REPTILE (version 1.0)

reptile_train: Learn a REPTILE enhancer model

Description

Learn a REPTILE enhancer model based on epigenomic signature of known enhancers.

Usage

reptile_train(epimark_region, label_region, epimark_DMR = NULL, label_DMR = NULL, family = "randomForest", ntree = 2000, nodesize = 1)

Arguments

epimark_region

data.frame instance from read_epigenomic_data, which containing intensity and intensity deviation values of each mark for each query region.

label_region

factor instance from read_label, containing the label of each query region. The possible values and their meanings of a label are: 0 (not enhancer), 1 (enhancer) and NA (unknwon and it will be ignored).

epimark_DMR

data.frame instance from read_epigenomic_data, which containing intensity and intensity deviation values of each mark for each DMR. If either this value or label_DMR is NULL, the output enhancer model will not inlclude a classifier for predicting the enhancer activities of DMRs. Default: NULL

label_DMR

factor instance from read_label, containing the label of each DMR. The possible values and their meanings of a label are: 0 (not enhancer), 1 (enhancer) and NA (unknwon and it will be ignored). If either this value or label_DMR is NULL, the output enhancer model will not inlclude a classifier for predicting the enhancer activities of DMRs. Default: NULL

family

classifier family used in the enhancer model

Default: RandomForest

Classifiers available:

- RandomForest: random forest

- Logistic: logistic regression

ntree

Number of tree to be constructed in the random forest model. See the function randomForest() in "randomForest" package for more information. Default: 2000

nodesize

Minimum size of terminal nodes. See the function randomForest() in "randomForest" package for more information. Default: 1

Value

A list containing two objects of class randomForest.

D

Classifier for DMRs. It is an randomForest object or glm object when family is set to be "Logistic".

R

Classifier for query regions. It is an randomForest object or glm object when family is set to be "Logistic".

References

Breiman, L. (2001), Random Forests, Machine Learning 45(1), 5-32.

A. Liaw and M. Wiener (2002), Classification and Regression by randomForest, R News 2(3), 18--22.

See Also

read_epigenomic_data, read_label, reptile_predict

Examples

library("REPTILE")
data("rsd")

## Training
rsd_model <- reptile_train(rsd$training_data$region_epimark,
                           rsd$training_data$region_label,
                           rsd$training_data$DMR_epimark,
                           rsd$training_data$DMR_label,
                           ntree=5)

print(rsd_model$D)
print(rsd_model$R)