Late-stage solid tumors often lack effective treatments, but T-cell receptor (TCR) based drugs, which can target intracellular cancer markers, are a promising approach. However, identifying potent and specific TCRs for cancer antigens is challenging. Vcreate's platform uses advanced machine learning and screening assays to rapidly identify T-cell receptors (TCRs) for important cancer targets. Their technology allows for high-throughput screening of over 100,000 TCRs against up to 1,000 targets in a single experiment, offering over 100 times greater efficiency than existing methods. Their proprietary machine learning models can predict TCR interactions with cancer targets, including those with limited or no training data, resulting in the discovery of novel TCRs against valuable cancer targets.
What is the problem?
Most late-stage solid tumors still lack effective treatment options for patients. T-cell receptor (TCR) based drugs offer a unique modality to target cancer cells via their ability to recognize intracellular cancer targets (e.g. mutated KRAS and PIK3CA). T-cell receptors can be used for T-cell therapies or engineered into T-cell engagers (e.g. BiTE) to treat cancer without cell manufacturing. The FDA approved the first TCR-based T-cell engager in 2022, and most clinically valuable cancer antigen targets remain undrugged. It is challenging to identify TCRs that interact with cancer antigens with high potency and specificity.
What is their solution?
Vcreate's platform combines novel machine learning and screening assays to rapidly identify TCRs against clinically valuable cancer targets. Their wet lab platform can interrogate 100k+ TCRs against up to 1000 targets in a single experiment, with paired TCR-target readouts. This is >100x higher throughput than other existing screening methods. Their proprietary machine learning models can predict highly presentable cancer targets (via MHC antigen presentation) and TCR interactivity against targets of interest, including targets with extremely limited or no training data, something not possible with previous work (e.g., TCRAI, DeepTCR, NetTCR, etc.). Advancements here come from both novel machine learning architecture and data generated by their proprietary assay. They've already discovered novel TCRs against several valuable targets.
