SynMix: Improving mechanistic understanding of chemical mixtures using large-scale cell profiling in an automated laboratory

There is an urgent need to develop strategies on how exposure to multiple chemicals and the combination of their effects can be assessed. The vast number of potential chemical combinations in chemical mixtures calls for novel approaches. In this project we aim to improve the identification and mechanistic understanding of synergistic chemical mixtures.

We propose a new in vitro method based on morphological cell profiling, producing rich data at reasonable costs that is highly suitable for use with artificial intelligence methods.

We will start by establishing reference datasets for single compounds and their combinations, and based on this data we will develop AI methods to select the most interesting combinations of chemicals to test for synergistic effects on cell morphology.

The selected experiments will then be carried out in our automated laboratory, and the system will learn with every iteration.

From the identified synergistic combinations we will make a selection to be profiled with arrayed CRISPR screening combined with morphological profiling to study changes in morphology given a certain treatment and genetic knockouts. The rich datasets will be deposited publicly.

This data will, together with the identified synergistic combinations and information from arrayed CRISPR screens, contribute with new and important mechanistic understanding for a large set of compounds and combinations, relevant for future chemical mixture analysis.