Multivariate mutual information based feature selection for predicting histone post-translational modifications in epigenetic datasets

Mutual information (MI) has been traditionally employed in many areas including biology to identify the non-linear relationships between features. This technique is particularly useful in the biological context to identify features such as genes, histone post-translational modifications (PTMs), transcriptional factors, etc. In this work, instead of evaluating the more conventional pairwise MI between PTM features, we compute the multivariate mutual information (MMI) between PTM triplets on some publicly available ChIP-seq datasets. This enables us to form a small subset of PTMs that serve as principal features for predicting the intensity values of any other histone PTM across the epigenome. This prediction step employs machine learning using the chosen MMI-based features. The selected principal features based on MMI are global and not specific to the target being predicted. The predictive performance of this technique is demonstrated on the ChIP-seq datasets from the yeast and the human epigenomes and compared with other feature selection techniques.