Autonomic platform for drug discovery and lead optimization

The development of a new drug is a years-long, expensive process with estimated costs of 1.5 billion USD. Early-stage drug discovery and lead identification and optimization contribute about a quarter of these costs. Drug development can be viewed as a multidimensional challenge in which various characteristics of compounds - including efficacy, pharmacokinetics, and safety - need to be optimized in parallel to provide lead drug candidates. The quality of early-stage lead compounds determines the success rate of later stages in the process. Increasing R&D productivity and efficiency is critical for controlling timeframes and costs in the pharmaceutical industry. Recently, various artificial intelligence (AI) and machine learning (ML) tools have emerged in the scope of drug development with significant potential. Yet, these tools are limited by low quality data, resulting from human-bias, surpassing the biological signal.

Here, we present a setup for an autonomic platform integrating automatic cell culture liquid handling and screening apparatuses with AI and ML tools using multiple feedback and decision loops. Combined with in-development custom algorithms, the infrastructure can utilize machine learning capabilities to allow an autonomous analysis of experimental data and objective decision-making to be made without intervention. We show that high-throughput screenings performed using this platform in cell culture and microtissue models allow real-time data acquisition of biological activity with minimal bias resulting from human intervention. Also, using this approach we can identify new novel molecules that modulate biological pathways with an optimized therapeutic window.