Improving and evaluating deep learning models of cellular organization.

Journal: Bioinformatics (Oxford, England)
Published Date:

Abstract

MOTIVATION: Cells contain dozens of major organelles and thousands of other structures, many of which vary extensively in their number, size, shape and spatial distribution. This complexity and variation dramatically complicates the use of both traditional and deep learning methods to build accurate models of cell organization. Most cellular organelles are distinct objects with defined boundaries that do not overlap, while the pixel resolution of most imaging methods is n sufficient to resolve these boundaries. Thus while cell organization is conceptually object-based, most current methods are pixel-based. Using extensive image collections in which particular organelles were fluorescently labeled, deep learning methods can be used to build conditional autoencoder models for particular organelles. A major advance occurred with the use of a U-net approach to make multiple models all conditional upon a common reference, unlabeled image, allowing the relationships between different organelles to be at least partially inferred.

Authors

  • Huangqingbo Sun
    Computational Biology Department, Carnegie Mellon University, Pittsburgh, PA 15213, USA.
  • Xuecong Fu
    Department of Biological Sciences, Carnegie Mellon University, Pittsburgh, PA 15213, USA.
  • Serena Abraham
    Computational Biology Department, Carnegie Mellon University, Pittsburgh, PA 15213, USA.
  • Shen Jin
    Computational Biology Department, Carnegie Mellon University, Pittsburgh, PA 15213, USA.
  • Robert F Murphy
    Computational Biology Department, Center for Bioimage Informatics, and Departments of Biological Sciences, Biomedical Engineering and Machine Learning, Carnegie Mellon University, 5000 Forbes Avenue, Pittsburgh, PA, USA; Freiburg Institute for Advanced Studies and Faculty of Biology, Albert Ludwig University of Freiburg, Germany. Electronic address: murphy@cmu.edu.