Syntensor

To guide drug discovery and design, the Syntensor platform allows the user to simulate a variety of preclinical assays. We have prioritized predicting potency and toxicity endpoints, as collectively, these two areas account for 60-70% of clinical trial failures. The predictions of these assays are contextualized with genome-wide predictions of changes in gene expression and pathway activation.

In drug discovery and development, the decision to progress an asset to clinical trials is made by considering the results of many preclinical assays. A compound is iteratively modified to improve its efficacy while reducing its toxicity. One of the biggest challenges in drug discovery is the lack of explainability from preclinical assays. In these assays, a compound is either a "hit" or not. In no other field of engineering would this lack of resolution be acceptable. If the output of Cadence’s Virtuoso was “this chip does/doesn’t work” that would not be helpful, we would need to mechanistically understand why it doesn’t work so that we can address the issue.

The Syntensor platform predicts and explains the outcomes of preclinical assays.

Preclinical assays have low information content and often yield conflicting results, leading to high uncertainty about the mechanism of action of a drug beyond its primary target.

If scientists were able to understand all of the pathways that were perturbed by a drug, they could design a more informative assay strategy and anticipate issues with efficacy and toxicity before progressing a candidate to clinical trials.

To meet this challenge, the Syntensor Platform leverages predictive models trained on mutlti-omics and experimental data to simulate cellular assays at scale and contextualize assay outcomes with predicted genome-wide changes in gene expression levels for novel compounds across cell lines. These changes are further summarized at the pathway level so that users can understand how perturbing the primary target of the drug leads to changes in the state of the cell. Users can visualize these predictions in the Pathway Explorer to generate testable hypotheses about how their drugs are perturbing cellular signaling pathways downstream of binding the primary target of the drug.

The platform allows users to simulate three types of experiments and contextualize the outcomes of assays with changes at the gene and pathway level:

• Mechanism of Action: Generates a ranked list of potential MoA targets by taking into account all perturbed pathways
• Potency: Predicts growth inhibition and sensitivity in cancer cell lines and ties these to changes in apoptotic pathways
• Toxicity: Characterizes the potential for hepatotoxicity by predicting the possibility of drug-induced liver injury and results for 14 assay endpoints including CYP450 inhibition and mitochondrial toxicity