You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Hi, thank you developers for this amazing tool. After processing necessary inputs and fixing all the technical issues, I was able to run the LD score regression. However, I got a genetic correlation of 1.0187 between the two traits I analyzed, which doesn't make sense. I suspect there might be sample overlap in my cohorts, as both summary statistics were downloaded from Pan UKB, but this still shouldn't result in a correlation > 1. I’m wondering if something might be wrong in my code. Any insights would be appreciated.
Beginning analysis at Tue Oct 15 14:38:59 2024
Reading summary statistics from ../ukbb_ptsd.sumstats.gz ...
Read summary statistics for 1130078 SNPs.
Reading reference panel LD Score from ../eur_w_ld_chr/[1-22] ... (ldscore_fromlist)
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] ... (ldscore_fromlist)
Read regression weight LD Scores for 1290028 SNPs.
After merging with reference panel LD, 1100373 SNPs remain.
After merging with regression SNP LD, 1100373 SNPs remain.
Computing rg for phenotype 2/2
Reading summary statistics from ../ukbb_adjd.sumstats.gz ...
Read summary statistics for 1217311 SNPs.
After merging with summary statistics, 1100373 SNPs remain.
1100373 SNPs with valid alleles.
/Users/ys/Downloads/LDSC/ldsc/ldscore/sumstats.py:532: FutureWarning: Method .as_matrix will be removed in a future version. Use .values instead.
ref_ld = sumstats.as_matrix(columns=ref_ld_cnames)
/Users/ys/Downloads/LDSC/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)
Heritability of phenotype 1
---------------------------
Total Observed scale h2: 0.0012 (0.0012)
Lambda GC: 1.0075
Mean Chi^2: 1.0029
Intercept: 0.9941 (0.0061)
Ratio < 0 (usually indicates GC correction).
Heritability of phenotype 2/2
-----------------------------
Total Observed scale h2: 0.001 (0.001)
Lambda GC: 1.0075
Mean Chi^2: 1.0072
Intercept: 0.9989 (0.0057)
Ratio < 0 (usually indicates GC correction).
Genetic Covariance
------------------
Total Observed scale gencov: 0.0011 (0.0009)
Mean z1*z2: 0.5466
Intercept: 0.5379 (0.0047)
Genetic Correlation
-------------------
Genetic Correlation: 1.0187 (0.4699)
Z-score: 2.1677
P: 0.0302
Summary of Genetic Correlation Results
p1 p2 rg se z p h2_obs h2_obs_se h2_int h2_int_se gcov_int gcov_int_se
../ukbb_ptsd.sumstats.gz ../ukbb_adjd.sumstats.gz 1.0187 0.4699 2.1677 0.0302 0.001 0.001 0.9989 0.0057 0.5379 0.0047
Beginning analysis at Tue Oct 15 15:02:21 2024
Reading summary statistics from ../ukbb_ptsd.sumstats.gz ...
Read summary statistics for 1130078 SNPs.
Reading reference panel LD Score from ../eur_w_ld_chr/[1-22] ... (ldscore_fromlist)
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] ... (ldscore_fromlist)
Read regression weight LD Scores for 1290028 SNPs.
After merging with reference panel LD, 1100373 SNPs remain.
After merging with regression SNP LD, 1100373 SNPs remain.
Computing rg for phenotype 2/2
Reading summary statistics from ../ukbb_adjd.sumstats.gz ...
Read summary statistics for 1217311 SNPs.
After merging with summary statistics, 1100373 SNPs remain.
1100373 SNPs with valid alleles.
/Users/ys/Downloads/LDSC/ldsc/ldscore/sumstats.py:532: FutureWarning: Method .as_matrix will be removed in a future version. Use .values instead.
ref_ld = sumstats.as_matrix(columns=ref_ld_cnames)
/Users/ys/Downloads/LDSC/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)
Heritability of phenotype 1
---------------------------
Total Liability scale h2: 0.5133 (0.4971)
Lambda GC: 1.0075
Mean Chi^2: 1.0029
Intercept: 0.9941 (0.0061)
Ratio < 0 (usually indicates GC correction).
Heritability of phenotype 2/2
-----------------------------
Total Liability scale h2: 0.1863 (0.1808)
Lambda GC: 1.0075
Mean Chi^2: 1.0072
Intercept: 0.9989 (0.0057)
Ratio < 0 (usually indicates GC correction).
Genetic Covariance
------------------
Total Liability scale gencov: 0.315 (0.2576)
Mean z1*z2: 0.5466
Intercept: 0.5379 (0.0047)
Genetic Correlation
-------------------
Genetic Correlation: 1.0187 (0.4699)
Z-score: 2.1677
P: 0.0302
Summary of Genetic Correlation Results
p1 p2 rg se z p h2_liab h2_liab_se h2_int h2_int_se gcov_int gcov_int_se
../ukbb_ptsd.sumstats.gz ../ukbb_adjd.sumstats.gz 1.0187 0.4699 2.1677 0.0302 0.1863 0.1808 0.9989 0.0057 0.5379 0.0047
The text was updated successfully, but these errors were encountered:
Hi, thank you developers for this amazing tool. After processing necessary inputs and fixing all the technical issues, I was able to run the LD score regression. However, I got a genetic correlation of 1.0187 between the two traits I analyzed, which doesn't make sense. I suspect there might be sample overlap in my cohorts, as both summary statistics were downloaded from Pan UKB, but this still shouldn't result in a correlation > 1. I’m wondering if something might be wrong in my code. Any insights would be appreciated.
Here's the code I used:
Some logs for this:
And here's for conversion to liability scale
and some logs:
The text was updated successfully, but these errors were encountered: