The genetic material of different microorganisms, even those from the same species, can vary by 5% to 30% so Metagenomic species-based colorectal cancer (CRC) microbial biomarkers may mislead diagnosis. For a study, a total of 25 species were analyzed from 3 CRC (n = 249) studies with 7549 non-redundant single nucleotide variants (SNVs). The random forest technique, which was employed to analyze the entire dataset, identified 22 microbial SNV markers that were essential for differentiating subjects with CRC from healthy individuals. The accuracy scores of this predictive model were surprising with AUC = 75.35% in training and further, AUC = 73.08%-88.02% in validation cohorts. A meta-analysis on 4 metabolic disease groups was performed to observe the distinctiveness of these SNV markers in a larger context. Among the various SNVs, 3 that were more common in colorectal cancer patients and present in Eubacterium rectale and Faecalibacterium prausnitzii bacteria were predictive (AUC = 72.51% – 94.07%). The metagenomic microbiome sequencing and single nucleotide marker variants (SNVs) provide a novel and accurate predictive model for colorectal. Establishing a novel colorectal cancer predictive model based on unique gut microbial single nucleotide variant markers.
Link:www.tandfonline.com/doi/full/10.1080/19490976.2020.1869505
