Fudan University

Vision-language foundation model for coronary angiography that aligns six-view cine studies with procedural reports for zero-shot lesion assessment.

Single-cell metabolome inference from scRNA-seq, learned from spatially paired Visium and MALDI-MSI sections by multiple-instance learning.

Renal tumor histopathology model that detects tissue regions, classifies nine subtypes, grades nuclei and scores prognosis from a single H&E slide.

Protein language model for microbial smORF-encoded small proteins, pairing multi-scale convolutions with transformer layers in a compact encoder.

Genomic foundation model on a Mamba, attention, and mixture-of-experts backbone, with 1M-token context for variant scoring and regulatory DNA design.

Antibody CDR design model post-trained by on-policy distillation, cutting RAbD CDR-H3 backbone RMSD from 2.37 Å to 1.95 Å.

Reinforcement-learning generative framework for multi-objective RNA codon optimization that generalizes across six species and five RNA types.

Protein-text foundation model placing amino acid sequences and natural language in one token space for protein understanding and de novo design.

Dirichlet flow-matching model for protein design that generates family-aware sequences from ancestral-reconstruction priors, not random noise.

Codon-level mRNA language model adapted from ESM-2 650M by swapping amino-acid tokens for codon tokens, transferring protein knowledge to mRNA tasks.

Pathology foundation model that infers spatial transcriptomics and proteomics directly from routine H&E whole-slide images, with no spatial assay.

Phosphopeptide detectability prediction for mass spectrometry, rescoring DDA identifications and pruning DIA spectral libraries to cut search time.

Connects a frozen protein language model to a general LLM via a cross-modal projector, adding protein reasoning without catastrophic forgetting.

De novo peptide sequencing transformer that reads modified and unmodified peptides directly from tandem mass spectra without a reference database.

Cross-species RBP-RNA binding site predictor that turns RNA-binding protein conservation into label smoothing, reaching 0.85 AUC from human to mouse.

Generative antibody and nanobody design model that co-designs CDR sequences and antigen-bound structures for de novo design and affinity maturation.

Vision-language foundation model for kidney cancer CT, covering zero-shot malignancy diagnosis, report generation, and recurrence risk prediction.

Dermatology foundation model pretrained on 432,776 skin images, covering malignancy classification, severity grading, and lesion segmentation.

Molecular reasoning language model for molecule captioning and text-to-molecule generation, trained by chain-of-thought distillation then reward RL.

Histopathology segmentation model aligning SAM to clinical intent through direct preference optimization, tested zero-shot on 12 external datasets.

Structure-based virtual screening model that jointly predicts protein-ligand complex structures and binding fitness from sequence and SMILES.

DNA foundation model for germline variant pathogenicity, pretrained on 27 mammalian genomes and fine-tuned on ClinVar and HGMD for SNVs and indels.

Tandem mass spectrum prediction for intact N- and O-glycopeptides that localizes O-glycosylation sites from HCD spectra alone, without ETD.

Structure-based drug design model that generates 3D ligands inside a protein pocket by interpolating distribution parameters instead of samples.

Protein foundation model built on Bayesian Flow Networks, prompted with MSA profiles for structure- and function-preserving sequence design.

De novo peptide sequencing from tandem mass spectra with a non-autoregressive Transformer trained on 100 million peptide-spectrum matches.

Medical multimodal LLM (2B and 8B) trained for generalizable, step-by-step clinical reasoning via Mentor-Intern Collaborative Search.

De novo peptide sequencing from tandem mass spectra, using curriculum learning and iterative self-refinement to stabilize non-autoregressive decoding.

Text-to-protein design retrieving natural protein fragments as a dynamic vocabulary, matching larger baselines on under 0.04% of their training data.

Bidirectional DNA foundation model with a Mamba-attention-mixture-of-experts design, reading 1 million base pairs at single-nucleotide resolution.

Single-cell chromatin accessibility foundation model with genome-aware tokenization, pretrained on 1.97 million scATAC-seq cells across 30 tissues.

Single-cell foundation model inferring cis-regulatory relationships from scRNA-seq and scATAC-seq, pretrained on an atlas of 1.3 million cells.

General medical vision-language model trained with reinforcement learning to reason step by step over medical images for diagnosis and visual QA.

Whole-body CT segmentation covering 235 fine-grained anatomies: 193 organs, 33 lymph node stations, and 9 lesion types from one unified network.

RNA inverse folding model that designs nucleotide sequences for a target 3D backbone by running discrete diffusion in hyperbolic space.

Secondary metabolite structure prediction from microbial biosynthetic gene clusters, generating SMILES strings from Pfam functional-domain tokens.

Protein conformational ensemble generator that denoises backbone geometry under language-model sequence conditioning with locality-aware attention.

Promptable 3D segmentation foundation model for whole-body PET, delineating organs and lesions from one or a few clicked points.

Multi-omics instruction-tuned LLM that reads DNA, RNA, protein, and multi-molecule sequences and answers natural-language questions about them.

Protein and RNA sequence annotation that also names the residues driving each label, learned from sequence-level supervision alone.

Pathology image embeddings supervised by spatial transcriptomics instead of text captions, aligned over 697K image-gene expression pairs.

TCR-antigen binding affinity prediction for unseen peptides, pairing an ESM-2-initialized receptor encoder with a peptide cross-attention module.

Protein conformational ensemble generator using SE(3) flow matching from a perturbed ESMFold prior, sampling MD-like dynamics from sequence alone.

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

Protein subcellular localization from sequence, returning both a text label and a synthetic fluorescence image of the protein inside a given nucleus.

Cryo-EM and cryo-ET heterogeneity analysis that separates subunit rigid-body motion from compositional change into distinct latent spaces.

RNA language model adapted from ESM-2 by cross-modality transfer learning, matching RNA-native baselines with 1/8 the trainable parameters.

Ultrasound foundation model pretrained on over two million multi-organ images, transferring to segmentation, classification, and image enhancement.

Molecular property prediction model pretrained by masked reconstruction of SMILES and graph, with disjoint masks that force cross-modal recovery.

Unified DNA, RNA, and protein foundation model with 1.8B parameters, pretrained across 169,861 species to learn the central dogma from sequence.

Swin transformer foundation model for fluorescence microscopy image restoration, unifying denoising, super-resolution, and volumetric reconstruction.

Protein language model-based sequence search that detects remote homologs with threefold higher sensitivity than MMseqs2 at comparable speed.

RNA foundation model pretrained on 23.7 million non-coding RNA sequences, producing embeddings for structure prediction, annotation, and RNA design.