identifying dysregulated genes in autoimmune disease
TRANSCRIPT
Identifying Dysregulated Genes in Autoimmune Disease
Chris Cotsapas PhD Yale Neurology/Genetics
Broad Institute [email protected]
Causall Identifying Dysregulated Genes
in Autoimmune Disease
Chris Cotsapas PhD Yale Neurology/Genetics
Broad Institute [email protected]
Multiple sclerosis GWAS
47 new hit15,000Immunochip200,000 SNPs (targeted) 10,000 25 new hits
WTCCC28000 650,000 SNPs
MS 6000Subjects Meta-‐Analysis 3 new hits
2.6 million SNPs 4000 Meta v2.56 new hits 2.6 million SNPs IMSGC GWAS 1 new hit
2000 345,000 SNPs ANZ GWAS2 hits
550,000 SNPs IMSGC NEJM 2007
De Jager et al. Nat Genet 2009 2007 2008 2009 2010 2011 2012 Rubio et al. Nat Genet 2009
IMSGC Nature 2011
Patsopoulos et al. Ann Neurol 2011
Date of comple/on
Meta-‐Analysis v3.0… and that’s not all! 16K MS cases / 26K controls
Replica;on18K MS cases / 18K controls
100 new hits
47 new hit15,000Immunochip200,000 SNPs (targeted) 10,000 25 new hits
WTCCC28000 650,000 SNPs
MS 6000Subjects Meta-‐Analysis 3 new hits
2.6 million SNPs 4000 Meta v2.56 new hits 2.6 million SNPs IMSGC GWAS 1 new hit
2000 345,000 SNPs ANZ GWAS2 hits
550,000 SNPs IMSGC NEJM 2007
De Jager et al. Nat Genet 2009 2007 2008 2009 2010 2011 2012 Rubio et al. Nat Genet 2009
IMSGC Nature 2011 Date of comple/on Patsopoulos et al. Ann Neurol 2011
GWAS signals are enriched in regulatory DNA
Maurano et al Science 2012
MS GWAS hits enriched in transcription factor binding sites
Farh et al Nature 2015
Plotted SNPs
− lo
g 10(p−v
alue
)
NFKB1 locus in MS GWAS 10
8
6
4
2
0
103.4 103.5 103.6 103.7
Position on chr4 (Mb) IMSGC, Nat Genet, 2013 Housley et al STM 2015
0
20
40
60
80
100
Recom
bination rate (cM/M
b)
rs7665090
●
●
●
● ● ●
●● ●
●●● ●
● ●
● ●
●
● ●
●
●
0.2 0.4 0.6 0.8
r2
NFKB1 MANBA UBE2D3
72
A) Regional Association and B) Forest Plot
Supplementary File
Supplementary Figure 42. Discovery phase rs12946510. A
B
IKZF3/ORMDL3 locus in MS GWAS
IMSGC, Nat Genet, 2013
Gene
Gene-DHS correlation
posterior CP x PPA
DHS
Posterior probability
of association PPA
Approach Total gene posterior GP = Σ(CPDHS,gene x PPADHS)
Regulatory posterior RP = Σ(PPADHS)
SNP
Parisa Shooshtari
Problem 1: DHS-gene correlations
CD 3 CD 14 CD 34 Thymus Kidney Lung Heart Brain
Gene 1 Gene 2
CD 3 CD 14 CD 34 Thymus Kidney Lung Heart Brain
Parisa Shooshtari
Aligning DHSs Over Samples
Parisa Shooshtari
Identify detectable DHS clusters Scenario 2
Align over 57 tissue replicates from REP
C1 1 C1 2 Tissues Hotspot peaks C2
C3 1 1
C2 C3
2 1
T1-Rep1 C4 0 C4 0 T1-Rep2 2 C5 0 C5 0
T2-Rep1 T2-Rep2 1
C6 1 C6 2
T3-Rep1 T3-Rep2 2 Unreliable Cluster Reliable Cluster
T4-Rep1 T4-Rep2 2
T5-Rep1 1 1,079,138/1,994,675 (54.1%) clusters pass
Cover 8% of genome (cf. 14% of all DHS) T5-Rep2
T6-Rep1 2T6-Rep2 NB singletons, low power
Scenario 1
1 1 1
1 1 1 1 1
1 1
0
0
Parisa Shooshtari
0.0
0.2
0.4
Proportion of Heritability QC+ DHS clusters capture most MS heritability
