Wenxin Long

This paper introduces PertOmni, a multimodal framework that aligns transcriptomic data with text and image embeddings to improve perturbation effect prediction.

spVelo infers RNA velocity in multi-batch spatial transcriptomics data, showing improved velocity consistency, transition accuracy, and direction correctness compared to existing methods.

CosGeneGate selects non-redundant, cell-type-specific marker genes for single-cell analysis, outperforming existing methods in downstream applications.