Omics Integration

Integrative causality analysis of genetic, epigenetic, andtranscriptomic data in a large cohort

Rosemary McCloskey and Sara Mostafavi

[email protected]

http://slideshare.net/rmcclosk/omics-integration

March 27, 2015

R. McCloskey & S. Mostafavi () Omics data integration March 27, 2015 1 / 12

Motivation

genetic, epigenetic, and transcriptomic data provide snapshots ofcellular processes

usually one data type is studied at a time, in relation to a phenotypeor disease

GATTACA

?

geneexpression

methylation

histoneacetylation

genotype

how do these data fit together?

R. McCloskey & S. Mostafavi () Omics data integration March 27, 2015 2 / 12

Motivation

genetic, epigenetic, and transcriptomic data provide snapshots ofcellular processes

usually one data type is studied at a time, in relation to a phenotypeor disease

GATTACA

?

geneexpression

methylation

histoneacetylation

genotype

how do these data fit together?

R. McCloskey & S. Mostafavi () Omics data integration March 27, 2015 2 / 12

Motivation

genetic, epigenetic, and transcriptomic data provide snapshots ofcellular processes

usually one data type is studied at a time, in relation to a phenotypeor disease

GATTACA

?

geneexpression

methylation

histoneacetylation

genotype

how do these data fit together?

R. McCloskey & S. Mostafavi () Omics data integration March 27, 2015 2 / 12

The data

large cohort designedto study cognitivedecline andAlzheimer’s disease

genotype, geneexpression, DNAmethylation, andhistone acetylation(CHiP-seq) data

392 individuals withall four data typeswere used for thisanalysis

2

19

1080

0

3

392

152

20

0

140 61

47

17

11

expression methylation

acetylation genotype

R. McCloskey & S. Mostafavi () Omics data integration March 27, 2015 3 / 12

The data

large cohort designedto study cognitivedecline andAlzheimer’s disease

genotype, geneexpression, DNAmethylation, andhistone acetylation(CHiP-seq) data

392 individuals withall four data typeswere used for thisanalysis

2

19

1080

0

3

392

152

20

0

140 61

47

17

11

expression methylation

acetylation genotype

R. McCloskey & S. Mostafavi () Omics data integration March 27, 2015 3 / 12

The data

large cohort designedto study cognitivedecline andAlzheimer’s disease

genotype, geneexpression, DNAmethylation, andhistone acetylation(CHiP-seq) data

392 individuals withall four data typeswere used for thisanalysis

2

19

1080

0

3

392

152

20

0

140 61

47

17

11

expression methylation

acetylation genotype

R. McCloskey & S. Mostafavi () Omics data integration March 27, 2015 3 / 12

Quantitative trait loci (QTLs)

a QTL is a genetic locuscorrelated with aphenotype

we are interested inQTLs for geneexpression (eQTLs),histone acetylation(aceQTLs), andmethylation (meQTLs)

QTLs provide a tool tostudy interactionbetween other molecularphenotypes

-2-10123

-2-1012

-1

0

1

expressionacetylation

meth

ylation

0 1 2genotype

R. McCloskey & S. Mostafavi () Omics data integration March 27, 2015 4 / 12

Quantitative trait loci (QTLs)

a QTL is a genetic locuscorrelated with aphenotype

we are interested inQTLs for geneexpression (eQTLs),histone acetylation(aceQTLs), andmethylation (meQTLs)

QTLs provide a tool tostudy interactionbetween other molecularphenotypes

-2-10123

-2-1012

-1

0

1

expressionacetylation

meth

ylation

0 1 2genotype

R. McCloskey & S. Mostafavi () Omics data integration March 27, 2015 4 / 12

Quantitative trait loci (QTLs)

a QTL is a genetic locuscorrelated with aphenotype

we are interested inQTLs for geneexpression (eQTLs),histone acetylation(aceQTLs), andmethylation (meQTLs)

