Show simple item record

Machine-Learning-Based Interpretation of Rare Disease Variants Leveraging Genomics and Computational Structural Biology

dc.contributor.advisorCapra, John Anthony
dc.contributor.advisorMeiler, Jens
dc.creatorMukherjee, Souhrid
dc.date.accessioned2022-07-12T16:47:39Z
dc.date.available2022-07-12T16:47:39Z
dc.date.created2022-06
dc.date.issued2022-06-20
dc.date.submittedJune 2022
dc.identifier.urihttp://hdl.handle.net/1803/17539
dc.description.abstractRare genetic diseases affect more than 300 million people around the world; however, the causative genes and variants have not been identified for most. The Undiagnosed Diseases Network (UDN) was established to help elucidate the molecular mechanisms underlying rare and undiagnosed diseases. Although this approach has yielded much success, more than half of all UDN cases remain undiagnosed. During my doctoral dissertation I have collaborated extensively with the UDN to help resolve rare disease mechanisms by developing computational tools and techniques for rare variant interpretation. My hypothesis was that in some of the unresolved UDN cases, the phenotypes resulted from effects of rare variants in more than one gene, and I have designed a machine learning classifier (DiGePred) that accurately predicts digenic disease-related gene pairs in individuals suffering from rare diseases. Several novel candidate digenic gene pairs have been identified. Subsequently, I devised a “Personalized Structural Biology” approach to predict the molecular mechanisms underlying diseases caused by individual rare missense variants. I analyzed a couple of proximal rare de novo variants (V469L, V471L) in the potassium ion channel KCNC2 (Kv3.2), suspected to cause developmental epileptic encephalopathy (DEE)-like symptoms in two patients. The two proximal variants were found to have drastically different molecular effects, with V471L being a gain-of-function, while the V469L was a loss-of-function, and I was able to predict heterogenous mechanisms associated with the two variants, using computational structural biology and molecular dynamics (MD) simulations. My goal for my work during my PhD was to study rare diseases and variants, and develop computational methods that facilitate discovery of rare disease molecular mechanisms. This would lead to an improved development of various intervention strategies for individuals suffering from rare diseases.
dc.format.mimetypeapplication/pdf
dc.language.isoen
dc.subjectcomputational biology, machine learning
dc.titleMachine-Learning-Based Interpretation of Rare Disease Variants Leveraging Genomics and Computational Structural Biology
dc.typeThesis
dc.date.updated2022-07-12T16:47:39Z
dc.contributor.committeeMemberHamid, Rizwan
dc.contributor.committeeMemberCreanza, Nicole
dc.contributor.committeeMemberLopez, Carlos
dc.type.materialtext
thesis.degree.namePhD
thesis.degree.levelDoctoral
thesis.degree.disciplineBiological Sciences
thesis.degree.grantorVanderbilt University Graduate School
dc.creator.orcid0000-0002-8355-3000
dc.contributor.committeeChairRokas, Antonis


Files in this item

Icon
Icon
Icon
Icon
Icon
Icon
Icon

This item appears in the following Collection(s)

Show simple item record