Using studies of gene expression to investigate species radiations in the New Zealand alpine flora

Using studies of gene expression to investigate species radiations in the New Zealand alpine flora

Southern Connection 2010

Claudia Voelckel, Peter B Heenan, Peter J Lockhart

DNA

mRNA

proteins

Transcription

Translation

provide structure & drive metabolism

substrate product

Why Gene Expression Studies?

Genomics

Transcriptomics

Proteomics

Metabolomics

*Comparative transcript profiling within & between species

* Evolutionary

2

1. Transcriptomics and species radiation – a case study

2. New tool in town – sequencing based methods replace microarrays

3. Putting the new tool to the test – case study revisited

4. Systems biology and species radiation

Outline

3

1. Transcriptomics and species radiation – a case study

2. New tool in town – sequencing based methods replace microarrays

3. Putting the new tool to the test – case study revisited

4. Systems biology and species radiation

Outline

4

Pachycladon (Brassicaceae)

Pachycladon super-network, S. Joly, unpubl.

stellatum

fastigiatum

enysii

enysii

cheesemanii exile

novae-zealandiae

wallii

latisiliqua

Diversification in New Zealand Alpine Cress

Habitat Rosette

Flowering Fruiting

Habitat Rosette

Flowering Fruiting

vs.

6

Pachycladon fastigiatum Pachycladon enysii

Sampling in the New Zealand Southern Alps

7

P. enysiiP. fastigiatum

8

DNA chip

with gene probesAAAAAA3’TTTTTT5’

TTTTTT5’green-labeled cDNA

AAAAAA3’TTTTTT5’

TTTTTT5’ red-labeled cDNA

Microarrays (DNA chips)

Sample 1

AAAAAA3’mRNA

Sample 2

AAAAAA3’ mRNA

DATA ANALYSISintensity 1intensity 2

Expression ratio: log

P. enysiiP. fastigiatum

Prob

abili

ty o

f diff

eren

tial e

xpre

ssio

n ( l

og o

dds

ratio

)

Magnitude of differential expression (log fold change)

ESM1 ESP

Arabidopsis microarray (20,468 genes)

310 genes (1.5%) up in P. fastigiatum 324 genes (1.6%) up in P. enysii

up-regulation of ESM1 and ESP predict P. fastigiatum to produce isothiocyanates and P. enysii to produce nitriles

Results

Voelckel et al. 2008, Molecular Ecology, 17: 4740–4753 9

Methionine Chain elongation pathway

Homomethionine (C3 GLS)Dihomomethionine(C4 GLS)

Methylthioalkyl GLS

Methylsulfinylalkyl GLS

Alkenyl GLS Hydroxalkyl GLS

Hydroxalkenyl GLS

GLS core pathway

Glucosinolate hydrolysis

Thiocyanates Nitriles (Eithionitriles)

Isothiocyanates Oxazolidine-2-thione

Side

cha

in m

odifi

catio

n

(Aliphatic) Glucosinolates (GLS) – Synthesis and hydrolysis genes

MAM, MAM-I, MAM-D, BCAT4

CYP79, CYP83, C-S lyase, SGT, SOT

FMO

AOP2 AOP3

GS-OH

myrosinase

ESM1 ESP

0

2

4

6

8

10

12

14

Isothiocyanates Nitriles/Epithionitriles

Allyl 3MTP

01234567

Isothiocyanates Nitriles/Epithionitriles

3MSOP

ESP (At1g54040)

ESM 1 (At3g14210)

6.29

- 4.62

Nitriles in P. enysii

Isothiocyanatesin P. fastigiata

P. enysii

P. fastigiata

HP

(μ m

ol/g

fw)

HP

(μ m

ol/g

fw)

Gene Prediction Regulation (log ratio)

Test

HPLC Test of Microarray Prediction

Voelckel et al. 2008, Molecular Ecology, 17: 4740–4753

Hypothesis: Role for herbivory in species diversification?

