GEO dataset study population characteristics. The GEO microarray datasets included in this study are indicated. Of those which met the quality criteria, four datasets were available for HNSCC, four for PDAC, and five for gastric cancer (referred to as STAD in the manuscript). Samples comprised tumor and adjacent tissue or healthy biopsy. For the analysis of the dataset GSE138206, control samples consisted of contralateral normal samples; in this study the tissue adjacent to cancer was excluded. Abbreviations: GEO, Gene Expression Omnibus database; HNSCC, head and neck squamous cell carcinoma; PDAC, pancreatic ductal adenocarcinoma; STAD, stomach adenocarcinoma; GC, gastric cancer; OSCC, oral squamous cell carcinoma.