The authors developed AfterQC ( Chen et al., 2017) to integrate quality control, adapter trimming, data filtering and other useful functions into one tool. Yet the tools must be used in combination because no single tool currently exists that can effectively address all these problems. The requirement to read and load data multiple times made preprocessing slow and I/O inefficient. A typical combination was the use of FASTQC for quality control, Cutadapt for adapter trimming and Trimmomatic for read pruning and filtering. In the past, multiple tools were employed for FASTQ data quality control and preprocessing. SOAPnuke ( Chen et al., 2018) is a recently published tool for adapter trimming and read filtering with the implementation of MapReduce on Hadoop systems. Trimmomatic ( Bolger et al., 2014), another popular trimming adapter tool, can perform quality pruning using algorithms such as sliding window cutting. Cutadapt ( Martin, 2011) is a commonly used adapter trimmer, which also provides some read-filtering features. For instance, FASTQC ( Andrews, 2010) is a Java-based quality control tool providing per-base and per-read quality profiling features. Quality control and preprocessing of FASTQ data could be considered resolved given the availability of several relevant tools. ctDNA sequencing data are usually highly noisy, and detected mutations often exhibit ultra-low mutation allele frequencies (MAF) quality control and data preprocessing are especially important for detecting low-MAF mutations to eliminate false positives and false negatives. As a major technology in liquid biopsy, cell-free tumor DNA (ctDNA) sequencing is used to detect tumor-derived DNA fragments from plasma, urine and other circulating liquids. For example, liquid biopsy technology ( Esposito et al., 2017), which seeks out cancer-related biomarkers in the circulatory system, can be used to facilitate cancer diagnosis and personalized treatment regimens. Sequencing technologies, especially next-generation sequencing (NGS), have been broadly used in clinical applications in recent years, particularly for noninvasive prenatal testing ( Bianchi et al., 2015) and cancer diagnosis. Even worse, library preparation and sequencing steps always involve errors and can cause inaccurate representations of original nucleic acid sequences. Data can suffer from adapter contamination, base content biases and overrepresented sequences.

Quality control and preprocessing of sequencing data are critical to obtaining high-quality and high-confidence variants in downstream data analysis.

Aio pack run times