Kinship in scan1snps

[Harry C.]

Dear Karl,

I have encountered an issue with including kinship in scan1snps. The function works correctly when kinship is not provided, but gives the following error with kinship. I can't work out why because the kinship matrix "k" exists in the R environment.

Would be very grateful for any help.

Kind regards,

Harry

[Karl Broman]

We’ve seen similar problems before, but unfortunately I have no clue what’s going wrong: 

   https://github.com/rqtl/qtl2/issues/177

All I can suggest is to use cores=1.

karl

On Jul 20, 2022, at 1:19 AM, Harry C. wrote:

Dear Karl,

--
You received this message because you are subscribed to the Google Groups "R/qtl2 discussion" group.
To unsubscribe from this group and stop receiving emails from it, send an email to rqtl2-disc+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/rqtl2-disc/04fcedd5-2cc7-4ef4-b3a8-79badaa04db5n%40googlegroups.com.

[Dan Gatti]

I agree. I think that reducing the number of cores helped me in a similar situation. Maybe it’s an out-of-memory error masquerading as a variable-not-found error.

To view this discussion on the web visit https://groups.google.com/d/msgid/rqtl2-disc/A0A5FD78-817C-4369-B64F-E60FBAC16409%40gmail.com.

---

The information in this email, including attachments, may be confidential and is intended solely for the addressee(s). If you believe you received this email by mistake, please notify the sender by return email as soon as possible.

[Harry C.]

Hi both,

Thanks for your suggestions.

I tried running this with only 3 cores and the RAM usage was ~20% of the 256GB capacity. Seems unlikely to be a memory issue for this setup. When I first run the code, the RAM usage increases up to a plateau as data is exported to the cores, and only then the code produces an error. My guess would be that there is a bug in the specific export of the kinship matrix to the cores, as the export of geno probs works fine.

Kind regards,

Harry