Skip to contentSkip to filters

We use cookies for analytics.

Understanding transcription factor binding determinants using neural networks

Shaun Mahony

Penn State University

Abstract

The DNA-binding activities of transcription factors (TFs) are influenced both by intrinsic sequence preferences and by extrinsic interactions with cell-specific chromatin landscapes and other regulatory proteins, and disentangling the roles of sequence and chromatin determinants of TF-DNA binding remains challenging. So-called pioneer factors, for example, can bind relatively inaccessible chromatin sites during development, yet often vary their binding targets across cell types, pointing to a combination of intrinsic and extrinsic forces guiding their binding. The Mahony lab studies how sequence and chromatin features combine to influence TF binding. This talk describes neural network-based approaches that estimate the relative contributions of intrinsic DNA sequence preference and cell-specific chromatin environments to a TF's DNA-binding activities, and demonstrates a new concept activation vector approach for interpreting the predictive features that neural networks learn about TF binding sites. It focuses on a selection of Forkhead domain (Fox) TFs that have similar in vitro DNA-binding preferences: despite the association between Forkhead domains and pioneering activity, the selected Fox TFs display a wide range of affinities for preexisting chromatin states and bind different DNA targets even when induced in identical chromatin settings.

Topics