PGC Worldwide Lab Call Details

Transcription

1 PGC Worldwide Lab Call Details DATE: Friday, September 11, 2015 PRESENTER: TITLE: START: DURATION: Nancy J. Cox, PhD Mary Phillips Edmonds Gray Professor of Genetics Director, Vanderbilt Genetics Institute Director, Division of Genetic Medicine at Vanderbilt University Medical Center Novel Methods for Discovering and Characterizing Genome Variation Affecting Neuropsychiatric Disorders in BioVU, a 200,000 Member Biobank We will begin promptly on the hour EDT US East Coast 0700 PDT US West Coast 1500 BST UK 1600 CEST Central Europe 0000 AEST Sydney, Australia (Saturday, September 12, 2015) 1 hour TELEPHONE: US Toll free: International direct: Toll-free numbers: Press *0 to get assistance from an operator. PASSCODE: then #

2 Lines are Muted NOW Lines are automatically muted by the operator to prevent unnecessary background noise. Dial *1 if you would like to ask a question of the presenter. Presenter will respond to questions as time allows. Operators announce callers one at a time during question and answer sessions. Dial *0 if you need operator assistance at any time during the call.

3 UPCOMING PGC Worldwide Lab The October PGC Worldwide Lab call is cancelled due to the World Congress of Psychiatric Genetics being held the same month. The next call will be in November. Mark your calendars! DATE: Friday, November 13, 2015 PRESENTER: TITLE: START: DURATION: Pamela Sklar, MD, PhD Mount Sinai Medical Center To Be Announced We will begin promptly on the hour EST US East Coast 0700 PST US West Coast 1500 BT UK 1600 CET Central Europe 0000 AEST Sydney, Australia (Saturday, November 14, 2015) 1 hour

4 Novel Methods for Discovering and Characterizing Genome Variation Affecting Neuropsychiatric Disorders in BioVU, a 200,000 Member Biobank Fun with BioVU! anderbilt Genetics Institute! Nancy J. Cox, Ph.D. 1

5 Overview Our new approach to genome analysis (PrediXcan) and why we think it has value What we are doing in BioVU with PrediXcan Results of preliminary studies small vignettes and big picture overviews Future plans 2

6 3

7 GTEx: Genotype Tissue Expression Subjects at autopsy, tissue / organ donation Whole genome sequencing of subjects; RNAseq on ~50 tissues eqtl studies within and across tissues and a variety of applications 4

8 Type 1 Diabetes Crohns Disease Overall

9 6

10 A Missing Data Problem? If a substantial fraction of the genome variation affecting risk of common disease is regulatory, why not alter focus of analyses to endophenotypes that are more direct measures of what we really want the genetically determined part of transcript levels Instead of testing individual variants, aggregate variants into SNP-based predictors of transcript levels and test those directly for association with disease 7

11 PrediXcan Haky Im Published in Nat Genet (online) as GTEx companion paper 8

12 PrediXcan Other factors GReX Genetically regulated expression Traitaltered component Figure 1: Regulatory Mechanism Tested by PrediXcan Trait Gene Expression Decomposition 9

13 PrediXcan Other factors GReX Genetically regulated expression Traitaltered component Figure 1: Regulatory Mechanism Tested by PrediXcan Trait Gene Expression Decomposition 10

14 Analogous to Imputation Learn relationship of genome variation to transcriptome in reference sample (GTEx) Store weights from prediction equations Apply to any dataset with genome interrogation 11

15 Prediction Performance R 2 by Heritability 12

16 Prediction in an Independent Sample Significance of correlation between predicted and directly measured expression levels: q- value < 0.05 for 40-50% of genes, < 0.1 for 60-70% 80% of genes have correlation between predicted and measured expression > 0.1, 50% > 0.2 Polygenic prediction < {lasso, elastic net} genetic architecture Lasso prediction includes ~60-80 SNPs (CONTEXT!) 13

17 Advantages of Framework We iteratively use more and more of what we do know to figure out what we most want to learn for new discovery Signals come at the levels of the gene, improving the ability to do pathway/ network analyses Sets up a natural framework for a unified analysis of whole genome sequence data, and for combining results of sequencing and GWAS 14

18 Resources for EMR-based research at Vanderbilt The Synthetic Derivative A de-identified and continuously-updated image of the EMR: 2,358,760 subjects BioVU Subjects with DNA: >200,000 Dense (GWAS-level) genotypes: ~20,000 Exome chip data: 42,000 15

19 Resources for EMR-based research at Vanderbilt end 2016 The Synthetic Derivative A de-identified and continuously-updated image of the EMR: >2,500,000 subjects BioVU Subjects with DNA: ~225,000 Dense genotypes: >100,000 Whole genome sequencing: ~1000 s 16

20 The genome-wide association study Target phenotype associatio n P value chromosomal location The phenome-wide association study Target genotype association P value diagnosis code PheWAS requirement: A large cohort of patients with genotype data and many diagnoses 17

21 BioVU X PrediXcan: Gene-based PheWAS BioVU An in silico Discovery Engine 18

22 PrediXcan X BioVU Preliminary studies on those genes in which all SNPs included in the prediction equation are included in the Illumina 1M SNP set (~5000 BioVU subjects) Studies completed for whole blood (built in >900 from DGN) and cardiac tissue (built from > 300 in GTEx) 125 genes with whole blood predictors; 300+ genes with cardiac tissue 19

23 20

24 Continuum from Mendelian to Complex Continuum from LOF to deleterious to expression 21

25 What Phenotypes Are Associated with Reduced GReX of Mendelian Genes? 22

26 PrediXcan in BioVU for PEX19 Mutations in PEX19 lead to a peroxisomal biogenesis disorder, Zellweger Syndrome spectrum of Mendelian phenotypes Hypotonic, seizures, bony stippling at the patella and other long bones, kidney and liver cysts, coagulopathies, stone formation and renal failure What BioVU phenotypes are associated with reduced GReX of PEX19? 23

