disentangling the gene-by- environment interaction

1

~~Laura K. Reed ([email protected])

Department of Biological SciencesUniversity of AlabamaTuscaloosa, AL USA

flygxe.ua.edu

Disentangling the gene-by-environment interaction architecture of metabolic disease through the lens

of evolutionary genetics and metabolomics

Reed et al, Current Opinion in Chemical Biology 2017

Mechanisms of Phenotypic Variation

Genome

Phenotype

1

2

2

Source: CDC Behavior Risk Factor Surveillance System

Obesity Trends* Among U.S. AdultsBRFSS, 1999 - 2010

(*BMI ≥30, or ~ 30 lbs. overweight for 5’ 4” person)

2960 Genetic Associations for Obesity in Humans* Only a tiny proportion of the genetic variation is explained by a specific

association.

*Retrieved 3/29/2018, NHGRI-EBI GWAS Catalog

Obesity and its comorbidities (metabolic syndrome- MetS) have both genetic and environmental influences

Outline

• The Fly as a MetS model

• Unlocking the black box of MetS using systems biology

• Determining genetic basis of Genotype x Diet Interactions

• Other environmental effects on MetS

Tori Nelko

3

4

3

~~

Fly-Human Homology

• Insulin signaling and other metabolic pathways largely homologous

• 77% OMIM human disease genes have cognates in Drosophila

Rulifson, et al. (2002). Science 296 , 118-20Reiter et al. (2001) Genome Research 11: 1114-1125

Hotamisligil (2006) Nature 444, 860-867O’Neill et al.,(2012) Biochemical Society Transactions 40:721-727;

Environmental Heterogeneity

How to Measure GxE

Wild Population

Inbred Lines

EnvironmentalVariation

Genetic VariationGenetic Variation

5

6

4

Normal Control High Glucose High FatDiets

PhenotypesWeight

Blood SugarLipid Storage

Survival

MetabolomicsGC-MS

quantification of over 180

metabolites

Gene ExpressionWhole Genome

Expression Profiling

146 Isolines in initial screen

20 Isolines for Systems Analysis

20 Isolinesfor Systems

Analysis

20 Isolinesfor Systems

Analysis

Stephanie Williams, Kelly Dew-Budd, Kristen Davis, Julie Anderson, Kenda Freeman, Mastafa Springston

Wei

ght

Possible Outcomes

7

8

5

Weight Genetic Variation

Reed et al., Genetics, 2010

Weight GxD

Reed et al., Genetics, 2010

Proportion of Variance Explained

18.9% Genetic1.9 % Dietary

12% GxDn = 14800

Significant decanalization on the high fat diet

9

10

6

0.0

0.1

0.1

0.2

0.2

0.3

0.3

0.4

Pro

port

ion

of V

aria

nce

Exp

lain

ed

Genetic x Diet

Diet

Genetic

MetS-like traitsGenetic and GxD effects greater than diet alone

Expression and metabolite principal

components do NOT correlate

GxD > diet for most MetS traits

Reed et al., Genetics 2014

Transcriptome and metabolome

significant GxD effects diet more important for

metabolome

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

Expression Metabolites

WeightTAG

Sugar

Weight

Williams et al, G3 2015

3

• Lack of conservation of correlated gene transcripts by diet

Ruth Bishop, Dana Davis, Katie Bray, Lauren Perkins, Joana Hubickey

Are Molecular Mechanisms for Phenotypes Conserved Across Environments?

11

12

7

Metabolite PC2 highly predictive of MetS-like phenotypes

Arrhythmia Index correlated with MetPC2 and MetS

phenotypesReed et al. 2014 Genetics

MetPC2L-dopa

N-arachidonoyldopamineglucosevaline

leucineisoleucine

glycine methionine

phenylalanine

Branched Chain AA

Ocorr et al. Trends in Cardiovasc Med. 2007 17:177-182

What Links MetS Phenotypes?

Eigenvector Metabolite Analysis

(EvMA)

Clare Scott ChialvoRonglin CheDavid ReifAlison Motsinger-Reif

Scott Chialvo, Che, Reif, Motsinger-Reif, Reed 2016 Metabolomics

What Links MetS Phenotypes?

