Review Article Study Designs in Genetic Epidemiologygenetic epidemiology is one of the main di erences between classic and genetic epidemiology. In Table , main designs in genetic

Hindawi Publishing CorporationISRN GeneticsVolume 2013 Article ID 952518 8 pageshttpdxdoiorg1054022013952518

Review ArticleStudy Designs in Genetic Epidemiology

Leyla Sahebi1 Saeed Dastgiri2 Khalil Ansarin1

Roya Sahebi3 and Seyyed Abolghasem Mohammadi4

1 Tuberculosis and Lung Disease Research Center Tabriz University of Medical Sciences Iran2 Tabriz Health Services Management Research Centre Department of Community and Family Medicine School of MedicineTabriz University of Medical Sciences Tabriz 5166615739 Iran

3 Tabriz University of Medical Sciences Iran4Department of Plant Breeding and Biotechnology School of Agriculture University of Tabriz Iran

Correspondence should be addressed to Saeed Dastgiri saeeddastgirigmailcom

Received 16 March 2013 Accepted 15 April 2013

Academic Editors M M DeAngelis J Moreaux and X-X Zhang

Copyright copy 2013 Leyla Sahebi et al This is an open access article distributed under the Creative Commons Attribution Licensewhich permits unrestricted use distribution and reproduction in any medium provided the original work is properly cited

Genetic epidemiology as a relatively new issue aims to explore the independent role of genetic-environmental determinants ofdiseases Genetic epidemiology studies depending on the objective encompass the most preliminary surveys from the attempts tofind family history in the occurrence of diseases to the most advanced surveys including specific strategies by clinical trials in theprevention of genetic diseases Different objectives in genetic epidemiology studies require special methods and study designs Inthis review chief designs including familial aggregation heritability segregation linkage and association are evaluated likewise thepurpose of diverse kinds of studies and analyses is briefly discussedThe utilization of study designs and related analyses accordingto the aims are the main issues and necessary in the accurate implementation of the study Some methodological issues in relationto studies on tuberculosis are also reported Attention to these issues might be useful in the implementation of these methods inthe studies designed for the prevention and treatment of genetic disorders

1 Introduction

Epidemiology is the study of distribution and determinantsof disease frequency in human populations and the use ofthis information to control and promote health [1] Thegoal of epidemiologic research is to collect valid and preciseinformation on the causes prevention and treatment ofdisease [1] Genetic epidemiology is the study of the roleof genes and their interaction with environmental factorsin the occurrence of disease in human populations [2] Thebranch of genetic epidemiology is still quite young althoughthe parents of that (epidemiology and genetics) have ratherlong history [3] The objectives of epidemiological studies ingenetic science are to determine the risks related to allelicvariants of candidate genes to map more accurately regionsof the genome for which there is evidence of linkage todisease susceptibility and to contribute cases to a genome-wide search for susceptibility genes [4]

2 Study Designs in Classic Epidemiology

The selection of one design over another in studies dependson the particular research question [3] and also on costtime and ethical considerations The most common types ofstudies are listedwith brief explanations about them inTable 1[1 5ndash8]

3 Study Designs in Genetic Epidemiology

Similar to classical epidemiology observational studies ingenetic epidemiology are divided into descriptive and ana-lytical studies In descriptive studies the pattern of variationin disease or behavior among immigrants familial groupsas well as racialethnic groups social classes and temporalage and gender variations is surveyed and can provide cluesto whether genetic or environmental factors are involved[8] In analytical studies the effect of genetic exposure on

2 ISRN Genetics

Table 1 Main study designs in classic epidemiology

Case reports Case series Ecologicalstudies

Cross sectionalstudies Case control Cohort Intervention

trials

Unit of study (Single case)individual

(gt1 cases)individual Population Individual Individual Individual Individual

Description

Describesunusual

characteristicsof a case

A group of cases

Comparingpopulations indifferent placesat the same timeor in a time

series

Study ofpopulation atldquopoint-in-timerdquo

Study of twogroups of

subjects (casedisease ofinterest andcontrol

disease-free)

Study of twogroups ofsubjects

(exposure andnon-exposure

groups)

Study andexamine twogroups ofsubjects

(interventionand controlgroups)

Direction Present Present Present Present Reverse Forward Forward

Type ofmeasurement

Reportingdescription

Reportingdescription Correlation Prevalence

association Odds ratio

Prevalenceincidencerelative risk

attributable risk


attributable risk

Advantages

Quick havingclinical

importanceopportunities

for physicians toexchange ofthoughts




Quickinexpensivegroup-level

studies may alsobe the only wayto study theeffects of

group-levelconstructs forexample laws

Easyinexpensiveuseful for

investigatingfixed exposuressuch as bloodgroup mostconvenient inoutbreaks of

disease

Relativelyinexpensivegood for rare

diseasesEfficient in

resources andtime

Better for rareexposuresability todeterminecausalityrelations

The strongestevidence forcausalitycontrol ofunknown

confoundersfulfils the basicassumption ofstatistical

hypothesis tests

limitation

Inability todeterminestatistical

relations andanalysis



Ecologicalfallacy

Susceptible toselection biasand misclassifi-cation difficultto establish a

putative ldquocauserdquoNot good forrare diseases orrare exposures

Susceptible toselection bias

andmisclassificationbias may bedifficult to

establish thatldquocauserdquo

preceded effect

Costly and timeconsumingsusceptible toselection biasRelativelystatisticallyinefficient

unless disease iscommon

Expensive timeconsumingsometimesethically

generalizabilityproblem

disease is analyzed and if it is proved that there are causalrelationships the genetic region responsible for the diseaseis identified Therefore relying on familial relationship ingenetic epidemiology is one of the main differences betweenclassic and genetic epidemiology In Table 2 main designsin genetic epidemiology and process of surveys in geneticdiseases are summarized [8ndash16]

