Huan Xu

scEvolver enables continual cell-type annotation across expanding single-cell atlases, revealing cellular state transitions in disease contexts.

This paper proposes a framework that transfers knowledge from multi-modal genomic data to gene expression-only models to improve drug response prediction.