Bayes Optimal Informer Sets for Early-Stage Drug Discovery

TL;DR

This paper introduces BOISE, a Bayesian optimal method for selecting informer compounds in early-stage drug discovery, improving prediction accuracy over existing strategies especially with incomplete data.

Contribution

The paper formalizes the informer set selection as a two-stage decision problem and proposes the BOISE method, which outperforms existing strategies in retrospective studies.

Findings

BOISE achieves better predictive performance than existing methods.

BOISE performs well with missing data, outperforming matrix completion approaches.

Empirical studies on kinase inhibition and cancer drug sensitivity validate BOISE's effectiveness.

Abstract

An important experimental design problem in early-stage drug discovery is how to prioritize available compounds for testing when very little is known about the target protein. Informer based ranking (IBR) methods address the prioritization problem when the compounds have provided bioactivity data on other potentially relevant targets. An IBR method selects an informer set of compounds, and then prioritizes the remaining compounds on the basis of new bioactivity experiments performed with the informer set on the target. We formalize the problem as a two-stage decision problem and introduce the Bayes Optimal Informer SEt (BOISE) method for its solution. BOISE leverages a flexible model of the initial bioactivity data, a relevant loss function, and effective computational schemes to resolve the two-step design problem. We evaluate BOISE and compare it to other IBR strategies in two retrospective studies, one on protein-kinase inhibition and the other on anti-cancer drug sensitivity. In both empirical settings BOISE exhibits better predictive performance than available methods. It also behaves well with missing data, where methods that use matrix completion show worse predictive performance. We provide an R implementation of BOISE at https://github.com/wiscstatman/esdd/BOISE