QTLs provide a tool tostudy interactionbetween other molecularphenotypes

-2-10123

-2-1012

-1

0

1

expressionacetylation

meth

ylation

0 1 2genotype

R. McCloskey & S. Mostafavi () Omics data integration March 27, 2015 4 / 12

Identifying QTLs

↓ SNPs in 200 kb windowSpearman’s ρ

↓ Holm-Bonferroni correctionbest SNP per feature

↓ FDR correction

R. McCloskey & S. Mostafavi () Omics data integration March 27, 2015 5 / 12

Identifying QTLs

↓ SNPs in 200 kb windowSpearman’s ρ

↓ Holm-Bonferroni correctionbest SNP per feature

↓ FDR correction

R. McCloskey & S. Mostafavi () Omics data integration March 27, 2015 5 / 12

Identifying QTLs

↓ SNPs in 200 kb windowSpearman’s ρ

↓ Holm-Bonferroni correctionbest SNP per feature

↓ FDR correction

R. McCloskey & S. Mostafavi () Omics data integration March 27, 2015 5 / 12

Identifying QTLs

↓ SNPs in 200 kb windowSpearman’s ρ

↓ Holm-Bonferroni correctionbest SNP per feature

↓ FDR correction

R. McCloskey & S. Mostafavi () Omics data integration March 27, 2015 5 / 12

Removing Principal Components

technical, environmental,and biological covariatescan swamp out QTLeffects

correct by removingprincipal components

number of peaks with aQTL plateaus at 10 PCs,while genes and CpGscontinue to increase

for this analysis, removed10 PCs from all data

40004500500055006000

3000

3500

4000

7500080000850009000095000

genes

peaks

CpGs

0 5 10 15 20PCs removed

R. McCloskey & S. Mostafavi () Omics data integration March 27, 2015 6 / 12

Removing Principal Components

technical, environmental,and biological covariatescan swamp out QTLeffects

correct by removingprincipal components

number of peaks with aQTL plateaus at 10 PCs,while genes and CpGscontinue to increase

for this analysis, removed10 PCs from all data

40004500500055006000

3000

3500

4000

7500080000850009000095000

genes

peaks

CpGs

0 5 10 15 20PCs removed

R. McCloskey & S. Mostafavi () Omics data integration March 27, 2015 6 / 12

Removing Principal Components

technical, environmental,and biological covariatescan swamp out QTLeffects

correct by removingprincipal components

number of peaks with aQTL plateaus at 10 PCs,while genes and CpGscontinue to increase

for this analysis, removed10 PCs from all data

40004500500055006000

3000

3500

4000

7500080000850009000095000

genes

peaks

CpGs

0 5 10 15 20PCs removed

R. McCloskey & S. Mostafavi () Omics data integration March 27, 2015 6 / 12

Removing Principal Components

technical, environmental,and biological covariatescan swamp out QTLeffects

correct by removingprincipal components

number of peaks with aQTL plateaus at 10 PCs,while genes and CpGscontinue to increase

for this analysis, removed10 PCs from all data

40004500500055006000

3000

3500

4000

7500080000850009000095000

genes

peaks

CpGs

0 5 10 15 20PCs removed

R. McCloskey & S. Mostafavi () Omics data integration March 27, 2015 6 / 12

Identifying multi-QTLs

By intersecting QTL sets, found240 gene, CpG, and peak tripleswhich shared the same QTL

29841799

50981

127

240

1604

2129

eQTL meQTL

aceQTL

29841799

50981

127

240

1604

2129

eQTL meQTL

aceQTL

Also assessed QTL overlap usingπ0 approach

100 %

46 %

14 %

31 %

100 %

11 %

83 %

84 %

100 %

eQTLs

aceQTLs

meQ

TLs

eQTLs

aceQTLs

meQTLs

R. McCloskey & S. Mostafavi () Omics data integration March 27, 2015 7 / 12

Identifying multi-QTLs

By intersecting QTL sets, found240 gene, CpG, and peak tripleswhich shared the same QTL

29841799

50981

127

240

1604

2129

eQTL meQTL

aceQTL

29841799

50981

127

240

1604

2129

eQTL meQTL

aceQTL

Also assessed QTL overlap usingπ0 approach

100 %

46 %

14 %

31 %

100 %

11 %

83 %

84 %

100 %

eQTLs

aceQTLs

meQ

TLs

eQTLs

aceQTLs

meQTLs

R. McCloskey & S. Mostafavi () Omics data integration March 27, 2015 7 / 12

Bayesian networks

Bayesian networks are directed graphical models, where the directededges represent causal relationships

We use conditional Gaussian networks

Score = likelihood of data given network

temperature precipitation

Pr(temp) ∼ N(0, 1) Pr(precip | temp) ∼ N(0, 1)

