Background:
Technological advances have greatly reduced the cost of whole genome sequencing. For single individuals clinical application is apparent, while exome sequencing in tens of thousands of people has allowed a more global view of genetic variation that can inform interpretation of specific variants in individuals. We hypothesized that genome sequencing of patients with monogenic cardiomyopathy would facilitate discovery of genetic modifiers of phenotype.
Methods and Results:
We identified 48 individuals diagnosed with cardiomyopathy and with putative mutations in MYH7, the gene encoding beta myosin heavy chain. We carried out whole genome sequencing and applied a newly developed analytical pipeline optimized for discovery of genes modifying severity of clinical presentation and outcomes. Using a combination of external priors and rare variant burden tests we scored genes as potential modifiers. There were 96 genes that reached a modifier score of 6 out of 12 or better (9=2, 8=8, 7=17, 6=69). We identified NCKAP1, a gene that regulates actin filament dynamics, and CAMSAP1, a calmodulin regulate gene that regulates microtubule dynamics, as top scoring modifiers of hypertrophic cardiomyopathy phenotypes (score=9) while LDB2, RYR2, FBN1 and ATP1A2 had modifier scores of 8. Of the top scoring genes, 21 out of 96 were identified as candidates a priori. Our candidate prioritization scheme identified the previously described modifiers of cardiomyopathy phenotype, FHOD3 and MYBPC3, as top scoring genes. We identified structural variants in 21 clinically sequenced cardiomyopathy associated genes, 13 of which were at less than 10% frequency. Copy number variants in ILK and CSRP3 were nominally associated with ejection fraction (p=0.03), while 8 genes showed copy gains (GLA, FKTN, SGCD, TTN, SOS1, ANKRD1, VCL and NEBL). Structural variants were found in CSRP3, MYL3 and TNNC1, all of which have been implicated as causative for HCM.
Conclusion:
Evaluation of the whole genome sequence, even in the case of putatively monogenic disease, leads to important diagnostic and scientific insights not revealed by panel-based sequencing.
Text-mined phenotype annotation and vector-based similarity to improve identification of similar phenotypes and causative genes in monogenic disease patients
