Phylogenetic convolutional neural networks in metagenomics.

Journal: BMC bioinformatics
Published Date:

Abstract

BACKGROUND: Convolutional Neural Networks can be effectively used only when data are endowed with an intrinsic concept of neighbourhood in the input space, as is the case of pixels in images. We introduce here Ph-CNN, a novel deep learning architecture for the classification of metagenomics data based on the Convolutional Neural Networks, with the patristic distance defined on the phylogenetic tree being used as the proximity measure. The patristic distance between variables is used together with a sparsified version of MultiDimensional Scaling to embed the phylogenetic tree in a Euclidean space.

Authors

  • Diego Fioravanti
    Fondazione Bruno Kessler (FBK), Via Sommarive 18 Povo, Trento, I-38123, Italy.
  • Ylenia Giarratano
    Centre for Medical Informatics, Usher Institute, University of Edinburgh, 9 Little France Road, Edinburgh, EH16 4UX, UK.
  • Valerio Maggio
    Fondazione Bruno Kessler, Trento, Italy.
  • Claudio Agostinelli
    Department of Mathematics, University of Trento, Via Sommarive 14 Povo, Trento, I-38123, Italy.
  • Marco Chierici
    a Fondazione Bruno Kessler , Trento , Italy.
  • Giuseppe Jurman
    a Fondazione Bruno Kessler , Trento , Italy.
  • Cesare Furlanello
    a Fondazione Bruno Kessler , Trento , Italy.