1. Transcriptomics and species radiation – a case study

2. New tool in town – sequencing based methods replace microarrays

3. Putting the new tool to the test – case study revisited

4. Systems biology and species radiation

Outline

12

NEXT-GEN Sequencing

Inexpensive production of large volumes of sequence data

Several platforms (Roche/454, Illumina/Solexa, ABI/SOLiD)

Many applications (de-novo assembly, re-sequencing, epigenetics and chromatin structure, metagenomics)

Revolutionary tools for gene expression analysis (e.g. Tag profiling, RNA-seq)

14

Tag Profiling

12

21

11

count 1count 2

log

STATISTICAL ANALYSIS

Solexa Genome Analyzer

Sample 1 mRNA

AAA3’AAA3’

AAA3’AAA3’

Sample 2mRNAAAA3’

AAA3’AAA3’

AAA3’

18 bp tag library

AAA3’

AAA3’

AAA3’

AAA3’

18 bp tag library

AAA3’

AAA3’

AAA3’

AAA3’

Sample 1

Sample 2

Reference

TAG MAPPING

Advantages & Challenges of Tag Profiling

open to any organism

any expressed transcript detectable (1 copy/cell)

less RNA needed (tag profiling = 1µg, microarrays = 100 µg)

minor data normalization/no background

Advantages

Challenges

mapping 18 bp tags (sequence differences Pachycladon/Arabidopsis)

counting tags per gene (noise, location, abundance)

statistical analysis of differential expression (proportion data)

15

1. Transcriptomics and species radiation – a case study

2. New tool in town – sequencing based methods replace microarrays

3. Putting the new tool to the test – case study revisited

4. Systems biology and species radiation

Outline

16

Tag Profiling Results

17423 A. thaliana loci (noise filter 10, count most abundant tag per gene)

2654 genes (15.2%) up in P. fastigiatum 1857 genes (10.7%) up in P. enysii

(tagwise normalization, -log2(1.5) < logfc < log2 (1.5))

P. enysiiP. fastigiatum

17

Microarrays (MA) vs. Tag Profiling (TP)

more differentially expressed genes in TP (10.7-15.2% ) than with MA (1.5-1.6% )

310 up in PF324 up in PE

2654 up in PF1857 up in PE

PF

MA TP41269 2613

PE

50274 1807MA TP

13.2% (PF) and 15.4 % (PE) of MA results confirmed by TP results

18

biological inferences from both studies identical

Locus lfc MA lfc TPAT1G54040, ESP 6.3 7.0

Locus lfc MA lfc TPAT3G14210, ESM1 -4.6 -35.0

MA: 20,468 genes

TP: 17,423 genes116058863 5818

“...not a popular product, too expensive, tricky chemistry.. instead use:

RNA-Seq!”

Tag Profiling is dead, long live RNA-Seq!

2 Oct 09: “Illumina is discontinuing the support of Tag Profiling and will no longer be manufacturing the reagent kits for this application.”

One year later: Tag profiling works for a non-model plant with a distant reference transcriptome! Let’s do more experiments!

19

20

RNA-Sequencing

Sample 1

AAA3’AAA3’

AAA3’AAA3’

Sample 2mRNA mRNA

Solexa Genome Analyzer

AAA3’

AAA3’AAA3’

AAA3’

cDNA library cDNA library

Sample 1

Sample 2

ReferenceREAD MAPPING

12

21

11

count 1count 2

log

STATISTICAL ANALYSIS

gene length

read mapping (reference transcriptome)

quantification of reads (lack of software, but packages evolve: e.g. edgeR)

Advantages & Challenges of RNA-Seq

Advantages

Challenges

whole transcriptome coverage and longer reads

large dynamic range of expression levels

base-resolution expression profiles for each gene

multiplex-compatible

sequence variation in transcribed regions (e.g. SNPs)

splicing isoforms, gene boundaries, novel transcribed regions

21

Great for non-model organisms!

