snpRFImpute: Missing Value Imputations by snpRF

In snpRF: Random Forest for SNPs to Prevent X-chromosome SNP Importance Bias

Description

Impute missing values (in covariate predictor data only) using proximity from snpRF.

Usage

snpRFImpute(x.autosome=NULL,x.xchrom=NULL,x.covar, y, iter=5, ntree=300, ...)

Arguments

x.autosome	A matrix of autosomal markers with each column corresponding to a SNP coded as count of a particular allele (i.e. 0,1 or 2), and each row corresponding to a sample/individual.(NA's not allowed).
x.xchrom	A matrix of X chromosome markers, each marker coded as two adjacent columns, alleles of a marker are coded as 0 or 1 for carrying a particular allele. Since males only have one X-chromosome, their markers are 2 columns as well, the second column being a duplicate of the first. Each row of this matrix corresponsponds to a sample/individual. This data must be phased in chromosomal order. (NA's not allowed).
x.covar	A matrix of covariates, each column being a different covariate and each row, a sample/individual, some entries with NA.
y	Response vector, must be a factor (NA's not allowed).
iter	Number of iterations to run the imputation.
ntree	Number of trees to grow in each iteration of randomForest.
...	Other arguments to be passed to snpRF.

Details

It is assumed that the genetic data (autosome and x-chromosome) would have missing values imputed in a different manner prior to using snpRF (i.e. using a program specifically intended for imputing SNP data). Thus only missing values in the covariate matrix can be imputed using data from genetic and covariate matrices.

The algorithm starts by imputing NAs using na.roughfix. Then snpRF is called with the completed data. The proximity matrix from the randomForest is used to update the imputation of the NAs. For continuous predictors, the imputed value is the weighted average of the non-missing obervations, where the weights are the proximities. For categorical predictors, the imputed value is the category with the largest average proximity. This process is iterated iter times.

Note: Imputation has not (yet) been implemented for the unsupervised case. Also, Breiman (2003) notes that the OOB estimate of error from randomForest tend to be optimistic when run on the data matrix with imputed values.

Value

A matrix containing the completed covariate matrix, where NAs are imputed using proximity from randomForest. The first column contains the response.

Author(s)

Andy Liaw, with slight modifications by Greg Jenkins

References

Leo Breiman (2003). Manual for Setting Up, Using, and Understanding Random Forest V4.0. http://oz.berkeley.edu/users/breiman/Using_random_forests_v4.0.pdf

See Also

na.roughfix.

Examples

data(snpRFexample)
covar.na <- covariates
set.seed(111)
## artificially drop some data values.
for (i in 1:2) covar.na[sample(200, sample(20)), i] <- NA
set.seed(222)
eg.imputed <- snpRFImpute(x.autosome=autosome.snps,x.xchrom=xchrom.snps,
                          x.covar=covar.na, y=phenotype)

set.seed(333)
eg.rf <- snpRF(x.autosome=autosome.snps,x.xchrom=xchrom.snps,
               xchrom.names=xchrom.snps.names,x.covar=eg.imputed, 
               y=phenotype)
print(eg.rf)

snpRF documentation built on May 2, 2019, 6:51 a.m.