Hello,
FastQC ( http://www.bioinformatics.bbsrc.ac.uk/projects/fastqc/ ) provide an overview what is wrong with a Fastq file.
However, it does not provide any tools to fix the problems. What kind of tools have to be used to fix the problems?
Thank you in advance.
FastQC is just an informative tool, very useful in my opinion to evaluate how the lab sequencing jobs have performed before going into mapping. but as any informative tool, it just tells you what's going on down there, and of course you are the one ultimately having to curate those reads in case you need to. but you'll have to know that there are things that may be curated, and there are others that you won't be able to do anything with them.
since the available tests performed by FastQC are multiple, there isn't a particular to fix the reported errors if any. things like the "per base sequence quality" for instance may be solved by simple sequence trimming if the reported error is that the last 3 bases are always wrong, so they should always be removed from posterior analysis. other error reports may be also solvable by simple scripting, but most of them, such as "per sequence GC contect", are purely descriptive, and in case they fail to pass the FastQC thresholds there's nothing you can do about them but to repeat the lab sequencing if wanted.
The FASTX toolkit may help in some cases; there are tools to filter, trim and clip low-quality sequence. There are also some tools around to "heal" sequencing reads, which may be able to correct platform-dependent systematic errors.
However, as Jorge says, QC generally tells you whether or not something quite fundamental is wrong at the experimental level. If it is, you can't always just "fix" it computationally.
You could try https://github.com/optimuscoprime/autoadapt
It will remove contaminant adaptors and primers, as identified by FastQC, as well as removing low quality sequences
For Illumina Sequencing i would suggest Trimmomatic, it is kind of designed for illumina and has more specific options. However, fastX is very useful!
Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
version 2.3.6
+1 here. I guess I focused too much in how I think FastQC results should be taken into account, rather than in answering the question itself. yes, FASTX toolkit would be the approach I too would recommend for anyone willing to play around with fastq file, plus its integration in Galaxy may reduce the slope of its learning curve.