4 Analytical Study in Genetic Epidemiology

41 Familial Aggregation The first step in the study of a po-tentially genetic characteristic is to realize whether it tendsto aggregate in families without having any specific geneticmodel in mind [11] In other words familial aggregation isthe tendency for disease to cluster in families [17] A simpleapproach to assess aggregation in families is to identify agroup of affected case subjects and a group of healthy controlsubjects and compare the odds of a positive family history incase subjects who report a disease to the odds of a positivefamily history in the control group Finally an odds ratio canbe estimated according to Table 3 [1]

While this approach is valuable there is no way to controlfor individual or environmental risk factors for each relativewhichmight be driving the aggregation (eg cooccurrence ofsmoking behaviour and lung cancer aggregation in families)it is also imprecise since the probability of a positive familyhistory of disease increases with age and the number ofrelatives considered [17] One way to control confoundingvariables is to use categorizing methods such as logisticregression modeling [17]

42 Heritability Heritability is the proportion of variation ina trait that is attributable to genetic differences One of thesimplest designs in heritability study is the ldquotwin studyrdquo Thisstudy design uses the variance component framework and itsfrequency to estimate heritability in this way if the ratio ofgenetic variance to phenotypic variance (general heritability)is close to 1 this is evidence for a genetic component [18]

43 Segregation Analysis Segregation analysis aims to deter-minewhether the pattern of disease occurrence in families fitsa particular type of inheritance For example Huntingtonrsquos

ISRN Genetics 3

Table 2 Main study designs in genetic epidemiology

Research questionproblem Type of study Unit of study Type of measurement Aim

Is there evidence ofphenotypic aggregationwithin families

Familial aggregationCase-parentCase-parent-grandparent

Recurrence risk ratioCorrelationOdds ratioConditional regressionlogistic

Identifying new diseasegenes

Is the pattern ofaggregation reliable with aneffect of genes

Heritability Twin study Variance componentheritability


Is there evidence of a genewith subnational enougheffects to justify expensivestudies

Segregation Familial pedigreeMaximum likelihoodestimationBinomial distribution


Where in the genome is acausative gene most likelyto lie

LinkageCase disease of

interest and controlgroup disease-free

Parametric maximumlikelihoodNonparametric meanproportion chi squarelikelihood test


Is there a causativepolymorphism

AssociationLinkage disequilibrium

CohortCase only study

Relatedmdashcase controlUnrelated-case

controlExposure and

non-exposure groupCase groups

Transmissiondisequilibrium testChi squareindependenceOdds ratioConditional regressionlogistic CorrelationLewontinersquos D primeRelative riskAttributable riskExposure odds ratioestimate

Characterizing knowndisease genes

Estimate allele frequency Cross-sectional PrevalenceCorrelation


Estimating penetrance AssociationCohort

Relatedmdashcase controlFamilial cohort

Odds ratioConditional regressionlogisticRisk ratioAttributable risk


Evaluating strategies forprevention of geneticdiseases

ExperimentalRetrospective cohort

Clinical trial(intervention andcontrol group)Exposure and

non-exposure groups

Risk ratioAttributable riskRisk ratioAttributable risk


Table 3 OR (odds ratio) = (119886 lowast 119889)(119887 lowast 119888)

Family history Case ControlPositive 119886 119887

Negative 119888 119889

disease is controlled by an autosomal dominant allele Thusif one parent has ldquoAardquo genotype and another genotypeldquoAArdquo due to the possible genotypes of ldquoAArdquo and ldquoAardquo inthe offspring it is expected that all the children will beaffected Therefore if the risk of a heritable disease in a largepopulation is 50 among girls and boys it can be concludedthat heritability is of autosomal dominant type Binomialdistribution can be used for analysis Thus if 119899 is the sample

size 119909 is the number of affected offspring and 119901 is theprobability for a case to be affected (supposed to be 50) byassuming that H0 119901 = 12 we will have

(119883 = 119909) = (119899

119909)119901119909

(1 minus 119901)(119899minus119909)

(1)

44 Linkage Analysis The goal is to find the approximatelocation of the responsible gene or genes [10 19 20] In link-age analysis two-point LOD scores evaluate the evidence forlinkage between the disease locus and only a single markerwhen more than one marker is considered multipoint LODscores are reported [11]

Two broad types of linkage analysis exist parametric andnonparametric If there is enough information for knowledge

4 ISRN Genetics

of parameters (mode of inheritance and DNA from multiplemembers of informative families) it is possible to use model-based (parametric) linkage [17] however when the geneticmodel is unknown nonparametric analysis should be used[17] Parametric linkage analysis is a powerful strategy formapping genes with a simple Mendelian form of inheritance[11] A result of linkage analysis is usually expressed in termsof an LOD score (logarithms to the base 10)The LOD score isa function of the recombination fraction (120579) [21] Recombina-tion fraction estimates probability of recombination betweentwo markers Although a probability its maximum value is05 indicating a 50 50 chance of recombination or that twoloci sort independently and are unlinked but a recombinationfraction less than 05 indicates that two loci are not sortingindependently and there is linkage between them [22] TheLOD score is computed by comparing the likelihoods for arange of value 120579 and comparing with the likelihood when 120579 isequal to 05

