Hendrik Poorter,Frank Gilmer& Uli Schurr

JPPC, FZJ, DE

The yin and yang of meta-phenomics

JPPC: Jülich Plant Phenotyping Centre

1. Quality control before data enter the database

2. Extracting biological knowledge from the database

1. Environmental control

2. Data collection

3. Deduction of information

4. Deduction of knowledge

Two topics:

1. Quality control:

Quality targets for experiments:

• Setup SOPs (standard operation procedures).

• SOPs for – reproducible – growth environments.

• SOPs for – reproducible - plant analysis.

• Collect information (history, phenotype) for every

individual in a plant information database.

• Reliable, reproducible, transparent

The biologist’s perspective:plant size variability is an issue

80 100 120 140 160 180 200 220 2400

2

4

6

8

# of

obs

erva

tions

Water loss (rel. units) Plant variability across labs:Massonnet et al. (2010) Plant Phys.

What variability can one expect?

0.1 0.3 0.5 0.7 0.9 1.10

10

20

30

P75 = 0.46P50 = 0.30P25 = 0.21

%

slnW

n = 600 Mean = 1.41

n = 800

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.00

5

10

15

20

25

30

Fre

quen

cy (

%)

W700 / W 350

A meta-analysis of the effect of elevated CO2:350 experiments with 800 mean values

for 350 species

Poorter & Navas (2003) New Phytol.

600 estimates of variability in plant size(standard deviation ln-transformed weight)

Poorter & Garnier (1996) J. Exp. Bot.

Could the variation in growth response to elevated CO2 be explained simply by plant-to-plant variability?

SDlnW

# ofplants

P20 0.21 4

P50 0.31 8

P80 0.51 12

1. Assume a true W700/W350 of 1.41

2. Draw at random 4, 8 or 12 plants

from a population with 3 variabilities:

B

W700/W350 ?

90.000 simulations

Yes, all observed variation in growth response could just happen to be caused by sampling too few individuals from too variable experimental populations:

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.00

5

10

15

20

25

30

Fre

quen

cy (

%)

Weight Ratio

Poorter & Navas (2003) New Phytol.

Conclusion 1:

- Quality control in your procedures is an issue

- Biological variation is an equally important issue, and growth chambers are NOT solving this problem

2. Extracting biological knowledge from the database:

At the phenotypic level, there is – for plants –a lack of information structured in a database:

TAIR, PLEXdb, Genevestigator,Drastic, CSB,DB, Germinate

LedaGlopnetTRY

(TurboVeg)

? (Floral DB)

Chloroplast 2010, Germinate

How do plants respond to their environment?

Investigator A:Arabidopsis

Trait x

low light 20 units

high light 40 units

Investigator B:Brassica

Trait x

low light 60 units

high light 60 units

The 2 experiments may actually tell the same thing:

The classical dose-response curve:

Nutrient supply

Yie

ldMitscherlich (1909)

The example ofSLA vs Light:

Light intensity

Irradiance

PFDPPFD

PAR

PFR

µmol m-2 s-1

mol m-2 day-1

W m-2

luxcal cm-2 s-1

langley min-1

lumen foot-2

MJft-c

SLA

SLW

SLM

LSM

LSW

LMAg m-2

mg cm-2

µg cm-2

µg mm-2

m2 kg-1

dm2 g-1

mm2 mg-1cm2 mg-1

Ma

SLA: leaf area / leaf dry mass

How can we achieve a clear picture from fragmented data?:

SLA(m2 kg-1)

Daily Photon Irradiance(mol m-2 day-1)

A literature analysis of >1100 data points (mean values) from >150 experiments on >300 species:

DPI (mol m-2 day-1)

SLA

(m

2 kg

-1)

0 10 20 30 40 500

20

40

60

80

100

DPI (mol m-2 day-1)

Rice & Bazzaz (1989) Ingestad et al. (1994) Sharew et al. (1996) Evans & Poorter (2001)