13

14

8

Number of Correlated

Genes

Enriched GO Terms

Overt Phenotypes

0 -- Weight, Larval Survival

845 respiratory electron transport chain

Sugar, Weight, Pupal Survival

688 response to temperature stimulus, response to hypoxia

Weight

3134 Mitochondrion, generation of

precursor metabolites and energy, cellular

respiration, TCA cycle

Sugar, Lipid, Weight

9 -- Larval Survival, Pupal Survival, Development Time

3013 Mitochondrion, structural constituent

of ribosome

Weight, Development Time

0 -- --

227 ribosome biogenesis, rRNA metabolic

process

Larval Survival, Pupal Survival, Development Time

Fatty Acids (un- & saturated, l-dopa)

Sugars(disaccharides, maltose)

Amino Acids (BCAA, proline, serine)

Mixing Pot (monosaccharaides,

NADA, tyrosine)

Mixing Pot (unsaturated FA, AA)

Fatty Acids (unsaturated FA)

Amino Acids

Mixing Pot (sugars, FA, AA)

EvMA across all dietsWhat Links MetS Phenotypes?

Scott Chialvo, Che, Reif, Motsinger-Reif, Reed 2016 Metabolomics

High FatNormal

350 metabolites • 270 with IDs • Metabolon Inc.Top 30 diet differentiating metabolites• 9 unknowns • medium chain fatty

acids• short chain fatty

acids• dicarboxylic/

monohydroxy fatty acids

• Glycolysis/TCA cycle metabolites

Random forest identifies metabolites that differentiate based on diet

*metabolites potentially involved in omega FA oxidation

*

Oza, Aicher, & Reed, Metabolomics, 2018 Vishal Oza – now postdoc Lasseigne lab

15

16

9

Long/ Very long Chain Fatty Acids

Fatty Acids

Beta Fatty Acid Oxidation

TCA Cycle

Omega Fatty Acid Oxidation

Elo

ng

atio

n

Carnitineshuttle

?

?

No difference between ND and HFD

Elevated in HFD

? No evidence found in our study

Pathway suggested in this study

Hypothesized shift toward Omega fatty acid oxidation on a High Fat Diet

?

Oza, Aicher, & Reed, Metabolomics, 2018

King, McDonald, Long, 2012

Drosophila Synthetic Population Resource

raised larvae on normal

and high fat diets

Measure phenotypes and genetically map

association

What loci control genotype-by-diet interactions?

17

18

10

Dew-Budd et al. in revision

Main Genetic Effect QTLs


Male Pupae Weight


Genotype-by-Diet Interacting QTLs (Plastic QTLs)


Male Pupae Weight

19

20

11

Model●

●

Diet

MainGxDGenetic

●● ● ●● ● ● ● ●● ●● ● ●● ● ● ●

●●● ● ●

● ●

● ●

● ● ●● ●

●● ● ●● ● ●● ● ●

●●● ●

●● ● ●● ● ● ●

● ●●● ●

● ● ●

● ●

● ●

● ●

● ●● ● ● ● ●● ● ● ● ● ●● ● ● ● ● ● ● ● ● ●

●● ● ●

● ● ●

● ●●

●

●●

●●

●● ● ● ●● ●● ● ● ● ● ● ●

● ●● ●

●● ●

●

● ●● ● ●

● ●

●● ●● ●

●● ● ● ● ● ●● ● ●● ● ● ● ● ●

● ● ● ●●● ● ●●

● ● ● ●● ● ●● ● ● ● ●

● ● ● ● ●● ● ●● ● ● ● ●● ● ● ●

● ●●

● ●● ● ●● ●● ● ●

● ●● ●● ●

● ●● ●

2L 2R 3L 3R X

QTL Position (Mb)

Main Genetic Effect QTLs

and Plastic (GxD) QTLs are different!

Dew-Budd et al. in revisionWhat loci control genotype-by-diet interactions?

