Question: small RNA-seq pipelines
2
gravatar for Saad Khan
4.3 years ago by
Saad Khan330
United States
Saad Khan330 wrote:

Hi,

I have to analyze a small RNA-seq data. This is my first time working with small RNA. I wanted to do differential expression and micro-rna comparison along with GO annotation. Something that mirtools does. But mirtools takes fasta files and I have paired end fastq files. I was wondering which publicly available pipelines (based on ease of usage and installation) can anybody recommend to do a comprehensive small RNA-seq analysis. 

ADD COMMENTlink modified 4.3 years ago by Mike Axtell240 • written 4.3 years ago by Saad Khan330
1
gravatar for Brian Bushnell
4.3 years ago by
Walnut Creek, USA
Brian Bushnell16k wrote:

If you want something standalone, easy to use, and easy to install, I suggest the BBMap package - it's already compiled and will run on any platform that has Java.  It does not do annotation, but it does do adapter-trimming with bbduk.sh (which is important for micro-RNA) and alignment with bbmap.sh, or alignment-free RNA quantification against a transcriptome with seal.sh.  You could also use reformat.sh to convert your fastq files to fasta.

ADD COMMENTlink written 4.3 years ago by Brian Bushnell16k

How is the alignment of small rna reads going to be different from general RNA-seq reads? What bbmap options would be specific to small rna read mapping?

ADD REPLYlink written 4.2 years ago by Saad Khan330

Mainly BBMap is just easy to install and use, which is what you asked for.  And it supports direct output of coverage and RPKM, which is often useful for RNA-seq.

ADD REPLYlink written 4.2 years ago by Brian Bushnell16k

What I am really asking is what parameters should I change or specifically use for aligning small RNA reads when using bbmap?

ADD REPLYlink written 4.2 years ago by Saad Khan330

Oh, the defaults are usually fine.  You could set "maxindel=20" if you want to speed it up a little since the defaults are geared toward  long RNAs, and small RNAs don't contain long introns.  Just be sure to adapter-trim the reads first, because small RNAs are expected to have adapter sequence, which interferes with mapping.

ADD REPLYlink written 4.2 years ago by Brian Bushnell16k

So once I have mapped the reads I'll have to depend on other tools to do further analysis. Is there a way I can leverage other tools like srnabench once I have the mapped sam file or do any functionalities exist in BBtools to do so.

ADD REPLYlink written 4.2 years ago by Saad Khan330

BBTools can output coverage across the genome, or generate RPKMs if you are mapping to a transcriptome.  And it can produce various metrics on coverage, like the fraction of bases across a transcript that were covered, but it does not do most RNA-seq-specific things like differential expression analysis.  Many RNA analyses use mapped sam files, but I don't personally do small-RNA analysis so I'm not sure which downstream tools are best.

ADD REPLYlink written 4.2 years ago by Brian Bushnell16k
1
gravatar for Mike Axtell
4.3 years ago by
Mike Axtell240
Penn State University
Mike Axtell240 wrote:

You could try Shortstack ..https://github.com/MikeAxtell/ShortStack , http://www.ncbi.nlm.nih.gov/pubmed/23610128

ADD COMMENTlink written 4.3 years ago by Mike Axtell240

Does not support paired end reads!!

ADD REPLYlink written 4.2 years ago by Saad Khan330
0
gravatar for gaurav.singh
4.3 years ago by
gaurav.singh10
United States
gaurav.singh10 wrote:

http://mirnablog.com/web-based-tools-for-microrna-analysis-from-deep-sequencing-data/ 

ADD COMMENTlink written 4.3 years ago by gaurav.singh10

I am looking for standalone applications not webservers

ADD REPLYlink written 4.3 years ago by Saad Khan330
0
gravatar for David Langenberger
4.3 years ago by
Deutschland
David Langenberger8.8k wrote:

You could try sRNAbench.

But paired-end sequencing ist also not very common when doing small RNA-Seq. How long are your reads?

