res Int 34:461-71.
Muse MR, Hartel RW. 2004. Ice cream structural elements that affect melting rate and
hardness. J Dairy Sci 87:1-10.
Nechuta SJ, Caan BJ, Chen WY, Lu W, Chen Z. 2012. Soy food intake after diagnosis of
breast cancer and survival: an in-depth analysis of combined evidence from cohort
studies of US and Chinese women. Am J Clin Nutrit 96(1):123-32.
Ogawa T. 2006. Soybean. In: Mine Y and Shahidi F. Nutraceutical Proteins and Peptides
in Health and Disease. CRP press. New York. P461-75.
Ogawa T, Bando H, Tsuji H, Okajima H, Nishikawa K. 1991. Investigation of the IgE-
binding proteins in soybeans by immunoblotting with the sera of the soybean-sensitive
patients with atopic dermatitis. J Nutr Sci Vitaminol 37:555-65.
Ogborn MR, Nitschmann E, Bankovic-Calic N, Weiler HA, Aukema HM. 2010. Dietary
soy protein benefit in experimental kidney disease is preserved after isoflavone depletion
of diet. Exp Biol Med 235(11):1315-20.
96
Owusu M, Petersen MA, Heimdal H. 2011. Relationship of sensory and instrumental
aroma measurements of dark chocolate as influenced by fermentation method, roasting
and conching conditions. J Food Sci Technol. Published online.
Parsons CM, Zhang Y, Araba M. 2000. Nutritional evaluation of soybean meals varying
in oligosaccharide content. Poult. Sci. 18:581-4.
Pathomrungsiyounggul P, Lewis MJ, Grandison AS. 2010. Effects of calcium-chelating
agents and pasteurization on certain properties of calcium-fortified soy milk. Food Chem
118(8):8-14.
Peng CYC, Sankaran D, Ogborn MR, Aukema HM. 2009. Dietary soy protein selectively
reduces renal prostanoids and cyclooxygenases in polycystic kidney disease. Exp Biol
Med 234(7):737-43.
Pereira GDG, Resende JVD, Abreu LRD, Giarola TMDO, Perrone IT. 2011. Influence of
the partial substitution of skim milk powder for soy extract on ice cream structure and
quality. Eur Food Res Technol 232:1093-102.
Pham AT, Shannon JG, Bilyeu KD. 2012. Combinations of mutant FAD2 and FAD3
genes to produce high oleic acid and low linolenic acid soybean oil. Theor Appl Genet
125:503-15.
Reed SD, Lampe JW, Qu C, Gundersen G, Fuller S. 2013. Self-reported menopausal
symptoms in a racially diverse population and soy food consumption. Maturitas
75(2):152-8.
97
Resano H, Sanjuán AI, Cilla I, Roncalés P, Albisu LM. 2010. Sensory attributes that
drive consumer acceptability of dry-cured ham and convergence with trained sensory
data. Meat Sci 84(3):344-51.
Saba A, Rosati S. 2002. Understanding consumer perceptions of fermented yoghurt
products using conjoint and Generalized Procrustes Analysis. Italian J Food
Sci 14(4):339-50.
Sadowska-Krowicka H, Mannick EE, Oliver PD, Sandoval M, Zhang XJ. 1998.
Genistein and gut inflammation: role of nitric oxide. Proc Soc Exp Biol Med 217:351-7.
Savage JH, Kaeding AJ, Matsui EC, Wood RA. 2010. The natural history of soy allergy.
J Allergy Clin Immun 125:683-6.
Schmidt TB, Schilling MW, Behrends JM, Battula V, Jackson V. 2010. Use of cluster
analysis and preference mapping to evaluate consumer acceptability of choice and select
bovine M. Longissimus Lumborum Steaks cooked to various endpoint temperatures. Meat
Sci 84:46-53.
Sicherer SH, Sampson HA. 2006. 9. Food allergy. J Allergy Clin Immun 117(2): S470-5.
Siri-Tarino PW, Sun Q, Hu FB, Krauss RM. 2010. Saturated fat, carbohydrate, and
cardiovascular disease. Am J Clin Nutrit 91:502-9.
Slominski BA. 1994. Hydrolysis of galactoologosaccharides by commercial preparations
of alpha-galactosidase and beta-fructofuranose: potential for use as dietary additives. J
Sci Food Agric 65:323-30.
98
Sofjan RP, Hartel RW. 2004. Effects of overrun on structural and physical characteristics
of ice cream. Int Dairy J 14:255-62.
Soyfoods Association of North America. 2011. Sales and Trends. Available from:
http://www.soyfoods.org/soy-products/sales-and-trends. Accessed 2013 August 18.
SoyInfo Center. 2001. Chronology of Soymilk Worldwide 1500 A.D. to 1949. Available
from: http://www.soyinfocenter.com/chronologies_of_soyfoods-soymilk.php. Accessed
August 13.
Stone H, Sidel JL. 1998. Sensory Evaluation Practices. 2nd
Edition. Academic Press, Inc
San Diego, California. 338p.
Stone H, Sidel JL. 2004. Sensory Evaluation Practices. 3rd
Edition. Academic Press, Inc.
San Diego, California. 202p.
Stone H, Sidel JL, Bloomquist J. 1980. Quantitative descriptive analysis. Cereal Food
World 25:624-34.
Stone H, Sidel JL, Oliver S, Woolsey A, Singleton RC. 1974. Sensory evaluation by
quantitative descriptive analysis. Food Technol 28(11):24-34.
