Table 1 The estimated correlation for a pair of cohorts via their summary statistics given 30 000 independent loci
From: Across-cohort QC analyses of GWAS summary statistics from complex traits
| n1,2 | n1 | n2 |
|
|
|
|---|---|---|---|---|---|---|
0.25 | 100 | 1000 | 1000 | 0.1 | 0.1072±0.0064 | 0.101±0.0093 |
1000 | 2000 | 0.0707 | 0.0814±0.0054 | 0.0709±0.0088 | ||
1000 | 5000 | 0.0447 | 0.0615±0.0055 | 0.0425±0.0096 | ||
1000 | 10 000 | 0.0316 | 0.0556±0.0063 | 0.0325±0.0099 | ||
0.25 | 1 | 1000 | 1000 | 0.001 | 0.0092±0.0056 | 0.0017±0.0093 |
1000 | 2000 | 0.0007 | 0.0126±0.0053 | 0.0006±0.0079 | ||
1000 | 5000 | 0.000447 | 0.0189±0.0060 | 0.0016±0.0090 | ||
1000 | 10 000 | 0.000316 | 0.0259±0.0059 | 0.0008±0.0092 | ||
0 | 100 | 1000 | 1000 | 0.1 | 0.0996±0.0052 | 0.094±0.0085 |
1000 | 2000 | 0.0707 | 0.0704±0.0048 | 0.0712±0.0097 | ||
1000 | 5000 | 0.0447 | 0.0453±0.0057 | 0.0441±0.0090 | ||
1000 | 10 000 | 0.0316 | 0.0335±0.0057 | 0.0325±0.0079 |
- Notes: Heritability was simulated on 1000 QTLs. We also tried 100 QTLs, and results were nearly identical; n1, n2, and n1,2 represent the sample size for cohort 1, 2, and overlapping samples between them. γ1,2 represents the true correlation for a pair of summary statistics due to overlapping samples.
represents the estimated correlation estimated via direct correlation between summary statistics, the method proposed by Bolormaa et al.18 and Zhu et al.19. 
represents the estimated correlation estimate via λmeta, 
.
represents the estimated correlation estimated via direct correlation between summary statistics, the method proposed by Bolormaa et al.
represents the estimated correlation estimate via λmeta, 
.