Improving the identification of regulatory noncoding SNPs with genomic signals and CNN
Identification of causal noncoding single nucleotide polymorphisms (SNPs) is vital for maximizing the knowledge dividend from human genome-wide association studies (GWAS). Recently, diverse supervised learning methods have been used for functional SNP identification; however, only a few deep-learning-based methods take advantage of local patterns of sequential data around SNPs instead of only using features related to the specific locations of SNPs. Moreover, these deep-learning-based methods all rely on the flanking DNA sequence and did not take advantage of the rich resource of genomic signals (phylogenetic, epigenomic, chromatin structural, etc.) that could provide correlates of regulatory function. Here I propose a new CNN-based learning method that leverages genomic signals flanking each SNP to improve the prediction accuracy of regulatory SNPs (rSNPs) in the context of post-GWAS analysis.
Major Advisor: Stephen Ramsey
Committee: David Hendrix
Committee: Amir Nayyeri
Committee: Xiaoli Fern
Committee: Glencora Borradaile
GCR: Harold Bae
Monday, November 18, 2019 at 9:00am to 11:00am
Kelley Engineering Center, 1007
110 SW Park Terrace, Corvallis, OR 97331
Calvin Hughes
No recent activity