Tre-GTre-DTG-GTG-DmWt-GmWt-DfWt-GfWt-D

Epistatic Interactions Occur Throughout the Genome



49

36

18

45

26

17

40 5 43

2 61

5712

32

48

28

23

9

59

17

50 10

51

29

42

6434

30

22

15

253837

2765

3156

60

16

8

55

33

20

3 6

54 14

58

44

39

21

41

19

66

67

47

35

11

24

13

68

46

62

53

63

52

4

Male Weight- Main

1011 2

13 714 3

8

20

15

12

27

9

4

25

18

5

16 24

23

21

19

26

6 22 1

28

17

Male Weight- GxD

21

22

12

Diet 1 Diet 2

Main Genetic Effect (Gene A)

Main GxD Effect (Gene A)

EpistaticGenetic Effect

EpistaticGxD Effect

AA

Aa

aa

BB Bb bb

AA

Aa

aa

BB Bb bb

AA

Aa

aa

BB Bb bb

AA

Aa

aa

BB Bb bb

AA

Aa

aa

BB Bb bb

AA

Aa

aa

BB Bb bb

AA

Aa

aa

BB Bb bb

AA

Aa

aa

BB Bb bb



Gene ontology enrichment - Many loci linked to

metabolic functions- miR310 cluster

(regulation in hedgehog pathway)

- Pathways regulating growth and development (neurodevelopment)

R. Mather, A. Motsinger-Reif NCSU

NIH: West Coast Metabolomics Center

• Fiehn lab (GCTOF MS)

• 421 metabolites detected

• 169 with confirmed

chemical ids

raised larvae on normal

and high fat diets

What loci control metabolite profiles?

23

24

13

Metabolite 1

Gene 1Simple Genetic

Architecture

Architecture of Main Genetic Effect mQTLs

Hub Gene

Metabolite 2 Metabolite 3

Gene 1

Metabolite 1

Hub MetaboliteGene 1 Gene 2 Gene 3

Metabolite 1

Genetic red, GxD blue Genomic Position

Most mQTLs for Genetic effects are different from those with GxD effect

Met

abo

lite

Ind

ex

2L 2R 3L 3R X

What loci control metabolite profiles?

Hub Metabolite

Hub Gene

Simple QTL

Mather et al., in prep

Fiehn lab (GCTOF MS) 421 metabolites detected (169 chemical ids)

25

26

14

Genomic Position

NL

P

2L 2R 3L 3R X

Linking Metabolites through mQTLs

• mQTL shared between glycolic acid and citrulline

• pyruvate kinase rate limiting step in glycolysis

glycolate phosphoenolpyruvate pyruvate

pyruvate kinase

citrulline

Mather et al., in prep

Met 1 unknown

Met 2 known

Met 3 unknown

Met 4 known

Met 5 known

Gaussian Graphical Model

Correlation Structure

Conditional Dependence

27

28

15

Normal Diet GGM

High Fat Diet GGM

Phospholipid subnetworks

Dipeptide subnetworks

Fatty acid subnetworks

Overlaid Network

Gaussian graphical models by diet reveals distinct portions of sub-networks


Correlation networkGaussian graphical model

Edges in the GGM are between biologically (and chemically) similar metabolites, but correlation network misses known biological

relationships

R > 0.9235

C (8:0) C (10:0) C (12:0)

Fatty Acid Synthesis“positive control”


29

30

16

Unknown

Dipeptide “pathways” - GGM Dipeptide “hairball” - correlation

R > 0.9235

*GGM provides testable hypotheses


Treatment Control

TreadWheel

Sean Mendez

Uses negative geotaxis to exercise

adult flies

Rachel Hill

Exercise

Mendez et al. 2016 PlosOne

31

32

17

0.75

1.25

1.75

2.25

2.75G

lyco

gen

1.3

1.4

1.5

1.6

1.7

Tri

gly

ceri

de

0.11

0.12

0.13

0.14

0.15

0.16

Pro

tein

0.82

0.84

0.86

0.88

0.9

0.92

Wei

gh

t

Control ExerciseControl Exercise

p =10-7.52 p =10-2.10

p =10-4.27 p =10-46.66

Exercise Works in Flies

Control ExerciseControl Exercise


1.7

1.9

2.1

2.3

2.5

2.7

2.9

3.1

Clim

bin

g

Control Exercise

p =10- 36.87

Climbing Performance Improves

Nicole Riddle, Louis Watanabe, Maria DeLuca UAB


• Exercise interacts with genetic line and sex • Mitochondrial function genes show exercise and

genotype-by-exercise interactions• Adult exercise interacts with larval diet and genotype

ExerciseControl

Normal High Fat Normal High Fat

Exercise Works in Flies(for some better than others)

Nicole Riddle, Louis Watanabe, Maria DeLuca UAB

Clim

bin

g H

eig

ht

Lowman et al, in prep Genotype x Diet x Exercise p=5.7e-17

33

34

18

www.motherjones.com

Larval bacterial communities from natural conditions are interact with genetic and dietary effects

