biobankMeta-Oct232019-updated-genetica.pptx

Biobanco de metaanálisis global de los genes más conocido para un mejor tratamiento de diversas anomalías de causas congénitas

1. Global Biobank Meta-analysis Initiative: Powering genetic discovery across human diseases Wei Zhou, Ph.D. Post-doctoral Fellow Massachusetts general hospital Broad Institute Harvard Medical School ‹#›

2. Use of biobanks becomes an essential approach for genetic discovery ‹#›

3. Genotyping/Imputing 100s or 1000s GWASs Curating Phenotypes Phenome-wide GWASs can be conducted in biobanks ‹#›

4. Genotyping/Imputing 100s or 1000s GWASs Curating Phenotypes Phenome-wide GWASs can be conducted in biobanks ‹#› More cost-efficient than disease-based cohorts Tend to have relatively smaller case numbers

5. + + + Biobank 1 Biobank 2 Biobank 3 Biobank N Meta-analysis ‹#› Biobank meta-analysis has many potential benefits

6. + + + Biobank 1 Biobank 2 Biobank 3 Biobank N Meta-analysis ‹#› Better power for GWASs More accurate polygenic risk scores GWASs of understudied diseases Opportunity for cross-validation Improvements in fine-mapping (PgmNr 2283) Potential to explore subgroup analyses … Biobank meta-analysis has many potential benefits

7. China Kadoorie Biobank 100k UK Biobank 500k FinnGen 180k Biobank Japan 200k HUNT Study 70k BioVu 120k BioME 32 k UCLA Precision Health Biobank 14k Michigan Genomics Initiative 47k Partners Biobank 25k Colorado Biobank 30k Estonian Biobank 150k Generation Scotland 24k Million Veteran Program 500k deCODE Genetics 250k East London Genes & Health 10k LifeLines NL 52K *genotyped sample sizes ‹#› Global Biobank Meta-analysis Initiative > 2 million genotyped samples Netherlands twin register 24K

8. Asthma ‹#› Biobank Japan  8,204 cases, 10 loci BioMe-AA  842 cases, 0 loci BioMe-EA 455 cases, 0 loci FinnGen 15,926 cases, 14 loci Generation Scotland 196 cases, 1 locus HUNT 5,587 cases, 3 loci UK Biobank 26,332 cases, 31 loci deCODE  16,767 cases, 5 loci BioMe-HA  1,459 cases, 1 locus China Kadoorie Biobank  990 cases, 0 loci Michigan Genomics Initiative 6,079 cases, 0 loci

9. ‹#› Can different biobanks be meta-analyzed together?

10. China Kadoorie Biobank 100k UK Biobank 500k FinnGen 180k Biobank Japan 200k HUNT Study 70k BioVu 120k BioME 32 k UCLA Precision Health Biobank 14k Michigan Genomics Initiative 47k Partners Biobank 25k Colorado Biobank 30k Estonian Biobank 150k Generation Scotland 24k Million Veteran Program 500k deCODE Genetics 250k East London Genes & Health 10k LifeLines NL 52K *genotyped sample sizes ‹#› Netherlands twin register 24K Biobanks with different origins

11. China Kadoorie Biobank 100k UK Biobank 500k FinnGen 180k Biobank Japan 200k HUNT Study 70k BioVu 120k BioME 32 k UCLA Precision Health Biobank 14k Michigan Genomics Initiative 47k Partners Biobank 25k Colorado Biobank 30k Estonian Biobank 150k Generation Scotland 24k Million Veteran Program 500k deCODE Genetics 250k East London Genes & Health 10k LifeLines NL 52K *genotyped sample sizes ‹#› Netherlands twin register 24K Biobanks with different sampling Population-based, N = 1.2m Hospital-based, N = 0.74m Mixed, N = 180k

12. Genotyping/Imputing Curating Phenotypes Biobanks may have different genotyping and/or phenotyping strategies 13.1m variants 1000G + Japanese WGS 37.1m variants HRC 16.9m variants Finnish WGS 24.2m variants HRC+ HUNT WGS BBJ FinnGen HUNT UKBB Doctors’ diagnosis 8,204 cases 167,883 controls ICD codes, prescription, insurance reimbursement, etc. 15,926 cases, 113,428 controls ICD codes mapped to Phecode, self-report, doctors diagnosis 5,587 cases, 49,870 controls ICD codes mapped to Phecode 26,332 cases, 375,505 controls ‹#›

13. ‹#› Hospital-based Population-based Mixed Prevalence of asthma varies among biobanks Sampling

14. ‹#› European (N=1.02m) Ancestry Group Hispanic American (N=11k) East Asian (N=252k) African American (N= 7.5k) Prevalence of asthma varies among biobanks

15. Using genetic analysis to validate the integration of association results across biobanks ‹#› Can different biobanks be meta-analyzed together?

