Train a gender model using GMM on Y-fractions

Fits a 2-component Gaussian mixture model to the Y-chromosome read fractions of a set of reference samples. Uses mclust::Mclust() as the R-native replacement for sklearn.GaussianMixture. The local minimum of the fitted density is used as the male/female cutoff.

Usage

.train_gender_model(samples, yfrac = NULL)

Arguments

samples

List of sample objects.

yfrac

Optional numeric; if given, used as the manual cutoff instead of the GMM-derived one.

Value

A list with:

genders

Character vector of "M" / "F" per sample.

cutoff

Numeric; the Y-fraction cutoff used.

Details

Mirrors newref_tools.train_gender_model() in upstream WisecondorX.