Bombin et al, 2020, Microorganisms 8(12), 1972

P,NS

P,S

PA,SPA,NS

R,NS

R,S

100 most abundant microbial ZOTU

Discriminate analysis

1. Maternal microbes shape bacterial community2. Community on natural peach diets differ from lab and

autoclaved diets 3. Genotype-by-diet interactions on bacterial

communities4. Correlations between dominant taxa and metabolic

traits (consistent with the literature)

High Fat (NS) High Sugar (NS)

Bacterial communities more strongly influenced by environmental bacteria on a high fat diet

Bombin et al, in prep

Autoclaved Peach

Lab and non-autoclaved peach

The bacterial species correlated with metabolic phenotype vary with diet and maternal microbes

35

36

19

• Flies are great models!

• Genotype-by-diet interactions are a substantial contributor to MetS variation

• Metabolite profiles have predictive power for MetS phenotypes

• Quantitative data can be used to categorize unknown compounds and hypothesize new pathways

• Loci for genetic and genotype-by-diet interactions are different

• Additional environmental factors can be modeled in Drosophila Tori Nelko

The Take Home Messages

[email protected] flygxe.ua.edu

Flies, More Flies, and Statistics….

[email protected]

Funding:NSF 1915544 to LKRNIH R25GM130517 to LKR NIH-R01 GM098856 to LKRNIH-NRSA Fellowship to LKR

NSF DEB 1737869 to LKRAustralian ARC DP0880204 to GGNIH R01-GM61600 to GG UA Office of Research UA College of Arts and Sciences

Collaborators:Greg Gibson Alison Motsinger-ReifSiddharth RoyRavi MatherRonglin CheOliver FiehnRolf BodmerNorm GlassbrookMaria DeLucaDavid ReifRonglin CheOliver FiehnThomas WernerSong SongThomas MerrittKelly DyerArt EddisonNicole RiddleLouis Watanabe

Kelly Dew-BuddDana DavisKatie BrayNick IzorSean MendezLeah LeonardMatt KiefferCheyenne PaivaAndrea DavidsonSteph Williams Julie BrownRuth BishopLevi MillerOlivia SorrellAlison AdamsAshley GilchristJulie JarniganCigdem Tunckanat

Alison AdamsAshley GilchristJulie JarniganCigdem TunckanatJordyn MerriamChristie TalleyTanner HallmanRachel HillMeredith OwensJaron NixJohn HendersonJoseph AicherMichael MooreAndrew DavisZavier MasonBeau SchafferNelson BrownKara Macintyre

Kelsey LowmanLaura MaflaLeigh Ann PounceyAndrei BombinCourtney MooreClare Scott ChialvoVishal OzaTreavor HearingLogan GriffinCarson FosterBrandon MoyeOwen CunneelyKatie SandlinOlivia FishKira EickmanSarah-Ashley GiambroneMichelle TanRosemary HartlineTyler Hale

Reed Lab Motto: When it comes to genes, our flies are always open.

Jordan BeveridgeBailey LoseRyan O’RourkeAbigail RuesyLauren HaynesMengting SueRachel CowanMaddison SharpCole KiserAnna Grace PriceCaroline HartAbigail MyersMcKenzie ChamberlainGrace KeirnSam HoffmanSean Shelley-TremblayJohn YordyInes MartinandJordan AlbriechtAnnie BacklundJade Miller

37

38

20

The Genomics Education Partnership:

A community of practice that enhances research opportunities for students and faculty at diverse

institutions

Funded by the National Science Foundation, National Institutes of Health, HHMI

With continuing support from Washington University in St Louis

and the University of AlabamaCopywrite 2020, Genomics Education Partnership

thegep.org

Copywrite 2019, Genomics Education Partnership

GEP Membership Distribution

• nationwide collaboration of 160+ institutions

• > 2000 students participate annually

• core team provides computational, scientific, and pedagogical support - facilitates implementation

• uses comparative genomics to study the evolution of genes and genomes

• implementation of takes many forms

– Several-week modules to entire semester

• removes the technical barriers – all GUI and web-based

• students gain confidence in their ability to “do” science

Course-based Undergraduate Research

Experience (CURE)

thegep.org

thegep.org

39

40

21

Gene annotation workflow

thegep.org*looking for new education and scientific partners

41

42

disentangling the gene-by- environment interaction

Documents