Augustus: internal stop codons in transcripts

[mht]

Gene prediction of a draft assembly was done with Augustus, and some of my peptide sequences contain either single or strings of "X"s. From what I've read, "X"s could be caused by strings of N's in the assembly. However, I have also noticed that "X" may represent some internal stop codons in some cases of my transcripts.

This prediction was run without turning on the --noInFrameStop flag. How does one decide whether or not to report transcripts with internal stop codons? I'm curious to know whether these transcripts are of any use and how authetic they are.

I am getting errors flagged during submissions to public databases because of this. Is it advisable to just discard these transcripts? What would be the best way to handle transcripts with internal stop codons?

[JamieHeather]

I've not used Augustus, but either way if it were me I'd like to be sure those Xs were representing premature stop codons or something else.

Certainly even one N in your DNA will cause ambiguous amino acids on translation, but this is an easy thing to check for.

I would be surprised if your software would output the same character for unknown AA as for stop character - usually they use some punctuation, such as a * to represent stop codons.

[ADELALU]

please, i have problem in identifying internal stop codons in my sequence, the interval and the reading frame.
ai would appreciate any suggestion offered.
thank you