16. To validate the integration of association results across biobanks  Comparing of effect sizes at top established GWAS hits 27 index variants identified for Asthma identified by previous GWAS studies Moffatt MF et al. (2010), Ferreira MA et al. (2011), Torgerson DG et al. (2011), Ferreira MA et al. (2013), Pickrell et al, (2016), Vincente et al. (2017), Demaina et. al. (2017), Demaina et. al. (2019) ‹#›

17.  Comparing of effect sizes at top established GWAS hits 27 index variants identified for Asthma identified by previous GWAS studies All biobanks meta-analysis vs. previous GWAS Individual biobank vs. leave-one-biobank-out meta-analysis ‹#› To validate the integration of association results across biobanks

18. Effect sizes by all biobank meta-analysis show consistent directions with previous GWAS ‹#› 27 Asthma index variants x=y

19. All 27 asthma index variants have p-value < 5x10 -5 in all-biobank meta-analysis ‹#›

20. All 27 asthma index variants have p-value < 5x10 -5 in all-biobank meta-analysis ‹#› Biobank meta: larger sample = greater power

21. All 27 asthma index variants have p-value < 5x10 -5 in all-biobank meta-analysis ‹#› Biobank meta: larger sample = greater power Loci with comparable p-values: winners curse? mis-mapped original index variant? phenotypic variability?

22. Individual biobanks have highly correlated effect sizes with leave-one-biobank-out meta-analysis ‹#› 15,926 cases 26,332 cases 16,767 cases 27 Asthma index variants

23. Using genetic analysis to validate the integration of association results across biobanks Comparing of effect sizes at top established GWAS hits 28 index variants identified for Asthma identified by previous GWAS studies Moffatt MF et al. (2007), Moffatt MF et al. (2010), Ferreira MA et al. (2011), Torgerson DG et al. (2011), Ferreira MA et al. (2013), Pickrell et al, (2016), Vincente et al. (2017), Demaina et. al. (2017), Demaina et. al. (2019) Evaluating genetic correlation between biobanks ‹#›

24. Biobanks have high genetic correlation with leave-one-biobank-out meta-analysis ‹#› Endpoint: Asthma R 2 were estimated using LDSC (Bulik-Sullivan e t al., 2015) Trans-ethnic R 2 between BBJ and leave-out-out Meta-analysis was estimated using Popcorn (Brown et al., 2016)

25. + + + Biobank 1 Biobank 2 Biobank 3 Biobank N Meta-analysis ‹#› Better power for GWASs More accurate polygenic risk scores GWASs of understudied diseases Opportunity for cross-validation Improvements in fine-mapping (PgmNr 2283) Potential to explore subgroup analyses … Biobank meta-analysis has many potential benefits

26. Asthma 54 loci have not been seen in any individual biobanks ‹#› Biobank Japan  8,204 cases, 10 loci BioMe-AA  842 cases, 0 loci BioMe-EA 455 cases, 0 loci FinnGen 15,926 cases, 14 loci Generation Scotland 196 cases, 1 locus HUNT 5,587 cases, 3 loci UK Biobank 26,332 cases, 31 loci deCODE  16,767 cases, 5 loci BioMe-HA  1,459 cases, 1 locus China Kadoorie Biobank  990 cases, 0 loci Michigan Genomics Initiative 6,079 cases, 0 loci Combined 82,837 cases, 82 loci

27. Primary open-angle glaucoma 13 loci have not been seen in any individual biobanks ‹#› Biobank Japan   5,749 cases, 6 loci Combined 14,831 cases, 29 loci BioMe-AA  236 cases, 1 locus BioMe-HA 290 cases, 0 loci Estonian Biobank 2705 cases, 1 locus FinnGen 3,375 cases, 11 loci Michigan Genomics Initiative 307 cases, 2 loci UK Biobank 1,037 cases, 4 loci BioVU-EUR-mega 464 cases, 1 locus deCODE  596 cases, 1 locus BioMe-EA 72 cases, 0 loci

