Prediction of Deleterious Point Mutations of Disease-Related Proteins: Case Study of Hemophilia and Coagulation Factors

Friday, November 08, 2013 — Poster Session IV
2:00 p.m. – 4:00 p.m.	FAES Academic Center (Upper-Level Terrace)	FDA/CBER	COMPBIO-5

Authors

N. Hamasaki-Katagiri
R. Salari
A. Wu
Y. Qi
T. Schiller
A.C. Filiberto
E.F. Schisterman
A.A. Komar
T.M. Przytycka
C. Kimchi-Sarfaty

Abstract

Accurately predicting genotype-phenotype correlations is a fundamental goal of medical genetics. As an approach to develop more accurate in silico tools for prediction of disease-causing mutations of structural proteins, a gene- and disease-specific prediction tool was developed based on a large systematic analysis of the missense mutations from hemophilia A patients. This newly developed hemophilia A prediction tool showed disease-prediction performance comparable to other publicly available prediction software. The use of computational metrics for nucleotide sequences along with multiple amino acid sequence alignment classification improved overall predictive performance. Furthermore, this addition also made the prediction of the impact of synonymous mutations possible, a novel feature unavailable in current prediction software. Given the role of synonymous mutations in disease and drug codon optimization, we propose that utilizing a gene- and disease-specific approach can be highly useful for incorporating information at the nucleotide level to make functional predictions possible even for synonymous mutations.