MSIseq (version 1.0.0)

MSIseq.classify: Classify Microsatellite Instability with MSIseq Classifier

Description

This function gives a classification result for MSI status based on classifier with the mutation information in the mutationNum argument.

Usage

MSIseq.classify(mutationNum, classifier = NGSclassifier, cancerType = NULL)

Arguments

mutationNum

A data frame output from Compute.input.variables, which containing 9 variables: T.sns, S.sns, T.ind, S.ind, T, S, Ratio.sns, Ratio.ind, Ratio.

classifier

A RWeka J48 model returned from the function MSIseq.train. Notice that if this argument is missing, the function will use a default build-in classifier, NGSclassifier.

cancerType

A data frame with two columns: Tumor_Sample_Barcode (tumor ID) and the corresponding cancer_type. Check NGStraintype for detail. If the classifier used in this function is trained with cancerType argument, cancerType should be provided here.

Value

A data frame with three columns: Tumor_Sample_Barcode, the corresponding classified MSI_status, and a third column indicating whether the sample is likely POLE deficient.

Details

This function gives a classification of MSI status for the samples with their mutationNum information. The classification is made with the decision tree model in classifier. It also flags the samples with likely POLE deficiency based on the criteria of high T.sns (> 60/Mb) and low S.ind (< 0.18/Mb).

References

Kurt Hornik, Christian Buchta, Achim Zeileis (2009) Open-Source Machine Learning: R Meets Weka. Computational Statistics, 24(2), 225-232.

See Also

MSIseq.train, Compute.input.variables

Examples

## load sample data: test.mutationNum

data(test.mutationNum)
data(NGStestseqLen)


## classify on test data with NGSclassifier (the default classifier)

result <- MSIseq.classify(mutationNum = test.mutationNum)