All Competitors

Protein-protein interface prediction that summarizes molecular surface patches with persistent homology descriptors, at 0.77 test AUC.

RNA subcellular localization predictor that fuses physicochemical interaction graphs with frozen RiNALMo embeddings via a gated fusion layer.

Structure-based conformational B-cell epitope predictor that scores local antigen surface patches with ESM-2 embeddings and an ensemble MLP.

Gene Ontology prediction for malaria parasite proteins, trained on SAR supergroup structure embeddings with calibrated uncertainty and FDR control.

Sequence-based multitask model predicting covalently ligandable cysteines and reversible ligand-binding residues across the human proteome.

RNA small-molecule binding site prediction from sequence alone, pairing frozen RiNALMo embeddings with a lightweight MLP classifier.

B-cell epitope predictor fusing ESM-2 embeddings with residue contact and protrusion features to score linear and conformational epitopes.

Enzyme Commission number prediction that pools ESM Cambrian embeddings across unlabeled sequence homologs, scoring 0.788 F1 on full 4-digit EC.

Plant DNA-binding protein prediction that averages a ProtT5 sequence-embedding classifier with a SaProt structure-aware one at the score level.

Linear B-cell epitope prediction for cancer antigens, pairing ESM-2 embeddings with an MLP classifier; ROC-AUC 0.94 on a held-out IEDB benchmark.

Single-cell RNA-seq encoder trained with contrastive learning to merge plate- and droplet-based protocols, zero-shot on unseen tissues.

Bacterial protein-compound binding affinity prediction from amino acid sequence and SMILES, evaluated zero-shot on two species held out of training.

Histopathology model predicting gene expression and DNA methylation from H&E slides across 23 cancer types, fusing FFPE and fresh-frozen predictors.

Yeast gene regulatory network model with one pretrained subnetwork per gene, simulating target-gene response to transcription factor perturbation.

CRISPR-Cas9 repair outcome prediction from microhomology and sequence features, with transfer learning that adapts to a new cell line from 50 samples.

Per-residue prediction of where a receptor domain can be inserted into a protein without breaking it, for building allosteric and inducible switches.

Histopathology model predicting extrachromosomal DNA status from routine H&E slides by first inferring the tumor transcriptome from tile features.

Cell-type-specific gene expression prediction from DNA sequence, mapping Enformer epigenomic features to pseudobulk expression for cell-resolved TWAS.

CRISPR editing outcome prediction returning a probability over the near-full indel spectrum, with few-shot transfer to new cell types and to embryos.

Metagenomic read binning from tetranucleotide k-mer profiles, using a contrastive two-layer encoder trained on split halves of unlabeled reads.

Spatial transcriptomics prediction from H&E slides, inferring spot-level expression and tumor microenvironment composition in breast cancer.

Predicts the radius of gyration of intrinsically disordered proteins from 23 physics-derived sequence features, screening missense mutants in bulk.

Drug combination synergy prediction across cancer cell lines, from LLM text embeddings of drugs and cell lines rather than structures or expression.