Single-cell and bulk transcriptomics reveal immune-telomere biomarkers in colorectal cancer

Immune analysis revealed differences in 22 immune cell types between the two groups, with activated B cells showing significant differences (adj. p < 0.05). Drug sensitivity analysis covered 168 drugs, among which pyrimethamine exhibited significantly different IC50 values between the HRG and LRG (p < 0.05). Single-cell analysis annotated nine cell types, with PLCG2, CCNA2, and CHEK1 differentially expressed in T cells, which were identified as key cells.

Finally, RT-qPCR validation using clinical samples confirmed that the expression patterns of all seven biomarkers matched the bioinformatic predictions.The dataset consists of three parts: raw data, processed data, and analysis code. The raw data include transcriptomic expression matrices and sample metadata downloaded from public databases. The processed data include quality-controlled expression matrices, differentially expressed gene lists, risk score tables, immune infiltration results, drug sensitivity results, and single-cell annotation results.

The analysis code covers data preprocessing, differential analysis, risk model construction, functional enrichment, immune analysis, drug sensitivity analysis, and single-cell analysis workflows. In the tabular data, rows represent genes or cells and columns represent samples or cell types, with units depending on the specific file (e.g., TPM, log2 fold change, IC50 values). The number of records is detailed in each file's description.

Batch effects may exist due to the different public database sources, which have been corrected during processing, though residual errors should be interpreted with caution. All files are in common formats (CSV, TXT, R scripts) and can be viewed without specialized software; the analysis code requires an R environment to run.