All Competitors

Inverse RNA folding from contact maps: an axial-attention transformer designing sequences for pseudoknots, non-canonical pairs and multiplets.

De novo peptide sequencing from tandem mass spectra, re-weighting training losses by conditional mutual information to recover modified residues.

Foundation model for medical image registration that aligns CT and MRI across anatomies and modalities without per-pair optimization or retraining.

Deep network that predicts structures of full biological assemblies: proteins, nucleic acids, small molecules, metals, and covalent modifications.

Antimicrobial peptide generator running denoising diffusion in the continuous ESM-2 embedding space, validated in mouse infection models.

Bidirectional, reverse-complement equivariant DNA language models built on Mamba state space models for long-range variant effect prediction.

Target-aware peptide design by flow matching on torus manifolds, generating a peptide's torsion angles, pose and sequence inside a receptor pocket.

Neuroscience literature language model that predicts experimental outcomes from study methods, LoRA-tuned on 1.3 billion tokens of papers.

Self-supervised foundation model for functional brain network analysis from resting-state fMRI, pretrained across 30 datasets for disorder diagnosis.

RNA language model with 650M parameters pretrained on 36 million non-coding RNA sequences, generalizing structure prediction to unseen RNA families.

Discrete-diffusion protein language model pretrained on UniRef50 that generates sequences and gives embeddings that outperform ESM-2 on prediction.

Blind protein-ligand docking that transfers to binding domains absent from training, scoring 22.6% top-1 on DockGen and 50% on PoseBusters.

Generative co-design of a ligand-binding protein pocket's residue sequence and full-atom structure, conditioned on the target small molecule.

Self-supervised pretraining framework for 3D medical image encoders that learns anatomy by predicting where a sub-volume sits within a CT scan.

Generative pretrained transformer trained on 33 million human cells for single-cell annotation, batch correction, and perturbation prediction.

RNA foundation model trained on chemical-mapping data from millions of sequences, predicting reactivity, secondary structure, and degradation.

Protein large language model adapted from LLaMA-2 that unifies sequence generation and superfamily classification in one 7B-parameter framework.

Fine-grained cell-type abundance prediction from H&E histology, transferring to unseen cohorts and large slide archives without any retraining.

All-atom 3D molecular foundation model pretrained across small molecules, proteins, and complexes with an E(3)-equivariant denoising objective.

Foundation model spanning invasive SEEG/iEEG and non-invasive EEG in one backbone, with zero- and few-shot transfer across neurological disorders.

Compound-protein interaction prediction coupling a chemical language model to a protein language model via a cross-attention block.

DNA embedding model built on DNABERT-2, using contrastive learning to cluster sequences by species for metagenomic binning without labeled data.

Conversational T5-based framework that turns scRNA-seq data into cell sentences for cell type annotation and drug sensitivity prediction.