AUTHOR=Zhang Yu 

TITLE=Epigenetic Combinatorial Patterns Predict Disease Variants

JOURNAL=Frontiers in Genetics

VOLUME=Volume 8 - 2017

YEAR=2017

URL=https://www.frontiersin.org/journals/genetics/articles/10.3389/fgene.2017.00071

DOI=10.3389/fgene.2017.00071

ISSN=1664-8021

ABSTRACT=Most genetic variants identified in genome-wide association studies are non-coding, and they are likely tagging nearby causal variants. Pinpointing the precise locations of disease causal variants and understanding their functions in disease are challenging. A promising approach is to integrate functional data currently available in hundreds of human tissues and cell types to improve fine mapping. Although several methods have used functional data to prioritize disease variants, they mainly used linear models or equivalent naïve likelihood-based models for prediction. Here, we investigated whether the combinatorial patterns of functional data across cell types can improve the prediction accuracy of disease variants. Using functional annotation in 127 human cell types, we first introduce a Bayesian method to identify the recurring cell type specificity partitions in the genome-scale. We show that our de novo identification of the epigenome partition patterns agreed well with known cell type origins and were strongly enriched in disease variants. Using the cell type specificity patterns in addition to enrichment of functional elements, we further demonstrate that the power for predicting disease variants can be greatly improved over linear models. Our approach thus provides a new way to prioritize disease functional variants for testing.