Date sources

All expression profiling data used in the current study were downloaded from GEO database (https://www.ncbi.nlm.nih.gov/geo/) with the following inclusion criteria: (1) more than 30 peripheral blood samples from Homo sapiens; (2) concurrent inclusion of both IS patients and non-stroke healthy controls (HC). Accordingly, two GEO datasets were obtained, including {"type":"entrez-geo","attrs":{"text":"GSE16561","term_id":"16561"}}GSE16561 (IS = 39, HC = 24) from {"type":"entrez-geo","attrs":{"text":"GPL6883","term_id":"6883"}}GPL6883 platform of Illumina HumanRef-8 v3.0 expression BeadChip as the training set and {"type":"entrez-geo","attrs":{"text":"GSE58294","term_id":"58294"}}GSE58294 (IS = 69, HC = 23) from {"type":"entrez-geo","attrs":{"text":"GPL570","term_id":"570"}}GPL570 platform of Affymetrix Human Genome U133 Plus 2.0 Array as an external verification set. The “oligo” R package was used to preprocess (background correction, log transformation and normalization) the raw gene expression profiles for downstream analyses. Besides, 185 human CGs were sourced from AmiGO2 database (http://amigo.geneontology.org/amigo) [55].