Active learning for computational chemogenomics.

Journal Article (Journal Article)


Computational chemogenomics models the compound-protein interaction space, typically for drug discovery, where existing methods predominantly either incorporate increasing numbers of bioactivity samples or focus on specific subfamilies of proteins and ligands. As an alternative to modeling entire large datasets at once, active learning adaptively incorporates a minimum of informative examples for modeling, yielding compact but high quality models. Results/methodology: We assessed active learning for protein/target family-wide chemogenomic modeling by replicate experiment. Results demonstrate that small yet highly predictive models can be extracted from only 10-25% of large bioactivity datasets, irrespective of molecule descriptors used.


Chemogenomic active learning identifies small subsets of ligand-target interactions in a large screening database that lead to knowledge discovery and highly predictive models.

Full Text

Duke Authors

Cited Authors

  • Reker, D; Schneider, P; Schneider, G; Brown, JB

Published Date

  • March 2017

Published In

Volume / Issue

  • 9 / 4

Start / End Page

  • 381 - 402

PubMed ID

  • 28263088

Electronic International Standard Serial Number (EISSN)

  • 1756-8927

International Standard Serial Number (ISSN)

  • 1756-8919

Digital Object Identifier (DOI)

  • 10.4155/fmc-2016-0197


  • eng