28. + + + Biobank 1 Biobank 2 Biobank 3 Biobank N Meta-analysis ‹#› Better power for GWASs More accurate polygenic risk scores GWASs of understudied diseases Opportunity for cross-validation Improvements in fine-mapping (PgmNr 2283) Potential to explore subgroup analyses … Biobank meta-analysis has many potential benefits

29. Meta-analysis of biobanks with similar ancestry improves the prediction accuracy by polygenic risk scores ‹#› Test Cohort (Holding out 1,000 asthma cases and 1,000 controls)

30. ‹#› Test Cohort (Holding out 1,000 asthma cases and 1,000 controls) Martin, Alicia R., et al. 2019

31. ‹#› Test Cohort (Holding out 1,000 asthma cases and 1,000 controls) Martin, Alicia R., et al. 2019 Need more biobanks with non-European participants

32. + + + Biobank 1 Biobank 2 Biobank 3 Biobank N Meta-analysis ‹#› Better power for GWASs More accurate polygenic risk scores GWASs of understudied diseases Opportunity for cross-validation Improvements in fine-mapping (PgmNr 2283) Potential to explore subgroup analyses … Biobank meta-analysis has many potential benefits

33. Thyroid cancer 8 loci have not been seen in any individual biobanks Combined 3,613 cases, 13 loci ‹#› Biobank Japan   361 cases, 0 loci FinnGen 811 cases, 3 loci HUNT 144 cases, 3 loci Michigan Genomics Initiative 787 cases, 3 loci UK Biobank 358 cases, 1 locus deCODE  1,152 cases, 5 loci

34. + + + Biobank 1 Biobank 2 Biobank 3 Biobank N Meta-analysis ‹#› Better power for GWASs More accurate polygenic risk scores GWASs of understudied diseases Opportunity for cross-validation Improvements in fine-mapping (PgmNr 2283) Potential to explore subgroup analyses … Biobank meta-analysis has many potential benefits Masahiro Kanai

35. + + + Biobank 1 Biobank 2 Biobank 3 Biobank N Meta-analysis ‹#› Better power for GWASs More accurate polygenic risk scores GWASs of understudied diseases Opportunity for cross-validation Improvements in fine-mapping (PgmNr 2283) Potential to explore subgroup analyses … Biobank meta-analysis has many potential benefits

36. China Kadoorie Biobank 100k UK Biobank 500k FinnGen 180k Biobank Japan 200k HUNT Study 70k BioVu 120k BioME 32 k UCLA Precision Health Biobank 14k Michigan Genomics Initiative 47k Partners Biobank 25k Colorado Biobank 30k Estonian Biobank 150k Generation Scotland 24k Million Veteran Program 500k deCODE Genetics 250k East London Genes & Health 10k LifeLines NL 52K *genotyped sample sizes ‹#› Netherlands twin register 24K Global Biobank Meta-analysis Initiative globalbiobankmeta.org  

37. Welcome more biobanks to join us Website: globalbiobankmeta.org  Email to [email protected] If any question, please feel free to contact any of us Mark Daly [email protected] Cristen Willer [email protected] Benjamin Neale [email protected] Juha Karjalainen [email protected] Masahiro Kanai [email protected] Mitja Kurki [email protected] Wei Zhou [email protected] ‹#›

38. Meet-up at ASHG 2019 Room: Marriott Marquis - Galveston Room 2nd level Date: Thursday 10/17/2019 Time: 3pm local time ‹#›

39. Acknowledgements ‹#› Yukinori Okada Masahiro Kanai Ruth Loos Judy Cho Eimear Kenny Michael Preuss Nancy Cox Jibril Hirbo Robin Walters Kuang Lin Kathleen Barnes Kári Stefánsson Unnur Þorsteinsdóttir Caroline Hayward Marioni Riccardo Aarno Palotie Mark Daly Samuli Ripatti Juha Karjalainen Kristian Hveem Cristen Willer Sarah Graham Ben Brumpton  Serena Sanna Esteban Lopera Sebastian Zoellne Michael Boehnke Lars Fritsche Christopher J. O'Donnell Jordan Smoller Chia-Yen Chen Bogdan Pasaniuc David A van Heel Thank participants in all biobanks! Andres Metspalu Tõnu Esko  Priit Palta DI Boomsma MG Nivard