Functional enrichment methodology for analyzing omics data to study the aetiology of rare diseases

Abstract

Rare diseases (RDs) are a large and diverse group of disorders and defined by low prevalence, in other words, it is any disease that affects a small percentage of the population. According to OMIM and Orphanet, ~7000 different RDs have been estimated, but the number of phenotypes that remain to be defined could be considerably higher. The difficulty in obtaining the correct diagnosis is the most dramatic problem to be solved for the patients, about 30% still lack a diagnostic definition. The patients living with rare diseases visit an average of 7.3 physicians before receiving an accurate diagnosis and the mean length of time from symptom onset to accurate diagnosis is 4.8 years. Late diagnoses delay specific treatments and may have severe and life-threatening consequences. Molecular diagnosis is the most prominent way to facilitate earlier and accurate diagnosis, and hence an effective treatment for rare undiagnosed cases. In this dissertation project, a novel bioinformatics workflow is constructed for whole-exome/genome sequencing data analysis, variant prioritization and pathogenicity prediction from a cascade of different tools shading light into different aspects of the diagnostic process. The pathogenicity mechanisms of mutations are elucidated via molecular dynamics (MD) simulations. The newly developed pipeline is planned to be used for diagnosis of undiagnosed patients with a suspected genetic disorder, where other testing modalities have been inconclusive or noninformative. The workflow was tested on several undiagnosed clinical cases with their family members and achieved high success rates by identifying the causative variant. For two of these families, the pathogenicity mechanisms of mutations were described via MD simulations, and these findings have been submitted to two different SCI journals and passed the editorial approval. The diagnosis of one of these families was Periventricular Nodular Heterotopia, while the other was Nail Dysplasia-10. Both of the diseases are extremely rare that is seen in one in a million cases. In conclusion, we developed a unique workflow for molecular diagnosis of rare undiagnosed diseases. Our pipeline contributes to the already existing knowledge through the combination of population frequency, pathogenicity prediction tools, gene intolerance scores, and MD simulations for the first time.