RNA-seq and expression analysis with 3 iterations in each treatment - cuffmerge stage

[netpumber]

Hi!

First of all i have to draw the experiment.

There are 3 (4 with wt) treatments with 3 iterations each one which have been RNA sequenced.
Treatm_1 (x3)
Treatm_2 (x3)
Treatm_3 (x3)
WT (x3)

The goal is to calculate the differential expression between each treatment and the wt.

The protocol i chose to use is the one referred at stringtie's web page.

#1
So in the first stage i aligned all reads with the reference genome with tophat2

tophat2 -o Treatm1.1_tophat_output --microexon-search -G ~/Projects/RefGenes/genes.gtf ~/Projects/Sequence/Bowtie2Index/genome Treatm1.1_1.fastq Treatm1.1_2.fastq

Did that for each treatment and iteration.

#2

In the second step run the assembler stringtie for each treatment and iteration again.

stringtie -G ~/Projects/RefGenes/genes.gtf -o Treatm1.1_stringtie_output -p 3 Treatm1.1_tophat_output/accepted_hits.bam

Also off the record, in that step stringtie returned that single warning that i couldn't find what really means.

Warning: bundle chr12:70257058-70462447(2179480) (27js) reached coverage saturation (1000000) starting with read mapped at 70260345

#3

In that step my question was born. At stringtie's site they say:

Just like in the Cufflinks protocol, this Cuffmerge run will take as input all the assembled transcripts (GTF files) produced by StringTie for each sample, and also the reference annotation if available.

Should i run the cuffmerge for all treatments and their iterations at once or each iteration from each treatment every time ?

I mean should i run this command: (?)

cuffmerge -g ~/Projects/RefGenes/genes.gtf -p 3 -s ~/Projects/Sequence/Bowtie2Index/genome assembled_list.txt

where assembled_list.txt will have

Treatm1.1_stringtie_output
Treatm1.2_stringtie_output
Treatm1.3_stringtie_output
Treatm2.1_stringtie_output
Treatm2.2_stringtie_output
Treatm2.3_stringtie_output
Treatm3.1_stringtie_output
Treatm3.2_stringtie_output
Treatm3.3_stringtie_output
wt.1_stringtie_output
wt.2_stringtie_output
wt.3_stringtie_output

or this one: (?)

cuffmerge -g ~/Projects/RefGenes/genes.gtf -p 3 -s ~/Projects/Sequence/Bowtie2Index/genome assembled_list.txt

where assembled_list.txt will have

Treatm1.1_stringtie_output
Treatm2.1_stringtie_output
Treatm3.1_stringtie_output
wt.1_stringtie_output

and then run again with these in assembled_list.txt

where assembled_list.txt will have

Treatm1.2_stringtie_output
Treatm2.2_stringtie_output
Treatm3.2_stringtie_output
wt.2_stringtie_output

and finally run it with assembled_list.txt

Treatm1.3_stringtie_output
Treatm2.3_stringtie_output
Treatm3.3_stringtie_output
wt.3_stringtie_output

Thank you.