Elsevier and Iktos partner to deliver an AI-driven synthetic chemistry platform for drug discovery

Elsevier, a global leader in information and data analytics, has signed a multi-year agreement with Iktos, a company specializing in artificial intelligence for new drug discovery. The partnership will strengthen Elsevier’s flagship chemistry solution, Reaxys, by combining the company’s high-quality chemistry data with synthetic planning AI technologies developed by Iktos to accelerate chemistry research for pharmaceutical companies.

Mirit Eldor, Managing Director, Life Sciences Solutions, Elsevier said: “Our mission is to support research and development (R&D) organizations in fueling innovation and drug discovery with industry-leading predictive algorithms trained on our high-quality data. We are delighted to partner with Iktos to bring the valued industry insights that Reaxys offers powered by artificial intelligence, which will help re-shape the landscape of small molecule discovery,”

The partnership aims to support R&D organizations to critically decrease the time required to complete Design-Make-Test-Analyze cycles of small molecules, thus reducing the length and cost of early-stage drug discovery. New predictive models will be launched to support use cases including retrosynthesis, synthetic accessibility, and other reaction-based analyses. The new tools will be delivered via Reaxys’ user-friendly interface and Application Programming Interface (APIs), accelerating research for synthetic, medicinal, computational and process chemistry teams at pharmaceutical, agro-chemical and contract research organizations (CROs).

Yann Gaston-Mathé, CEO, Iktos said: “Data is the foundation of AI, and we are very excited to partner with Elsevier. We will combine the strength of Reaxys, a chemistry reaction database with unmatched volume, diversity, and quality, with Iktos’ leading edge retrosynthesis AI technology. We are committed to making our technology available to R&D organizations around the world to maximize the impact of AI on the productivity of drug discovery.”

The new predictive retrosynthesis and synthetic accessibility tools is based on Iktos’ proprietary retrosynthetic technology and trained on Reaxys’ reaction database. The tool uses a template-based disconnection prediction model, to then apply a proprietary filter that scores reaction feasibility and includes chemo and regioselectivity parameters. The solution space is explored using an advanced search functionality that delivers promising candidate routes in minutes. The new tools will offer multiple exciting new features including stereochemistry support, defining first disconnection, ability to include or exclude intermediates, avoiding regioselectivity conflicts and defining price limit of starting materials. In addition, customers will have the option to customize the underlying model with proprietary reactions and libraries of starting materials.