0.7 0.5

Pr(N(0, 1) = 0.7) Pr(N(0.7, 1) = 0.5)×

R. McCloskey & S. Mostafavi () Omics data integration March 27, 2015 8 / 12

Bayesian networks

Bayesian networks are directed graphical models, where the directededges represent causal relationships

We use conditional Gaussian networks

Score = likelihood of data given network

temperature precipitation

Pr(temp) ∼ N(0, 1) Pr(precip | temp) ∼ N(0, 1)

0.7 0.5

Pr(N(0, 1) = 0.7) Pr(N(0.7, 1) = 0.5)×

R. McCloskey & S. Mostafavi () Omics data integration March 27, 2015 8 / 12

Bayesian networks

Bayesian networks are directed graphical models, where the directededges represent causal relationships

We use conditional Gaussian networks

Score = likelihood of data given network

temperature precipitation

Pr(temp) ∼ N(0, 1) Pr(precip | temp) ∼ N(0, 1)

0.7 0.5

Pr(N(0, 1) = 0.7) Pr(N(0.7, 1) = 0.5)×

R. McCloskey & S. Mostafavi () Omics data integration March 27, 2015 8 / 12

Bayesian networks

Bayesian networks are directed graphical models, where the directededges represent causal relationships

We use conditional Gaussian networks

Score = likelihood of data given network

temperature precipitation

Pr(temp) ∼ N(0, 1) Pr(precip | temp) ∼ N(0, 1)

0.7 0.5

Pr(N(0, 1) = 0.7) Pr(N(0.7, 1) = 0.5)×

R. McCloskey & S. Mostafavi () Omics data integration March 27, 2015 8 / 12

Networks for QTLs

deal and CGBayesNets packages to construct one Bayesian networkfor each multi-QTL by exhaustive search

With deal, edges into genotype were blacklisted

Most common network structure was independence

Accounted for 42% of deal networks, 29% of CGBayesNets networks

genotypeexpression acetylation

methylation

R. McCloskey & S. Mostafavi () Omics data integration March 27, 2015 9 / 12

Networks for QTLs

deal and CGBayesNets packages to construct one Bayesian networkfor each multi-QTL by exhaustive search

With deal, edges into genotype were blacklisted

Most common network structure was independence

Accounted for 42% of deal networks, 29% of CGBayesNets networks

genotypeexpression acetylation

methylation

R. McCloskey & S. Mostafavi () Omics data integration March 27, 2015 9 / 12

Networks for QTLs

deal and CGBayesNets packages to construct one Bayesian networkfor each multi-QTL by exhaustive search

With deal, edges into genotype were blacklisted

Most common network structure was independence

Accounted for 42% of deal networks, 29% of CGBayesNets networks

genotypeexpression acetylation

methylation

R. McCloskey & S. Mostafavi () Omics data integration March 27, 2015 9 / 12

Networks for QTLs

deal and CGBayesNets packages to construct one Bayesian networkfor each multi-QTL by exhaustive search

With deal, edges into genotype were blacklisted

Most common network structure was independence

Accounted for 42% of deal networks, 29% of CGBayesNets networks

genotypeexpression acetylation

methylation

R. McCloskey & S. Mostafavi () Omics data integration March 27, 2015 9 / 12

Future Work

Expand the number of multi-QTLs

More that just the best SNP per featureIdentify overlapping QTLs intelligently

More rigourous criterion for number of PCs to remove

Try other packages for network learning (HyPhy)

Are QTLs enriched in SNPs identified in GWAS studies?

Correlations with phenotype (cognitive decline etc.)

R. McCloskey & S. Mostafavi () Omics data integration March 27, 2015 10 / 12

Thank you!

Harvard / Broad

Philip L. D. Jager

Lori Chibnik

Jishu Xu

Charles White

Cristin McCabe

Towfique Raj

Rush

David A Bennett

Chris Gaiteri

Lei Yu

Bioinformatics Training Program

All the students

Sharon Ruschkowski

R. McCloskey & S. Mostafavi () Omics data integration March 27, 2015 11 / 12

Software

QTL analysis

Matrix eQTL

qvalue

Bayesian networks

deal

CGBayesNets

Slides

beamer

TikZ

tikzDevice

Plots

pheatmap

ggplot2

VennDiagram

Colour Scheme

solarized

R. McCloskey & S. Mostafavi () Omics data integration March 27, 2015 12 / 12