Excited to share our recent work on cross-biobank and pan-ancestry genetic analyses of rare variants for human disease 🧬 published last week in @NatureGenet @xxinxwang @patrick_ellinor Link:

More and more case studies revealing lifelong effects of genetic variation mimicking pharmacological interventions. A new example for a Finnish-enriched missense variant in SCN5A that proxies cardiac sodium channel blockade👇

Two new papers in @NatureGenet from @patrick_ellinor and team Expanding genetic discovery of rare, structural and common variation in atrial fibrillation Sequencing >50K cases https://t.co/MjPcyQxcQY GWAS >180K cases https://t.co/u8ABesAwZ0

Atrial fibrillation is far more heritable than generally perceived. Two new reports @NatureGenet identify many more linked genomic variants and a refined polygenic risk score @patrick_ellinor https://t.co/u3MP4BKCNR https://t.co/07dtEIbJm3

Now on medRxiv, we explore the genetic similarity between dilated cardiomyopathy (DCM) and peripartum, alcohol-induced, and cancer therapy-related cardiomyopathies (PPCM, ALCM, CCM). Led by star trio @DimitriMaamari, @kiranbiddinger, and @JurgensSean. https://t.co/HP798ztfU8

In a new study, @MGHHeartHealth researchers and collaborators identified genetic variants associated with common cardiac rhythm disorders. The findings are published in @NatureGenet. Read more: https://t.co/gyUoGechRw https://t.co/yiLgqmKKeZ @luchenweng @joeltramo @jurgenssean

Is slow heart rate or conduction block heritable? Check out our new meta-analysis @NatureGenet incl common (~1.3m individuals) + rare (~400k) variants for bradyarrhythmias and conduction dz. Co-led w/ @LuChenWeng @joeltramo @JurgensSean @broadinstitute ! https://t.co/602MjN6iYk

Also, please also see amazing concurrent work from @SeanLZheng @AH_AlbertHenry @drjamesware @tomlumbers https://t.co/f26Kuz1hEk

To conclude, we hope that our work provides insights into the genetic architecture and mechanisms underlying DCM and myocardial resilience ♥️ @amsterdamumc @broadinstitute @Hartstichting @mghcvrc @FinnGen_FI @DeptVetAffairs @uk_biobank

Our polygenic scores were also associated with broad systolic HF in @AllofUsResearch and remained associated with systolic HF in various clinical settings: ▶️ after hypertension diagnosis ▶️ after AF diagnosis ▶️ after myocardial infarction ▶️ in carriers of rare DCM variants

In our clinical DCM dataset from @amsterdamumc we found that PRS values were higher in DCM patients without rare disease-causing variants 🧬 indicating a higher polygenic burden in genotype-negative patients.

Finally, we created polygenic scores from our GWAS data. Our MTAG scores showed strong prediction in three European datasets: ▶️ OR per SD: 1.75-1.95 ▶️ Liability-scale R^2: 8-11% and also transferred to African ancestry in @AllofUsResearch

We went on to perform Mendelian randomization, to find causal risk factors for DCM through genetics 🧬 We identified several potential causes of DCM, including AF, bodyweight, height and SBP with bodyweight and SBP remaining as independent and robust causal risk factors ♥️➡️❤️‍🔥

We scrutinized our prioritized genes further, by focussing on expression within the single-cell data. Many prioritized genes showed strong expression in cardiomyocytes and several showed significant ⬆️ or ⬇️ regulation in DCM cardiomyocytes eg CAMK2D, HSPB7, MTSS1, CRIM1

We then intersected our GWAS with multiple single-cell datasets, to identify causal cell types. Only cardiomyocyte-specific genes were strongly and robustly enriched for DCM heritability pointing to cardiomyocyte dysfunction as a primary disease driver ♥️➡️❤️‍🔥

So, which genes drive DCM? 🧬 We incorporated several lines of evidence to nominate 63 potentially causal genes from our GWAS data 🔍 including: ▶️ Mendelian DCM genes (eg TTN, FHOD3, ACTN2, PLN, ALPK3) ▶️ heatshock proteins (HSBP7, HSPB8, HSPA4, DNAJC18, FBXO32) ▶️ Many more

We replicated our loci in a set of independent samples from @HERMESgenetics, @AllofUsResearch and Million Veteran Program @DeptVetAffairs totalling 13,258 cases of DCM/NICM 👯🌎 We found broad replication of our loci ✅ with replication rates close to the expectation ✅

Our MTAG identified 65 genome-wide significant loci (50 novel) bringing the total number of genome-wide significant DCM loci to 70 (55 novel)

We then leveraged GWAS data for cardiac MRI traits from @uk_biobank. We found that global circumferential strain (r=0.75) and LVESVi (r=0.7) were strongly genetically related to DCM and we used these correlations to boost our discovery power in a multi-trait framework (MTAG)

We therefore used a strict case definition in our final GWAS which included 9,365 DCM cases and 946,368 controls👯🌎 and uncovered 38 significant loci of which 27 (!) were novel for DCM 🔍

We investigated both a strict 'nonischemic DCM' and a broader 'nonischemic cardiomyopathy' phenotype. Despite yielding less than half the number of cases, the strict nonischemic DCM phenotype yielded: ▶️ more significant GWAS loci ▶️ almost double the SNP-heritability