Modeling promiscuity based on in vitro safety pharmacology profiling data

Area of Science:

• Pharmacology
• Computational Chemistry
• Drug Discovery

Background:

• Distinguishing selective from promiscuous compounds is crucial in drug discovery.
• In vitro safety pharmacology data offers a rich source for assessing compound behavior across multiple targets.

Purpose of the Study:

• To develop and validate computational models for predicting compound promiscuity and selectivity.
• To leverage large-scale binding data for improved drug candidate evaluation.

Main Methods:

• Mining and modeling of binding data from extensive in vitro safety pharmacology assays.
• Generation and validation of two Naïve Bayes models for promiscuity and selectivity.
• Testing models against independent datasets and public drug databases.

Main Results:

• Models successfully differentiated between promiscuous and selective compounds.
• Marketed drugs exhibited higher selectivity scores compared to compounds in early development or those failing clinical trials.
• The models demonstrated robust performance on validation datasets.

Conclusions:

• Computational models based on in vitro safety pharmacology data can effectively predict compound selectivity.
• These models can aid in the early-stage triage of high-throughput screening hits.
• The approach supports lead optimization by identifying compounds with desirable selectivity profiles.