Caveat DHS clusters are
Status
DHS_Peaks_Hotspot
DHS_Clusters
Penis_Foreskin_Melanocyte_Prim
ary_Cells
Penis_Foreskin_Keratinocyte_Primary_C
ellsPenis_Foreskin_Fibroblast_Prim
ary_Cells
PancreasM
obilized_CD
4_Primary_C
ellsM
obilized_CD
34_Primary_C
ellsM
obilized_CD
3_Primary_C
ellsIM
R90
H9
H1_D
erived_Mesenchym
al_Stem_C
ellsH
1_BMP4_D
erived_Trophoblast_Cultured_C
ellsH
1_BMP4_D
erived_Mesendoderm
_Cultured_C
ellsH
1G
astricFibroblasts_Fetal_Skin_U
pper_BackFibroblasts_Fetal_Skin_Scalp
Fibroblasts_Fetal_Skin_Quadriceps_R
ightFibroblasts_Fetal_Skin_Q
uadriceps_LeftFibroblasts_Fetal_Skin_Biceps_R
ightFibroblasts_Fetal_Skin_Biceps_Left
Fibroblasts_Fetal_Skin_BackFibroblasts_Fetal_Skin_Abdom
enFetal_Thym
usFetal_Testes
Fetal_Stomach
Fetal_Spinal_Cord
Fetal_Renal_Pelvis_R
ightFetal_R
enal_Pelvis_LeftFetal_R
enal_PelvisFetal_R
enal_Cortex_R
ightFetal_R
enal_Cortex_Left
Fetal_Renal_C
ortexFetal_Placenta
Fetal_Muscle_Trunk
Fetal_Muscle_Leg
Fetal_Muscle_Back
Fetal_Muscle_Arm
Fetal_Lung_R
ightFetal_Lung_Left
Fetal_LungFetal_Kidney_R
ightFetal_Kidney_Left
Fetal_KidneyFetal_Intestine_Sm
allFetal_Intestine_Large
Fetal_Heart
Fetal_BrainFetal_Adrenal_G
landC
D8_Prim
ary_Cells
CD
56_Primary_C
ellsC
D4_Prim
ary_Cells
CD
3_Primary_C
ellsC
D19_Prim
ary_Cells
CD
14_Primary_C
ellsBreast_vH
MEC
Cell
wider than DHS peaks (250-400bp vs 150bp
Parisa Shooshtari Hilary Finucane Alkes Price
Prop
ortio
n of
h2g
Correction
Challenge 2: Gene expression correlation QQ plot for P Value of Correlation Between
One DHS and 14000 Genes
Correlation Structure of the Gene Expression Data
Before Correction After
Parisa Shooshtari
Gene
Gene-DHS correlation
posterior CP x PPA
DHS
Posterior probability
of association PPA
Approach Total gene posterior GP = Σ(CPDHS,gene x PPADHS)
Regulatory posterior RP = Σ(PPADHS)
SNP
Parisa Shooshtari
Application to MS GWAS
Chr 6 90.5-91.5Mb RP = 0.945
Gene MDN1 0.555 BACH2 0.162
GABRR2 0.106 RRAGD 0.065 GJA10 0.029
MAP3K7 0.028
GP
Parisa Shooshtari IMSGC NG 2013
72
B
A) Regional Association and B) Forest Plot
Supplementary File
Supplementary Figure 42. Discovery phase rs12946510. A
IKZF3/ORMDL3 locus Gene GP
Chr 17 ORMDL3 0.029 34.5-35.5Mb PIP4K2B 0.022 RP = 0.295 IGFBP4 0.018
IKZF3 0.015 GSDMB 0.014
SMARCE1 0.013 CCR7 0.013 TNS4 0.01
ZPBP2 0.009 MED1 0.009
MED24 0.009 KRT24 0.009 PNMT 0.008 CDK12 0.007 RPL23 0.007 PSMD3 0.007 PLXDC1 0.006 TOP2A 0.006 RARA 0.006
MS GWAS hits enriched in transcription factor binding sites
Farh et al Nature 2015
MS GWAS risk effect: NFKB1 locus
MS patients show altered NFκB signaling in CD4+ T cells
Figure 1. Naïve CD4 cells from patients with MS exhibit increased phospho-p65 NFκB. Flow cytometry of PBMCs from age-matched healthy control (HC) and relapsing-remitting MS (RRMS) patients stained for CD4, CD45RA, CD45RO, and pS529 p65 NFκB. MFI of p65 results are shown gated on naïve CD4+CD45RA+CD45RO- T-cells.
