Baylor College of Medicine

Context-specific protein embeddings across 286 liver disease and cell-type combinations, learned over interactomes built from a single-cell atlas.

Spatially aware transcriptomic foundation models for cancer, pairing 50um-Local and 250um-Extended views of spot-resolution spatial transcriptomes.

Chemical language model for fragment-based drug discovery, trained on all 62M ZINC-22 fragments. Samples 99.9% chemically valid fragment SMILES.

Single-cell foundation model for Drosophila that generates hierarchical cell-type annotations on new scRNA-seq datasets without any fine-tuning.

Chemical language model of the human metabolome that generates and ranks candidate structures for unidentified mass spectrometry peaks.

fMRI foundation model pretrained with masked autoencoding on roughly 6,700 hours of recordings for clinical prediction and network discovery.

Peptide retention time prediction for LC-MS/MS proteomics, from a genetic-algorithm search over convolutional and bidirectional GRU architectures.