When methods matter : disagreement between multiplex qPCR and metagenomic shotgun sequencing in defining the resistome of canine fecal samples

Abstract

Antimicrobial resistance (AMR) is a global health threat to humans and animals. It is of great importance to study and monitor the spread of AMR. This thesis intends to characterize the resistome in 35 canine fecal samples with bioinformatic tools using acquired shotgun metagenomic sequence data, and to compare the resistome results to those obtained using an extended multiplex qPCR method used in an earlier study. Fecal samples collected by dog owners for the HUNT-One Health project in addition to blank samples and mocks was analyzed. FastQC, MultiQC and Trim Galore were used for quality control and trimming of sequence data. AMRPlusPlus with MEGARes was used for AMR gene detection and resistome analysis. Only 8,2% of the AMR genes detected by qPCR were also detected in the metagenomic shotgun sequencing data. There were also AMR detections in the resistome that were not detected using qPCR as the corresponding detections accounted for 22,8% of the total detections made with the metagenomic shotgun sequencing data. Various factors such as extraction and detection method, sequencing depth, and different starting material could explain some discrepancy observed between the qPCR and resistome analysis. This master study does however tell a cautionary tale that resistome results vary depending on analysis method chosen, and that care should be taken when interpreting such results, especially if just one method is used. It would be valuable to have more starting material, increased sequence depth and use multiple resistome pipelines/databases in an effort to describe the “true” resistome better, as well as investigating taxonomic classification to study the bigger picture in the canine fecal microbiota.