ADD COMMENTlink written 4.3 years ago by David Langenberger8.8k

BTW: sRNAbench is also available as standalone: http://arn.ugr.es/sRNAbench/

ADD REPLYlink written 4.3 years ago by David Langenberger8.8k

I believe sRNAbench does not handles paired end reads atleast according to their documentation

ADD REPLYlink written 4.3 years ago by Saad Khan330

But I actually do not see any sense in using paired-end sequencing for molecules of ~18-30 nt in length. Why should someone do that? To waste money?

ADD REPLYlink modified 4.3 years ago • written 4.3 years ago by David Langenberger8.8k

I would have to ask my collaborator that :( . BTW the length of the sequences in the fastq files is 50nt I am sure it does includes some adapter sequences that I would have to remove.

ADD REPLYlink written 4.3 years ago by Saad Khan330

That is actually a very important tipp: Know your data, before you do any analysis!

ADD REPLYlink written 4.3 years ago by David Langenberger8.8k

Well it was just a one-off analysis I had to do for my collaborator to add on some results to a paper. So I don't know much about why they used paired end (Which I have read on Illumina website don't help much in case of small RNA or miRNA).

ADD REPLYlink written 4.2 years ago by Saad Khan330

BTW do you happen to know of any sirna pipeline that works with paired end data?

ADD REPLYlink written 4.2 years ago by Saad Khan330

No, since paired-end makes no sense in my eyes.

But if your paired-end data is strand-specific (what I really hope you checked), then you can just create a single-end file out of it. mate1 is in the correct direction and mate2 is directed in reverse direction. So, you can just take mate1.fastq and add the reverse complement of mate2.fastq. Assure to do that after clipping the adapters! 

I'm not sure, but probably there are some tools available which do the re-direction of mate2 for you. But I never needed it, so I don't know. But you can use the resulting file single-end file with all small RNA-Seq tools available.

ADD REPLYlink written 4.2 years ago by David Langenberger8.8k

What if they are not strand specific. What can I do with such data then??

ADD REPLYlink written 4.2 years ago by Saad Khan330

Not that much, I would say

ADD REPLYlink written 4.2 years ago by David Langenberger8.8k

What should I do then. I was wondering just aligning after removing the reads and using edger etc for differential expression.

ADD REPLYlink written 4.2 years ago by Saad Khan330

Just adapter-trim, then throw away read 2 and use read 1.

ADD REPLYlink written 4.2 years ago by Brian Bushnell16k

How about using the 2nd read pair as a replicate??

ADD REPLYlink written 4.2 years ago by Saad Khan330

It's not a replicate, it's the same; so no.

ADD REPLYlink modified 4.2 years ago • written 4.2 years ago by Brian Bushnell16k

Ask the lab, which preparation protocol they used!

ADD REPLYlink written 4.2 years ago by David Langenberger8.8k

Actually I just found out it is strand specific data. Do you know who can direct me to a package/tool that does the redirection of mate2?

 

ADD REPLYlink written 4.2 years ago by Saad Khan330

Hi David,

I have finally done the mapping etc for small rna. Since sRNAbench does not handle data which has no replicates (like in my case) what would be a good procedure to use in order to give the required results. Since both edgeR and DEseq handle replicate free data for RNA-seq do you think these techniques can be extended to small rna as well.

ADD REPLYlink written 4.2 years ago by Saad Khan330
0
gravatar for Vivek Todur
4.3 years ago by
Vivek Todur50
Bangalore
Vivek Todur50 wrote:

The UEA small RNA Workbench is another option, Its cross platform GUI (Graphical User Interface) application, You can run either on Linux or Windows seamlessly. It has different modules for different analysis such as Mirprof for Known miRNA Analysis and Mircat for Novel miRNA Analysis.

ADD COMMENTlink written 4.3 years ago by Vivek Todur50
Please log in to add an answer.

Help
Access

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 2.3.0
Traffic: 844 users visited in the last hour