Planned RNA Sequencing Projects

EST library for Pachcladon fastigiatum (31,116 genes, 79% of Arabidopsis)

22

Allopolyploidy and genome bias in Pachycladon

Adaptation to warmer climates in Pachycladon

SNP development in Pachycladon

1. Transcriptomics and species radiation – a case study

2. New tool in town – sequencing based methods replace microarrays

3. Putting the new tool to the test – case study revisited

4. Systems biology and species radiation

Outline

23

DNA

mRNA

proteins

Transcription

Translation

provide structure & drive metabolism

substrate product

How about System Biology?

Genomics

Transcriptomics

Proteomics

Metabolomics

* Evolutionary

24

* Evolutionary

* Evolutionary

*Comparative transcript, protein and metabolite profiling within & between species

Q: Ecological drivers of diversification?

A: Comparative gene and protein expression profiling in common gardens

FA

ST

EN

EN

LA

CHEX

NZ

WA

Questions & Approach

P. cheesemanii (CH)

P. exile (EX)

P. novae-zelandiae(NZ)

CHEX

NZ

LincolnPlant growth

Peter Heenan Murray Dawson

AucklandMicroarray analysis

Bart Janssen Luke Luo Silvia Schmidt

Jena Glucosinolate

analysis

Michael Reichelt

PalmyLink all data

Claudia VoelckelPete Lockhart

SydneyProtein analysis

Paul A. Haynes Mehdi Mirzai Dana Pascovici

People who helped:

Submitted

9601 loci 1489 loci

Overall correlation:

8527 4151074

T PTP

CH EX NZ

CH 0.52 0.43 0.30

EX 0.47 0.45 0.32

NZ 0.40 0.36 0.34

TP

T = transcript profiling, P = protein profiling

similar to other non-plant systems (0.2-0.5)

Interconversion of carbon dioxide and bicarbonate (carbonic anhydrase)

Interconversion of carbon dioxide and bicarbonate (carbonic anhydrase)

Draught response

Draught response

Serine racemase

Vegetative storage proteins

97 6129*

18 814

14 2288

T PTP

T PTP

T PTP

EX+NZ

CH+NZ

CH+EX

23%32%

18%4%

36%3%

CH

EX

NZ

Specific Genes Found by T AND P

EX+NZ

CH -

CH+NZ

EX iso

CH+EX

NZ -

EX+NZ

CH -

CH+NZ

EX -

CH+EX

NZ nitriles

Testing Predictions from T & P: Glucosinolate Hydrolysis

Prediction

P. cheesemanii

P. novae-zelandiae

Test

T CH EX NZ

CH 1 0.91 0.74

EX 1 0.83

NZ 1

P CH EX NZ

CH 1 0.75 0.59

EX 1 0.72

NZ 1

= =

Profiling Patterns Through the Phylogenetic Lens:

3MSO

P4M

SOB

3-Bu

teny

l4M

TB8M

SOO

6MSO

H

4OH

I3M

7MTH

1MO

I3M

4MO

I3M

7MSO

H

3MTP

S-2O

H3-

But.

Ally

l

EX

CH*

NZ*

CHEX

NZ≠

Glucosinolates

Thanks to:

YOU!

FundingMarsden & Humboldt Foundation

New ZealandLandcare: Peter Heenan, Kerry Ford, Murray Dawson, Kat Trought

Plant and Food: Bart Janssen, Luke Luo, Silvia SchmidtAWC Genome Service: Pete Lockhart, Patrick Biggs, Lorraine Berry, Lesley Collins, Maurice Collins

Students: Christine Reinsch, Hanna Daniel, Helene Kretzmer

GermanyMPICE: Michael Reichelt, Jonathan Gershenzon

AustraliaMacquarie University: Mehdi Mirzai, Dana Pascovici, Paul Haynes, Mark Westoby