Closed mgalardini closed 2 years ago
Yeah could certainly add this! It's also possible in fastlmm to use two kinship matrices, that might be another way of doing it
Closing this as we actually have this script already! See PR #211 for some fixes to it
A reviewer for our last paper suggested to compute the LMM Kinship matrix from unitigs presence/absence patterns, so that core/accessory population structure is fully account for. Could this be useful as part of pyseer's scripts? We have it here for the moment: https://github.com/microbial-pangenomes-lab/2021_ecoli_pathogenicity/blob/main/workflow/scripts/unitigs2covar.py