When "Quality trimming using Trimmomatic" is really needed?

[Farbod Emami]

Dear Brian, Hi.

First of all let me thank you for this new version of Trinity , it seems it is more easy to install and the documentation archiving seems better than the old one!

I am using the Trinity package for de novo assembly of the transcriptome of a non-model fish. So as I gave many reads in my illumina paired-end fastq, I usually use the "--normalize_reads " parameter.

Before running Trinity, I have checked the quality of my fastq files with fastQC program and there were no adapter or low quality situations.

Do I need to run "--quality_trimming_params " ? 

Does it has any effect of improving the assembly procedure (e.g creating  robust contigs or better isoforms or . . . ) or not ?

Thank you again

[Mark Chapman]

Hi Farbod,

My experience is that quality trimming can't hurt. You're not going to lose anything, and might gain a few more robust contigs at the expense of split or poorly supported ones. My advice would be to try a parallel assembly with trimming and see if the results look particularly different. If you're starting from scratch Id say definitely do it but if you're wondering about an older assembly then maybe see if the results are particularly different before thinking you need to redo every analysis.

BW, Mark

--
You received this message because you are subscribed to the Google Groups "trinityrnaseq-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to trinityrnaseq-u...@googlegroups.com.
To post to this group, send email to trinityrn...@googlegroups.com.
Visit this group at http://groups.google.com/group/trinityrnaseq-users.
For more options, visit https://groups.google.com/d/optout.

--

Dr. Mark A. Chapman

M.Ch...@soton.ac.uk

+44 (0)2380 594396

------------------------------------

Centre for Biological Sciences
University of Southampton

Life Sciences Building 85
Highfield Campus
Southampton
SO17 1BJ

[Farbod Emami]

Dear Dr. Mark

Hi .

Yes, I am trying the new version of Trinity to assembly again from beginning, :)

So, 

1- what parameter for "--trimmomatic" do you suggest?

2- for normalization I usually use just "--normalize_reads" , is it enough or I must add any extera parameter wit it?

Thank you and best wishes

Farbod

On Thursday, October 1, 2015 at 11:37:07 AM UTC+3:30, Mark Chapman wrote:

Hi Farbod,

My experience is that quality trimming can't hurt. You're not going to lose anything, and might gain a few more robust contigs at the expense of split or poorly supported ones. My advice would be to try a parallel assembly with trimming and see if the results look particularly different. If you're starting from scratch Id say definitely do it but if you're wondering about an older assembly then maybe see if the results are particularly different before thinking you need to redo every analysis.

BW, Mark

On 1 October 2015 at 08:45, Farbod Emami <farbo...@gmail.com> wrote:

Dear Brian, Hi.

First of all let me thank you for this new version of Trinity , it seems it is more easy to install and the documentation archiving seems better than the old one!

I am using the Trinity package for de novo assembly of the transcriptome of a non-model fish. So as I gave many reads in my illumina paired-end fastq, I usually use the "--normalize_reads " parameter.

Before running Trinity, I have checked the quality of my fastq files with fastQC program and there were no adapter or low quality situations.

Do I need to run "--quality_trimming_params " ? 

Does it has any effect of improving the assembly procedure (e.g creating  robust contigs or better isoforms or . . . ) or not ?

Thank you again

--
You received this message because you are subscribed to the Google Groups "trinityrnaseq-users" group.

To unsubscribe from this group and stop receiving emails from it, send an email to trinityrnaseq-users+unsub...@googlegroups.com.

To post to this group, send email to trinityrn...@googlegroups.com.
Visit this group at http://groups.google.com/group/trinityrnaseq-users.
For more options, visit https://groups.google.com/d/optout.

[Matthew MacManes]

Mark and others, 

You’ve just touched upon my pet-peeve issue.. trimming.. Quality trimming CAN and DOES (often) make assemblies worse. See http://journal.frontiersin.org/article/10.3389/fgene.2014.00013/abstract for a description of the effects. 

Trinity includes the trimming recommendations from the above paper. It would be surprising to me that raw Illumina data would NOT contain any adapters, even if FastQC says otherwise. I’d go ahead and trim using the Trinity default using the --trimmomatic` command without specifying anything else, which will take care of bases with a Phred score <5, and adapters. You can combine this with --normalize_reads also passed to Trinity without other modification. 

