Inability to generate custom-annotation files.

[Daniel Richard]

Hello,

I was running through the tutorial in the wiki which describes generating a custom annotation file for use in performing h2 partitioning. I've been able to generate custom annotation.gz files for my bed sets using the hapmap3 lists indicated in the wiki, however when I attempt to use these annotations along with the v2.2 baseline model for h2 estimation I get the following error:

"ValueError: LD Scores for concatenation must have identical SNP columns."

Looking into the logs of the baseline model it looks like the annotations here were made using a SNP file named: /n/groups/price/ldsc/reference_files/1000G_EUR_Phase3/list.txt

I have no idea what this SNP set is and where it came from, and unfortunately without knowing this I can't generate custom annotation files to run partitioning - is this file available, or can you describe how it was generated so I can use the baseline model with my custom annotations?

Thanks,

Daniel Richard

[Jory Schossau]

Hi Daniel, fellow user here.

I am also doing partitioned ld score regression using custom snp sets, and I found that the ldscore files must first be regenerated before regression can be performed.

If you look at the ldscore files you'll find they have category information in there that must match up to your new annot files.

Once you do "LD Score Estimation" then you can proceed to "Partitioned Heritability"

Good luck.

[Daniel Richard]

Hi Jory,

Thanks for the reply - are you saying that I need to run the --l2 command specifying multiple --annot files, my annotations and the baseline? I have L2 scores already calculated for my annotations. In the Github tutorial where they detail adding custom annotations ontop of the baseline they indicate just using the baseline along with your custom annotation (LD score files) in the --ref-ld-chr flag, which is where I'm tripping up.

This is mostly just for my understanding - I ended up just coding a workaround where I slot-in my annotations on top of the baseline files themselves and have everything work out.

Thanks,

Daniel

[Jory Schossau]

Hi Daniel,

I'm unsure exactly what the error is getting at, but I suspect it's because of either of these requirements:

* annot files and plink filfes must all list the same SNPs in the same order

* annot files and ldscore files have a requirement about the annotation columns and possibly the ordering.

Glad you got it working! I initially did the same as you and it seems to work okay.