Figures and data
Schematic overview of the study design.
Transcriptome profiling of circulating sEVs.
a) TEM images of circulating sEVs isolated from human plasma. b) NTA results of circulating sEVs enriched from plasma. c) WB results of sEV positive (Alix, TSG101, CD9) and negative (Calnexin) markers. d) The numbers of detected RNA species in different groups. e) The hierarchical clustering results of top 100 miRNAs (left panel), mRNAs (middle panel), and lncRNAs (right panel). f) t-SNE clustering by those 300 RNAs. g) A Venn diagram showed DEGs shared between different comparisons (CRC vs NC, AA vs NC, CRC vs AA). h) KEGG enrichment of all those DEGs identified. i-k) potential core regulatory networks between miRNAs and mRNAs in DEGs identified in three (i), two (j), and one (k) of all comparisons.
Cell-specific features of the sEV-RNA profile.
a) The hierarchical clustering heatmap of immune cell-specific features of each sample. b) The hierarchical clustering heatmap of stromal-related features of each sample. c) Boxplot of cell-specific features overexpressed in CC and RC patients (*P < 0.05, **P < 0.01, ***P < 0.001). d) Boxplot of cell-specific features overexpressed in NC participants (*P < 0.05, **P < 0.01, ***P < 0.001). e) The violinplot of the microenvironmental scores in different subgroups (*P < 0.05, **P < 0.01, ***P < 0.001). f) Correlation among cell-specific features differentially enriched among different groups.
WGCNA analysis of sEV-RNAs.
a) Gene coexpression module construction of all DEGs identified in sEV-RNAs. b) The heatmap exhibited Pearson correlations among different modules. c) Bar plot of module composition of different modules (all DEGs). d) Percentage bar plot of the RNA composition of different modules (all DEGs). e) A heatmap exhibited the expression levels of the top 10 DEGs in each module. f) t-SNE clustering by the top 10 DEGs in each module. g) t-SNE clustering by the top 5 DEGs in each module. h) t-SNE clustering by the top1 DEGs in each module.
The expression trends of sEV-RNA modules.
a-c) GSEA analysis of DEGs in different modules (a: CRC vs. NC; b: AA vs. NC; c: CRC vs. AA). d-i) The expression trends of the Top 10 DEGs of each module among NC, AA, and CRC (d: green module; e: red module; f: turquoise module; g: black module; h: blue module; i: brown module).
Module-trait correlation analysis of sEV-RNA modules.
a) The heatmap exhibited the correlation between modules and clinical traits. b) The RT-qPCR validation of representive module-trait correlation (left panel: correlation between MT-ND2 and Paris classification; right panel: correlation between HIST2H2AA4 and LST morphology). c) The heatmap exhibited the sEV-RNA expression levels of red, black, and green modules. d-f) Circos plot showed the inner correlations among sEV-RNAs in the module green (d), red (e), and black (f).
Selection of plasma sEVs-RNA candidates for AA and T1a stage CRC diagnosis
*+: significant difference found in this comparision. ** -: no significant difference found in this comparision.
The plasma sEVs-RNA signature to detect early CRC and AA.
a-d) The ROC analysis of different sEV-RNA signatures in the prediction of CRC patients by different algorithms (a: 5-gene panel; b: 6-gene panel; c: 7-gene panel; d: 8-gene panel). E-h) The ROC analysis of different sEV-RNA signatures in the prediction of AA patients by different algorithms (a: 6-gene panel; b: 7-gene panel; c: 8-gene panel; d: 9-gene panel). i) The QDA results of all 13 sEV-RNAs in classifying all samples. J) Statistical summary of QDA performance in each sample group.