SLA

(m2 k

g-1)

Four different experiments show that interspecific variation in SLA is large:

0 10 20 30 400

20

40

60

80

100

0 10 20 30 400

20

40

60

80

100

0 10 20 30 400

20

40

60

80

100

After scaling SLA relative to the (interpolated) value at a reference light intensity of 8 mol m-2 day-1:

0 10 20 30 400

20

40

60

80

100

0 10 20 30 400

20

40

60

80

100

0 10 20 30 400

20

40

60

80

100

0 10 20 30 40 500.0

0.5

1.0

1.5

2.0

DPI (mol m-2 day-1)

SLA

(scale

d)

>1000 data points from >150 experiments on >300 species:

0 10 20 30 400

20

40

60

80

100

0 10 20 30 400

20

40

60

80

100

0 10 20 30 400

20

40

60

80

100

10 20 30 40 500.25

0.5

1

2

4

DPI (mol m-2 day-1)

SLA

(scale

d)

Median and the interquartile range for 7 light classes:

0 10 20 30 400

20

40

60

80

100

0 10 20 30 400

20

40

60

80

100

0 10 20 30 400

20

40

60

80

100

0 10 20 30 40 500.25

0.5

1

2

4

SLA

(scale

d)

DPI (mol m-2

day-1

)P10

P90

P50

10th and 90th percentiles give norm values, by which you can compare new experiments: The red line is an example of an outlying experiment

0 10 20 30 40 500.25

0.50.5

11

22

44

88

P90

Ann. Bot. (1986)

SLA

(scale

d)

DPI (mol m-2 day-1)

P10

Terminalia ivorensis

1. Light quantity (DPI) 2. Light quality (R/FR) 3. UV-B 4. CO2

5. O3

6. Nutrient availability (N, P, G) 7. Drought stress 8. Waterlogging 9. Submergence10.Temperature11. Salinity12. Soil compaction

Stress box

Can we follow the same approach for other environmental factors?

SLA responses to light, gases, and nutrients:

0 10 20 30 400.5

1

1.5

2a

SLA

(sc

ale

d)

DPI (mol m-2 day-1)

0.2 0.4 0.6 0.8 1.0 1.2

b

R / FR

0 5 10 15 20

c

UV-B (kJ m-2 day-1)

0 200 400 600 800 10000.5

1

1.5

2

d

SLA

(sc

ale

d)

CO2 (µmol mol-1)

0 20 40 60 80 100

e

O3 (nmol mol-1)

0.2 0.4 0.6 0.8 1.0 1.2

f

Nutrient availability (rel. scale)

1000 50 30

700 150 600

SLA responses to water, temperature, salinity and soil compaction:

0.0 0.2 0.4 0.6 0.8 1.00.5

1

1.5

2g

SLA

(sc

aled

)

Water availability (rel. scale)0.2 0.4 0.6 0.8 1.0

h

Waterlogging0.0 0.2 0.4 0.6 0.8 1.0

i

Submergence

5 10 15 20 250.5

1

1.5

2

j

SLA

(sc

aled

)

Temperature (°C)0.0 0.2 0.4 0.6 0.8 1.0

l

k

Salinity (rel. to seawater)1.1 1.2 1.3 1.4 1.5 1.6

Soil Compaction (g cm-3)

300 90

70

300 200 70

An overall non-linear equation to describe the response:

0 10 20 30 400

20

40

60

80

100

0 10 20 30 400

20

40

60

80

100

0 10 20 30 400

20

40

60

80

100

0 20 40 60 800.2

0.5

1.5

4 S

LA

(sc

ale

d)

Poorter et al., Fig. 1

DPI (mol m-2

day-1

)

3202 640201 ...)(log DPISLAsc

r2 = 0.72; PI = 2.92

Plasticity index: highest divided by lowest fitted value across a predefined range

