[IBseq]

Hi Artur,
yes, that makes sense. Thus I shall not take in consideration the p/q values and their consequent significance.

I did follow the tophat/cufflinks/cuffcompare/cuffdiff pipeline, but I had a doubt and ran it twice with a difference. I shall outline that I have one control sample and 7 treated samples but I am not using them as replicates, thus I ran them separately each of them against the control.

1st run: tophat.cufflinks with refence annotation; then cuffcompare with the cufflinks of all 7 samples and control sample plus reference annotation. I used this output of couffcompare as input for 7 cuffdiffs (one for each of the 7 treated samples).

2nd run: tophat/cufflinks with reference annotation, then cuffcompare with cufflinks of each of the treated samples with reference annotation (thus I have 7 cuffcompare outputs). I used one by one each of these cuffcompare outputs to run 7 cuffdiffs.

My question: I have more consistent results when I run a cuffcompare which includes all my cufflinks (control + 7 treated) rather than cifflink control+cufflink treated one by one. Similarly, I do have more interesting results if I run 7 different cuffdiff rather thn one cuffdiff by taking my 7 treated samples as replicates (these 7 samples are 7 people biopsis of the same tumor)

which way is best? I can use the first run and take it as reliable?

sorry for long question...Hope you can help.

thanks,
ib

[tankman]

interpretation of FPKM tracking files at gene level

hi guys,

quick question for you cuffdiff-output experts:

I'm having trouble understanding why the gene level fpkm tracking file shouldn't be unique in the gene short name. That's what I thought the description entailed (sums over isoforms to get gene level FPKM matrix) from the cufflinks website:

genes.fpkm_tracking Gene FPKMs. Tracks the summed FPKM of transcripts sharing each gene_id


Instead, e.g., I get

users-mac-pro:cuffdiff_SE_plus_PE_trimmed_all_samples user$ grep ADAR genes.fpkm_tracking | cut -f 1-6
XLOC_000826 - - XLOC_000826 ADAR TSS1470
XLOC_002045 - - XLOC_002045 ADAR TSS3725,TSS3726,TSS3727,TSS3728
XLOC_003087 - - XLOC_003087 ADARB2 TSS5153
XLOC_003599 - - XLOC_003599 ADARB2 TSS6095,TSS6096,TSS6097,TSS6098
XLOC_019202 - - XLOC_019202 ADARB1 TSS32453,TSS32454,TSS32455
XLOC_019369 - - XLOC_019369 ADARB1 TSS32746


All the TSS id's correspond to the same physical locus. I am misunderstanding why gene short id is different than gene id. Can you please help me interpret this data?

thanks,
tankman