The 'isoform_exp.diff' file after Cuffdiff contains genes that don't have isoforms.

[mondongho]

I run cuffdiff and I'm looking for differentially expressed isoforms in my tissue samples, but my isoform_exp.diff file contain genes that don't have any isoform reported (e.g. cxcl10, cxcl9, isg15) mixed with genes that have isoforms (e.g. oas2, mx1, ifi27). I thought that this file would report only genes that have isoforms. I wonder if I did something wrong or this is normal.

Thaks for your help.

[pm2012]

They could be novel isoforms of the genes. You can check if they are novel by looking into cuffcompare files & checking if these isoforms have class code 'j'. The other reported isoforms are likely to be known isoforms.

[mondongho]

Thanks for the pointer.

I checked my cuffcompare.tracking file and looked for the genes that are in my isoform_exp.diff file and don't have any isoform reported; some of them have the class code 'j' as you said but others don't. For example, Gene A is in my isoform list, does not have any isoform reported but it shows code 'j' and I can see the (putative) new isoform using IGV. However, Gene B is also in my isoform list, does not have any isoform reported and its codes is '='.

All suggestions are welcome.