Therefore with assuming 119871 the likelihood probabilityand 120579

119894 recombination fraction and maximum recombina-

tion fraction denoted as 120579 (120579 = 12) LOD is assessed in thefollowing manner [17 23ndash25]

LOD = log10

119871 (family120579119894)

119871 (family120579 = 12) (2)

An LOD score of 3 (which represents odds of 1000 1) orgreater in favour of linkage is used to indicate statisticallysignificant linkage If this score is minus 2 or less linkage isunlikely [22]

45 Association Study Association studies are similar tocase-control studies except that the disease associated ldquoexpo-suresrdquo that one seeks to identify are variant alleles of genes[11] These kinds of studies are used to find more commongenetic variations that are highly prevalent in the general pop-ulation [17] In practice frequencies of variant alleles amongaffected individuals are compared to unaffected individuals[26]

There are two types of association studies The first oneis the candidate gene study [17 27] which focuses on thespecific gene in addition to what may be accomplished bythe identification of a gene product such as a specific proteinrather than the gene itself [27] The second one is genome-wide study (GWAS) comprising a wide search of the genomefor genes that are related to the disease [27] The genesdiscovered thus far using GWAS have been common loci ofsmall effects but many genetic epidemiologists believe thatwhen taken together these genes of small effects may cumu-latively have vital effects on the risk for complex diseases[28] Association studies are subdivided into two types ofanalysis direct and indirect [17] In direct association studiesthe candidate gene has been designated and association istested directly but in the indirect one candidate gene hasnot been identified and is linked to marker genes [16]The direct method uses an actual causative mutation at aparticular gene in the test for association with the diseasephenotype The indirect method uses linkage disequilibriummethod Linkage disequilibrium (LD) describes the strength

Table 4 OR = (119886 lowast 119889)(119887 lowast 119888)

Case ControlExposure 119886 119887

Not exposure 119888 119889

of a relationship between alleles at different loci [23 28]More specifically if knowledge of an allele at one locus canpredict the allele that will reside at a second locus thenlinkage disequilibrium exists between the alleles However ifknowledge of an allele at the first locus cannot help predictthe allele that is at the second locus then linkage equilibrium(LE) exists [11 16 29]There are many statistical measures forLD and the more common metrics are Lewontinrsquos 119863 primeindex odds ratio (logistic regression) correlation and 1205942independence [30ndash33]

451 Lewontinrsquos 119863 Prime Index (1198631015840) [30 34] In mathemat-ical terms if there is no association or dependence betweentwo alleles (eg C and G) then

119875 (haplotype c-g) = 119875 (allele c) lowast 119875 (allele g) (3)

If alleles C and G are associated (in LD) in other words theyare dependent on each other then

119875 (haplotype C-G) = 119875 (allele c) lowast 119875 (allele g) + 120575 (4)

where 120575 is the raw disequilibrium coefficient Therefore if twoalleles are in linkage equilibrium then 120575 = 0

The raw disequilibrium coefficient 120575 can be difficult tointerpret because it is dependent on allele frequencies at thetwo loci 119863

1015840

is a scaled version of 119889 that measures LD as aproportion of the maximum amount of LD possible for thespecific allele frequencies at the two loci It can take valuesfrom minus1 to +1 if 119863

1015840

is equals to 1 it means that there iscomplete LD as follows

1198631015840

=

120575

120575max (5)

120575max can be estimated by the following equation (119901 allelefrequency in a locus and 119902 allele frequency in another locus

120575max = min [119901 (1 minus 119902) (1 minus 119901) 119902] (6)

452 Odds Ratio [11 29] One of the other indicators for theanalysis of the relationship is comparing the odds of exposure(genotype or allele) in case group with the odds of exposurein controls groupwith assumingH0 odds ratio is = 1 we have(Table 4)

453 Correlation [31] 1199032 measures the correlation betweenalleles with a range of minus1 to +1 in a situation of linkagedisequilibrium 1199032 will be 1 and can be calculated as follows

1199032

=

1205752

119901 (1 minus 119901) 119902 (1 minus 119902)

(7)

ISRN Genetics 5

Table 5 OR (in matched paired studies) = 119886119889

Patient groupEminuslowast E+lowastlowast

Parents group E+ 119886 119887

Eminus 119888 119889

Eminuslowast non exposureE+lowastlowast exposure

454 Odds Ratio in Matched Studies An allelic variant of acandidate gene or of a genetic marker was associated withincreased risk of disease one would expect that variant tobe transmitted from a heterozygous parent to an affectedoffspring more often than the 50 frequency expected bychance Assuming a biallelic locus let b be the number oftimes then the A1 allele is transmitted from a heterozygousA1A2 parent to an affected offspring and 119888 is the number oftimes the A2 allele is transmitted from heterozygous A1A2parent It will be recognized as McNamararsquos test for a pair-matched case-control study The ldquocaserdquo is the transmittedallele and the ldquocontrolrdquo is the nontransmitted allele [35]

