Case Study: Genentech optimises kinase profiling with Intellegens Alchemite

Optimising kinase profiling programs with Optibrium and Genentech

Genentech worked with Intellegens to improve its kinase profiling programs, a challenging area where many kinase assays have sparse measurement data and traditional QSAR or similarity approaches struggle to scale across hundreds of targets. The company needed a way to use both chemical information and existing experimental data more effectively to narrow screening efforts and focus on the most informative kinases and compounds.

Intellegens applied its Alchemite™ machine learning platform to data covering approximately 650 kinases and 10,000 compounds. The solution learned relationships across kinase assays and delivered more accurate imputation models than traditional single-target QSAR and biological fingerprint similarity methods, as well as outperforming state-of-the-art multi-target deep learning. The improved accuracy and confidence supported practical use in active learning, assay prioritisation, and potential virtual screening.

Genentech

Fabio Broccatelli

Principal Scientist