gcae_project: Project

View source: R/gcae_project.R

gcae_projectR Documentation

Project

Description

Project, using the GCAE command-line interface

Usage

gcae_project(gcae_setup, gcae_options = create_gcae_options(), verbose = FALSE)

Arguments

gcae_setup

GCAE model setup, as created by create_gcae_setup, read from file by read_gcae_setup_file, checked by check_gcae_setup, and saved to file by save_gcae_setup.

gcae_options

options to run GCAE, as created by create_gcae_options

verbose

the verbosity of a function. Set to TRUE for more output. Use check_verbose to detect if this argument is valid.

Value

full paths to the created files. These are:

  • losses_from_project.csv and losses_from_project.pdf show the loss function value of the model per epoch. Note that this is the loss for the entire data set.

  • genotype_concordances.csv and genotype_concordances.pdf the genotype concordances of the model per epoch (this is the rate that the model output is equal to the model input). the black line shows the baseline genotype concordance, given by guessing the most frequently occurring genotype per marker.

  • A file called encoded_data.h5 containing the projected data (= the encoded data) for all samples at each epoch. This file is used by the plot, evaluate and animate commands.

  • for each saved epoch: a plot of the projected samples colored according to population, and if specified, superpopulation. A legend is written to a separate file.

  • A plot true_genotypes.pdf showing a histogram of the true (input) genotypes that the model is trained on

  • A plot output_as_genotypes.pdf showing a histogram of the model output interpreted as genotypes, for the last epoch

Author(s)

Richèl J.C. Bilderbeek


richelbilderbeek/gcaer documentation built on March 25, 2024, 3:08 p.m.