Filtering data after bmerge in plink

Question

[dirranrak]

Hi all,

After filtering and merging data from using --bmerge flag in Plink 1.9, I did use the same thresholds as before merging the dataset to test the quality of the merged dataset. Then I have this "error: All people removed due to missing genotype data (--mind)". Why I loose the quality that I have got from the separate datasets that I have applied the same filter thresholds?

Thank you

Answer 1

[chrchang523]

Suppose datasets A and B both have 1000 samples and 1 million variants each... but they're different samples, and different variants. Then every sample will suddenly have a 50+% missingness rate, since any sample in dataset A will have missing calls for all variants in dataset B, and vice versa.

[dirranrak]

Hi chrchang523,

Thank you for your answer. Does it mean that it is no need to test the quality of the merged dataset? and what should be the best way to minimize bias?