McNamararsquos Test [32] McNamararsquos test is a nonparametricmethod used on nominal data It is applied to 2 times 2contingency tables with matched pairs of subjects [32] In2 times 2McNamararsquos table subjects with offspring alleles Eminus andparent allele E+ will be in 119886 cell while subjects with offspringalleles E+ and parent allele Eminus will be in 119889 cell see (Table 5)

By assuming H0 it is expected that 119886 + 119889 = 12 then wehave

1205942

= sum(119874 minus E)2

E

=(119886 minus (119886 + 119889) 2)

2

(119886 + 119889) 2

+(D minus (119886 + 119889) 2)2

(119886 + 119889) 2

(8)

And briefly

1205942

=(119886 minus 119889)

2

119886 + 119889

119889119891 = 1 (9)

One potential limitation of association study is theprobability of a false positive relationship between markersand genes [33] Therefore it should be noted that significantcausation test in marker and alleles confirm causation anddo not confirm linkage disequilibrium [34 36] The two locimay tend to be inherited together without the causality of thediseaseThis condition is more common in small populations(ethnic or tribal assembly) that have a lot of shared traits[37] LD can be influenced by several factors includingchance selection migration and isolated populations [3839] Another restriction of association study is variation oftest power when the disease allele is recessive compared withdominant allele [11] Association studies (based on controlgroup) are classified in to two groups related case-control andunrelated case-control studies

In the related case-control studies relatives of casepatients are used as control subjects These designs can havevarious control groups such as sib-control cousin control

and pseudosibling Although the use of sibling control offersseveral advantages over population controls (unrelated case-control) such asmorewillingness to participate and reducingcost and time it has some disadvantages like the probabilityof overmatching and limitation in age and sex matching insmall families [39] The advantage of a cousin-control groupis that one may be able to obtain closer matching on age andyear of birth with less loss in efficiency because the case andcousin are not as closelymatched on genotypes in addition toless chance of overmatching because the case and cousin haveonly one side of their families shared [39] In pseudosiblingdesigns no actual controls are selected instead genotypicdata are obtained on the parents of the case and the genotypetransmitted to the case is then compared with the threegenotypes (pseudosibling) that were not transmitted to thecase The question this design seeks to address is whether aspecific allele or genotype occurs more commonly in casesthan in their pseudo-sibs [39]

46 Estimation of Interaction in Exposures One of the mostfunctional studies in genetic epidemiology is the case-onlywhich is used for cross-sectional cohort or case controlpatients [35] The case-only design which is to assess gene-environment interaction was presented by Aalen et al andthen byHamajima et al [35]The case-only study design usedto study gene-environment interaction has been criticized forits susceptibility to bias caused by nonindependence betweengenetic and environmental factors [39ndash41] Each person withdisease (D) that is coded as positive or negative for a geneticfactor (G) and environmental factor (E) can be located inone of the four situations of (D++) (Dndash) (Dplusmn) and (Dplusmn)It is noteworthy that when G and E are independent in thesource population the case-only OR is equivalent to theinteraction estimate based on RRs regardless of disease riskConceptually the interaction between G and E refers to theextent to which the joint effect of the two factors on diseasediffers from the independent effects (effect of each of thefactors on disease in the absence of the other factor) [42ndash44]

Multiplicative interaction is assessed by comparing thejoint effect (effect on D due to the presence of both factors)with the product of the independent effects (product of effecton D in the absence of other) For example if independenteffect of G equals 3 and independent effect of E equals 2 thenwe would expect the joint effect of G and E to be 6 if there isno multiplicative interaction [44]

Given that independence between genetic and environ-mental factors is critical to the validity of case-only estimatesof interaction [40 41] when G and E are independent theOR relating G and E are equivalent to the interaction effect ofG-E see Table 6

In practice if the sources of nonindependence are mea-sured classification or adjustment for using multiple modelscan be used to remove the bias in case-only analysis ofinteraction [41]

47 New Designs in Genetic Studies Although case-controlstudies are commonly used for genetic-epidemiologic stud-ies an increasing number of cohort studies have been

6 ISRN Genetics

Table 6 OR (G-E interaction effect) = (119886 lowast 119889)(119887 lowast 119888)

E+lowast Eminuslowastlowast

E+ 119886 119887

Eminus 119888 119889

E+lowast exposureEminuslowastlowast non exposure

established over the past decade [45 46] Two progressivestudies the nested case-cohort and nested case-control haverecently been suggested The major advantage of nesteddesigns is their ability to match controls with cases on follow-up duration [4]

