ValueError: Do you really want to compute whole-chromosome LD Score? (no! I don't think...)

[Jory Schossau]

Hi,

I'm using custom data (my own hg38 baseline, and constructed annotation files to match), but I'm now getting this warning/error, which is very different from the tutorial despite me using the same command. Why is this?

It's definitely running WAY longer if I force it to go ahead and compute whole-chromosome LD Score.

Tutorial Command (with no error):

	python ldsc.py\
		--l2\ 
		--bfile 1000G.EUR.QC.22\ 
		--ld-wind-cm 1\ 
		--annot Brain_DPC_H3K27ac.annot.gz\ 
		--out Brain_DPC_H3K27ac

My Command (results in error):

python ldsc.py\
 --l2\
 --bfile s3/subset.22\
 --ld-wind-cm 1\
 --annot s8/subset.22.annot.gz\
 --out s8/subset.22

So my question is what about my data would be causing this? I wonder if this is indicative of something I've constructed incorrectly.

From the ldsc github FAQ:

Q. Why am I getting the error message

Do you really want to compute whole-chomosome LD Score? If so, set the --yes-really flag (warning: it will use a lot of time / memory)"

A. Your --ld-wind-[snp | kb | cm] option told ldsc to compute whole-chromosome LD Scores, i.e., sum r2 with the sum taken over all the SNPs in your .bim file. This is more often a time-consuming mistake than the real intent of the user (whole-chromosome LD Score contains more information about population structure in the sample than anything else, and this information can be obtained much more rapidly via e.g., PCA). If you actually want to compute sum r2 with the sum taken over all SNPs in your .bim file, set the --yes-really flag, but be warned that if you .bed file is large, this could take a very long time to finish running.