SEQanswers

Go Back   SEQanswers > Bioinformatics > Bioinformatics



Similar Threads
Thread Thread Starter Forum Replies Last Post
Advice on RNA-Seq Depth KDS RNA Sequencing 3 02-19-2014 03:03 PM
Genomatix RNA-Seq workflow rebrendi Bioinformatics 1 02-13-2012 07:13 AM
RNA-seq Galaxy workflow for PE barcoded samples? jjw14 Bioinformatics 0 04-19-2011 12:58 PM
RNA-seq workflow foxyg Bioinformatics 1 11-11-2010 03:26 PM
Splicing Analysis in Partek RNA-Seq workflow xguo Bioinformatics 1 12-21-2009 09:06 AM

Reply
 
Thread Tools
Old 12-15-2011, 02:10 PM   #1
slavailn
Junior Member
 
Location: Lethbridge

Join Date: Mar 2011
Posts: 8
Default advice/critique on my RNA-seq workflow

Hi guys,

Would love some feedback on my RNA-seq workflow. I've got to analyze NGS dataset, and basically, I'm a noob in NGS - a bench researchers, who wants to move closer to bioinformatics.

The goal of the experiment is to find differentially expressed genes between two treatment groups, I've got ~ 10 mil reads per sample, 36 cycles, barcoded.

Here is the workflow:

1. Examined quality in FastQC;
2. Trimmed adapters (I used cutadapt program);
3. Aligned reads to the genome with bowtie: bowtie -v 2 --best -s <ebwt> <reads> <sam>

4. Converted sam to bam:
samtools -view -bS <sam> <bam>

5. Downloaded refseq gene annotation from UCSC in bed format from gene and prediction tracks using UCSC table browser;

6. This step bothers me a bit. As far as I understand UCSC views genes as transcripts, so there are multiple entries listed for the same or closely overlapping genomic coordinates. As the result I'll end up with the same read counts vs basically same gene, which is listed under multiple transcript names. I suspect this could mess with the size factors estimation in DESeq.
So I merged overlapping transcripts in order to avoid this ambiguity in read counting using bedtools software, which I really came to appreciate:
mergeBed -nms -i <bed> > output_bed: -nms option lists all of the merged entries opposite to new set of genomic coordinates

The problem with merging is that some genuinely different genes may have small overlaps and will be merged into one feature. But I hope that this is a minor risk and I'll be able to examine differentially expressed genes manually.

7. Finally, I've obtained read count for each of the merged features using bedtools:
intersectBed -abam <bam> -b refseq_exons_hg19.bed -bed -split -f 1.0 | intersectBed -a stdin -b merged_refseq_hg19.bed | coverageBed -a stdin -b merged_refseq_hg19.bed > file.coverage

This sequence of commands should obtain only those reads which have 100% overlaps with known refseq exons (uploaded from UCSC), intersect those with whole genes and count raw read coverage for each gene;

8. And lastly, I'm going to use DESeq to detect differential expression.

Thanks in advance!
Slava.
slavailn is offline   Reply With Quote
Old 12-15-2011, 02:32 PM   #2
biznatch
Senior Member
 
Location: Canada

Join Date: Nov 2010
Posts: 124
Default

You probably want to align your RNA-seq data with Tophat not Bowtie.

Assuming you're using something like mouse or human, ~10 million reads may not be enough to quantify lower expressing genes.

I found that trimming didn't really increase the percent of reads that align. I've only done a little RNA-seq analysis but you can see some of the stuff I've tried here.
biznatch is offline   Reply With Quote
Old 12-15-2011, 02:51 PM   #3
slavailn
Junior Member
 
Location: Lethbridge

Join Date: Mar 2011
Posts: 8
Default

Thanks for your suggestions biznatch!
I've tried to use Tophat, but it looks like its geared to align longer reads, since it wants to split them into multiple segments and align those separately. With about 29 nt I was getting really weird results, for instance my resulting bam files contained more reads aligned than I had initially in the fastq file. With bowtie I'm getting 80 - 90% alignment.
Adapter trimming in this case did not do much since very few reads contained adapters and reads were of high quality.
slavailn is offline   Reply With Quote
Old 12-15-2011, 04:20 PM   #4
zee
NGS specialist
 
Location: Malaysia

Join Date: Apr 2008
Posts: 249
Default

Take a look at our protocol at http://www.novocraft.com/wiki/tiki-i...page_ref_id=35 . The output is a SAM file that may be used for calculating expression, gene fusions and splice junctions.
zee is offline   Reply With Quote
Reply

Thread Tools

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off




All times are GMT -8. The time now is 04:35 PM.


Powered by vBulletin® Version 3.8.9
Copyright ©2000 - 2019, vBulletin Solutions, Inc.
Single Sign On provided by vBSSO