5 Study Designs inMycobacterium Tuberculosis

Twin studies are one of the primary and inexpensive heritablestudies on tuberculosis (TB) which provided valuable andimportant information about the etiology of TB Becausetwins theoretically share the same environment higher ratesof concordance for monozygous (identical) twins than fordizygous (fraternal) twins suggest that genetic factors areimportant in susceptibility to tuberculosis and provide anestimate of the magnitude of this effect [47 48] Duringthe past 15 years various surveys have been carried out onthe genetics of susceptibility to mycobacterial diseases [4950] Etiology effects on tuberculosis have been used in case-control studies too like the case-control study carried out inGambia which showed that polymorphisms in the NRAMP1gene were significantly associated with susceptibility totuberculosis [51 52] Another case-control study in Londonshowed VDR gene effect on susceptibility to TB [52] Usingassociation designs important pathogeneses of tuberculosishave been discovered too such as NRAMP1 vitamin D3receptor interferon-120574 interleukin-1120573 interleukin-12 tumornecrosis factor-120572 interleukin-4 and interleukin-10 [53ndash55]Linkage studies have also shown that there is disease suscepti-bility gene or genes in the neighbourhood of the marker anddetailed investigation of genes in the region is indicated Sucha genome-wide scan of affected sibling pairs from Gambiaand South Africa identified potential susceptibility loci onchromosomes 15q and Xq [49 56] Deng [56] have reviewedthe use of genetic linkage and association studies in theinvestigation of genetic susceptibility to infectious diseasesImplementation of such studies in developing countriespresents some particular challenges However it is obviousthat since tuberculosis occurs mainly in adults parents ofa case are frequently unavailable for genotyping But usingunaffected siblings as controls is possible [57] In the studyof complex diseases as TB because the effects of genesmay be modified by environmental (ie non-genetic) factorsgene-environment interactions may be explored in studydesigns such as case-only cross-sectional cohort and case-control studies and family-based designs such as case-parentaffected sibling pair and twin studies [57]

References

[1] R Bonita R Beaglehole and T Kjellstrom Basic EpidemiologyWHO Library Cataloguing-in-Publication Data 2nd edition2006

[2] J Last A Dictionary of Epidemiology Oxford University PressOxford UK 3rd edition 1993

[3] D C Thomas Statistical Methods in Genetic Epidemiology pp3ndash22 Oxford University Press New York NY USA 2004

[4] C Lienhardt S Bennett G Del Prete et al ldquoInvestigation ofenvironmental and host-related risk factors for tuberculosisin Africa I Methodological aspects of a combined designrdquoAmerican Journal of Epidemiology vol 155 no 11 pp 1066ndash10732002

[5] S Schwartz ldquoThe fallacy of the ecological fallacy the potentialmisuse of a concept and the consequencesrdquoAmerican Journal ofPublic Health vol 84 no 5 pp 819ndash824 1994

[6] F D K Liddell ldquoThe development of cohort studies in epidemi-ology a reviewrdquo Journal of Clinical Epidemiology vol 41 no 12pp 1217ndash1237 1988

[7] L Rodrigues and B R Kirkwood ldquoCase-control designs inthe study of common diseases updates on the demise of therare disease assumption and the choice of sampling scheme forcontrolsrdquo International Journal of Epidemiology vol 19 no 1 pp205ndash213 1990

[8] R Peto M C Pike and P Armitage ldquoDesign and analysisof randomized clinical trials requiring prolonged observationof each patient I Introduction and designrdquo British Journal ofCancer vol 34 no 6 pp 585ndash612 1976


[10] D C Thomas ldquoChapter 4 Basic epidemiologic and statisticalprinciplesrdquo in Statistical Methods in Genetic EpidemiologyOxford University Press New York NY USA 2004

[11] W J Gauderman J S Witte and D C Thomas ldquoFamily-basedassociation studiesrdquo National Cancer Institute Monograph vol26 pp 31ndash37 1999

[12] Q Yang and M J Khoury ldquoEvolving methods in genetic epi-demiology III Gene-environment interaction in epidemiologicresearchrdquo Epidemiologic Reviews vol 19 no 1 pp 33ndash43 1997

[13] J S Witte W J Gauderman and D C Thomas ldquoAsymp-totic bias and efficiency in case-control studies of candidategenes and gene-environment interactions basic family designsrdquoAmerican Journal of Epidemiology vol 149 no 8 pp 693ndash7051999

[14] J L Hopper G Chenevix-Trench D J Jolley et al ldquoDesignand analysis issues in a population-based case-control-familystudy of the genetic epidemiology of breast cancer and the Co-operative Family Registry for Breast Cancer Studies (CFRBCS)rdquoNational Cancer Institute Monograph vol 26 pp 95ndash100 1999

[15] P R Burton M D Tobin and J L Hopper ldquoKey concepts ingenetic epidemiologyrdquo The Lancet vol 366 no 9489 pp 941ndash951 2005

[16] M Cote ldquoStudy designs in genetic epidemiologyrdquo in TumourBiomarker Discovery vol 520 ofMethods in Molecular BiologyHumana Press New Jersey NJ USA 2009

[17] M Korkeila J Kaprio A Rissanen andM Koskenvuo ldquoEffectsof gender and age on the heritability of body mass indexrdquoInternational Journal of Obesity vol 15 no 10 pp 647ndash654 1991

[18] J Akey L Jin and M Xiong ldquoHaplotypes versus single markerlinkage disequilibrium tests what do we gainrdquo EuropeanJournal of Human Genetics vol 9 no 4 pp 291ndash300 2001

ISRN Genetics 7

[19] M D Teare Genetic Epidemiology pp 49ndash78 Springer NewYork NY USA 2006

[20] A Kong and N J Cox ldquoAllele-sharing models LOD scores andaccurate linkage testsrdquo American Journal of Human Geneticsvol 61 no 5 pp 1179ndash1188 1997

[21] S E Hodge and D A Greenberg ldquoSensitivity of lod scoresto changes in diagnostic statusrdquo American Journal of HumanGenetics vol 50 no 5 pp 1053ndash1066 1999

[22] K Yoonhee ldquoGenetic epidemiology 101 a review of the statisti-cal basisrdquo BioWave Journal vol 10 pp 16ndash45 2008

