segmentation error while running in genomegenerate mode

[xlo...@uic.edu]

Hello, I'm having problem while running star in genomegenerate mode, the program would just stop while having the message "segmentation error". Attached is the Log.out file.   Could anyone help me figure out what's the problem. Is it the memory is not large enough? 

Regards,

Xiyao

[Alexander Dobin]

Hi Xiyao,

the Log.out file did not get attached, please re-send it.

Also, please use the latest patch https://github.com/alexdobin/STAR/releases/tag/STAR_2.4.0i, it fixed a seg-fault in the genome generation - but maybe yours is different.

Cheers

Alex

[Ricardo]

Hello,

I'm getting a similar problem with genome generate, though I am using STAR_2.4.0j.

Feb 13 14:12:26 ..... Started STAR run

Feb 13 14:12:26 ... Starting to generate Genome files

Segmentation fault

Any ideas on what should be done?

Thanks,

Ricardo

[Alexander Dobin]

Hi Ricardo,

it seems that you are using a non-standard annotation file, it has ~1.5M junctions.

Could you send this file to me - I will try to replicate the problem.

Cheers

Alex

[Ricardo]

Hello Alex,

Thanks for your reply. The gtf I used is too large to attach here, but I originally downloaded it from gencode:

ftp://ftp.sanger.ac.uk/pub/gencode/Gencode_mouse/release_M4/gencode.vM4.annotation.gtf.gz

I modified the header in this file, which had several lines commented out that also seemed to cause trouble with genomegenerate. In lieu of the entire file, I've attached the first 500 entries which also cause a segmentation fault.

Thanks for your help,

Ricardo

[Alexander Dobin]

Hi Ricardo,

I think the problem is with the --sjdbFileChrStartEnd  option, which requires a list of junction coordinates, not GTF file.

If you want to use GTF file, you need --sjdbGTFfile <annot.gtf> option.

Cheers

Alex