Dr. Shaun Mahony
Dr. Shaun Mahony, Assistant Professor, Biochemistry & Molecular Biology Department, Penn State University, USA, will present the lecture "Using neural networks to characterize the predeterminants of induced transcription factor binding sites."
The goal of my computational biology research is to understand how cellular identities are controlled by the dynamic regulatory activities of transcription factors (TFs). The typical vertebrate TF should have binding affinity for millions of sites along the genome, yet only a small fraction appears to be bound in a given cell type. Furthermore, a TF’s binding sites can vary broadly across cell types. While TF regulatory specificity is key to understanding the establishment of cell fates, we still know little about how such specificity results from the interplay between a TF’s sequence preference and cell-specific chromatin environments. My lab at Penn State aims to build machine learning applications that yield deeper insight into TF regulatory specificity.
To understand the determinants of TF binding specificity, we need to examine how newly activated TFs interact with sequence and preexisting chromatin landscapes to determine their binding sites. In this talk, I'll present a principled approach to model the sequence and preexisting chromatin determinants of TF binding. Specifically, we develop a bimodal neural network that jointly models sequence and prior chromatin data to interpret the binding specificity of TFs that have been induced in well-characterized chromatin environments. The bimodal network architecture allows us to quantify the degree to which sequence and prior chromatin features explain induced TF binding, both at individual sides and genome-wide. Our approach thus provides a framework for modeling, interpreting, and visualizing the joint sequence and chromatin landscapes that determine in vivo TF binding dynamics. I'll demonstrate our approach by using it to analyze the binding specificity of the two main proneural TFs in vertebrates, Ascl1 and Neurog2, when each is expressed in mouse embryonic stem cells. I'll show that the pre-existing chromatin landscape is an important determinant of these TFs' binding specificity. Furthermore, the binding patterns of the proneural TFs result in differential chromatin accessibility and activity landscapes that affect the genomic binding of shared downstream neurogenic TFs, and thereby have a profound impact on the subtypes of neurons that can be generated.
Zoom connection information
Meeting ID: 676 4965 9713
Rafel Riudavets Puig, PhD candidate at the Centre for Molecular Medicine Norway (NCMM), UiO, will present his work on UniBind: maps of high confidence direct TF-DNA interactions across nine species.