Next Article in Journal
Facile Synthesis of 3-(Azol-1-yl)-1-adamantanecarboxylic Acids—New Bifunctional Angle-Shaped Building Blocks for Coordination Polymers
Previous Article in Journal
Enhanced Nanoencapsulation of Sepiapterin within PEG-PCL Nanoparticles by Complexation with Triacetyl-Beta Cyclodextrin
Previous Article in Special Issue
High Impact: The Role of Promiscuous Binding Sites in Polypharmacology
Open AccessArticle

Applicability Domain of Active Learning in Chemical Probe Identification: Convergence in Learning from Non-Specific Compounds and Decision Rule Clarification

1
Kyoto University Graduate School of Medicine, Department of Molecular Biosciences, Life Science Informatics Research Unit, Kyoto, Sakyo, Yoshida, Konoemachi, Kyoto 606-8501, Japan
2
Kyoto University Graduate School of Medicine, Department of Radiation Genetics; Kyoto, Sakyo, Yoshida, Konoemachi, Kyoto 606-8501, Japan
*
Author to whom correspondence should be addressed.
Molecules 2019, 24(15), 2716; https://doi.org/10.3390/molecules24152716
Received: 30 June 2019 / Revised: 19 July 2019 / Accepted: 24 July 2019 / Published: 26 July 2019
(This article belongs to the Special Issue Computational Chemical Biology)
Efficient identification of chemical probes for the manipulation and understanding of biological systems demands specificity for target proteins. Computational means to optimize candidate compound selection for experimental selectivity evaluation are being sought. The active learning virtual screening method has demonstrated the ability to efficiently converge on predictive models with reduced datasets, though its applicability domain to probe identification has yet to be determined. In this article, we challenge active learning’s ability to predict inhibitory bioactivity profiles of selective compounds when learning from chemogenomic features found in non-selective ligand-target pairs. Comparison of controls versus multiple molecule representations de-convolutes factors contributing to predictive capability. Experiments using the matrix metalloproteinase family demonstrate maximum probe bioactivity prediction achieved from only approximately 20% of non-probe bioactivity; this data volume is consistent with prior chemogenomic active learning studies despite the increased difficulty from chemical biology experimental settings used here. Feature weight analyses are combined with a custom visualization to unambiguously detail how active learning arrives at classification decisions, yielding clarified expectations for chemogenomic modeling. The results influence tactical decisions for computational probe design and discovery. View Full-Text
Keywords: chemical probes; compound specificity; ligand-target interactions; chemogenomics; active learning; active projection; decision tree; molecular representation chemical probes; compound specificity; ligand-target interactions; chemogenomics; active learning; active projection; decision tree; molecular representation
Show Figures

Figure 1

MDPI and ACS Style

Polash, A.H.; Nakano, T.; Takeda, S.; Brown, J. Applicability Domain of Active Learning in Chemical Probe Identification: Convergence in Learning from Non-Specific Compounds and Decision Rule Clarification. Molecules 2019, 24, 2716.

Show more citation formats Show less citations formats
Note that from the first issue of 2016, MDPI journals use article numbers instead of page numbers. See further details here.

Article Access Map by Country/Region

1
Back to TopTop