[23] D A Greenberg ldquoInferringmode of inheritance by comparisonof lod scoresrdquoAmerican Journal of Medical Genetics vol 34 no4 pp 480ndash486 1989

[24] N M Gatto U B Campbell A G Rundle and H AhsanldquoFurther development of the case-only design for assessinggene-environment interaction evaluation of and adjustment forbiasrdquo International Journal of Epidemiology vol 33 no 5 pp1014ndash1024 2004

[25] L R Cardona and J I Bell ldquoAssociation study designs forcomplex diseasesrdquo Nature Reviews Genetics vol 2 pp 91ndash992001


[27] K G Ardlie L Kruglyak andM Seielstad ldquoPatterns of linkagedisequilibrium in the human genomerdquoNature Reviews Geneticsvol 3 no 4 pp 299ndash309 2002

[28] S A Bacanu B Devlin and K Roeder ldquoAssociation studies forquantitative traits in structured populationsrdquo Genetic Epidemi-ology vol 22 no 1 pp 78ndash93 2002

[29] B Devlin and N Risch ldquoA comparison of linkage disequilib-riummeasures for fine-scale mappingrdquoGenomics vol 29 no 2pp 311ndash322 1995

[30] J K Pritchard and M Przeworski ldquoLinkage disequilibriumin humans models and datardquo American Journal of HumanGenetics vol 69 no 1 pp 1ndash14 2001

[31] R Bellamy N Beyers and K P McAdam ldquoA genome-widesearch for tuberculosis susceptibility genes in Africansrdquo Pro-ceedings of the National Academy of Sciences vol 97 pp 8005ndash8009 2000

[32] D C Thomas and J S Witte ldquoPoint population stratificationa problem for case-control studies of candidate-gene associa-tionsrdquo Cancer Epidemiology Biomarkers and Prevention vol 11no 6 pp 505ndash512 2002

[33] SWacholder S ChanockMGarcia-Closas L El Ghormli andN Rothman ldquoAssessing the probability that a positive report isfalse an approach for molecular epidemiology studiesrdquo Journalof the National Cancer Institute vol 96 no 6 pp 434ndash4422004

[34] R C Lewontin ldquoOn measures of gametic disequilibriumrdquoGenetics vol 120 no 3 pp 849ndash852 1988

[35] N Hamajima H Yuasa K Matsuo and Y Kurobe ldquoDetectionof gene-environment interaction by case-only studiesrdquo JapaneseJournal of Clinical Oncology vol 29 no 10 pp 490ndash493 1999

[36] J H Zar Bio Statistical Analysis Prentice Hall New York NYUSA 5th edition 2009

[37] H Campbell and I Rudan ldquoInterpretation of genetic asso-ciation studies in complex diseaserdquo The PharmacogenomicsJournal vol 2 pp 349ndash360 2002

[38] H H H Goring J D Terwilliger and J Blangero ldquoLarge up-ward bias in estimation of locus-specific effects from genome

wide scansrdquoAmerican Journal of Human Genetics vol 69 no 6pp 1357ndash1369 2001

[39] P S Albert D Ratnasinghe J Tangrea and S WacholderldquoLimitations of the case-only design for identifying gene-environment interactionsrdquo American Journal of Epidemiologyvol 154 no 8 pp 687ndash693 2001

[40] M J Khoury and W D Flanders ldquoNontraditional epidemio-logic approaches in the analysis of gene-environment interac-tion case control studies with no controlsrdquo American Journalof Epidemiology vol 144 pp 207ndash213 1996

[41] A M Goldstein and N Andrieu ldquoDetection of interactioninvolving identified genes available study designsrdquo Journal ofthe National Cancer Institute vol 26 pp 49ndash54 1999

[42] C L Saunders C Gooptu and D T Bishop ldquoThe use of case-only studies for the detection of interactions and the non-independence of genetic and environmental risk factors fordisease (Abstract)rdquo Genetic Epidemiology vol 21 p 174 2001

[43] C L Saunders and J H Barrett ldquoFlexible matching in case-control studies of gene-environment interactionsrdquo AmericanJournal of Epidemiology vol 159 no 1 pp 17ndash22 2004

[44] T A Manolio ldquoCohort studies and the genetics of complexdiseaserdquo Nature Genetics vol 41 no 1 pp 5ndash6 2009

[45] D W Haas ldquoGenetic studies in clinical trials and observationalcohortsrdquo HIV PGX vol 1 pp 1ndash4 2006

[46] B S Hulka and B H Margolin ldquoMethodological issues in epi-demiologic studies using biologicmarkersrdquoAmerican Journal ofEpidemiology vol 135 no 2 pp 200ndash209 1992

[47] RG Loudon and SK Spohn ldquoCough frequency and infectivityin patients with pulmonary tuberculosisrdquo American Review ofRespiratory Disease vol 99 no 1 pp 109ndash111 1969

[48] M Moller E de Wit and E G Hoal ldquoPast present and futuredirections in human genetic susceptibility to tuberculosisrdquoFEMS Immunology and Medical Microbiology vol 58 no 1 pp3ndash26 2010

[49] B Simmonds Tuberculosis in Twins Pitman Medical LondonUK 1963

[50] G Madico R H Gilman W Checkley et al ldquoCommunityinfection ratio as an indicator for tuberculosis controlrdquo TheLancet vol 345 no 8947 pp 416ndash419 1995

