Hello all,

Quick question about clumping using the TwoSampleMR package and MR Base database.

If I executed the extract_instruments function with clump = TRUE, should I still also run the clump_data function?

I see that the use of both functions is included stepwise in the vignette tutorial, but I am having trouble understanding why extract_instruments does not already select for instruments that are both highly correlated with the trait of interest and independent of one another, with clump = TRUE.

Further, I have found that running these functions in succession (such as exampled below with GWAS study finn-b-L12_ATOPIC) can further change results -- that is, clump_data does indeed drop an additional result, even after running LD clumping with extract_instruments.

dat <- extract_instruments(outcomes = 'finn-b-L12_ATOPIC ', force_server = TRUE, clump = TRUE)
dat <- clump_data(dat)

Any and all help is greatly appreciated!