CD4+ T cells from MS patients proliferate more rapidly after
stimulus (Kofler et al JCI 2014)
ex vivo CD4+ T cells show higher p-p65 (Housley et al, STM 2015)
0 15 300
10
20
30
Minutes
Nu
cle
ar
localiza
tio
n
GGAA
p= 0.05
p50
NFk
B
0
10
20
30
GG AA
p= 0.037
0
20
40
60
80
100
% Iκ
Bα
p= 0.0091 p= 0.019
GG AAAG AAGG
TNF-αNo Stim PMA
AG AAGG AG0
500
1000
1500
2000
pNFκ
B
p= 0.029 p= 0.027
GG AG AA GG AG AA
TNF-α PMA
a b
rs228614
Housley et al, STM 2015
MS risk effect near NFKB1 alters signaling in CD4+ cells
Housley, unpublished
MS variant in TNFRSF1A alters TNFα-dependent NFκB signaling
GWAS loci harbor many NFκB genes
Housley, unpublished
Model: NFκB signaling variation
p50
P-p50 External stimulus
p65 *NFκB
Activation Proliferation Survival
p65 *NFκB
Activation Proliferation Survival Broader phenotype?
p50
P-p50
GV in NFκB pathway
GV in NFκB TFBS
New gene activation patterns by NFκB
Systematic dissection I
CD4+ T cells Timing Event
0 exposure
TNFα
15m signaling
Phospho-flow NFκB
30m translocation
NFκB CHiP-seq
2h gene activation
H3K27Ac-seq RNA-seq Brad Bernstein
Chris Cotsapas David Hafler Will Housley
3d cell phenotype
αCD3/CD28
CyTOF
25 NFKB1 risk variant homozygotes 25 NFKB1 non-risk variant homozygotes
Acknowledgements • IMSGC
– David Hafler – Phil De Jager – Steve Hauser – Adrian Ivinson – Nikos Patsopoulos – Many, many others
• Partners – David Hafler – Phil De Jager – Brad Bernstein – John Stamatoyannopoulos
• Yale labs – Parisa Shooshtari – Mitja Mitrovic – Alex Casparino – Will Housley
8.6e-09 1.6e-08 3.2e-09 1.5e-08 1.5e-08
Gen
es
Burdened DHSs DHS1 DHS2 DHS3 DHS4 DHS5
●
●
●
●
●
●
●
●
●
●
●
●
●●
● ●
●
●
●
●
●●
●
●
●
●
●
●
●
●●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●●●
●
●
●
●●
●
●
●●●
●
●●
●
●●●
●
●
●
●●
●●
● ●●
●
●
●●
●
●●●
●
●
●
●
●
●
●
●● ●●
●
●
●
●
●
●●
●
●
●
●
●
●
●●●
●
●
●
●
●
●
● ● ●
●●
●●
●
●●●
●
●
●
●●
●●
●
●● ●●
●
●
●●
●
●
●
●
●
●
●●
●●
●
●
●●
● ●
●
●
●●●●
● ●●
●●
●
●●
●
●
●
●●●
●
●● ●●
●
●
●●
●
● ●
●●
●
● ●
●●
●
●
●
●
● ●
●
●
●
●
●
●
●
●
●