[51] C M Stein ldquoGenetic epidemiology of tuberculosis susceptibil-ity impact of study designrdquo PLoS Pathogens vol 7 no 1 ArticleID e1001189 pp 1ndash8 2011

[52] X Ma R A Reich J A Wright et al ldquoAssociation betweeninterleukin-8 gene alleles and human susceptibility to tubercu-losis diseaserdquo Journal of Infectious Diseases vol 188 no 3 pp349ndash355 2003

[53] D Lopez-Maderuelo F Arnalich R Serantes et al ldquoInterferon-120574 and interleukin-10 gene polymorphisms in pulmonary tuber-culosisrdquo American Journal of Respiratory and Critical CareMedicine vol 167 no 7 pp 970ndash975 2003

[54] S Ryu Y K Park G H Bai S J Kim S N Park and S Kangldquo3rsquoUTR polymorphisms in the NRAMP1 gene are associatedwith susceptibility to tuberculosis in Koreansrdquo InternationalJournal of Tuberculosis and Lung Disease vol 4 no 6 pp 577ndash580 2000

[55] N Risch ldquoLinkage strategies for genetically complex traitsII The power of affected relative pairsrdquo American Journal ofHuman Genetics vol 46 no 2 pp 229ndash241 1990

8 ISRN Genetics

[56] L Abel and A J Dessein ldquoGenetic epidemiology of infec-tious diseases in humans design of population-based studiesrdquoEmerging Infectious Diseases vol 4 no 4 pp 593ndash603 1998

[57] H W Deng ldquoPopulation admixture may appear to maskchange or reverse genetic effects of genes underlying complextraitsrdquo Genetics vol 159 no 3 pp 1319ndash1323 2001

Submit your manuscripts athttpwwwhindawicom

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Anatomy Research International

PeptidesInternational Journal of


Hindawi Publishing Corporation httpwwwhindawicom

International Journal of

Volume 2014

Zoology


Molecular Biology International

GenomicsInternational Journal of


The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014


BioinformaticsAdvances in

Marine BiologyJournal of



Signal TransductionJournal of


BioMed Research International

Evolutionary BiologyInternational Journal of



Biochemistry Research International

ArchaeaHindawi Publishing Corporationhttpwwwhindawicom Volume 2014


Genetics Research International


Advances in

Virolog y

Hindawi Publishing Corporationhttpwwwhindawicom

Nucleic AcidsJournal of

Volume 2014

Stem CellsInternational



Enzyme Research



Microbiology

2 ISRN Genetics

Table 1 Main study designs in classic epidemiology

Case reports Case series Ecologicalstudies

Cross sectionalstudies Case control Cohort Intervention

trials

Unit of study (Single case)individual

(gt1 cases)individual Population Individual Individual Individual Individual

Description

Describesunusual

characteristicsof a case

A group of cases

Comparingpopulations indifferent placesat the same timeor in a time

series

Study ofpopulation atldquopoint-in-timerdquo

Study of twogroups of

subjects (casedisease ofinterest andcontrol

disease-free)

Study of twogroups ofsubjects

(exposure andnon-exposure

groups)

Study andexamine twogroups ofsubjects

(interventionand controlgroups)

Direction Present Present Present Present Reverse Forward Forward

Type ofmeasurement

Reportingdescription

Reportingdescription Correlation Prevalence

association Odds ratio


attributable risk


attributable risk

Advantages







Quickinexpensivegroup-level

studies may alsobe the only wayto study theeffects of

group-levelconstructs forexample laws

Easyinexpensiveuseful for

investigatingfixed exposuressuch as bloodgroup mostconvenient inoutbreaks of

disease

Relativelyinexpensivegood for rare

diseasesEfficient in

resources andtime

Better for rareexposuresability todeterminecausalityrelations

The strongestevidence forcausalitycontrol ofunknown

confoundersfulfils the basicassumption ofstatistical

hypothesis tests

limitation





Ecologicalfallacy

Susceptible toselection biasand misclassifi-cation difficultto establish a

putative ldquocauserdquoNot good forrare diseases orrare exposures

Susceptible toselection bias

andmisclassificationbias may bedifficult to

establish thatldquocauserdquo

preceded effect

Costly and timeconsumingsusceptible toselection biasRelativelystatisticallyinefficient

unless disease iscommon

Expensive timeconsumingsometimesethically

generalizabilityproblem

disease is analyzed and if it is proved that there are causalrelationships the genetic region responsible for the diseaseis identified Therefore relying on familial relationship ingenetic epidemiology is one of the main differences betweenclassic and genetic epidemiology In Table 2 main designsin genetic epidemiology and process of surveys in geneticdiseases are summarized [8ndash16]

4 Analytical Study in Genetic Epidemiology

41 Familial Aggregation The first step in the study of a po-tentially genetic characteristic is to realize whether it tendsto aggregate in families without having any specific geneticmodel in mind [11] In other words familial aggregation isthe tendency for disease to cluster in families [17] A simpleapproach to assess aggregation in families is to identify agroup of affected case subjects and a group of healthy controlsubjects and compare the odds of a positive family history incase subjects who report a disease to the odds of a positivefamily history in the control group Finally an odds ratio canbe estimated according to Table 3 [1]

