Dev

.Rbuildignore

@@ -1,5 +1,4 @@
 ^docs$
-^_pkgdown\.yml$
 ^.*\.Rproj$
 ^\.Rproj\.user$
 ^cran-comments\.md$
@@ -12,3 +11,5 @@
 ^doc$
 ^Meta$
 ^CRAN-RELEASE$
+pkgdown
+.travis.yml

DESCRIPTION

@@ -1,6 +1,6 @@
 Package: PhenotypeSimulator
 Title: Flexible Phenotype Simulation from Different Genetic and Noise Models
-Version: 0.3.3
+Version: 0.3.4
 Authors@R: c(
   person("Hannah", "Meyer", email = "hannah.v.meyer@gmail.com", role = c("aut", "cre"),
     comment = c(ORCID = "0000-0003-4564-0899")),
@@ -33,6 +33,7 @@ LinkingTo:
 Imports:
     methods,
     optparse,
+    Hmisc,
     R.utils,
     mvtnorm,
     snpStats,

INDEX.Rmd

@@ -53,7 +53,7 @@ knitr::include_graphics("docs/simulatedPhenotypes.png")
 
 ## <i class="fa fa-rocket" aria-hidden="true"></i> Installation
 
-The current github version of *PhenotypeSimulator* is 0.3.3 and can be
+The current github version of *PhenotypeSimulator* is 0.3.4 and can be
 installed via:
 ```{r, eval=FALSE}
 library(devtools)

INDEX.md

@@ -43,7 +43,7 @@ customised.
 
 ## <i class="fa fa-rocket" aria-hidden="true"></i> Installation
 
-The current github version of *PhenotypeSimulator* is 0.3.3 and can be
+The current github version of *PhenotypeSimulator* is 0.3.4 and can be
 installed via:
 
 ``` r

NEWS.md

@@ -1,3 +1,18 @@
+# PhenotypeSimulator 0.3.4
+## Minor changes
+1. Fixed missing --genotypefile flag [issue 27](https://github.com/HannahVMeyer/PhenotypeSimulator/issues/27)
+2. Update vignettes with new location for impute files and commands to get the
+CEU samples [issue 24](https://github.com/HannahVMeyer/PhenotypeSimulator/issues/24),
+thanks to @zfuller5280 for the suggestion!)
+3. Standardise genotypes on row with major alleles [issue 21](https://github.com/HannahVMeyer/PhenotypeSimulator/issues/21).
+Thank you for the detailed bug report by @alanw1!
+4. Add option to imput missing genotypes to standardise genotype function;
+otherwise, if genotypes are missing, function will fail
+[issue 17](https://github.com/HannahVMeyer/PhenotypeSimulator/issues/17)
+5. Fix function description and passing of SNP IDs in readStandardGenotypes with
+delim option [issue 25](https://github.com/HannahVMeyer/PhenotypeSimulator/issues/25),
+thanks @BSchmidt1.
+
 # PhenotypeSimulator 0.3.3
 ## Minor changes
 1. Fixed bug that failed to return causal SNP name when only one SNP was chosen

R/commandlineFunctions.R

@@ -524,6 +524,7 @@ simulatePhenotypes <- function() {
                                      header=args$header,
                                      sampleID=args$sampleID,
                                      phenoID=args$phenoID,
+                                     genotypefile = args$genotypefile,
                                      genoFilePrefix=args$genoFilePrefix,
                                      genoFileSuffix=args$genoFileSuffix,
                                      SNPfrequencies=SNPfrequencies,

