SNP Heritability for Phenotype 40001_C509

Estimated SNP heritability: -0.6504 (se=0.302, p=0.984)

Significance level: none (insufficient confidence)

Confidence rating: none

Our confidence in the LD Score regression result for this phenotype is reduced due to:

• Insufficient statistical power for meaningful estimation
• Sex bias in the phenotype may impact interpretability

Note: SNP heritability for this binary outcome is reported on the liability scale, assuming that the population prevalence matches the prevalence in the UK Biobank analysis set (0.0562). This may be unreliable if the outcome was limited to a subset of individuals, or if the UK Biobank study population is not representative for this phenotype (which is likely in many cases). The estimated observed-scale heritability is: -0.1570 (se=0.0729).

Confounding and model misspecification

In addition to SNP heritability, LD score regression also estimates an intercept term that indexes population stratification, other counfounding, and potential misspecification in the partitioned LD score model for the distribution of genetic effects genome-wide.

• Intercept: 1.0109 (se=0.00799, p=0.0865)
• Mean \(\chi^2\): 0.9928
• Ratio: NA (se=NA)
• \(\lambda_{GC}\): 0.9956

Intercept values near 1 indicate little or no confounding. The reported LDSR ratio compares the intercept estimate and the mean \(\chi^2\) statistic to provide a rough index for how much of the polygenic signal in the GWAS may be due to confounding rather than genetic effects (assuming the LD score model is well specified). Note that the intercept, mean \(\chi^2\), and genomic control \(\lambda_{GC}\) are all expected to scale with sample size, making the ratio better suited for comparisons between different GWAS.