Frequentmers - a novel way to look at metagenomic Next Generation Sequencing data and an application in detecting liver cirrhosis

Abstract

Early detection of human disease is associated with improved clinical outcomes. However, many diseases are often detected at an advanced, symptomatic stage where patients are past efficacious treatment periods and can result in less favorable outcomes. Therefore, methods that can accurately detect human disease at a presymptomatic stage are urgently needed. Here, we introduce frequentmers; short sequences that are specific and recurrently observed in either patient or healthy control samples, but not in both. We showcase the utility of frequentmers for the detection of liver cirrhosis using metagenomic Next Generation Sequencing data from stool samples of patients and controls. We develop classification models for the detection of liver cirrhosis and achieve an AUC score of 0.91 using ten-fold cross-validation. A small subset of 200 frequentmers can achieve comparable results in detecting liver cirrhosis. Finally, we identify the microbial organisms in liver cirrhosis samples, which are associated with the most predictive frequentmer biomarkers.

A survey of k-mer methods and applications in bioinformatics

“…Similarly, Mouratidis et al introduced “frequentmers,” short sequences that are specific and recurrently observed in either patient or healthy control samples, but not in both [146] . Using metagenomic NGS data from liver cirrhosis patients and healthy controls, machine learning models trained with frequentmers (k = 16) outperformed previous models and achieved an AUC of 0.91 in liver cirrhosis detection.…”