Detecting Microbiome in Human Colorectal Cancers from High Throughput Sequencing Data

Abstract

The microbiome is widely distributed on the human body and has a complex relationship with the host, through active metabolic interactions. The gut microbiome has shown to influence the risk and progression of colorectal cancer (CRC). Several tools are available for detecting microbiome from high throughput sequencing data. However, little is studied using the Cancer Genome Atlas (TCGA) CRC RNA-seq datasets. Here, we use the MOCAT2 software to detect bacterial sequences from the TCGA-CRC RNA-seq data. We analyzed a total of 55 cases, with 41 tumor samples and 14 control samples. The RNAseq data were first aligned to the human genome (hg19) and the reads that were not aligned to the human genome were used to detect bacterial sequences using MOCAT2. We found that Escherichia coli and Propionibacterium acnes were two bacterial species most commonly detected in the tumor samples (61% and 55% respectively), however, they were only detected in 14.29% and 7.14% of control samples, respectively. The Cupriavidus necator bacteria were most commonly detected in control samples (71.43%), while only present in 2.44% of tumor samples. Cupriavidus necator is the only bacteria species show significantly different presence between the tumor and control samples (chi-square test, p = 2.16E-07). Due to the limited sample size and reads detected by MOCAT2, future studies will use other software options such as PathSeq to detect the bacterial sequences.