bootstrap_prediction: BootStrap runs for both scGPS training and prediction

In scGPS: A complete analysis of single cell subpopulations, from identifying subpopulations to analysing their relationship (scGPS = single cell Global Predictions of Subpopulation)

Description

ElasticNet and LDA prediction for each of all the subpopulations in the new mixed population after training the model for a subpopulation in the first mixed population. The number of bootstraps to be run can be specified.

Usage

bootstrap_prediction(nboots = 1, genes = genes,
  mixedpop1 = mixedpop1, mixedpop2 = mixedpop2, c_selectID = NULL,
  listData = list(), cluster_mixedpop1 = NULL,
  cluster_mixedpop2 = NULL, trainset_ratio = 0.5, LDA_run = TRUE,
  verbose = FALSE)

Arguments

nboots	a number specifying how many bootstraps to be run
genes	a gene list to build the model
mixedpop1	a SingleCellExperiment object from a mixed population for training
mixedpop2	a SingleCellExperiment object from a target mixed population for prediction
c_selectID	the root cluster in mixedpop1 to becompared to clusters in mixedpop2
listData	a list object, which contains trained results for the first mixed population
cluster_mixedpop1	a vector of cluster assignment for mixedpop1
cluster_mixedpop2	a vector of cluster assignment for mixedpop2
trainset_ratio	a number specifying the proportion of cells to be part of the training subpopulation
LDA_run	logical, if the LDA prediction is added to compare to ElasticNet
verbose	a logical whether to display additional messages

Value

a list with prediction results written in to the index out_idx

Author(s)

Quan Nguyen, 2017-11-25

See Also

bootstrap_parallel for parallel options

Examples

day2 <- day_2_cardio_cell_sample
mixedpop1 <-new_scGPS_object(ExpressionMatrix = day2$dat2_counts, 
    GeneMetadata = day2$dat2geneInfo, CellMetadata = day2$dat2_clusters)
day5 <- day_5_cardio_cell_sample
mixedpop2 <-new_scGPS_object(ExpressionMatrix = day5$dat5_counts, 
    GeneMetadata = day5$dat5geneInfo, CellMetadata = day5$dat5_clusters)
genes <-training_gene_sample
genes <-genes$Merged_unique
cluster_mixedpop1 <- colData(mixedpop1)[,1]
cluster_mixedpop2 <- colData(mixedpop2)[,1]
c_selectID <- 2
test <- bootstrap_prediction(nboots = 1, mixedpop1 = mixedpop1, 
    mixedpop2 = mixedpop2, genes=genes, listData =list(), 
    cluster_mixedpop1 = cluster_mixedpop1, 
    cluster_mixedpop2 = cluster_mixedpop2, c_selectID = c_selectID)
names(test)
test$ElasticNetPredict
test$LDAPredict

scGPS documentation built on Nov. 8, 2020, 5:22 p.m.