Guangzhou National Laboratory

Virtual-cell model that compresses a transcriptome into eight discrete tokens in a reasoning LLM's vocabulary, predicting module-level drug response.

Single-cell foundation model that tokenizes scRNA-seq into 10 tokens in a Qwen3-4B vocabulary for cell type annotation and perturbation prediction.

Multimodal generative model predicting viral antigenic change zero-shot from disentangled evolutionary, physicochemical, and structural signals.

Cross-species-pretrained CNN that predicts single-CpG DNA methylation from genomic sequence and interprets the cis-regulatory motifs that govern it.

RNA language model transfer-trained from ESM-2 via a pseudo-protein alphabet mapping, outperforming 12 RNA language models in zero-shot evaluation.

Multimodal single-cell foundation model pretrained on 4M+ co-assayed cells, predicting 382 surface proteins from transcriptomes alone, zero-shot.

RNA-protein contact prediction from sequence, built on ERNIE-RNA and ESM-2 embeddings. Reaches 0.77 auROC where AlphaFold 3 reaches 0.61.

Single-cell DNA methylation foundation model capturing genome-wide CpG dependencies in whole-genome bisulfite sequencing across tissues and species.

Multiomic foundation model for zero-shot in silico perturbation, predicting gene regulation and cell fate transitions from DNA and ATAC signal.

Vision-language model for annotation-free pathology localization, marking the finding a text prompt names in X-ray, histology and fundus images.

Protein language model ranking stabilizing mutations with no assay data, trained jointly to predict the growth temperature of each sequence's host.