Loading…
Every biological foundation model, evaluated and ranked by the bio.rodeo team
Showing 49–72 of 2565 models
Drug repurposing model that propagates transcriptional signatures through a pretrained graph autoencoder of the human protein interaction network.
Differentiable surrogate for near-infrared light transport in the human head, predicting time-resolved fluence 466x faster than Monte Carlo.
Per-residue intrinsic disorder prediction from sequence, with a window-wide output encoding that lets a convolutional network span 91 residues.
Conditional flow matching model that reconstructs 4D left-ventricular geometry and motion from a 12-lead ECG and basic demographic variables.
Genomic variant effect prediction from a bidirectional absorbing-diffusion model over human DNA, pretrained with entropy-guided span masking.
Structure-based drug design model building 3D ligands atom by atom inside a protein pocket, with explicit hydrogens and fragment completion.
Protein language model adding latent N-gram memory and sparse expert routing to an ESM-2 150M backbone for mutation-effect prediction.
Developmental time inference for single-cell and spatial transcriptomes, regressing absolute embryonic age from frozen Geneformer embeddings.
Protein domain embeddings that assign CATH homologous superfamilies by nearest-centroid lookup, staying accurate below 20% sequence identity.
Cancer-registry encoder pretrained by masked field-value modelling on 9.4M SEER tumour records, carrying survival signal into unseen rare cancers.
Single-cell foundation model that learns cell and gene embeddings by latent-space self-distillation rather than reconstructing masked expression.
Joint sequence and backbone co-design from a single-stage transformer trained directly on 3D coordinates, with no structure tokenizer.
Drug-target interaction model that retains intermediate encoder states and probes them per atom and residue, so weak functional-group signals persist.
Pediatric brain MRI segmentation model trained on synthetic scans, so a single 3D U-Net labels clinical acquisitions from neonates to young adults.
Weakly supervised deep learning system that scores 17 histological lesions continuously from kidney transplant biopsy whole-slide images.
Graph-grounded language model for antimicrobial resistance reasoning that judges gene-drug links from serialized multi-hop knowledge-graph paths.
Instruction-driven latent diffusion that segments polyps and skin lesions by painting a color-coded overlay on the image, not predicting a mask.
Mammography vision-language model that samples sparse high-res patches instead of downscaling, for zero-shot BI-RADS, density and cancer readout.
Single-cell foundation model fusing scRNA-seq with protein structure embeddings and subcellular localization priors through cross-attention.
Image-to-transcriptome diffusion model generating scGPT-space single-cell RNA embeddings from Cell Painting fluorescence microscopy features.
Dense self-supervised pretraining that learns pixel- and voxel-level features for segmenting cardiac MRI and abdominal CT from very few labels.
Slide-level histopathology model that flags endocervical gastric-type adenocarcinoma on routine H&E, built on frozen Virchow2 tile features.
Skeletal CT reconstruction from two orthogonal X-ray projections, using conditional latent diffusion that classifies scoliosis as the volume forms.