Wanli Ouyang

NeuroPack compresses 3D neuroimaging data with high fidelity and ultra-high compression ratios using a hyperprior-driven deep learning toolkit.

CENO is a long-context generative model for genomic sequence understanding and design, showing improved variant prediction and regulatory sequence generation across species.

scMOBA is a multi-omics brain agent that integrates single-cell and spatial data across species, achieving state-of-the-art cell type classification and improving downstream tasks like aging clock construction and disease prediction.

This paper introduces scMRDR, a framework for integrating unpaired single-cell and spatial multi-omics data, showing strong performance in batch correction and spatial mapping.

This paper introduces deepGPS, a deep generative model that predicts protein subcellular localization using both sequence and image data, providing textual and image outputs.

ProDMM is a multimodal sequence modeling framework that captures interdependencies between protein motifs, coding DNA, and regulatory elements to predict phenotypes and generate optimized regulatory sequences.

scDNAm-GPT is a foundation model for single-cell DNA methylation analysis that improves clustering, trajectory inference, and deconvolution using context-aware tokenization and attention mechanisms.

The paper introduces Venus-MAXWELL, a framework for predicting protein mutant stability using sequence-to-landscape learning, achieving higher accuracy and faster inference than existing methods.

UniSPAC improves neuronal segmentation in connectomics by combining 2D interactive proofreading and 3D tracing, reducing manual effort and enabling semi-automatic labeling.

Genomics-FM is a foundation model that reduces reliance on labeled data and outperforms existing models in genome annotation, epigenomic prediction, and variant effect assessment.

BEACON introduces a comprehensive RNA benchmark to evaluate deep learning models, finding that single nucleotide tokenization and ALiBi improve performance.

π-PrimeNovo is a non-autoregressive deep learning model for de novo peptide sequencing that achieves higher accuracy and 69x faster performance than existing methods.

ProSST integrates protein sequences and structures into a Transformer-based model, achieving state-of-the-art performance in mutation effect prediction and downstream tasks.