Mendelian Randomization Analysis
The random-effects inverse-variance
weighted (IVW) method was used to estimate the bidirectional causality
between childhood asthma and T1DM by meta-analyzing the SNP specific
Wald ratio estimates.22 The causal effect estimates
(beta coefficients) were transformed to odds ratios (OR). The weighted
median, MR-Egger methods were also implemented to assess the causality.
Since many IVs are associated with multiple traits (pleiotropy), we used
MR-Egger regression23 and MR pleiotropy residual sum
and outlier (MR-PRESSO)24 method to examine the
robustness of the main findings. When the MR-Egger intercept differs
from zero (at P < 0.05), horizontal pleiotropy or
violation of the MR assumption is indicated.25MR-PRESSO method was used to correct the IVW-estimate via outlier
removal, when MR-PRESSO outlier test suggested the presence of outlying
variants. All MR analysis were performed using R version (4.0.3) with
the “TwoSampleMR” and “MRPRESSO” package.26