predict_suretype: Run classification model using combination of Random Forest...

In Meiomap/SureTypeSCR: Library for (single-cell) SNP array processing

Description

Cascade classification model with Random Forest in the first layer and Gaussian Discriminant Analysis in the second layer. This function analyzes the whole dataset comprised in .data in one batch. To minimize bias in the model, it is recommended to run suretype_model(.) and build a model per sample basic instead of the whole dataset.

Usage

predict_suretype(.data, rf_clf)

Arguments

.data	Input dataframe coming from scbasic() or compatible
rf_clf	Instance of classifier loaded using scload currently embodied in RF

Value

data frame decorated with SureTypeSC genotyping score

Examples

setwd(system.file(package='SureTypeSCR'))
samplesheet=system.file('files/GSE19247_example.csv',package='SureTypeSCR')
manifest=system.file('files/HumanCytoSNP-12v2_H.bpm',package='SureTypeSCR')
cluster=system.file('files/HumanCytoSNP-12v2_H.egt',package='SureTypeSCR')
clf=system.file('files/rf.clf',package='SureTypeSCR')


#Load data
df=scbasic(manifest,cluster,samplesheet)

# The Random Forest classifier
clf_instance <- scload(clf) 

#assign prediction results back to the original dataframe using margittr %<>% operator
#df %<>% predict_suretype(clf_instance) 

Meiomap/SureTypeSCR documentation built on Dec. 17, 2021, 3:22 a.m.