Question about the reference panel and # of SNPs used to estimate heritability.

[Jiayi Xu]

Hi LDSC support group,

     I have been trying to estimate a phenotype heritability using its GWAS summary statistics. For the munge_sumstats.py, I ran it without the specification of the merge-alleles w_hm3.snplist given only <10% of the SNPs would be kept if this specification is included. And then I moved on to run the ldsc.py. And I used the eur_w_ld_chr as the LD score reference panel (the GWAS was based on a European ancestry population). Please see below for the log information. The GWAS summary statistics file included ~4.6M SNPs (based on 1000_genomes_project_phase3_CEU/ALL.chr_merged.phase3_shapeit2_mvncall_integrated_v5.2 0130502.genotypes); whereas the reference panel LD has ~1.3M SNPs. And after merging the GWAS SNPs with the reference panel LD, only ~0.6 M SNPs remained to estimate the heritability. Since this only used ~12% of the SNPs out of the 4.6M SNPs, I was not sure whether I have used the correct reference panel. Do you have any suggestion on this? Thanks so much! 

Read summary statistics for 4558621 SNPs.

Reading reference panel LD Score from eur_w_ld_chr/[1-22] ...

Read reference panel LD Scores for 1290028 SNPs.

Removing partitioned LD Scores with zero variance.

Reading regression weight LD Score from eur_w_ld_chr/[1-22] ...

Read regression weight LD Scores for 1290028 SNPs.

After merging with reference panel LD, 553140 SNPs remain.

After merging with regression SNP LD, 553140 SNPs remain.

Using two-step estimator with cutoff at 30.

/hpc/packages/minerva-common/ldsc/1.0.1/ldsc/ldscore/irwls.py:161: FutureWarning: `rcond` parameter will change to the default of machine precision times ``max(M, N)`` where M and N are the input matrix dimensions.

To use the future default and silence this warning we advise to pass `rcond=None`, to keep using the old, explicitly pass `rcond=-1`.

  coef = np.linalg.lstsq(x, y)

Total Observed scale h2: 0.1564 (0.4166)

Lambda GC: 1.1082

Mean Chi^2: 1.3187

Intercept: 1.3112 (0.0136)

Ratio: 0.9765 (0.0426)

Best,

Jiayi