The R script named as "Script_BLin_JAMA.R" conducts replication efforts of a part of Hartwig et al’ s study for Two-sample MR based analyses to estimate the true causal role of CRP levels in risk of schizophrenia.
The script contains two parts: 1) Using “TwosampleMR”, “MendelianRandomlization” package and our own script to replicate the Hartwig et al’ s MR analyses; and 2) Use the data from etable 2 of Hartwig and repeat the study.
We conclude that: 1) We disagree with the authors’ choice for the effect allele (G instead of A) at rs9987289 as their input data; 2) and that our results point to weaker causality of CRP in risk of schizophrenia; and 3) using the exactly same input data as Hartwig et al (G instead of A at rs9987289), some of our estimates are still different from theirs, so we invite the authors to also make their scripts publicly available.