Fig. 4

Interrogation on different cellular layers further classifies and defines leukemia in single-cell studies. (a) AML heterogeneity was better defined with deconvolution of bulk data by single-cell referencing. By deconvolution > 1000 AML patients bulk RNA-seq data using single-cell referencing, AML composition was converged into four overall classes, Primitive (LSPC-enriched), Mature (Mono-like and cDC-like blasts), GMP and Intermediate (balanced dstribution) and used as references for patient sample [59]. (b) Ribosomal protein expression levels are indicative of heterogeneity in prognosis in different leukemia subtypes. Higher expression of RPs may associated with poor outcomes [60, 61]. (c) Different malignant epigenetic layers indicate leukemia-specific modifications and provide references for subtyping [67, 68]. (d) Multi-omics framework (CITE-seq, scATAC-seq and scRNA-seq) defined the normal epigenetic baseline of healthy blood development and were used to deconvolve aberrant molecular features of MPAL patients [35]. (e) Multiple frameworks combining the mitochondrial mutational landscape with transcriptome and genetic mutation information (MutaSeq + mitoClone; CloneTracer) more confidently differentiated HSCs from LSCs [75]