data.surrogateProfiles {CTD}

Generate surrogate profiles

Description

Fill in a data matrix rank with surrogate profiles., when your data is low n, high p.

Usage

data.surrogateProfiles(data, std = 1, ref_data)

Arguments

data

- Data matrix with observations (e.g., patient samples) as columns, features (e.g., metabolites or genes) as rows

std

- The level of variability (standard deviation) around each observed feature's z-score you want to add to generate the surrogate profiles

ref_data

- Data matrix for healthy control "reference" samples, observations (e.g., patient samples) as columns, features (e.g., metabolites or genes) as rows

Value

data_mx_surr - Data matrix with added surrogate profiles

Examples

data("Miller2015")
data_mx=Miller2015[-1,grep("IEM_", colnames(Miller2015))]
data_mx=apply(data_mx, c(1,2), as.numeric)
diags=unlist(Miller2015["diagnosis",grep("IEM_", colnames(Miller2015))])
refs=data_mx[,which(diags=="No biochemical genetic diagnosis")]
ref_fill=as.numeric(Miller2015$`Times identifed in all 200 samples`[-1])/200
refs2=refs[which(ref_fill>0.8),]
diag_pts=names(diags[which(diags==unique(diags)[1])])
diag_data=data_mx[which(rownames(data_mx) %in% rownames(refs2)), 
                    which(colnames(data_mx) %in% diag_pts)]
data_mx_surr=data.surrogateProfiles(data=diag_data, std=1, ref_data=refs2)