Study on potential differentially expressed genes in stroke by bioinformatics analysis