Histone-Net: a multi-paradigm computational framework for histone occupancy and modification prediction

Abstract Deep exploration of histone occupancy and covalent post-translational modifications (e.g., acetylation, methylation) is essential to decode gene expression regulation, chromosome packaging, DNA damage, transcriptional activation. Existing computational approaches are unable precisely predict mainly due the use sub-optimal statistical representation sequences. For establishment an improved modification landscape for multiple markers, paper in hand presents end-to-end multi-paradigm framework “Histone-Net”. To learn local global residue context aware sequence representation, Histone-Net generates unsupervised higher order embeddings (DNA2Vec) a different application language modelling, where it encapsulates information while generating (SuperDNA2Vec) supervised manner. We perform intrinsic extrinsic evaluation both presented distributed learning schemes. A comprehensive empirical over ten benchmark markers data sets three analysis tasks indicates that SuperDNA2Vec softmax classifier-based approach outperforms state-of-the-art by average accuracy 7%. eliminate overhead training separate binary classifiers all evaluated multi-label classification paradigm, produces decent performance simultaneous prediction occupancy, methylation.