Abstract
Variability in blood RNA-Seq data can obscure transcriptional changes that reflect tumor responses, and conventional normalization methods such as RLE/DESeq2 or TMM/edgeR often fail to capture these changes consistently. To address this challenge, we applied DiRT (DEG-by-index Ratio Transformation), a normalization and analysis strategy previously used in insect models, to 111 blood RNA-Seq datasets from mouse tumorigenesis models. DiRT achieved clearer separation between tumor and control samples as early as three days after tumor induction and maintained consistent marker signals across all stages of disease progression. In contrast, standard methods typically revealed differences only at later or scattered time points. KEGG pathway analysis further showed that DiRT-derived differentially expressed genes (DEGs) were enriched in the platelet activation signaling pathway, a pathway not identified using RLE/DESeq2 or TMM/edgeR. These findings demonstrate that DiRT enhances both sensitivity and reproducibility, enabling earlier and more consistent detection of transcriptional changes in blood during tumor development.