Seesaw signatures capture trajectory-like transcriptomic shifts and enable compact tumour cell classification across cancers

Accurate identification of tumor cells remains a major challenge in single-cell cancer research because malignant and normal cells often differ only subtly and vary greatly across datasets. Here we show that Seesaw pairs, defined by consistent reversals in the relative expression ranks of gene pairs between normal and tumor cells, provide compact and informative markers of malignant transformation. Across 44 single-cell RNA sequencing datasets spanning 22 cancer types, a classifier based on just three Seesaw pairs achieves a median area under the receiver operating characteristic curve of 0.93 in independent test sets and outperforms CellTypist, CTISL, ikarus, and scMalignantFinder. Recurrent Seesaw pairs reveal shared cancer-associated programs across malignancies, and many associated genes are linked to poor prognosis in The Cancer Genome Atlas cohorts. Because the framework is simple and interpretable, it may facilitate practical low-dimensional assays and may also be extendable to other data types in which coordinated rank disruption is informative. Seesaw signatures reveal trajectory-like transcriptomic shifts and enable compact, interpretable tumor cell classification across 22 cancer types in single-cell RNA-seq data.