Closed yidingdd closed 2 years ago
import xarray as xr dset_region = xr.open_zarr("/u/scratch/y/yiding/admixture-finemapping/data/ukb_eur_afr_imputed/1/1:1000000-2000000") X = dset_region.geno.values X[:,3,:].T @ X[:,3,:]
array([[ -7, 21], [ 21, -12]], dtype=int8)
The data type int8 leads to numerical overflow when computing XTX matrix
This is fixed because genotype will be read as float now.
The data type int8 leads to numerical overflow when computing XTX matrix