Multitask matrix completion for learning protein interactions across diseases

Meghana Kshirsagar, Jaime G. Carbonell, Judith Klein-Seetharaman, Keerthiram Murugesan

Research output: Chapter in Book/Report/Conference proceedingConference contribution

3 Scopus citations

Abstract

Disease causing pathogens such as viruses, introduce their proteins into the host cells where they interact with the host’s proteins enabling the virus to replicate inside the host. These interactions between pathogen and host proteins are key to understanding infectious diseases. Often multiple diseases involve phylogenetically related or biologically similar pathogens. Here we present a multitask learning method to jointly model interactions between human proteins and three different, but related viruses: Hepatitis C, Ebola virus and Influenza A. Our multitask matrix completion based model uses a shared low-rank structure in addition to a task-specific sparse structure to incorporate the various interactions. We obtain upto a 39% improvement in predictive performance over prior state-of-the-art models.We show how our model’s parameters can be interpreted to reveal both general and specific interactionrelevant characteristics of the viruses. Our code, data and supplement is available at: http://www.cs.cmu.edu/∼mkshirsa/bsl mtl.

Original languageEnglish (US)
Title of host publicationResearch in Computational Molecular Biology - 20th Annual Conference, RECOMB 2016, Proceedings
EditorsMona Singh
PublisherSpringer Verlag
Pages53-64
Number of pages12
ISBN (Print)9783319319568
DOIs
StatePublished - 2016
Externally publishedYes
Event20th Annual Conference on Research in Computational Molecular Biology, RECOMB 2016 - Santa Monica, United States
Duration: Apr 17 2016Apr 21 2016

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume9649
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference20th Annual Conference on Research in Computational Molecular Biology, RECOMB 2016
Country/TerritoryUnited States
CitySanta Monica
Period4/17/164/21/16

Keywords

  • Host-pathogen protein interactions
  • Matrix completion
  • Multitask learning

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Fingerprint

Dive into the research topics of 'Multitask matrix completion for learning protein interactions across diseases'. Together they form a unique fingerprint.

Cite this