Fig. 2From: Location and condition based reconstruction of colon cancer microbiome from human RNA sequencing dataValidation of the reconstructed microbiome profiles from The Cancer Genome Atlas (TCGA) with the European Institute of Oncology (IEO) cohort. a Principal component analysis of microbiome profiles from the TCGA (primary tumour and non-pathological solid tissue normal samples) and the IEO cohort shows cross-cohort clustering of colon adenocarcinoma samples. b Total number of normal and tumour samples. c Density plot of the Spearman coefficients describing the correlation between the 76 most abundant bacterial genera (present in more than 20% of the samples) profiled in both RNA sequencing (RNA-Seq) and ribosomal RNA 16S gene (16S) sequencing data. The majority of correlation coefficients are significantly larger than zero (p < 0.005, one-sample Wilcoxon test), meaning a similar tendency of detecting bacteria by the two approaches. d Correlation between the reconstructed microbiome (RNA-Seq bacterial relative abundances) and fluorescence in situ hybridisation (FISH) quantification of two bacteria, Faecalibacterium prausnitzii and Akkermansia muciniphila. Pearson (rp) and Spearman (rs) coefficients are indicated. e Representative FISH images of A. muciniphila and F. prausnitzii. GBM, glioblastoma multiforme; LUAD, lung adenocarcinoma; LUSC, lung squamous cell carcinoma; HNSC, head and neck squamous cell carcinoma; OV, ovarian serous cystadenocarcinoma; READ, rectum adenocarcinoma; SKCM, skin cutaneous melanoma, BRCA, breast invasive carcinoma. Number of samples analysed in bracketsBack to article page