Biologically relevant transfer learning improves transcription factor binding prediction.

Journal: Genome biology
Published Date: Sep 27, 2021

Abstract

BACKGROUND: Deep learning has proven to be a powerful technique for transcription factor (TF) binding prediction but requires large training datasets. Transfer learning can reduce the amount of data required for deep learning, while improving overall model performance, compared to training a separate model for each new task.

Authors

  • Gherman Novakovsky
    Centre for Molecular Medicine and Therapeutics, BC Children's Hospital Research Institute, Vancouver, BC, V5Z 4H4, Canada.
  • Manu Saraswat
    Institut de Biologie Computationnelle, Montpellier, France.
  • Oriol Fornes
    Centre for Molecular Medicine and Therapeutics, BC Children's Hospital Research Institute, Vancouver, BC, V5Z 4H4, Canada. oriol@cmmt.ubc.ca.
  • Sara Mostafavi
    Department of Statistics, University of British Columbia, Vancouver, BC V6T 1Z4, Canada; cb@hms.harvard.edu saram@stat.ubc.ca.
  • Wyeth W Wasserman
    Centre for Molecular Medicine and Therapeutics, Child and Family Research Institute, Department of Medical Genetics, University of British Columbia Vancouver, British Columbia V5Z 4H4, Canada. Electronic address: wyeth@cmmt.ubc.ca.

Keywords

External Resources

Popular Topics
Recent Journals