What are good means for performing quality control (QC) or NGS reads?

I'm aware of:

  • FastQC
  • NGS Screen
  • Kraken (e.g., for screening against contaminants)

What are other popular means for such QC?

  • 5
    $\begingroup$ This question is too broad. There are many available QC programs which all have their own advantages/disadvantages. Do you have a specific use case that you would like the QC to be applied to? $\endgroup$
    – gringer
    Jun 1, 2017 at 2:05
  • 1
    $\begingroup$ agree with @gringer. Right now the answer to this question is, roughly, whatever you get by googling 'NGS QC'. If you can provide a detailed use case - e.g. maybe you have 200GB of paired-end illumina data from 12 individuals of a non-model species, and you have a distant reference available - it will be a much more useful question and one that google couldn't solve. $\endgroup$
    – roblanf
    Jun 4, 2017 at 10:30
  • 1
    $\begingroup$ Agreed. I'd be happy to rescind my downvote if the question text were updated. $\endgroup$ Jun 8, 2017 at 4:32

5 Answers 5


MultiQC can merge all your different reports into a single one. Which could be useful once you manage to know which QC tools to use.

  • 2
    $\begingroup$ FWIW, Multiqc is also really pretty... $\endgroup$
    – winni2k
    Jun 10, 2017 at 15:18

We routinely run both FastQC and FastQ Screen on all of our raw sequencing reads. FastQ Screen is a tool for detecting cross-species contamination. MGA is another similar tool.

There are then lots of QC tools specific to different types of data, most of which run after alignment. For example RSeQC (RNA data), Qualimap and many many others. Without specifying what kind of data you have this is a bit difficult to make recommendations for though.



The quality control of ngs reads is heavily dependent on type of the project.

For genome assembly projects based on short reads, beside already covered checking quality of sequencing, you would like to look at the kmer spectra to find out, if your reads are going to make sense when they will be translated to De Brujin graph. You will get also a clues about genome coverage, genome size, repetitive content and small clue about heterozygocity. A lot of useful info about interpretation you can find in the README of GenomeScope.

A list tools I used:

  • jellyfish - to count k-mer frequencies
  • GenomeScope - a package for analysis and visualisation of k-mer frequencies (they recommend to use jellyfish for counting k-mer frequencies)
  • kmergenie - for prediction of optimal kmer for assembly

Using these tools can save you a lot of frustration if you accidentally sequence a contaminated sample or if your colleagues / a sequencing company have sent you a wrong file!


The kind of QC you do routinely depends on what your lab's focus is. We do a lot of low-quality, multiplexed DNA and RNA. If you routinely do fresh frozen whole genomes, your QC will be different.

Weighing in from the resequencing side of things (i.e. sequencing an organism that has a good reference), there are several things we are testing with quality control:

  • Did the library preparation succeed? Was the capture successful? Is our input material of high enough quality?
  • Did the sequencing work? For Illumina, did enough clusters pass filter? What's our Q30? Did we sequence into adapters?
  • Can we proceed with analysis? Are the indices on multiplexed samples correct? Do we have the correct reference sequence? Do we have the right targets?

You may use a combination of tools:

  • bcl2fastq : if you multiplexed, check the undetermined indices file after converting to fastq to make sure you caught all of the indexed reads
  • FastQC: General sequencing quality, sequencing into adapters, handy flagging for obvious problems
  • Illumina interop/SAV: Cluster pass filter, Q30
  • Alignment (with BWA, bowtie, etc), bedtools, samtools, Picard: check for contamination, make sure we're using the right reference, overlap with target regions, depth of coverage, soft-clipping for low-quality sequence

If you are looking for NGS QC for your fastq, bam, bed and vcf files I would suggest a commercial tool called omnomicsQ.

It automates coverage analysis and sample pass, warn fail according according to your defined SOPs. Unlike fastqc it also comes with a database so that you can compare protocols, samples, and your performance to peer organisations. It also charts performance over time, exceptions,and correlations between metrics


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.