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

Molecular modeling frequently constructs classification models for the prediction of two‐class entities, such as compound bio(in)activity, chemical property (non)existence, protein (non)interaction, and so forth. The models are evaluated using well known metrics such as accuracy or true positive rates. However, these frequently used metrics applied to retrospective and/or artificially generated prediction datasets can potentially overestimate true performance in actual prospective experiments. Here, we systematically consider metric value surface generation as a consequence of data balance, and propose the computation of an inverse cumulative distribution function taken over a metric surface. The proposed distribution analysis can aid in the selection of metrics when formulating study design. In addition to theoretical analyses, a practical example in chemogenomic virtual screening highlights the care required in metric selection and interpretation.

PURPOSE: Homologous recombination DNA repair deficiency (HRD) is associated with sensitivity to platinum and poly (ADP-ribose) polymerase inhibitors in certain cancer types, including breast, ovarian, pancreatic, and prostate. In these cancers, BRCA1/ 2 alterations and genomic scar signatures are useful indicators for assessing HRD. However, alterations in other homologous recombination repair (HRR)-related genes and their clinical significance in other cancer types have not been adequately and systematically investigated. METHODS: We obtained data sets of all solid tumors in The Cancer Genome Atlas and comprehensively analyzed HRR pathway gene alterations, their loss-of-heterozygosity status, per-sample genomic scar scores, ie, the HRD score and mutational signature 3 ratio, DNA methylation profiles, gene expression profiles, somatic TP53 mutations, sex, and clinical information including chemotherapeutic regimens. RESULTS: Biallelic alterations in HRR genes other than BRCA1/ 2 were also associated with elevated genomic scar scores. The association between HRR-related gene alterations and genomic scar scores differed significantly by sex and the presence of somatic TP53 mutations. HRD cases determined by a combination of these indices also showed HRD features in gene expression analysis and were associated with better survival when treated with DNA-damaging agents. CONCLUSION: This study provides evidence for the usefulness of HRD analysis in all cancer types, improves chemotherapy decision making and its efficacy in clinical settings, and represents a substantial advancement in precision oncology.

PURPOSE: Homologous recombination DNA repair deficiency (HRD) is associated with sensitivity to platinum and poly (ADP-ribose) polymerase inhibitors in certain cancer types, including breast, ovarian, pancreatic, and prostate. In these cancers, BRCA1/2 alterations and genomic scar signatures are useful indicators for assessing HRD. However, alterations in other homologous recombination repair (HRR)-related genes and their clinical significance in other cancer types have not been adequately and systematically investigated. METHODS: We obtained data sets of all solid tumors in The Cancer Genome Atlas and Cancer Cell Line Encyclopedia, and comprehensively analyzed HRR pathway gene alterations, their loss-of-heterozygosity status, and per-sample genomic scar scores, that is, the HRD score and mutational signature 3 ratio, DNA methylation profiles, gene expression profiles, somatic TP53 mutations, sex, and clinical or in vitro response to chemical exposure. RESULTS: Biallelic alterations in HRR genes other than BRCA1/2 were also associated with elevated genomic scar scores. The association between HRR-related gene alterations and genomic scar scores differed significantly by sex and the presence of somatic TP53 mutations. HRD tumors determined by a combination of indices also showed HRD features in gene expression analysis and exhibited significantly higher sensitivity to DNA-damaging agents than non-HRD cases in both clinical samples and cell lines. CONCLUSION: This study provides evidence for the usefulness of HRD analysis in all cancer types, improves chemotherapy decision making and its efficacy in clinical settings, and represents a substantial advancement in precision oncology.