An atlas of non-redundant sequences and structures of transcription factor assemblies across domains of life.

Abstract

Transcription factors (TFs) regulate gene expression by controlling the recruitment of transcriptional machinery to regulatory regions of the genome. Nearly 10% of the human genome encodes TFs, making them one of the largest protein families. Despite their central roles in gene regulation, TFs are historically considered challenging therapeutic targets due to their complex interactions with DNA, RNA and associated proteins. Although recent progress in studying TFs both at molecular and structural level excels our understanding on their function, yet a universal rule decoding their recognition process remains elusive. Here, we present a curated non-redundant dataset of TFs with 3570 sequences and 377 structures. We further characterize "unique interfaces" by quantifying interface identity across interacting chains in TF assemblies. Surprisingly, our data shows that the "unique interfaces" have optimal size ranging from 2000 Å2 to 4000 Å2 irrespective of their quaternary assembly. To understand the functional diversity, we integrate sequence motifs, structural domains, subcellular localization and functional enrichment of TFs. We have also catalogued association of TFs with various human diseases. Our dataset provides a comprehensive platform to perform large scale analysis of TF-assemblies and aid in computational methods for their prediction across domains of life.