ZSeq 2.0: A fully automatic preprocessing method for next generation sequencing data

Abstract

Preprocessing is a critical step in next generation sequencing (NGS) data analysis, since any error or artifact in library preparation and the sequencing process can affect subsequent steps, leading to possibly erroneous biological conclusions. In this work, we propose ZSeq 2.0, a fully automatic NGS preprocessing method, which combines the strength of the original ZSeq method with a free of parameters scheme that automatically detects and filters out low complexity and highly biased regions, without any need for parameter adjustment. We estimate parameters by applying dynamic penalty rates to high and low GC-content sequences. We also use a labeling rule method to detect outlier sequences that have very low NUS. Some other preprocessing features have been added to ZSeq2.0, including adapter detection and low-quality nucleotides trimming at each side of the sequence. ZSeq2.0 is publicly available and can be downloaded from http://sourceforge.net/p/ZSeq/wiki/Home/.