Question: Differential Expression of Single Gene
0
gravatar for gtasource
6 weeks ago by
gtasource20
gtasource20 wrote:

Hi all,

I have 6 aligned BAM files for a species I am working with. We have a single gene we want to see if it's differential. I'd rather not have to run an entire DE analysis when we just want to see if one specific gene is differential. My plan was to extract the gene coordinates and run BedTools coverage to get raw counts. From there, I'd do some type of statistical test (not sure which one) to see if there is a significant difference in the counts.

ADD COMMENTlink modified 6 weeks ago by geek_y9.3k • written 6 weeks ago by gtasource20
2

This does not really sound like a good approach to me.

For any DEG analysis, the mapping step is the largest part of the analysis and since you apparently already have that I don't think you will gain much be filtering before running DEG. It might even be so that your statistics will turn out unreliable (eg. the library normalisation).

Personally I would run the whole DEG and only then select your gene of interest.

ADD REPLYlink written 6 weeks ago by lieven.sterck4.1k
1

How were the BAMs aligned? I would not run any statistical test on raw counts. Why not take the opportunity to do a 'robust' (yep, that magical word again...) differential expression analysis for all genes and have that resource for future purposes? What if, in 1 month, you want to go back and check another gene?

With an aligned BAM from, e.g., STAR, you could run featureCounts or HTseq for the purposes of obtaining raw counts over transcripts listed in an input GTF. You could then normalise these raw counts in EdgeR or DESeq2.

ADD REPLYlink modified 6 weeks ago • written 6 weeks ago by Kevin Blighe39k

this could have been a (bad though) solution for the problem, but you have to normalize by sequencing depth. Otherwise the counts for 1 gene don't tell you anything.

ADD REPLYlink written 6 weeks ago by grant.hovhannisyan1.4k
0
gravatar for geek_y
6 weeks ago by
geek_y9.3k
Barcelona/CRG/London/Imperial
geek_y9.3k wrote:

Best is to run differential analysis on all the data and check for your gene of interest in the DE results.

ADD COMMENTlink written 6 weeks ago by geek_y9.3k
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: 1894 users visited in the last hour