While this approach is valuable there is no way to controlfor individual or environmental risk factors for each relativewhichmight be driving the aggregation (eg cooccurrence ofsmoking behaviour and lung cancer aggregation in families)it is also imprecise since the probability of a positive familyhistory of disease increases with age and the number ofrelatives considered [17] One way to control confoundingvariables is to use categorizing methods such as logisticregression modeling [17]

42 Heritability Heritability is the proportion of variation ina trait that is attributable to genetic differences One of thesimplest designs in heritability study is the ldquotwin studyrdquo Thisstudy design uses the variance component framework and itsfrequency to estimate heritability in this way if the ratio ofgenetic variance to phenotypic variance (general heritability)is close to 1 this is evidence for a genetic component [18]

43 Segregation Analysis Segregation analysis aims to deter-minewhether the pattern of disease occurrence in families fitsa particular type of inheritance For example Huntingtonrsquos

ISRN Genetics 3






















controlExposure and













non-exposure groups





Negative 119888 119889



(119883 = 119909) = (119899

119909)119901119909

(1 minus 119901)(119899minus119909)

(1)



4 ISRN Genetics





LOD = log10

119871 (family120579119894)

119871 (family120579 = 12) (2)














1015840


1015840


1198631015840

=

120575

120575max (5)





1199032

=

1205752

119901 (1 minus 119901) 119902 (1 minus 119902)

(7)

ISRN Genetics 5




Eminus 119888 119889





1205942


E

=(119886 minus (119886 + 119889) 2)

2

(119886 + 119889) 2

+(D minus (119886 + 119889) 2)2

(119886 + 119889) 2

(8)

And briefly

1205942

=(119886 minus 119889)

2

119886 + 119889

119889119891 = 1 (9)









6 ISRN Genetics



E+ 119886 119887

Eminus 119888 119889





References



















ISRN Genetics 7







































8 ISRN Genetics










Volume 2014

Zoology






















Advances in

Virolog y



Volume 2014




Enzyme Research



Microbiology

ISRN Genetics 3






















controlExposure and













non-exposure groups





Negative 119888 119889



(119883 = 119909) = (119899

119909)119901119909

(1 minus 119901)(119899minus119909)

(1)



4 ISRN Genetics





LOD = log10

119871 (family120579119894)

119871 (family120579 = 12) (2)














1015840


1015840


1198631015840

=

120575

120575max (5)





1199032

=

1205752

119901 (1 minus 119901) 119902 (1 minus 119902)

(7)

ISRN Genetics 5




Eminus 119888 119889





1205942


E

=(119886 minus (119886 + 119889) 2)

2

(119886 + 119889) 2

+(D minus (119886 + 119889) 2)2

(119886 + 119889) 2

(8)

And briefly

1205942

=(119886 minus 119889)

2

119886 + 119889

119889119891 = 1 (9)









6 ISRN Genetics



E+ 119886 119887

Eminus 119888 119889





References



















ISRN Genetics 7







































8 ISRN Genetics










Volume 2014

Zoology






















Advances in

Virolog y



Volume 2014




Enzyme Research



Microbiology

4 ISRN Genetics





LOD = log10

119871 (family120579119894)

119871 (family120579 = 12) (2)














1015840


1015840


1198631015840

=

120575

120575max (5)





1199032

=

1205752

119901 (1 minus 119901) 119902 (1 minus 119902)

(7)

ISRN Genetics 5




Eminus 119888 119889





1205942


E

=(119886 minus (119886 + 119889) 2)

2

(119886 + 119889) 2

+(D minus (119886 + 119889) 2)2

(119886 + 119889) 2

(8)

And briefly

1205942

=(119886 minus 119889)

2

119886 + 119889

119889119891 = 1 (9)









6 ISRN Genetics



E+ 119886 119887

Eminus 119888 119889





References



















ISRN Genetics 7







































8 ISRN Genetics










Volume 2014

Zoology






















Advances in

Virolog y



Volume 2014




Enzyme Research



Microbiology

ISRN Genetics 5




Eminus 119888 119889





1205942


E

=(119886 minus (119886 + 119889) 2)

2

(119886 + 119889) 2

+(D minus (119886 + 119889) 2)2

(119886 + 119889) 2

(8)

And briefly

1205942

=(119886 minus 119889)

2

119886 + 119889

119889119891 = 1 (9)









6 ISRN Genetics



E+ 119886 119887

Eminus 119888 119889





References



















ISRN Genetics 7







































8 ISRN Genetics










Volume 2014

Zoology






















Advances in

Virolog y



Volume 2014




Enzyme Research



Microbiology

6 ISRN Genetics



E+ 119886 119887

Eminus 119888 119889





References



















ISRN Genetics 7







































8 ISRN Genetics










Volume 2014

Zoology






















Advances in

Virolog y



Volume 2014




Enzyme Research



Microbiology

ISRN Genetics 7







































8 ISRN Genetics










Volume 2014

Zoology






















Advances in

Virolog y



Volume 2014




Enzyme Research



Microbiology

8 ISRN Genetics










Volume 2014

Zoology






















Advances in

Virolog y



Volume 2014




Enzyme Research



Microbiology








Volume 2014

Zoology






















Advances in

Virolog y



Volume 2014




Enzyme Research



Microbiology

Documents

Review Article Study Designs in Genetic Epidemiologygenetic epidemiology is one of the main di erences between classic and genetic epidemiology. In Table , main designs in genetic