Predictive biophysical neural network modeling of a compendium of in vivo transcription factor DNA binding profiles for Escherichia coli
Lally, Patrick (Cummington Mall. Department of Biomedical Engineering, Boston University)
Gómez-Romero, Laura (Ciudad de México. Escuela de Medicina y Ciencias de la Salud, Tecnológico de Monterrey)
Tierrafría, Víctor H. (Cuernavaca. Centro de Ciencias Genómicas, Universidad Nacional Autónoma de México, Avenida Universidad s/n)
Aquino, Patricia (Cummington Mall. Department of Biomedical Engineering, Boston University)
Rioualen, Claire (Cuernavaca. Centro de Ciencias Genómicas, Universidad Nacional Autónoma de México, Avenida Universidad s/n)
Zhang, Xiaoman (Cummington Mall. Department of Biomedical Engineering, Boston University)
Kim, Sunyoung (Regina. Department of Biochemistry, University of Regina)
Baniulyte, Gabriele (New York State Department of Health. Wadsworth Center)
Plitnick, Jonathan (New York State Department of Health. Wadsworth Center)
Smith, Carol (New York State Department of Health. Wadsworth Center)
Babu, Mohan (Regina. Department of Biochemistry, University of Regina)
Collado-Vides, Julio (Universitat Pompeu Fabra (UPF). Centre for Genomic Regulation (CRG), The Barcelona Institute of Science and Technology, Dr. Aiguader 88, Barcelona 08003)
Wade, Joseph T. (SUNY. Department of Biomedical Sciences, University at Albany)
Galagan, James E. (Cummington Mall. Bioinformatics Program, Boston University)

Data:	2025
Resum:	The DNA binding of most Escherichia coli Transcription Factors (TFs) has not been comprehensively mapped, and few have models that can quantitatively predict binding affinity. We report the global mapping of in vivo DNA binding for 139 E. coli TFs using ChIP-Seq. We use these data to train BoltzNet, a novel neural network that predicts TF binding energy from DNA sequence. BoltzNet mirrors a quantitative biophysical model and provides directly interpretable predictions genome-wide at nucleotide resolution. We use BoltzNet to quantitatively design novel binding sites, which we validate with biophysical experiments on purified protein. We generate models for 124 TFs that provide insight into global features of TF binding, including clustering of sites, the role of accessory bases, the relevance of weak sites, and the background affinity of the genome. Our paper provides new paradigms for studying TF-DNA binding and for the development of biophysically motivated neural networks. The authors describe BoltzNet, a neural network that learns the energy of transcription factor (TF)-DNA binding from genomic data and can be used to design new binding sites. They report the in vivo mapping and BoltzNet modeling of 139 E. coli TFs.
Ajuts:	Consejo Nacional de Ciencia y Tecnología (CONCYT) 929687 Universidad Nacional Autónoma de México (National Autonomous University of Mexico) Gouvernement du Canada | Natural Sciences and Engineering Research Council of Canada (Conseil de Recherches en Sciences Naturelles et en Génie du Canada) DG-20234
Drets:	Aquesta url de drets no existeix a la base de dades.
Llengua:	Anglès
Document:	Article ; recerca ; Versió publicada
Matèria:	Machine learning ; Gene regulation ; Thermodynamics
Publicat a:	Nature communications, Vol. 16 (May 2025), art. 4255, ISSN 2041-1723

DOI: 10.1038/s41467-025-58862-8
PMID: 40335485

El registre apareix a les col·leccions:
Articles > Articles de recerca
Articles > Articles publicats