A frozen host-anchored framework identifies a host-coupled Klebsiella pneumoniae-centered ecological signal in colorectal cancer

This dataset provides reanalysis data and scripts for host–microbe coupling analysis in colorectal cancer (CRC), based on previously published, publicly available, de-identified human datasets. It contains host transcriptome expression matrices, paired microbial abundance tables, external stool microbiome benchmarking cohorts, and analysis scripts used for host anchoring, Kp-centered ecological score (KpNS) calculation, paired host–microbe coupling analysis, robustness analyses, and figure reproduction. Data sources include TCGA_CRC, ERP155818, PRJEB27928, SRP114892, PRJNA404030, and PRJNA941834. The dataset is intended to support reproducibility of a frozen host-anchored framework in CRC and to facilitate recalculation of the 10-gene IMI, 5-bug KpNS, paired host-coupling, bootstrap, leave-one-out, permutation, and technical adjustment results. It includes processed expression matrices, microbial tables, and analysis scripts, and may be used for method reproduction, secondary analysis, and teaching.