add_deconvolution_training_model_bseqsc: add_deconvolution_training_model_bseqsc

In RaikOtto/ArtDeco: Alikeness Based on Regression on Transciptome DECOnvolution

Description

add_deconvolution_training_model_bseqsc adds a new model

Usage

add_deconvolution_training_model_bseqsc(
    transcriptome_data,
    model_name,
    subtype_vector,
    marker_gene_list,
    training_p_value_threshold,
    training_nr_permutations,
    training_nr_marker_genes
)

Arguments

transcriptome_data	Path to transcriptomic data to be used for training. Has to contain the cell subtypes to which the similarity will be calculated. Note that the row names have to contain the HGNC symbols and the column names the sample names.
model_name	Name of the model.
subtype_vector	Character vector containing the subtype labels of the training data samples (transcriptome_data).
marker_gene_list	List that contains the marker genes for each subtype. Has to be in the type of list() with each subtype being an entry.
training_p_value_threshold	P-value at which a training is deemed successfull. Default value 0.05.
training_nr_permutations	Amount of perturbation which results in a p-value. Higher number of perturbation generally improves the p-value estiamtes. Default 100.
training_nr_marker_genes	How many genes should be utilized as list of marker genes. Default 100.

Value

Stores a new model in the package directory

Examples

data("Lawlor") # Data from Lawlor et al.
data(meta_data)

# extract the training sample subtype labels
subtype_vector = as.character(meta_data$Subtype)

add_deconvolution_training_model_bseqsc(
    transcriptome_data = Lawlor,
    model_name = "my_model",
    subtype_vector = subtype_vector,
    training_nr_permutations = 10,
    training_nr_marker_genes = 100
)

RaikOtto/ArtDeco documentation built on Oct. 30, 2021, 6:20 p.m.