Compare with regular prediction: PLINK assoc_qt and betas

AJResearchGroup / nsphs_ml_qt

R package for nsphs_ml_qt

GNU General Public License v3.0

0 stars 1 forks source link

Open richelbilderbeek opened 2 years ago

richelbilderbeek commented 2 years ago

See #42 that M1 and p1 give the best prediction

> knitr::kable(t |> dplyr::filter(model_id == "M1" & pheno_model_id == "p1"))

model_id	pheno_model_id	window_kb	genotype_concordance	nmse
M1	p1	1	0.4773399	1.0988598
M1	p1	10	0.2206714	0.9964508
M1	p1	100	0.1404004	1.6963628
M1	p1	1000	0.0453623	1.0317558

However, how well does a simple PLINK run fare?