p2cor generating NaNs

[Fnyasimi]

@tshmak thanks for the nice tool I have been using for a while

I have these two snps in my summary stats which have really low pvalue and end up generating NaN correlations when using p2cor function. The sample size is above 1000

snp	af	effectsize	se	pvalue
snp1	NA	0.6433	0.08915	4.940656e-324
snp2	NA	0.6434	0.08916	4.940656e-324

What is the best approach to handle such snps in this function/tool because the NaNs present issues downstream. I would really appreciate your feedback on this issue.

[privefl]

You could probably derive these correlations yourself using the formula from p2cor(): https://github.com/tshmak/lassosum/blob/master/R/p2cor.R#L20

where t is basically effectsize / se and n is the sample size.

[Fnyasimi]

Thank you so much.

Does the package work on R 4.0.5 on macOS BigSur 11.6? I am getting this error here when I try to load it

Error: package or namespace load failed for ‘lassosum’:
 package ‘lassosum’ was installed before R 4.0.0: please re-install it

I installed the package using the instructions here

[privefl]

Restart and reinstall?

[Fnyasimi]

I have done that still getting the same error

[nini-tech23]

Hello, did you solve the problem? I have same problem here but didn't know how to solve it

[tshmak]

Have you tried uninstalling lassosum and re-installing?

[nini-tech23]

My package is working well, but I have a problem of getting NaN correlations values when running pcor2. There are 48 SNPs which have very low p-value, so I want to ask whether it is okay for excluding those SNPs from the analysis

[tshmak]

You mean the p-value is less than the 2.2e-308 (i.e. double.xmin)? If so, I think a better idea is to replace these with double.xmin, since low p-values are supposed to indicate important SNPs.