Background: Precision medicine approaches require genotype-phenotype associations that have translational utility and hence can impact disease management and outcomes. To date approximately one quarter of patients with pulmonary arterial hypertension harbour rare mutations in disease-causing genes. We hypothesised that integrating deep phenotyping data with whole genome sequencing data will reveal additional rare disease associated variants with smaller effect sizes.
Methods: We analysed whole genome sequencing data from 13,037 participants enrolled in the NHIR Bioresource-Rare Diseases study, of which 1148 were recruited to the PAH domain. In order to test for genetic associations between genes and selected phenotypes of pulmonary hypertension (PH) we deployed the Bayesian statistics based rare variant association test BeviMed. By assigning phenotypic tags based on the current diagnostic classification of PH, stratification by age at diagnosis and transfer coefficient (KCO) we defined the groups for comparison.
Results: A strong association was identified between protein truncating variants (PTV) in KDR and significantly reduced KCO (PP=0.985). KDR PTV carriers also presented significantly later in life (PP=0.889). None of the patients harbouring PTV in KDR(n=4) had significant parenchymal lung changes that could explain the reduced KCO. We identified four additional individuals with PTVs in KDR in the US PAH Biobank. We also confirmed associations between previously reported genes and PAH and corroborated association between preserved KCO and BMPR2 mutations as well as the association between significantly decreased KCO and biallelic EIF2AK4 variants.
Conclusions: We have found that PTVs in KDR, the gene encoding VEGFR2, are associated with significantly reduced KCO and delayed onset of disease suggestive of a critical role of VEGF signalling in the pathogenesis of PAH. With this we also demonstrate that deep clinical phenotyping can advance the identification of novel pathogenic variants and disease subtypes of smaller effect sizes and/or lower frequency with prognostic and therapeutic implications.