Predict the values of the fitted model to an object. — growth

This function fits a Gaussian Process model to a DGrowthR object and then generates predictions from the fitted model. The function supports "alternative" and "null" models. The "alternative" model includes both time and contrast as predictors, while the "null" model only includes time as a predictor.

Usage

growth_comparison(
  object,
  comparison_info,
  predict_n_steps = 100,
  downsample_every_n_timepoints = 1,
  save_gp_data = FALSE,
  permutation_test = FALSE,
  n_permutations = NULL,
  n_cores = 1
)

# S4 method for class 'DGrowthR'
growth_comparison(
  object,
  comparison_info,
  predict_n_steps = 100,
  downsample_every_n_timepoints = 1,
  save_gp_data = FALSE,
  permutation_test = FALSE,
  n_permutations = NULL,
  n_cores = 1
)

Arguments

object: A DGrowthR object to which the model should be fitted and then predictions should be generated.
comparison_info: character vector describing the contrast to be made. The first value indicates the column in metadata where the contrast takes place. The third value is taken as the reference condition.
predict_n_steps: A numeric value indicating the number of timepoints to make a prediction for with the fitted GP model. More points increases estimate accuracy, but adds compute time.
downsample_every_n_timepoints: A numeric value indicating that the OD from every n timepoint should be used for GP fit. Might speed up fitting.
save_gp_data: A logical indicating whether the fitted GP models should be stored for future analysis.
permutation_test: A logical value indicating if a permutation test should be performed.
n_permutations: A numerical values indicating the number of permutations to build in order to gather the null distribution of test statistics.
n_cores: A numeric value indicating the number cores to be used. Useful when performing a permutation test.

Value

An object with the updated growth_comparison slot.