27 Reduced GReX of PEX19 24

28 Increased GReX PEX19 25

29 PrediXcan in BioVU for TK2 TK2-related mitochondrial DNA depletion syndrome, myopathic form (TK2-MDS) is an inherited condition that causes progressive muscle weakness (myopathy). The signs and symptoms of TK2-MDS typically begin in early childhood. Development is usually normal early in life (but often born with ptosis of eyelid), but as muscle weakness progresses, people with TK2- MDS lose motor skills such as standing, walking, eating, talking even eye movement. Finnish study noted increase in fractures. Due to increase in falls caused by myopathy? Bones weak? If weak, is that a consequence of mitochondrial depletion? Or lack of muscle opposition? 26

30 Reduced Predicted Expression TK2 27

31 Increased Predicted Expression TK2 Note number and significance of phenotypes associated with increased GReX of Mendelian genes 28

32 Decreased Predicted Expression of HFE Renal failure NOS E Hypertensive heart and/or renal disease Pulmonary congestion and hypostasis Kidney replaced by transplant Hypertensive chronic kidney disease Renal failure Osteitis deformans and osteopathies associated w/ other disorders classified 731 elsewhere Acute renal failure Viral infection Chronic renal failure [CKD] Cardiomegaly Nephritis; nephrosis; renal sclerosis Hypertensive heart disease Renal dialysis Disorders of menstruation and other abnormal bleeding from female genital 626 tract Irregular menstrual cycle/bleeding

33 Does Altered (Increased or Decreased) Expression of Mendelian Disease Genes Contribute Disproportionately to Common Disease? 30

34 BioVU X PrediXcan: Gene-based PheWAS BioVU An in silico Discovery Engine And a phenome reference panel for validating results of rare variant discovery 31

35 Reduced Predicted Expression GRIK5 An Eye Super Gene? 32

36 33

37 Reduced GReX for FBXO5 112 Candidiasis E Gingival and periodontal diseases E Infection with drug-resistant microorganisms Late effects of cerebrovascular disease Bacteremia Intervertebral disc disorders Cerebral atherosclerosis Epilepsy Methicillin sensitive Staphylococcus aureus Bacterial infection NOS Acute osteomyelitis Dental caries Partial epilepsy Back pain Diseases of hard tissues of teeth Sprains and strains of back and neck Intervertebral disc disorder with myelopathy Chronic kidney disease, Stage I or II Osteomyelitis, periostitis, and other infections involving bone GERD Intracranial hemorrhage Pyelonephritis Esophagitis, GERD and related diseases Cerebrovascular disease Septicemia Other disorders of synovium, tendon, and bursa Displacement of intervertebral disc Iron deficiency anemias, unspecified or not due to blood loss Staphylococcus infections

38 Increased GReX of FBXO Lymphosarcoma Neutropenia Decreased white blood cell count

39 Reduced GReX of ZNF Cellulitis and abscess of trunk E Acid-base balance disorder Mental retardation Acidosis Sexually transmitted infections (not HIV or hepatitis) Thrombocytopenia Purpura and other hemorrhagic conditions Chromosomal anomalies and genetic disorders Complications of gastrostomy, colostomy and enterostomy Lack of normal physiological development Posttraumatic wound infection not elsewhere classified Diseases of pulp and periapical tissues Localized adiposity Mycoses Peritonitis and retroperitoneal infections Protein-calorie malnutrition Hypotension NOS Infantile cerebral palsy Respiratory failure, insufficiency, arrest Failure to thrive Periapical abscess Hemorrhage of gastrointestinal tract Poisoning by agents primarily affecting the cardiovascular system

40 Increased GReX of ZNF MRSA pneumonia Degenerative skin conditions and other 702 dermatoses Hyperparathyroidism

41 Reduced GReX of CCL Coronary atherosclerosis Ischemic Heart Disease Other chronic ischemic heart disease, unspecified Angina pectoris Myocardial infarction Enthesopathy

42 Increased GReX of CCL5 Malignant neoplasm of retroperitoneum and peritoneum Hammer toe (acquired) Other arthropathies Internal derangement of knee Other specified disorders of liver Ingrowing nail Acute appendicitis

43 Reduced Predicted Expression CCKBR cholecystokinin B receptor CCKBR is a receptor for regulatory peptides of the brain and gastrointestinal tract; No phenotypes have p<.01 for increased GReX 40

44 Reduced Predicted Expression ST6GALNAC4 41

45 Results on ~500 genes in 5000 individuals Results on all genes in 13,000 Results on all genes in 25,000 Results in 100,000+, 200,000+, 42

46 Big Picture Results Mendelian Genes GReX Associations All other genes GReX Associations 43

47 Big Picture Results Reduced GReX Gene Associations Increased GReX Gene Associations 44

48 Future Plans Phenome reference panel for validating rare variant discoveries Systematic discovery of genes where increased GReX is associated with phenotype =? Rare variant protective Comprehensive assessment of GReX for all genes that are drug targets Drug repurposing Previews of adverse events 45

49 Future Plans Multivariate heritability across EMR phenotypes Collaboration with Jackson labs to iterate between mouse/human phenotypes Use the Synthetic Derivative to identify patterns of health care usage and biomarkers predictive of later disease diagnoses Can we combine these biobank predictions with polygenic risk scores to achieve clinical utility and reduce diagnostic odysseys? 46

50 Our GTEx Team Dan Nicolae Eric Gamazon Hae Kyung Haky Im Lin Chen Anuar Konkashbaev Richard Jones Barbara Stranger Younghee Lee 47

51 48