Suratman LLI, Jeon IJ, Schmidt KA. 2004. Ability of cyclodextrin to entrap volatile
beany flavor compounds in soymilk. J Food Sci 69:109-13.
Tabachnik L, Fidell B. 2006. Using Multivariate Statistics. 5th
Edition. Allyn and Bacon.
p1-5.
99
Taku K, Melby MK, Nishi N, Omori T, Kurzer MS. 2011. Soy isoflavones for
osteoporosis: An evidence- based approach. Maturitas 70(4):333-8.
Tikkanen MJ, Wahala K, Ojala S, Vihma V, Aldercreutz H. 1998. Effect of soybean
phytoestrogen intake on low density lipoprotein oxidation resistance. Proc Soc Exp Biol
Med 95:3106-10.
The Ohio State University Wexner Medical Center. 2007. Lactose Intolerance. Available
from:http://medicalcenter.osu.edu/patientcare/healthcare_services/digestive_disorders/lac
tose_intolerance/Pages/index.aspx. Accessed 2013 August 18.
Torres-Penaranda AV, Reitmeier CA, Wilson LA, Fehr WR, Narvel JM.1998. Sensory
characteristics of soymilk and Tofu Made from lipoxygenases-free and normal soybeans.
J Food Sci. 63(6):1084-7.
Torres-Penaranda AV, Reitmeier CA. 2001. Sensory descriptive analysis of soymilk.
Sens Nutr Qual Food 66(2):352-6.
United Soybean Board. Soyfoods Guide. 2012. United Soybean Board. Chesterfield. p 1-
28.
USDA database for the isoflavone content of selected foods. 2008. Nutrient Data
Laboratory, Beltsville Human Nutrition Research Center, Agricultural Research Service.
p 17.
Wang ZH, Dou J, Macura D, Durance TD, Nakai S. 1998. Solid phase extraction for GC
analysis of beany flavours in soymilk. Food Res Int 30:503-11.
100
Wang T, Briggs JL. 2002. Rheological and thermal properties of soybean oils with
modified FA compositions. J Am Oil Chem Soc 79(8):831-6.
Wei P, Liu M, Chen Y, Chen DC. 2012. Systematic review of soy isoflavone
supplements on osteoporosis in Systematic review of soy isoflavone supplements on
osteoporosis in women. APJTM 5(3):243-8.
Warner K, Fehr W. 2008. Mid-oleic/ultra low linolenic acid soybean oil: a healthful new
alternative to hydrogenated oil for frying. J Am Oil Chem Soc 85:945-51.
Wei P, Liu M, Chen Y, Chen DC. 2012. Systematic review of soy isoflavone
supplements on osteoporosis in women. Asian APJTM 5(3):243-8.
Wensing M, Knulst AC, Piersma S, O’Kane F, Knol EF, Koppelman SJ. 2003. Patients
with anaphylaxis to pea can have peanut allergy caused by cross-reactive IgE to vicilin
(Ara h 1). J Allergy Clin Immunol 111:420-4.
Wilson S, Blaschek K, Mejia EG. 2005. Allergenic protein in soybean: Processing and
reduction of P34 allergenicity. Nutr Rev 63(2):47-58.
Wood RA. 2003. The natural history of food allergy. Pediatrics 111:1631-7.
Wu AH, Yang D, Pike MC. 2000. A meta-analysis of soyfoods and risk of stomach
cancer: the problem of potential confounders. Cancer Epidemiol Biomarkers Prev
9:1051-8.
Wu AH, Ziegler RG, Horn-Ross PL. 1996. Tofu and risk of breast cancer in Asian-
Americans. Cancer Epidemiol Biomarkers Prev 5:901-6.
101
Xiao CW. 2008. Health effects of soy protein and isoflavones in humans. J Nutr
138(6):1244s-9s.
Yan L, Spitznagel EL. 2005. Meta-analysis of soy food and risk of prostate cancer in men.
In J Cancer 117:667-9.
Yang DS, Shewfelt RL, Lee KS, Kays SJ. 2008. Comparison of odor-active compounds
from six distinctly different rice flavor types. . J Agr Food Chem 56:2780-7.
Yang G, Shu XO, Li HL, Chow WH, Cai H. 2009 Prospective cohort study of soy food
intake and colorectal cancer risk in women. Am J Clin Nutrit 89:577-83.
Yang G, Shu XO, Li HL, Chow WH, Cai H, Zhang XL, Gao YT, Zheng W. 2009.
Prospective cohort study of soy food intake and colorectal cancer risk in women. Am J
Clin Nutrit 89:577-83.
Yang G, Shu XO, Li HL, Chow WH, Wen WQ. 2013. Prediagnosis soy food
consumption and lung cancer survival in women. JCO 31(12):1548-53.
Yang WS, Va P, Wond MY, Zhang HL, Xiang YB. 2011. Soy intake is associated with
lower lung cancer risk: results from a meta-analysis of epidemiologic studies. Am J Clin
Nutrit 94(6):1575-83.
Yuan SH, Chang SKC. 2007. Selected odor compounds in soymilk as affected by
chemical composition and lipoxygenases in five soybean materials. J Agr Food Chem
55:426-31.
102
Zimmermann C, Cederroth CR, Bourgoin L, Foti M, Nef S. 2012. Prevention of diabetes
in db/db mice by dietary soy is independent of isoflavone levels. J Endocrinol
153(11):5200-11.