Crowdsourced mapping of unexplored target space of kinase inhibitors

Abstract

Despite decades of intensive search for compounds that modulate the activity of particular protein targets, a large proportion of the human kinome remains as yet undrugged. Effective approaches are therefore required to map the massive space of unexplored compound–kinase interactions for novel and potent activities. Here, we carry out a crowdsourced benchmarking of predictive algorithms for kinase inhibitor potencies across multiple kinase families tested on unpublished bioactivity data. We find the top-performing predictions are based on various models, including kernel learning, gradient boosting and deep learning, and their ensemble leads to a predictive accuracy exceeding that of single-dose kinase activity assays. We design experiments based on the model predictions and identify unexpected activities even for under-studied kinases, thereby accelerating experimental mapping efforts. The open-source prediction algorithms together with the bioactivities between 95 compounds and 295 kinases provide a resource for benchmarking prediction algorithms and for extending the druggable kinome.

Systematic review of computational methods for drug combination prediction

“…However, 20% of drug combinations remained poorly predicted by all the methods, regardless of the method category. This is consistent with other DREAM Challenges, where it has been observed that the specific prediction algorithm per se is not the critical component, rather how the algorithm is used in practice and based on what information often defines the best-performing computational methods for many prediction tasks in biomedical research [18] , [19] .

Fig.

…”

A community challenge for a pancancer drug mechanism of action inference from perturbational profile data

“…( Table 1 ). In general, consistent with prior results on the value of evidence integration, 25 overall performance was positively correlated with the number of additional databases utilized in the analysis, accounting for 27% of the variance in SC1 and a remarkable 82% of the variance in SC2.…”

“…Teams were further encouraged to use public data sources, such as the Cancer Cell Line Encyclopedia, 21 the Genomics of Drug Sensitivity in Cancer database, 22 and the CMap L1000 database, 23 and to leverage insights and models developed in previous DREAM Challenges. 8 , 19 , 24 , 25 …”

“…Previous projects, such as the IDG-DREAM Drug Kinase Binding Challenge 25 and the Multi-targeting Drug DREAM Challenge, 26 challenged the community to develop computational methods that leveraged publicly available chemical (e.g., chemical fingerprints, protein structures) and kinase binding data to predict drug-target interactions without using compound treatment data in a biological context. In contrast, this challenge asked the community to develop methods that could rank the proteins most affected by a compound using publicly available biochemical data.…”

DrugRepo: A novel approach to repurpose a huge collection of compounds based on chemical and genomic features

“…Several artificial intelligence (AI) methods for drug repurposing are based on DTIs as well as chemical structural similarities [13][14] [15]. However, these methods are applied only on selected set of compounds resulting in limited prediction outcomes [13].…”