Additional file 1 of Identification of an immune classification for cervical cancer and integrative analysis of multiomics data

Additional file 1. Fig. S1. Flow chart of the study. A total of 542 human cervical cancer samples were analysed in this study. A training cohort including 293 samples was virtually microdissected to identify an Immune class. Validation was then performed in 2 independent datasets. Fig. S2. Identification of an immune expression pattern. (A) We used Nonnegative Matrix Factorization (NMF, k=3 factors) to analyze the microarray-based gene expression data of 293 samples. One of the 3 expression factors showed the lowest intra-tumoral immune cells (indicated in blue) as shown in the heatmap. High and low gene set enrichment scores are represented in red and blue, respectively. Fig. S3. Identification of the immune class. Heatmap indicates NMF consensus-clustering on exemplar genes and the Immune class. High and low gene set enrichment scores are represented in red and blue, respectively. The tumor purity, ESTIMATES score, immune enrichment score and stromal enrichment score is also indicated. Fig. S4. Molecular characterization of the immune class. Gene set enrichment analysis (GSEA) between the immune class and non-immune class confirmed enrichment of inflammation–related pathways, signatures of immune cells (p<0.05, FDR<0.05). Fig. S5. Molecular characterization of the two subtypes of the tumour microenvironment in the immune class: active immune and exhausted classes. Gene set enrichment analysis (GSEA) between the exhausted class and active immune subgroups (p<0.05, FDR<0.05). Fig. S6. Correlation between SNP data and transcriptomic data. KRAS, ITGAX and MCF2 mutation were significant for up-regulation of gene expression (p<0.05). Fig. S7. Differentially expressed miRNAs between immune class and non-immune class. The network summarizes complex connections between differentially expressed miRNAs (green dots) and targeted gene (yellow dots) (-1>log2 FC>1, FDR < 0.05). Fig. S8. Differentially expressed lncRNAs between exhausted and active immune subgroups. The network summarizes complex connections between differentially expressed lncRNAs (pink dots), targeted miRNAs (green dots) and targeted gene (yellow dots) (-1>log2 FC>1, FDR < 0.05). Fig. S9. Disease-specific survival and progression-free intervals according to immune classes before and after adjustment for risk factors. Disease-specific survival and progression-free intervals were the worse in the exhausted subgroup before and after PSM. Table S1 DEGs (Immune-vs-non-immune). Table S2 GSEA (Non-immune-vs-Immune). Table S3 Univariate analysis for overall survival (OS), disease-specific survival (DSS) and progression-free interval (PFI). Table S4 Multivariate analysis for for overall survival (OS), disease-specific survival (DSS) and progression-free interval (PFI). Table S5 DEGs (Exhausted-vs-Active immune). Table S6 GSEA (Exhausted-vs-Active immune). Table S7 GSEA Validation set 1 (Non-immune-vs-Immune). Table S8 GSEA Validation set 2 (Nonimmune-vs-Immune). Table S9 CNV (Immune vs Non-immune). Table S10 CNV (Exhausted vs Active immune). Table S11 The pathway enrichment of genes involving in the amplification and deletion (Immune vs Non-immune). Table S12 SNP (Immune vs Non-immune). Table S13 SNP (Exhausted vs Active immune). Table S14 connect KRAS mutation status to gene ontology - Biological process (upregulation). Table S15 connect KRAS mutation status to gene ontology - Biological process (downregulation). Table S16 DEFM (Immune vs Non-immune). Table S17 Differentially expression lncRNAs (Immune vs Non-immune). Table S18 Differentially expression lncRNAs (Exhausted vs Active immune). Table S19 Differentially expressed proteins (Immune vs Non-immune). Table S20 Differentially expressed proteins (Active immune vs Exhausted). Table S21 Sensitive chemotherapeutic drugs in subgroups.