. 2021 Jan 28;11:2518. doi: 10.1038/s41598-021-82000-1

Table 2.

Details of the estimation of the trait relationship using our method and GCTA_mtCOJO.

	Our method	GCTA_mtCOJO
Parameters	Not required	Significance level for selecting GWAS signals for instrumental variable. Typically set to 5 × 10^–8 for most analysis. We reduce this threshold to 5 × 10^–6 for the BMI GWAS adjusted for ever-smoking
External data	FHS phenotype at the time closest to DNA draw	Genotype data from FHS unrelated individuals are used as LD reference panel LD-score regression results from European population based on 1,000 Genomes for the outcome

Our method

GCTA_mtCOJO

Parameters

Not required

Significance level for selecting GWAS signals for instrumental variable. Typically set to 5 × 10^–8 for most analysis. We reduce this threshold to 5 × 10^–6 for the BMI GWAS adjusted for ever-smoking

External data

FHS phenotype at the time closest to DNA draw

Genotype data from FHS unrelated individuals are used as LD reference panel

LD-score regression results from European population based on 1,000 Genomes for the outcome