About the the summary resut suitable for BHR

[Russian-1895]

Hi Ajay and Dan,

I have some confused about the BHR analysis. My summary level results afre very big, which need many storage space. So I wonder to know that If can I only keep the result that P value < 0.01, and then make it as the input of BHR to calculate the heritability.

Looking forward to your reply.

Best wishes,
Boxing

[ajaynadig]

Hi Boxing,
Thank you for your question and for your interest in BHR.

Unfortunately, BHR requires all of the input columns in the wiki. BHR examines the full spectrum of effect sizes, and does not apply a significance threshold, so a binary significance test is not sufficient.

In our analyses of the UK BIobank for 37 traits, the resulting summary level results were large but manageable, around 500 MB - 1 GB gzipped. How large are your tables? You may consider using strategies such as compressing your results and using a reader in R that can read compressed files, dividing your summary statistics into individual traits so that you only need to load one at a time, or other efficiency strategies.

Best,
Ajay

[ajaynadig]

Closing this issue for now, feel free to reopen if any additional questions arise.