create_gcae_setup: Create the GCAE setup

In richelbilderbeek/gcaer: Interface to GCAE

Create the GCAE setup

Description

Create the GCAE setup

Usage

create_gcae_setup(
  datadir,
  data,
  superpops,
  model_id,
  train_opts_id,
  data_opts_id,
  pheno_model_id,
  trainedmodeldir = get_gcaer_tempfilename(pattern = "ae_out")
)

Arguments

datadir	directory where sample data is stored. If not absolute: assumed relative to ⁠GenoCAE/⁠ directory. A valid datadir ends with a slash (/) and does not start with a full stop (.), as to mimic the GCAE interface
data	file prefix, not including path, of the data files. The data files must be in EIGENSTRAT or PLINK binary (.bed, .bim, .fam) format)
superpops	path to the superpops file, which contains the labels table and can be read using read_labels_file. This argument is named after the GCAE --superpops CLI flag.
model_id	model id, corresponding to a file models/model_id.json
train_opts_id	train options id, corresponding to a file train_opts/train_opts_id.json
data_opts_id	data options id, corresponding to a file data_opts/data_opts_id.json The data options affect how data is input to the model.
pheno_model_id	phenotype model ID, as checked by check_pheno_model_id
trainedmodeldir	name of the folder to save the trained model to. A valid trainedmodeldir ends with a slash (/) as to mimic the GCAE interface

Value

full paths to the created files

Author(s)

Richèl J.C. Bilderbeek

Examples

create_gcae_setup(
  datadir = "my_datadir/",
  data = "HumanOrigins249_tiny",
  superpops = "HO_superpopulations",
  model_id = "M1",
  train_opts_id = "ex3",
  data_opts_id = "b_0_4",
  pheno_model_id = "p2",
  trainedmodeldir = "ae_out/"
)

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