The Cancer Microbiome Atlas (TCMA): A Resource for Querying Host-Microbe Interactions

Despite the rapid expansion of microbiome research, studying the microbial composition of internal human organs and their association with disease states remains challenging due to the difficulty of acquiring clinical biopsies for microbial profiling. Next-generation sequencing data contain reads that map to both human and microbial genomes, but analyses of these datasets are plagued by contamination, which is relatively abundant in low-biomass tissue samples. We designed a statistical comparative model to analyze the prevalence of microbial species across sample types and sequencing centers from The Cancer Genome Atlas (TCGA). This revealed that species detected at equal rates across tissue and blood samples are predominantly contaminants, bearing unique signatures from each TCGA-designated sequencing center. Removal of such equiprevalent species from the dataset mitigated center-related batch effects and allowed isolation of the tissue-resident microbiome. The analytical results were further validated by metagenomic sequencing of original matched TCGA colorectal tumor and blood samples that were sequenced independently at each designated center. We thus present The Cancer Microbiome Atlas (TCMA), a collection of curated, decontaminated microbial compositions for TCGA sequencing data on oropharyngeal, esophageal, gastrointestinal, and colorectal cancer tissues with tissue-resident populations. The creation of TCMA led to several findings, including prognostic species, tumor-associated coabundance groups, and blood signatures of mucosal barrier injuries. Finally, we demonstrate that TCMA enables systematic matched microbe-host omics (epigenetic, transcriptomic, and proteomic) analyses, which will help guide future studies of the microbiome’s role in human health and disease.

Users can explore the TCMA database here: https://tcma.pratt.duke.edu