期刊
BRIEFINGS IN BIOINFORMATICS
卷 20, 期 1, 页码 26-32出版社
OXFORD UNIV PRESS
DOI: 10.1093/bib/bbx094
关键词
complex diseases; machine learning; SNPs; regulatory feature data; missing heritability; FDSP
资金
- National Natural Science Foundation of China [31371278, 31471188, 81573241, 31511140285]
- China Postdoctoral Science Foundation [2016T90902, 2016M602797]
- Natural Science Basic Research Program Shaanxi Province [2016JQ3026]
- Fundamental Research Funds for the Central Universities
Genome-wide association studies (GWASs) are an effective strategy to identify susceptibility loci for human complex diseases. However, missing heritability is still a big problem. Most GWASs single-nucleotide polymorphisms (SNPs) are located in noncoding regions, which has been considered to be the unexplored territory of the genome. Recently, data from the Encyclopedia of DNA Elements (ENCODE) and Roadmap Epigenomics projects have shown that many GWASs SNPs in the noncoding regions fall within regulatory elements. In this study, we developed a pipeline named functional disease-associated SNPs prediction (FDSP), to identify novel susceptibility loci for complex diseases based on the interpretation of the functional features for known disease-associated variants with machine learning. We applied our pipeline to predict novel susceptibility SNPs for type 2 diabetes (T2D) and hypertension. The predicted SNPs could explain heritability beyond that explained by GWAS-associated SNPs. Functional annotation by expression quantitative trait loci analyses showed that the target genes of the predicted SNPs were significantly enriched in T2D or hypertension-related pathways in multiple tissues. Our results suggest that combining GWASs and regulatory features data could identify additional functional susceptibility SNPs for complex diseases. We hope FDSP could help to identify novel susceptibility loci for complex diseases and solve the missing heritability problem.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据