Hello,
I recently made a new annotation using a well known annotation from a paper (Sessions) and another from a well known database (Xenbase/Taejoon Lab). I used Stringtie to organize the transcriptome and I have found many unusual transcripts. In the new annotation, most of the transcripts are normal, they go right to left or 3' to 5' (the way they're transcribed). However when I expand them in iGV I get one or two transcripts for every other gene that goes the opposite direction. Sometimes these transcripts code for an additional number of bases that weren't annotated before.
Here is an example:
As you can see the first one is going to the right while the others are going to the left. Is this a potential case for a "Natural antisense transcript" which is a type of regulatory RNA or is this a mistake by my pipeline? Thank you.
first ensure that you assembled the data correctly. I assume you have used stranded RNA-Seq but the nomenclature is confusing (and in the TrueSeq protocols the actual sequence data ends up being generated from the reverse complement) hence it is easy to get the results backward if your parameters are not set correctly.
I agree with Istvan. You can check the read directionality by aligning a subset to the genome and checking the SAM flags for the alignment orientation.
I believe that I assembled it correctly since I used the --rf or that I have a stranded library fr-firststrand in my alignment software (StringTie). When using htseq to retrieve counts I also put in paired end and reverse to account for the directionality. But my data matches up with the database well when i check for +/- strand and the blast results always pull up the same gene. I also used the same reference annotation for both sets of data that are stranded the same. Is there anything else that could have potentially caused this single transcript that goes in the opposite direction?
Natural antisense transcript are very common, but they are usually not defined as coming from the same gene as the sense transcript. In a "normal" annotation, you should have the main gene composed of only sense transcripts isoforms, and a possible anti-sense gene made of only anti-sense transcripts.