Jichun Xie

xVERSE is a transcriptomics-native foundation model that improves representation learning and generates virtual cells, outperforming existing methods in preserving biological heterogeneity and reducing batch effects.

DeepScence identifies senescent cells in single-cell and spatial transcriptomics data, outperforming existing methods.

B-Lightning identifies marker genes and cell subpopulations for specific sources of heterogeneity while filtering out confounding factors, improving sensitivity and robustness in single-cell data analysis.

SifiNet identifies gene sets and annotates cells without clustering, showing better performance than existing methods on multiomic data.

DYNATE identifies rare-variant association regions by dynamically aggregating genomic regions and controlling false discovery rate, finding a new ALS-associated gene.

This paper introduces CDI, a new unsupervised measure to evaluate scRNA-seq clustering accuracy and select optimal parameters.