Where target discovery starts when the patient is the unit of modeling
A technical session for computational biologists and translational R&D teams, followed by live Q&A.
Leads the research behind Converge Bio's foundation models for the life sciences, from antibody design to patient-level cell-state simulation.
A patient-level foundation model, not a cell-level one
Entire patient samples live in one embedding space, trained on 46.5M cells across 9,843 patients, 927 studies and 400+ diseases.
In-silico perturbation
Simulate gene and drug perturbations to generate mechanism-of-action hypotheses, and anticipate drug response and resistance before wet-lab work.
End-to-end discovery
From raw single-cell or bulk RNA-seq to cell-type-specific gene attribution, in one pass.
Results on held-out cohorts
Validated on lupus, multiple myeloma and sepsis, matching or beating differential-expression, ML and PaSCient foundation-model baselines.
The model release
What's shipping, which modalities are supported, and how to run it on your own cohorts.