jrs95
commented
2 years ago Hi Alex,
Apologies for the delay in answering your question.
The LD matrix is used to compute variant-specific priors when investigating correlated traits measured in overlapping samples, where the trait.cor matrix also has to be defined as non-identity. We still recommend using the standard approach (i.e. assuming non-overlapping samples) in most cases, as this model performed well even for traits that were highly correlated in overlapping samples and is much more computationally efficient.
Best wishes,
James
Hi,
I had a quick question regarding the use of your tool - I was playing around with the parameters and found that supplying your own LD matrix with correlation values (instead of the default
diag(1, dim(effect.est)[1], dim(effect.est)[1])) makes no difference to the output with a number of different snp sets. All other parameters are set at default, perhaps I am missing something?hyprcoloc(x3, y3, trait.names = traits, snp.id = variants,snpscores = TRUE)versus:
hyprcoloc(x3, y3, trait.names = traits, snp.id = variants,snpscores = TRUE,ld.matrix=ldmat)Thanks in advance!
Alex