Unlock Biology

At Unlock Biology we’re building models of cell behavior in various contexts. For example, using the features of microenvironments, we predict genes associated with T cell exhaustion and NK cell dysfunction. Data sparsity is the main hurdle for foundation models in biology. Our approach transforms static biological data into dynamic, allowing us to infer temporal information not explicitly present in the original data. This increased sample efficiency allows us to work with much smaller datasets whilst gaining more valuable insights. My co-founder was in the Harvard group that developed the RNA velocity in 2018 and received Swiss Bioinformatics Institute 2023 award for advancing spatial transcriptomics methods. Another co-founder previously led a Berkeley-backed immunology startup.