train_RFmodel: RF Model Training

In rr-2/PeptideRanger:

RF Model Training

Description

Trains a random forest model using a peptidome with reference information

Usage

train_RFmodel(peptidome, reference_name, total_trainingSet)

Arguments

peptidome	List of all possible peptide sequences for relevant proteome (total_trainingSet = FALSE) OR a list of peptide sequences in a custom training set (total_trainingSet = TRUE). Must also include peptide n of observations (n_obs_pep) and parent protein n of observations (n_obs_prot) in a desired reference, and peptide missed cleavages.
reference_name	Name of reference (same as column names)
total_trainingSet	If TRUE, model will be trained on all peptides in a custom input peptidome. If FALSE, peptides will be selected from whole input peptidome to create a training set of 10k peptides (same method as models provided in package).

Value

RF model for predicting peptide detectability

Examples

## Not run: 
CPTAC_peptidome <- peptides_inReference(peptidome = SwissProt2018_peptidome,
                                        reference_name = "CPTAC",
                                        pep_reference = CPTAC_exp_counts,
                                        exp_counts_col = "n_obs_pep",
                                        detection_ratio = TRUE)

train_RFmodel(peptidome = CPTAC_peptidome,
              reference_name = "CPTAC")

## End(Not run)

rr-2/PeptideRanger documentation built on May 27, 2023, 4:43 p.m.