Loading…
Every biological foundation model, evaluated and ranked by the bio.rodeo team
Self-supervised pathology foundation model with a 300M-parameter vision transformer tile encoder, trained on a multi-stain whole-slide image corpus.
Cryo-electron tomography membrane analysis pipeline pairing generalizable U-Net membrane segmentation with mesh-based particle localization.
Vision-language model for annotation-free pathology localization, marking the finding a text prompt names in X-ray, histology and fundus images.
Generative RNA design model that samples family sequences from a VAE latent space constrained by Rfam covariance models and consensus structure.
Deep graph contrastive learning framework for single-cell proteomics embedding, handling peptide uncertainty, missingness, and batch effects.
Protein function prediction models that assign Gene Ontology terms using language model embeddings and neuro-symbolic reasoning over GO axioms.
Peptide language model that generates antimicrobial, anticancer, and target-binding sequences, adapted per task by Mixture-of-Experts plugins.
Codon-level BERT model that captures genomic signals invisible to amino acid models, outperforming billion-parameter PLMs with just 86M parameters.
Diffusion model for synthesizing single-cell RNA-seq data, with guided generation of specific cell types, rare cells, and developmental trajectories.
Nucleotide language model for prokaryotic promoter design, fine-tuned from one pretrained base into 27 species-specific generative checkpoints.
Google's dermatology image embedding model that produces 6144-dimensional embeddings for data-efficient skin-condition classifiers.
Histopathology foundation model that encodes 224x224 H&E patches into compact 384-dimensional embeddings for tumor and biomarker classifiers.
Cell type annotation for single-cell RNA-seq that builds a graph per signaling pathway, learning across pathway views with graph neural networks.
Inverse protein folding from backbone coordinates, chaining a pretrained structure encoder into a pretrained sequence autoencoder on small data.
Binding energy for protein-ligand, protein-protein, and antibody-antigen complexes is read off an energy model trained on crystal structures alone.
500M-parameter transformer model pretrained on intracranial SEEG recordings for neural signal forecasting, imputation, and seizure detection.
Asymmetric encoder-decoder transformer for single-cell RNA-seq that encodes only non-zero genes, cutting FLOPs 10-100x versus standard transformers.
Self-supervised foundation model for wearable electrocardiograms, trained with participant-level contrastive learning on 141,000 participants.
Self-supervised foundation model for wearable photoplethysmography, trained with participant-level contrastive learning on 141,000 participants.
Controllable protein design by prefix-tuning a protein language model with learned virtual tokens that combine for multi-property generation.