Grid size error

116 views
Skip to first unread message

Anna Page

unread,
Jan 23, 2017, 6:58:44 AM1/23/17
to OmegaPlus
Hi all,

Thanks for everyone's questions and answers on this user group, it has been a great resource for troubleshooting my analyses. However I'm struggling with one issue right now.

I'm trying to run both Omega and SweeD on the same dataset, then combine the output from the two.
SweeD seems to be running ok - it hasn't finished yet, but no errors so far.
However when I try to run Omega I get this error

OmegaPlus-M: OmegaPlus_omega.c:86: boundaries: Assertion `endsnp-startsnp >= grid' failed.

It looks like I have made an error in choosing the grid size, but can't work out exactly what it wrong

My genome is ~833 Mbp, so based on this thread where it was suggested to choose a grid size allowing the score to be calculated ~5kb, I chose a grid size of 160000.

For SweeD and Omega I am using the same input: I used bwa to map GBS reads to the genome, converted the SAM output from bwa to fasta, then concatenated the fasta files for the individuals of my population, separating each alignment with '//'

My commands:

Omega:

./OmegaPlus-M -name meltest -input /path/to/concatenated_mel.fasta -minwin 5000 -maxwin 100000 -grid 160000 -threads 2

SweeD 

./SweeD-P -name meltest -input /path/to/concatenated_mel.fasta -grid 160000 -threads 16


Any advice on where I could be going wrong, or what I should investigate further would be much appreciated. Thank you!

Best,
Anna

Nikolaos Alachiotis

unread,
Jan 23, 2017, 7:10:57 AM1/23/17
to OmegaPlus
Hi Anna,

The error you are getting is due to the fact that the grid size you chose is larger than the number of SNPs in your dataset. We have this restriction for performance reasons. What you can do is let OmegaPlus run until it prints the number of SNPs. Then rerun by using a grid value thats is smaller than the number of SNPs. 

Note also that the OmegaPlus-M implementation might turn out to be slower than other implementations when you launch a small number of threads, such as 2 in your case. Either use a larger number of threads or switch to OmegaPlus-F or OmegaPlus-C.

Nikos

--
You received this message because you are subscribed to the Google Groups "OmegaPlus" group.
To unsubscribe from this group and stop receiving emails from it, send an email to omegaplus+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
Reply all
Reply to author
Forward
0 new messages