How to get counts for transcripts for all replicates

Question: How to get counts for transcripts for all replicates

7 months ago by

j.m.fustin • 10 wrote:

Hi BIOSTARS,

I have lots of samples that have been aligned to mm10 using RNA STAR. I have successfully assembled transcripts and merge the assembly for all samples using cufflinks. Although Cuffdiff has also worked I would to analyze data with a different statistical test that is not limited to pairwise comparisons. So, I would like to get a table similar to the "transcripts FPKM tracking" output of cudiff but, instead of having only the average for each group of replicates, I would like to have the FPKM value for each replicate. This table will then be analysed for splicing differences, so ideally the table should show fpkm values for isoforms of each gene.

Any suggestions?

Regards

assembly cufflinks fpkm rna-seq cuffdiff • 481 views

ADD COMMENT • link •

modified 7 months ago • written 7 months ago by j.m.fustin • 10

7 months ago by

Jennifer Hillman Jackson ♦ 25k

United States

Jennifer Hillman Jackson ♦ 25k wrote:

Hello,

Cufflinks will report FPKM for each isoform in the given input sample/replicate.

NGS: RNA Analysis >> Cufflinks transcript assembly and FPKM (RPKM) estimates for RNA-Seq data (Galaxy Version 2.2.1.0)

Please note that these tools are considered older and newer RNA-seq methods are described in the Galaxy tutorials here: https://galaxyproject.org/learn/

Thanks! Jen, Galaxy team

ADD COMMENT • link written 7 months ago by Jennifer Hillman Jackson ♦ 25k

7 months ago by

j.m.fustin • 10

j.m.fustin • 10 wrote:

Thanks Jen, but I believe this is not correct. Cufflinks will indeed generate a FPKM output for each input BAM file but then will generate a separate output file for each input, rather than one table containing all samples analyzed.

Currently trying stringtie with stringmerge...

ADD COMMENT • link written 7 months ago by j.m.fustin • 10

Right, This won't produce the same type of merged output as Cuffdiff and is what I mean by "the given input". Cufflinks processes each sample/replicate distinctly, not as a group. Reference annotation could be included to label the data and common transcripts between datasets joined.

ADD REPLY • link written 7 months ago by Jennifer Hillman Jackson ♦ 25k

7 months ago by

j.m.fustin • 10

j.m.fustin • 10 wrote:

Thank you for your answer. I am now joining the output tables from stringtie as you suggested! I was afraid that the lines of each sample table would be different but after stringtie merge on all transcript assemblies, then a second round of stringtie using the gtf output from stringtie merge gave tables that are joinable.

I hope it is ok!

ADD COMMENT • link written 7 months ago by j.m.fustin • 10

Please log in to add an answer.

Similar posts • Search »