Data-smoothing methods adjust all gene expression levels based on expression levels in “similar” cells, aiming to “denoise” the values (Fig.2). Several such methods have been proposed to handle imputation problems (Table2(B)). To take a simplified example (Fig.2), we might imagine that...