Question: multiple conditions analysis workflow rna seq
0
gravatar for novicebioinforesearcher
3.7 years ago by

Hello All, I need to analyze certain rna seq data. In the past I have done a simple wt(n=3) vs mutant(n=3) analysis following these steps(please correct me if this is not right):

  1. Trimming: trimmomatic
  2. Quality : fastqc, picard
  3. alignment : Hisat2 align each replicate as is
  4. quatification: stringtie
  5. differential gene expression: use the -Aflag in stringtie and prepDE.py python file to build a matrix compatible with deseq2 and design would be WT,WT,WT,MT,MT,MT
  6. gsea and functional analysis based on significant gene from deseq2

Where as now i need to perform RNAseq analysis where my design in something like this

condition/timepoints    seq-lib
1   paired-end
1   paired-end
1   paired-end
2   paired-end
2   paired-end
2   paired-end
3   paired-end
3   paired-end
3   paired-end
4   paired-end
4   paired-end
4   paired-end
5   paired-end
5   paired-end
5   paired-end
6   paired-end
6   paired-end
6   paired-end
7   paired-end
7   paired-end
7   paired-end
8   paired-end
8   paired-end
8   paired-end

what steps would vary and do i need to perform for such a experimental design , do i need to merge transcripts in stringtie?

How do I go about deseq2 design I may be wrong here but is it right to say

1vs2,3,4,5,6,7,8
 2vs1,3,4,5,6,7,8

so on and so forth and if yes how do i achieve this?

Many thanks

rna-seq • 1.6k views
ADD COMMENTlink modified 3.7 years ago • written 3.7 years ago by novicebioinforesearcher60

Hi there!

I think you should look at the edgeR manual, they have examples similar to what you have using generalized linear model and test DE genes with a likelihood ratio test, but really is deppending on what you've got and what questions you want to answer.

ADD REPLYlink written 3.7 years ago by biofalconch470

Hi there,

I was hoping to see differences between each group and was wondering what would be a best way to do it.

ADD REPLYlink written 3.7 years ago by novicebioinforesearcher60
1
gravatar for biofalconch
3.7 years ago by
biofalconch470
Mexico
biofalconch470 wrote:

So it will be All vs All in a pairwise manner? What I think it would be best is to load all the data into a DGE object using edgeR package, afterwards fit your data using glmFit(), and after that use the function glmLRT() to test for DE, using contrast to indicate which samples you want to compare. For example, if you wish to compare sample 2 with 4, you should make the contrast c(0,-1,0,1). I suggest reading section 2.10 of the edgeR manual :)

ADD COMMENTlink written 3.7 years ago by biofalconch470

Thanks will try it and hope i an make it work

ADD REPLYlink written 3.7 years ago by novicebioinforesearcher60

can you repost your reply as an answer so that I can accept it (sorry just following biostar best practice)

ADD REPLYlink written 3.7 years ago by novicebioinforesearcher60

Moved comment to answer per request. Accept only after testing to make sure solution answers your question.

ADD REPLYlink modified 3.7 years ago • written 3.7 years ago by GenoMax96k
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: 2604 users visited in the last hour
_