Question: How Can I Find Transcripts Which Omit Specific Exons Of A Gene?
1
gravatar for Dan D
4.9 years ago by
Dan D6.2k
Tennessee
Dan D6.2k wrote:

Most of my work has been on NGS QC, but lately I've been getting approached with queries for downstream analysis. An investigator was interested in evaluating differential expression of a small set of genes. I ran the investigator's RNASeq datasets through the Tuxedo suite and also through a BWA -> DESeq pipeline to get familiar with established techniques.

Now the investigator wants to focus on expression of a particular murine gene, but only on transcripts which are missing two specific exons. I have the consensus sequences of these exons.

What I'm not sure about is the general strategy to obtain this particular expression information. Having some sort of idea about which tools to investigate would be especially helpful.

ADD COMMENTlink written 4.9 years ago by Dan D6.2k
1

Just a clarifying question: are you trying to identify genes that include/exclude exons between two conditions? Or are you saying that you know the gene you want to look at which expressed several isoforms, and you are trying to deconvolve the expression of each isoform from the "total" expression of the gene?

ADD REPLYlink written 4.9 years ago by Steve Lianoglou4.8k

Good question, Steve. In my case, it's the latter of the two possibilities. Thanks!

ADD REPLYlink written 4.9 years ago by Dan D6.2k
3
gravatar for JC
4.9 years ago by
JC6.2k
Mexico
JC6.2k wrote:

You can try to infer the transcript expressions for each allele with Cufflinks or check differential exon expression with DEXseq in R:Bioconductor.

ADD COMMENTlink written 4.9 years ago by JC6.2k

Thanks! Looks like DEXseq may be the ticket. I'll have a look right now.

ADD REPLYlink written 4.9 years ago by Dan D6.2k
1

FWIW, if you are really looking to do the "latter" part of what I said as a comment to your original question, then DEXSeq won't be the ticket. DEXSeq is appropriate when you want to identify exons that are differentially expressed (the first part of what I said in the comment)

ADD REPLYlink written 4.9 years ago by Steve Lianoglou4.8k

Cool, thanks for the heads up. I'm deep in the documentation right now. I'm also reading through the Tuxedo suite docs, but I haven't yet determined a strategy.

ADD REPLYlink written 4.9 years ago by Dan D6.2k
1

if you want to deconvolve isoform expressions, then Cufflinks it's the ticket

ADD REPLYlink written 4.9 years ago by JC6.2k

Excellent. I got sidetracked on other projects but I will revisit the cufflinks output to see what I need to do. I'll report back on this thread when I've made progress. Many thanks!

ADD REPLYlink written 4.9 years ago by Dan D6.2k

I have the same problem, DEXseq will tell which exons are different but not which junctions are different. Other than manually looking at the bam files in IGV, is there a computational way to see how many reads span 1-3 rather than 1-2-3?

ADD REPLYlink written 4.9 years ago by Dave Bridges1.3k
1

maybe you can extract the reads in the region and separating them by the CIGAR information with some script

ADD REPLYlink written 4.9 years ago by JC6.2k
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: 651 users visited in the last hour