The Effect Of Quality Trimming On Joining Paired-End Reads In Microbiome Data Analysis

Abstract

The main goal of microbiome research is to determine the microbial composition of atarget sample. In successfully performed research, a cause of a disease can be found or pathogensin an environmental sample can be revealed. The correct evaluation of microorganisms present inthe analysed sample is therefore required, but unfortunately not that often reached. The accuracy ofthe resulting composition can be affected already in pre-processing steps of the analysis. The frequentissue that can negatively affect the reliability of research is data loss. This loss is most commonin the case of the paired-end reading method. The occurrence of data loss can be caused by theneed of joining this pair of sequences into one continuous sequence. The need of joining reads withlow-quality ends together can lead to counts reduction of sequences in an available dataset, andtherefore unrealistic results can be obtained. This paper shows, how quality trimming can affect thenumber of lost sequences during the reads joining step.