Quality-, adapter- and RRBS-trimming with Trim Galore!

[fkrueger]

That path seems to be lacking /cutadapt at the end. Could you try that?

[mvijayen]

That worked just fine! Thanks, Felix!

[maria.gr]

Hi guys !
I m new to trim_galore and I have a small issue introducing cutadapt in the command line of trim_galore :
Could you tell me how I should modify this line so that I introduce the cutadapt directory?

>>
./trim_galore --paired --retain_unpaired --quality 15 --fastqc_args $path_to_cutadapt='build/cutadapt' --fastqc my1.fastq.gz my2.fastq.gz

[fkrueger]

To supply the path to cutadapt you need to edit trim galore in a text editor and change the path as one of the first lines.

[maria.gr]

Thanks ! Found it , but it gave again error of line 471...
Actually I realised that I had to make the file executable ('chmod a+x build/cutadapt/bin/cutadapt' )...
Don't know if it's needed for everyone after downloading the cutadapt, but I say it in case sb has the same problem
So now it runs normally !
Thanks again!

[Kmok]

New Bee on Trim Galore

I use Trim Galore to trim an exome seq data captured with Illumina Nextera. The script used is
$myTrimGalore -q 15 -a CTGTCTCTTATACACATCT --stringency 3 --length 20 -e 0.1 -o $myoutDir --fastqc_args "--outdir $myoutDir" --dont_gzip --paired $myfastq1 $myfastq1.

The Fastqc results after running Trim Galore show there are bias in the nucleotides in the first 15bp (perbaseSequence). I guess this may be related to the non-random binding of transposase. There are over-representations of Kmer also at the 5' as well as in the middle of the sequence. Can anyone help in telling me what is the cause of the Kmers ( ? adapters ?indexes)? How should these be trimmed if they are adapters or indexes?

Thanks in advance

[Brian Bushnell]

I don't know about the later anomalies, but in my tests, Nextera seems to have highly irregular base frequencies for the first ~20bp (as you say, probably due to non-random binding). They are still fairly accurate and do not need to be trimmed.

It's possible that the later peaks are due to primer-dimers or other such artifacts. What is the insert-size distribution of the library?

[Kmok]

Hi Brain

Thanks very much.
I need to ask our lab on the QC of the library. How can we guess it is dimer from the library insert distribution? Is it the a peak of same size as we seen in the later peaks?

Kin

[Brian Bushnell]

Well... if you plot the insert size histogram, and see very sharp peaks at certain lengths, those may be some kind of non-genomic molecules. And once you know the length, you might be able to guess what they are considering all the different reagents that were used. Or you could look at reads with those specific insert sizes and see what the sequence is, to determine what they are. Once you know, you can easily filter them out (digitally). That is of course IF there are sharp peaks in the insert size histogram.

If they are non-genomic artifacts, you won't find them in the insert size histogram you would get from mapping, because they won't map. But (if you have paired reads) you can generate an insert size histogram by overlapping them with BBMerge.

[Kmok]

Thanks Brian
Shall try to run with BBMerge and see

[fkrueger]

First of all apologies for not having released Trim Galore updates lately, I seem to have somehow always postponed and then forgotten them entirely...

A new version of Trim Galore (v0.3.6) is now available from its project page (http://www.bioinformatics.babraham.a...s/trim_galore/), which adds several features and fixes:

- Added the new options '--three_prime_clip_r1' and '--three_prime_clip_r2' to clip any number of bases from the 3' end after adapter/quality trimming has completed
- Added a check to see if Cutadapt exits fine. Else, Trim Galore will bail a well
- The option '--stringency' needs to be spelled out now since using -s was ambiguous because of '--suppress_warn'
- Added the Trim Galore version number to the summary report
- Added single-end or paired-end mode to the summary report
- In paired-end mode, the Read 1 summary report will no longer state that no sequence have been discarded due to trimming. This will be stated in the trimming report of Read 2 once the validation step has been completed

(Edit: The manual needs a little updating, too, I'll work on that...)

[fkrueger]

I have just released a small fix to Trim Galore (v0.3.7) that makes paired-end trimming work again (which I had accidentally broken by introducing a small change...). The manual has now also been updated.

Please find the latest release here: https://www.bioinformatics.babraham....s/trim_galore/

[shawpa]

I have used trim galore before using Bismark many times. It just occurred to me though that there might be a problem in my use of the pipeline. What I usually do is trim the reads (both adaptor and quality trimming), then align with Bismark, remove duplicate reads using the deduplicatebismark script provided, then proceed with methylation calling. However, if I am trimming for quality, I am changing the start and end coordinates of the read, which I think would affect the detection of duplicate reads. Could someone please let me know if this is correct? Is trimming for quality, going to adversely affect the detection of duplicate reads?

[fkrueger]

I have used trim galore before using Bismark many times. It just occurred to me though that there might be a problem in my use of the pipeline. What I usually do is trim the reads (both adaptor and quality trimming), then align with Bismark, remove duplicate reads using the deduplicatebismark script provided, then proceed with methylation calling. However, if I am trimming for quality, I am changing the start and end coordinates of the read, which I think would affect the detection of duplicate reads. Could someone please let me know if this is correct? Is trimming for quality, going to adversely affect the detection of duplicate reads?

No, trimming should not affect the deduplication:

Single-end deduplication uses the chromosome, the start coordinate and the orientation of a read. Since you are trimming from the 3' end of a read this has no influence on the start coordinate. (for reverse reads the start coordinate is calculated by adding the read length (using the CIGAR string for gapped alignments if required)).

Paired-end deduplication uses the chromosome, the start coordinate of read 1, the end coordinate of read 2 and the orientation of the read pair (determined by read 1). Again, since you are trimming from the 3' end of both reads the relevant parameters are not affected.

[rzwu0721]

Hi, I am using the software named CLC Genomics Workbench, and it can trim the adapter for just need several minutes, eg. CTGTCTCTTATACACATCT you have mentioned above.So I would recommend you can try to use it.

Best Wishes!
Renzhi Woo,
Guangxi Academy of Sciences