Matt

______________________________________________
Matthew MacManes, Ph.D. 
University of New Hampshire  I  Assistant Professor of Genome Enabled Biology
Department of Molecular, Cellular, & Biomedical Sciences
Durham, NH  03824
Phone: 603-862-4052  I  Twitter: @macmanes | Web: genomebio.org

Office: 189 Rudman Hall | Laboratory: 145 Rudman Hall

To unsubscribe from this group and stop receiving emails from it, send an email to trinityrnaseq-u...@googlegroups.com.

[Tiago Hori]

I'd say it also depends on how complex your transcriptome is. Our experience with some tetraploid fish genomes is that if your error rate increases, you get more mis-assembled chimeric contigs. We did trim more aggressively in that instance and it did help.

One thing to remember is that Trinity does not greedily extend the reads but rather the identified kmers and in the de-brujin graph the kmer is really the assembly unit not the read and that greatly reduces the impact of lower quality bases in the assembly.

T.

"Profanity the is the only language all programmers understand" 

Sent from my iPhone, the universal excuse for my poor spelling.

[Farbod Emami]

Thank you all,

So it seems that it is a "YES" answer to perform trimming.

and the preferred way is to just use "--trimmomatic" without any extra parameters in the script, right?

[Farbod Emami]

Dear Mark and other friends, Hi

I have finished a de novo assembly with --normalize and without trimmomatic first, and now I am performing another assembly with BOTH --normalize and --trimmomatic on the same data (it will be finish tomorrow).

So, what parameters (e.g in  TrinityStat output)  I must compare to see which assembly is much better ? N50 ? number of contigs ? number of Longest Isoform Per Gene ? 

which parameter?

Thank you in advance

Farbod

[Mark Chapman]

Hi Farbod,

As discussed frequently on here there's no single statistic from Trinity that says assembly A is better or worse than assembly B. You can use detonate to compare your assemblies if you want to do this comprehensively. Some information from the trinity output might be useful but doesn't prove anything - eg if the N50 of one assembly was 500 and the N50 of another based on the same data was 20 then i would presume there's a problem with the latter, but slight differences are hard to interpret. If your N50 goes up a bit does this mean that you're assembling contigs better and longer? or does it mean you have some erroneous contigs where paralogous loci have been joined?

So I'd say interpret your Trinity output with a pinch of salt...

BW, Mark

Farbod

Matt

To unsubscribe from this group and stop receiving emails from it, send an email to trinityrnaseq-u...@googlegroups.com.

To post to this group, send email to trinityrn...@googlegroups.com.
Visit this group at http://groups.google.com/group/trinityrnaseq-users.
For more options, visit https://groups.google.com/d/optout.

--

Dr. Mark A. Chapman

M.Ch...@soton.ac.uk

+44 (0)2380 594396

------------------------------------

Centre for Biological Sciences
University of Southampton

Life Sciences Building 85
Highfield Campus
Southampton
SO17 1BJ

--
You received this message because you are subscribed to the Google Groups "trinityrnaseq-users" group.

To unsubscribe from this group and stop receiving emails from it, send an email to trinityrnaseq-u...@googlegroups.com.

To post to this group, send email to trinityrn...@googlegroups.com.
Visit this group at http://groups.google.com/group/trinityrnaseq-users.
For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "trinityrnaseq-users" group.

To unsubscribe from this group and stop receiving emails from it, send an email to trinityrnaseq-u...@googlegroups.com.

To post to this group, send email to trinityrn...@googlegroups.com.
Visit this group at http://groups.google.com/group/trinityrnaseq-users.
For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "trinityrnaseq-users" group.

To unsubscribe from this group and stop receiving emails from it, send an email to trinityrnaseq-u...@googlegroups.com.

To post to this group, send email to trinityrn...@googlegroups.com.
Visit this group at http://groups.google.com/group/trinityrnaseq-users.
For more options, visit https://groups.google.com/d/optout.

[Tiago Hori]

The new Trinity wiki has a real nice section in assembly quality assessment, I suggest reading it.

T.

"Profanity the is the only language all programmers understand" 

Sent from my iPhone, the universal excuse for my poor spelling.

[Ken Field]

I would also add BUSCO to that list. It is a great way to compare the representation of certain highly conserved genes in the two assemblies.

http://busco.ezlab.org/

Ken

Ken Field, Ph.D.

Associate Professor of Biology

Program in Cell Biology/Biochemistry

Bucknell University

Room 203A Biology Building

[Tiago Hori]

Ken,

I was taking a quick peek at BUSCO and I could not get a clear picture of wether you do or do not need a reference transcriptome to use it, can give us a brief blurb of how it works?

T.

"Profanity the is the only language all programmers understand" 

Sent from my iPhone, the universal excuse for my poor spelling.