●
●
●
●●●
●
●
●
●
●
●●
●
●●
●
●●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●●
●●●●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
● ●●
●
●●
●
●●
●
●
●
●
●●
● ●●
●●
●● ●
●●
●
●●
●
●
●●●
●
●● ●● ●●
●
●
●
●●
●
●
●
●
●
●●
●
●
●●
●
●
●●
●
●
●
●
●
●●
●
●
●
● ●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●●
●●●●
●
●
●
●
●
●●●●
●
●
●
●
●
●●●
●
●
●●●
●●
●●●
●●●●●
●
●
●●●●●
●●●●●
●
●●
●
●●●●●
●
●
●●●
●
●●
●
●
●
●
●●●
●
●
●●●●
●●
●
●
●●●●
●
●●●
● ●
● ●
●
●
●
● ●
●
●●
● ●
●
●
●●
● ● ●
●
●
●
● ●
●
●
●●●
●●●
●
●●
●
●
●
●●
●●●
●
●
●●
●
●●
●
● ●
●
●
●
●
●
●
●● ●
●
●
●
●
●
●●●
●
●
●
●
●●
●●
●
●
●
●
●
●
●
●
●
●●
●
●
●●● ●●
●
●
●
●
●
●
●●
● ● ●
●●●
●●
●
●
●
●
●●
●
●
●
●
●●
●
●●●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●●
●●
●
●
●
●●
●
●
●
●
●
●
●
●
●●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●●●
●
●●●
●●
●
●
●
●●
●
●
●
●
●
●
●●
●●
●
●
●
●●●●
●●
●
●
●●●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
● ●
●
● ●
●
● ●
●
●
●
●
●●
●●
●
●
●
●●
●●●●
●●
●
●
●●
●
●●
●●
●●
●
●
●●
● ●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●●●
●
●
●
●
●
●
●●
●
●
●●
●
●●
●
●
●●●●●
●
●●
●●●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●●
●
● ●
●
●
●●●
●
●
●
●
●
●
●●●
●
●●
●●
●
●
●●
●
●
●
●
●
●●
●
●●
●●●●●●
●
●
●
●
●
●
●
●●
●●
●
●
●
●
●
●
●●
●
●
●
● ●
●●
●●
●
●
●
● ●●
●●
●
●
●
●●
●
●
●
●
● ●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
● ●
●
●
●
●●
●
●
●
● ●
●
●
●
●●
●
●
●
●
●
●
●
●●
●
●
●
●●
●
●
●
●
●
●●
●
●
●
●●
●
●
● ●●● ●
●
●
●
●
●
●
●
●
●
● ●●●
●●
●●
●
●
●
●
●●
●●
●
●
●
●
●
●●
●●
●
●
●●
●
●
●
●●
●
●
●●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●●
●
●
●●●●●
●
●
●
●
●●
●●
●
●
●
●
●
●
●●
●●
●
●
●
●
●
●
●
●
●
●●
●
●● ●
●
●
●
●
●
●
● ●●
●
●
●
●
●
●
●
●
●
●
●
●●
●●
●
●●
●
●
●●
●
●●
●●
●
●●●●
●
●
●
●●●
●●
●
●
●
●
●
●
●
●0
2
4
6
8
42.5 43.0 43.5 44.0Chromosome 1 position (Mb)
−log
10(P)
In credible interval ● ●No Yes
Credible interval plot
THADA
PLEKHH2
LRPPRC
DYNC2LI1
ABCG8
ABCG5 ZFP36L2 KCNG3 OXER1
EML4 MTA3 HAAO
Position on Chromosome 2