Yu Kang

LucaCell is a sequence-centric foundation model that improves cross-species single-cell analysis by using mRNA sequence embeddings instead of fixed gene identifiers.

This paper benchmarks how reference genome choice affects human genome interpretation by comparing T2T references and finding ancestry-matched references provide more accurate variant discovery.

PRIS improves protein-RNA interaction prediction and RNA library screening by combining structure selection and binding preference inference.

T2T-HG002v1.1, a complete diploid HG002 assembly lacking detectable errors over 99.4% of its length, enables genome-based benchmarking of assemblies, sequencing reads, and phased variant calls.

BBBP-Atlas predicts blood-brain barrier permeability for small molecules and peptides using a cross-modal dataset and structure-aware model, achieving high accuracy and outperforming existing methods.

ECloudGen improves molecular design by using electron clouds as latent variables, generating more potent binders and exploring broader chemical space than existing methods.

This paper presents a telomere-to-telomere rabbit genome assembly using haploid embryonic stem cells, achieving high completeness and enabling new annotations.