[BioC] Re： DESeq warning: "Dispersion fit did not converge" withreplicates counts data

[Jiewencai]

Hi Mike,


Thanks for your help. Do you mean "plotDispEsts", Maybe it's a good fit, but I am not sure.
I attach the counts file to the email, Could you help me to make the judgment?
And, could you explain why the DESeq fail to fit the parametric curve through the dispersion
estimates over the mean counts. Is't because the counts are too much difference between the
two conditions? There are about 9% of isoforms's log2FC great than 2 or less than -2 between
CK and T.


Wencai