Shandong University

Virtual spatial transcriptomics model that infers spot-level gene expression from H&E slides by fusing tile, slide, and spatial-position features.

Signal peptide design framework that generates, filters, and ranks cargo-specific secretion signals using evolution-constrained discrete diffusion.

RNA modification profiling from nanopore direct RNA-seq signal; self-supervised pretraining resolves 11 modification types and extends to new ones.

Uncertainty-aware diffusion model that enhances cryo-EM density maps while estimating voxel-wise confidence via Monte Carlo sampling.

Protein complex quality assessment model that scores docked and predicted dimers with separate chain graphs and an interface contact graph.

Evolution-guided diffusion model that generates temporal protein folding pathways, from unfolded chain to native state, rather than static structures.

Deep learning framework that predicts DNA methylation from genomic sequence across 39 human tissues, with an scRNA-seq variant for unseen cell types.

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

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

Single-cell multi-omics foundation model whose three-stage pretraining and distillation yield RNA-and-ATAC-aware embeddings from RNA-only input.

Quality assessment model for predicted RNA 3D structures, estimating local and global lDDT from predicted contact and distance-deviation maps.

Species-conditioned codon language model that jointly reads 5'UTRs, coding sequence, and RNA secondary structure to design native-like genes.

Intrinsically disordered region function prediction, scoring every residue for five binding subtypes plus disordered flexible linkers.

Single-cell type annotation pairing a CellMarker-derived knowledge graph with multi-agent LLM retrieval, generalizing across 11 tissue types.

Grounded multimodal language model for endoscopic surgery, supporting visual dialogue, region-based question answering, and bounding-box grounding.

De novo atomic model building from cryo-EM density maps, adapting AlphaFold2 with local attention and a 3D rotary position embedding.

Protein complex interface quality assessment model that pairs persistent homology barcodes with a graph attention network to predict DockQ scores.

Sequence-conditioned diffusion model that samples backbone conformational ensembles for intrinsically disordered proteins without MSAs or simulation.

DNA language model for interpretable prediction of 4mC, 5hmC, and 6mA methylation sites across species, using multi-scale k-mer BERT encoders.