gene prediction without GT-AG splice site bias

Question

[praasu]

What would be the best approach to predict introns or gene model from RNA-seq/est and genome seq without GT-AG splice site rules?

[Eric Lim]

Would you mind elaborating further on that rationale? U2 and U12 confer to a small class of dinucleotide pairs at the splice sites and are extremely well conserved from plants to animals. While there has been a very small number of non-consensus splice sites reported in the literature, many would agree that these variants are likely due to errors in annotations/interpretations, polymorphic difference between cDNA/genomic, or maybe pseudogene variants.

[praasu]

There is a chance that some of the annotated non-consensus splice sites are due to error in annotation. However, there are several studies which suggests that non-canonical splice-sites really exist. I am working on the species where is possiblity that they have predominantly non-consensus splice site.

Some References, https://pubmed.ncbi.nlm.nih.gov/11058137-analysis-of-canonical-and-non-canonical-splice-sites-in-mammalian-genomes/

https://pubmed.ncbi.nlm.nih.gov/25123659-a-comprehensive-survey-of-non-canonical-splice-sites-in-the-human-transcriptome/

Answer 1

[caggtaagtat]

You could align your reads with STAR this directly gives you information about gaps in the alignment of your reads and therefor introns.