Baseline model in LDSC-SEG

[jie.q...@gmail.com]

Hi

I want to evaluate heritability enrichment of cell types from in-house transcriptome annotation, using LDSC-SEG in Finucane HK et al., Nature Genetics 2018.

In my understanding, Methods section in the paper said that baseline model should include:

 (i)  52 functional categories (NOT 53 functional categories in Nature Genetics 2015; all SNPs were excluded from this)

 (ii) All genes category (all analyzed genes in transcriptome data, to which 100-kb windows were added.)

Total 53 categories.

In that case, is it acceptable that there are several SNPs which do not belong to any category due to removing all SNPs category?

I think z-scores of the coefficient in ldsc depends on whether to include these SNPs in the analysis, but there is no clear description.

Sincerely,

Jie