In Silico Target Identification Highlights IL12B as a Candidate for Small Molecule Drug Development

ABSTRACT Drug discovery’s rising costs and complexities require innovative strategies to identify viable therapeutic targets. We developed a computational pipeline to pinpoint protein targets lacking known small molecule probes, focusing on sites traditionally considered challenging for small molecule intervention but validated by FDA-approved biologics. Our approach integrates machine learning, public databases, structural modeling, and functional annotations to prioritize novel binding pockets that overlap biologically validated interfaces. This method identified IL12B as a promising candidate, revealing a previously unexploited surface pocket that overlaps part of the briakinumab epitope. Static protein solvent mapping and dynamic fragment simulations provide convergent evidence of druggability, including fragment-binding clusters and chemically diverse hotspots. While not yet experimentally validated, this site represents a plausible target for orally available IL12B inhibitors. Such compounds could address current clinical limitations of antibody therapies - such as prolonged systemic exposure and infection risk - by enabling a shorter half-life and improved mucosal penetration in diseases like inflammatory bowel disease. Competing Interest Statement Trevor J. Tanner, the founder of Epilog, L.L.C., received computational resources through the Google Cloud for Startups program. Some of these resources, along with academic resources and affiliations, contributed to the research process used in the identification of the IL12B target described in this work. Epilog, L.L.C. asserts no rights associated with any of the findings presented herein. Footnotes Title refined; abstract rewritten for clarity and clinical context; methods updated with CatBoost model for interpretability instead of balanced random forest, added SiteMap feature names, added SILCS HotSpots analysis, and streamlined filtering; results and discussion revised to reflect updated analyses with a focus on prioritizing Site B; Figures 1-3 and Table 1 replaced or added; acknowledgements and terminology standardized; reference list expanded; minor editorial corrections made.