Prediction of Cell Type Specific Transcription Factor Binding Site Occupancy

Abstract

We propose a. machine learning approach to predict the particular cell type where a given transcription factor can bind a DNA sequence. The learning models are trained on the DNA sequences provided from the publicly available ChIPseq experiments of the ENCODE project for 52 transcription factors across the GM12878, K562, HeLa, H1-hESC and HepG2 cell lines. Three different feature extraction methods are used based on k-mer representations, counts of known motifs, and a new model called the skip gram model, which has become very popular in the analysis of text. The logistic regression with e1 penalty is used for the classification task. We find that predictors based on known motifs counts detect cell-type specific signatures better than a previously published method, with mean AUC improvement of 0.18 and can be used to identify t he interaction of transcription factors. Remarkably, the skip gram approach, which can be used without of any prior knowledge about transcription factor binding sit es, performs almost as well as the motif-based method. Overall, our family of predictors will be useful to both better predict cell-type specific TF occupancy and understand the mechanisms underlying this phenomenon.