DEL+ML paradigm for actionable hit discovery – a cross DEL and cross ML model assessment.

• DOI: 10.26434/chemrxiv-2024-2xrx4
• Publication Date: 2024-07-24T12:57:26Z
• Authors: Sumaiya Iqbal; Wei Jiang; Eric Hansen; Tonia Aristotelous; Shuang Liu; Andrew Reidenbach; Cerise Raffier; Alison Leed; Chengkuan Chen; Lawrence Chung; Eric Sigel; Alex Burgin; Sandy Gould; Holly Soutter
• Keywords: Data set

DNA-Encoded Library (DEL) technology allows the screening of millions, or even billions, encoded compounds in a pooled fashion which is faster and cheaper than traditional approaches. These massive amounts data related to DEL binders not-binders target interest enable Machine Learning (ML) model development large, readily accessible, drug-like libraries an ultra-high-throughput fashion. Here, we report comparative assessment DEL+ML pipeline for hit discovery using three DELs five ML models (fifteen combinations two different feature representations). Each was used screen diverse set compound collections identify orthosteric therapeutic targets, Casein kinase 1𝛼/δ (CK1𝛼/δ). Overall, 10% 94% predicted were confirmed biophysical assays, including nanomolar (187 69.6 nM affinity CK1𝛼 CK1δ, respectively). Our study provides insights into paradigm discovery: importance ensemble approach identifying binders, usefulness large training chemical diversity DEL, significance generalizability over accuracy. We shared our results via open-source repository further use similar efforts.