AI for Health

Iktos AI

Iktos specializes in the development of AI technology for de novo drug design. More specifically, we are experts in deep generative models applied to chemistry, a recent technology (first academic paper published in Oct 2016), which we believe brings major value in de novo molecule design compared to previous technologies (molecule enumeration and/or genetic algorithms). Since our incorporation in 2016, we have focused our initial efforts on addressing multi-parameter optimization (MPO) in lead optimization projects, with a ligand-based approach. We have published an experimental validation of the Iktos technology applied to MPO in a real-life drug discovery project, in collaboration with Servier (poster presented at EFMC meeting in Sep 2018). In a few weeks, we were able to identify compounds that met the project Target Product Profile (TPP) after synthesis and tests, outperforming by far the compounds from the initial data set. We are now have expanded our technology to address other challenges such as scaffold hopping for fast follower projects, and structure-based de novo design. Iktos has also developed Spaya, a web-based application using a data-driven AI approach to discover and prioritize synthetic routes. While Spaya will be freely available on the web, the full power of Spaya is realized when it is integrated into a company's internal systems (database of in-stock chemical products and database of approved starting materials providers).