java -jar trimmomatic-0.35.jar PE -phred33 input_forward.fq.gz input_reverse.fq.gz output_forward_paired.fq.gz output_forward_unpaired.fq.gz output_reverse_paired.fq.gz output_reverse_unpaired.fq.gz ILLUMINACLIP:TruSeq3-PE.fa:2:30:10 LEADING:3 TRAILING:3 SLIDINGWINDOW:4:15 MINLEN:36

This will perform the following:

• Remove adapters (ILLUMINACLIP:TruSeq3-PE.fa:2:30:10)
• Remove leading low quality or N bases (below quality 3) (LEADING:3)
• Remove trailing low quality or N bases (below quality 3) (TRAILING:3)
• Scan the read with a 4-base wide sliding window, cutting when the average quality per base drops below 15 (SLIDINGWINDOW:4:15)
• Drop reads below the 36 bases long (MINLEN:36)

More at http://www.usadellab.org/cms/?page=trimmomatic

Address of the bookmark: http://www.usadellab.org/cms/?page=trimmomatic

More at http://www.ncbi.nlm.nih.gov/pubmed/23303509

Address of the bookmark: http://sc932.github.io/ALE/about.html

Address of the bookmark: http://www.cbcb.umd.edu/software/phymm/

Address of the bookmark: http://www.niehs.nih.gov/research/resources/software/biostatistics/art/

CABOG is one of many software programs called genome assemblers. These programs exist to overcome the fundamental limitation of all sequencing machines, namely, that they read out very few DNA letters at a time. These programs reconstruct genomes that are billions of letters long from the hundreds of letters per read that modern sequencers provide. What these programs do is often described as a scaled up version of a family solving a jigsaw puzzle.

The CABOG software was the first to accomplish many scientific goals. It was the first to assemble the genome of a multicellular organism (Drosophila melanogaster, 2000). It was the first to assemble both parental haplotypes of one human genome (J. Craig Venter, 2007). It was the first to assemble environmental sequence from the oceans (Sargasso Sea in 2004 and Global Ocean Sampling in 2007). It was first to combine reads from first-generation Sanger sequencing machines and second-generation pyrosequencing machines (Marine microbes, 2006). Today, CABOG is one of the leading assembly programs for data sets that include paired end data from the Roche 454 line of sequencing machines.

Address of the bookmark: http://www.jcvi.org/cms/research/projects/cabog/overview/

https://f1000research.com/articles/6-100/

https://github.com/jason-weirather/AlignQC

Address of the bookmark: https://www.healthcare.uiowa.edu/labs/au/AlignQC/

The "Ifs" of NGS QC and Trimming

1. Ensures Data Integrity
   If you want to minimize errors in downstream analyses, QC and trimming remove low-quality reads and bases, ensuring high-confidence data. This step is essential for reliable variant calling, assembly, and other applications.

2. Removes Contaminants
   If adapter sequences or contaminants are present in the raw reads, trimming can eliminate them. This prevents issues like misalignment or incorrect biological interpretations, ensuring cleaner data for analysis.

3. Improves Mapping and Assembly
   If your goal is better alignment to a reference genome or improved de novo assembly, trimming low-quality bases and adapters is critical. High-quality reads map more efficiently and generate more accurate assemblies.

4. Reduces Computational Load
   If you want to save computational resources, trimming reduces the dataset size, which speeds up processing and analysis. Clean datasets mean less computational time spent on processing low-quality data.

5. Prepares for Standardized Analyses
   If your project involves multiple datasets, QC and trimming ensure uniformity across them. This standardization makes comparisons valid and reproducible, particularly in large collaborative studies.

The "Buts" of NGS QC and Trimming

1. Risk of Over-Trimming
   But excessive trimming can lead to the loss of informative sequences, reducing read depth and potentially discarding biologically relevant data. This is especially critical in studies with limited sequencing depth.

2. Bias Introduction
   But trimming algorithms might introduce biases, especially if they inadvertently remove sequences with specific biological patterns. This can skew results and compromise biological insights.

3. Loss of Context in Paired-End Reads
   But trimming one read in a pair more than the other can lead to loss of pairing information. This complicates downstream analyses that rely on paired-end data, such as structural variant detection.

4. Time and Resource Intensive
   But running QC and trimming for large datasets can be computationally expensive and time-consuming. As sequencing depth increases, preprocessing becomes a bottleneck in the analysis pipeline.

5. Variable Standards
   But the criteria for trimming (e.g., quality threshold, minimum read length) can vary between tools and datasets. This variability may affect reproducibility and comparability of results across studies.

Balancing the "Ifs" and "Buts"

To maximize the benefits of QC and trimming while mitigating the challenges, consider the following best practices:

• Use QC Tools Wisely: Start with tools like FastQC to identify quality issues in your raw data. Visualizing quality metrics helps tailor your trimming parameters.

• Choose Reliable Trimming Tools: Tools like Trimmomatic, Cutadapt, and BBduk offer adaptive and customizable trimming options. Select one that aligns with your dataset and project goals.

• Set Reasonable Parameters: Avoid over-trimming by setting quality thresholds and minimum read lengths that balance data retention and quality improvement.

• Test Downstream Effects: Validate the impact of QC and trimming on downstream analyses, such as alignment efficiency, variant calling accuracy, or assembly quality.

• Document Your Workflow: Maintain detailed records of the parameters and tools used for QC and trimming. This ensures reproducibility and enables better troubleshooting.

Conclusion

NGS quality control and trimming are essential steps to ensure reliable and accurate data for analysis. While the "ifs" highlight the clear benefits of these steps, the "buts" remind us of the potential pitfalls. By adopting best practices and carefully balancing these considerations, you can optimize your preprocessing workflow and unlock the full potential of your sequencing data.

The prepint version is found on http://www.biorxiv.org/content/early/2017/08/09/173872

It uses the pyPaSWAS framework for sequence alignment (https://github.com/swarris/pyPaSWAS)

Address of the bookmark: https://github.com/swarris/Pacasus

git clone --recursive https://github.com/tobiasrausch/alfred.git

cd alfred/

make all

https://github.com/tobiasrausch/alfred/

Address of the bookmark: https://gear.embl.de/alfred

per-base depth about 2x as fast samtools depth--about 25 minutes of CPU time for a 30X genome.
mean per-window depth given a window size--as would be used for CNV calling.
the mean per-region given a BED file of regions.
a distribution of proportion of bases covered at or above a given threshold for each chromosome and genome-wide.
quantized output that merges adjacent bases as long as they fall in the same coverage bins e.g. (10-20)
threshold output to indicate how many bases in each region are covered at the given thresholds.
A summary of mean depths per chromosome and within specified regions per chromosome.

Address of the bookmark: https://github.com/brentp/mosdepth