Range PI

Irradiance 1 – 50 mol m-2 day-1 2.92CO2 200 – 1200 µmol mol-1 1.41Salinity 0 – 100 % seawater 1.23Waterlogging - – + 1.12Compaction 1.0 – 1.6 g cm-3 1.05

R : FR 0.2 – 1.2 mol mol-1 1.00

UV-B 1 – 20 kJ m-2 day-1 1.00

O3 5 – 100 nmol mol-1 1.00

Nutrients 0.05 – 1 rel. units 1.13Water 0.05 – 1 rel. units 1.25

Submergence - – + 1.95Temperature 5 – 35 oC 2.24

- Species family / name- woody / herbaceous- deciduous / evergreen- shrub / tree- annual / perennial- N2 fixing- C3 / C4 / CAM

Species traits

- Glasshouse, Growth chamber, OTC, Garden- Light (DPI)- Temperature (24h-average)- Substrate (hydroponics / soil, pot volume)

Growth environment

- Shade / Sun- Dry / Wet- Cold / Warm- Non-saline / Saline

Environmental niche

Are there differences between subgroups?

An example: tropical species are more plastic than boreal species

5 10 15 20 250.5

1

2 low-temp species intermediate high-temp species

S

LA (sc

aled

)

Temperature (°C)

Growth chamber

Glasshouse

OTC, shade house

Functional groups

Most experiments with herbs were in growth chambers,most with trees were outside in shade houses:

SLA is just one trait, can we do the same for other traits?:

Env. Factor SLA Trait 2 Trait 3 Trait 4 .... Trait n

1 2 3 4 5 6 ... 12

0 10 20 30 400.5

1

2

LMF

(sc

aled

)

DPI (mol m-2 day-1)

PI = 1.22n = 400

Yes, for example the % allocation of biomass to leaves as dependent on light intensity:

- Yield- RGR, ULR, LAR- SLA- LMF, SMF, RMF, (HI)

Growth box (> 4)

- PHOT actual- PHOT capacity (/m2, /g, /N)- gs,Transpiration, - ci/ca

- J / Vmax

- RESP leaf, stem, root, fruit (/g)

Gas exchange (> 3)

- [C], [N], [P] leaf, stem, root, fruit- Starch, Fructan- Nitrate- Sol. Sugars- Lignin- (Hemi-)Cellulose- Protein, Org. N- Org. acids- Minerals, Ash- Sol. Phenolics- Tannin- Construction costs- Delta 13C

Chem. comp. (> 5)

- leaf size- plant height- leaf thickness- leaf density (or FW / DW)- vol / % epidermis, mesophyll air spaces, sclerenchyma- cell size

Morphology / anatomy (> 3)

- Rubisco capacity- PEP carboxylase- SBPase- AGPase- NR- etc

Enzyme box (> 4)

- link to mRNA

Do the same for these plant traits:

Conclusions:

► Is able to summarise data across many experiments

► Yields quantitative response curves

► As well as normal limits

This meta-phenomics approach :

► Is applicable to (almost all) environmental factors

► Is applicable to all plant traits

► Will be very useful for modeling (global change,

limiting factors)

Frank Gilmer, FZJ

Uli Schurr, FZJ

Thanks to:

for more info see: - J. Exp. Bot. (2010) 61: 2043-2055 - www.metaphenomics.org

Ismael ArandaOwen AtkinCorine de GrootYulong FengJurg FranzaringKeith FunnellYaskara HayashidaVaughan HurryKen Krauss

Dina RhonzinaFrancesco RipulloneCatherine RoumetPeter RyserDylan SchwilkSusanne TittmannJan Henk VenemaRafael Villar

Dina RhonzinaFrancesco RipulloneCatherine RoumetPeter RyserDylan SchwilkSusanne TittmannJan Henk VenemaRafael Villar

Gerard Bönisch, MPI-Jena

Benjamin Bruns, FZJ