Explaining decisions of graph convolutional neural networks: patient-specific molecular subnetworks responsible for metastasis prediction in breast cancer.

Journal: Genome medicine
Published Date:

Abstract

BACKGROUND: Contemporary deep learning approaches show cutting-edge performance in a variety of complex prediction tasks. Nonetheless, the application of deep learning in healthcare remains limited since deep learning methods are often considered as non-interpretable black-box models. However, the machine learning community made recent elaborations on interpretability methods explaining data point-specific decisions of deep learning techniques. We believe that such explanations can assist the need in personalized precision medicine decisions via explaining patient-specific predictions.

Authors

  • Hryhorii Chereda
    Medical Bioinformatics, University Medical Center Göttingen, Göttingen, Germany.
  • Annalen Bleckmann
    Dept. of Medicine A (Hematology, Oncology, Hemostaseology and Pulmonology), University Hospital Münster, Münster, Germany.
  • Kerstin Menck
    Dept. of Medicine A (Hematology, Oncology, Hemostaseology and Pulmonology), University Hospital Münster, Münster, Germany.
  • Júlia Perera-Bel
    Hospital del Mar Medical Research Institute (IMIM), Barcelona, Spain.
  • Philip Stegmaier
    geneXplain GmbH, Wolfenbüttel, Germany.
  • Florian Auer
    IT Infrastructure for Translational Medical Research, University of Augsburg, Augsburg, Germany.
  • Frank Kramer
    IT-Infrastructure for Translational Medical Research, Faculty of Applied Computer Science, Faculty of Medicine, University of Augsburg, Augsburg, Germany.
  • Andreas Leha
    Department of Medical Statistics, University Medical Center Göttingen, Humboldtallee 32, 37073 Göttingen, Germany.
  • Tim Beissbarth
    3 Department of Medical Statistics, University Medical Center Goettingen, Goettingen, Germany.