annotated and unaanoted splice sites

[Varun Gupta]

Hi Alex,
Hope you are doing well.
I am interested in new introns which are present in the sample. I generated the yeast genome with gtf file of S.cerevisiea. Then I mapped the reads to this genome. From the SJ.out file, I counted all those splice sites which are annotated(column 6 == 1)

I used a script from hisat2 to extract all the splice sites from my gtf file. The 5' end was off by 2 so I added it so as to make comparable with star. According to my understanding, if it is annotated, only then we can see 1 in column 6. If this is true, I compared the splice sites from my sample which have column 6 as 1 with hisat2 splice sites file and 9 junctions were still outputted by star as annotated(column 6 as 1) but they are not in gtf file. Can there be something wrong with hisat2 script to extract splice sites. Since Star creates splice sites from the gtf file, do you have a command I can use to generate splice sites from GTF file which STAR uses??

Hope to hear from you soon.

Thanks

Regards
Varun

[Alexander Dobin]

Hi Varun,

in the genome directory, there is a file "sjdbList.fromGTF.out.tab", which contains the junctions extracted from the GTF file.

If you cannot reconcile the differences, please send me the GTF file, the sjdbList.fromGTF.out.tab, and the hisat list.

Cheers

Alex

[Varun Gupta]

Hi Alex,
I did not provide the gtf file at the genome generation step but I did provide it at the mapping step. I was interested in getting annotated junctions but since I did not provide it at the genome generation step, I cannot find  sjdbList.fromGTF.out.tab. Is there a way I can know what are my annotated junctions from the gtf file? Otherwise I will regenerate the genome with gtf file.

This is human hg19

Thanks
Regards
Varun

[Alexander Dobin]

Hi Varun,

there should be _STARgenome/ directory inside your mapping run directory  it contains the sjdbList* files.

Cheers

Alex