search

cuelee / pleio

15 stars 6 forks source link

direction effect and results pleio #7

Open sevi2018 opened 3 years ago

sevi2018 commented 3 years ago

Hi, I'm running pleio including more than 15 GWASs, I have noticed that if I flip the beta of some of the GWASs i get different results, I still did confirm the loci detected in the first analysis realized using the original beta sign, but on top of that I get more loci that reach genome-wide significance. My question is, is this what you expect to see? And in case it is can you explain me why?

Thanks in advance for your help,

cuelee commented 3 years ago

Yes, reversing the beta of some (but not all) of the inputs will produce a different result, but in a PLEIO analysis, changing the direction of the input beta is against the model assumption.

As for the results of the analysis: very interesting, but I did not expect this result.