A question about the accuracy of AUGUSTUS

[yjiangnan]

Sorry for bothering you but I have an urgent question about the accuracy of the gene prediction of AUGUSTUS. I am trying to express segments of the new coronavirus SARS-CoV-2 from DNA. However, since this virus is an RNA virus and its life cycle does not contain a DNA stage, its genome sequences may have problems to be expressed from DNA. For example, it may be spliced.

So, I tried to use AUGUSTUS at http://bioinf.uni-greifswald.de/augustus/submission.php to predict whether the transcription will be successfully. However, the prediction often show some “introns” even though I have mutated the 5’-consensus sequence for splicing. It seems that AUGUSTUS can even use some highly mutated sequences, such as GCAGG|GTGCTGT and AGTG|GTAG and TG|GTGAAT and CATGT|AGTTG as the 5’-splice site. “|” shows the splice site.

So, my question is, are these predictions real? How accurate is such predictions? I want to get the full transcript without any splicing to be able to study viral functions. Should I be worried about such predictions? I also have to order the DNA sequences for the whole genome urgently given the current situation of the epidemic. So, if you can answer me as soon as possible, that will be highly appreciated.

Thank you very much!