Gaussian Mixture Model

A variant of DLBCLone that doesn’t use K-nearest neighbors relies instead on a gaussian mixture model and UMAP embeddings. This functionality is experimental. It’s important not to confuse this modeling with the previously introduced DLBCLone functions.

DLBCLone_train_mixture_model

Training a Gaussian Mixture Model for DLBCLone Classification, this function fits a supervised Gaussian mixture model (GMM) to UMAP-projected data for DLBCLone subtypes, excluding samples labeled “Other”. It assigns class predictions and optionally reclassifies samples as “Other” based on probability and density thresholds.

umap_out output from make_and_annotate_umap, containing a data frame with UMAP coordinates and truth labels

probability_threshold minimum posterior probability required to assign a class (default: 0.5)

density_max_threshold minimum maximum density required to assign a class (default: 0.05)

library(mclust) # <- necessary for running the mixture models

mixture_result <- DLBCLone_train_mixture_model(umap_out = mu_everything)

DLBCLone_predict_mixture_model

Applies a previously trained supervised Gaussian mixture model (GMM) to UMAP-projected data for DLBCLone subtypes. Assigns class predictions and optionally reclassifies samples as “Other” based on probability and density thresholds.

model fitted model object, as returned by DLBCLone_train_mixture_model

umap_out output from make_and_annotate_umap, containing a data frame with UMAP coordinates for the samples to be classified with the model. This must be projected using the same UMAP model that was generated using the training data.

probability_threshold minimum posterior probability required to assign a class (default: 0.5)

density_max_threshold minimum maximum density required to assign a class (default: 0.05)

predict_mixture_result <- DLBCLone_predict_mixture_model(
  model = mixture_result$gaussian_mixture_model, 
  umap_out = mu_everything
)