Validation by Gene Expression Omnibus
11824 studies were retrieved using the keyword “tuberculosis” in the GEO database (https://www.ncbi.nlm.nih.gov/geo/). Set the study type as expression profiling by array, species as mus musculus and homo sapiens, and sample type as the whole blood or PBMCs. The expression profile data of the GSE140943 dataset (36)composed of three uninfected mice and five 56 days infected mice and the expression profile data of six uninfected C57BL/6 mice and five 21 days infected C57BL/6 mice in the GSE89389 dataset(37). In the GSE102459 dataset(33), there were PBMC gene expression profile data of the M72/AS01 vaccine at 0 and 31 days of treatment. The DE genes were filtered using the GEO2R online analysis tool (set |log2FC|>1 and P < 0.05 as thresholds)(38). We compared the DE genes in the three datasets with our research results in the TB model vs. normal group, 100μg DNA IM vs. TB model group, and 50μg DNA EP vs. TB model group. The GSE89403 dataset using peripheral blood samples identified biomarkers for clinically relevant responses to TB treatment. Selected the DE genes both in the TB model vs. the normal group and in the GSE89403 dataset. The correlation between 6 DE genes’ fold change values at 4 and 24 weeks and anti-tuberculosis treatment response were analyzed by multivariate logistic regression analysis with SPSS version 26. P value<0.05 was considered to have a significant difference.