infer N from beta [SE or P] and reference panel MAF

[bulik]

probably better than using max N for studies that don't provide N???

(in munge_sumstats.py)

[bneale]

Makes sense. If frequency is supplied then this would capture the imputation uncertainty too (although I'm not sure how often you get frequency without N).

Best Ben

On Monday, May 4, 2015, Brendan Bulik-Sullivan notifications@github.com wrote:

probably better than using max N for studies that don't provide N???

— Reply to this email directly or view it on GitHub https://github.com/bulik/ldsc/issues/32.

For scheduling matters, please contact Carla Hammond < hammond@atgu.mgh.harvard.edu>

Benjamin Neale, Ph.D. Assistant Professor Analytic and Translational Genetics Unit Massachusetts General Hospital Institute Member Stanley Center for Psychiatric Disease and Program in Medical and Population Genetics Broad Institute

[bulik]

almost never, but the idea here is to use reference panel (e.g., 1kG) MAF. not as good as getting true N directly, but probably better than setting N_j := max(N) as reported in the paper for all SNPs j