R/genotypeFunctions.R

@@ -30,8 +30,15 @@ getAlleleFrequencies <- function(snp) {
 #' Genotypes are standardised as described in Yang et al:
 #' snp_standardised = (snp - 2 * ref_allele_freq)/
 #' sqrt(2 * ref_allele_freq * alt_allele_freq).
+#' 
+#' Missing genotypes can be mean-imputed and rounded to nearest integer
+#' before standardisation. If genotypes contain missing values and impute is set
+#' to FALSE, \code{standardiseGenotypes} will return an error.
 #'
 #' @param geno [N x NrSNP] Matrix/dataframe of genotypes [integer]/[double].
+#' @param impute [logical] Indicating if missing genotypes should be imputed; if
+#' set FALSE and data contains missing values,  \code{standardiseGenotypes} will
+#' return an error.
 #' @return [N x NrSNP] Matrix of standardised genotypes [double].
 #' @seealso \code{\link{getAlleleFrequencies}}
 #' @export
@@ -40,11 +47,19 @@ getAlleleFrequencies <- function(snp) {
 #' @examples
 #' geno <- cbind(rbinom(2000, 2, 0.3), rbinom(2000, 2, 0.4),rbinom(2000, 2, 0.5))
 #' geno_sd <- standardiseGenotypes(geno)
-standardiseGenotypes <- function(geno) {
+standardiseGenotypes <- function(geno, impute=FALSE) {
+    if (any(is.na(geno)) & !impute) {
+        stop("Missing genotypes found and impute=FALSE, cannot standardise",
+             "genotypes; remove missing genotypes or set impute=TRUE for mean",
+             "imputation of genotypes")
+    }
+    if (any(is.na(geno)) & impute) {
+        geno <- round(apply(as.matrix(geno), 2, Hmisc::impute, fun=mean))
+    }
     allele_freq <-  sapply(data.frame(geno),  getAlleleFrequencies)
     var_geno <- sqrt(2*allele_freq[1,]*allele_freq[2,])
     var_geno[var_geno == 0] <- 1
-    geno_mean <- sweep(geno, 2, 2*allele_freq[1,], "-")
+    geno_mean <- sweep(geno, 2, 2*allele_freq[2,], "-")
     geno_sd <- sweep(geno_mean, 2, var_geno, "/")
     return (geno_sd)
 }
@@ -177,12 +192,12 @@ simulateGenotypes <- function(N, NrSNP=5000, frequencies=c(0.1, 0.2, 0.4),
 #' with minor allele first. The remaining columns are the mean genotypes of
 #' different individuals – numbers between 0 and 2 that represents the
 #' (posterior) mean genotype, or dosage of the minor allele.
-#' \item delim: a [delimter]-delimited file of [NrSNPs x NrSamples] genotypes
-#' with the snpIDs in the first column and the sampleIDs in the first row and
-#' genotypes encoded as numbers between 0 and 2 representing the (posterior)
-#' mean genotype, or dosage of the minor allele. Can be user-genotypes or
-#' genotypes simulated with foward-time algorithms such as simupop
-#' (\url{http://simupop.sourceforge.net/Main/HomePage}) or MetaSim
+#' \item delim: a [delimter]-delimited file of [(NrSNPs+1) x (NrSamples+1)]
+#' genotypes with the snpIDs in the first column and the sampleIDs in the first
+#' row and genotypes encoded as numbers between 0 and 2 representing the
+#' (posterior) mean genotype, or dosage of the minor allele. Can be
+#' user-genotypes or genotypes simulated with foward-time algorithms such as
+#' simupop (\url{http://simupop.sourceforge.net/Main/HomePage}) or MetaSim
 #' (\url{project.org/web/packages/rmetasim/vignettes/CreatingLandscapes.html}),
 #' that allow for user-specified output formats.
 #' }}
@@ -218,6 +233,12 @@ simulateGenotypes <- function(N, NrSNP=5000, frequencies=c(0.1, 0.2, 0.4),
 #' filename_plink <- gsub("\\.bed", "", filename_plink)
 #' data_plink <- readStandardGenotypes(N=100, filename=filename_plink,
 #' format="plink")
+#' 
+#' filename_delim  <- system.file("extdata/genotypes/",
+#' "genotypes_chr22.csv",
+#' package = "PhenotypeSimulator")
+#' data_delim <- readStandardGenotypes(N=50, filename=filename_delim,
+#' format="delim")
 readStandardGenotypes <- function(N, filename, format = NULL,
                                   verbose=TRUE, sampleID = "ID_",
                                   snpID = "SNP_", delimiter = ",") {
@@ -300,11 +321,11 @@ readStandardGenotypes <- function(N, filename, format = NULL,
         }
         id_samples <- paste(sampleID, 1:N, "_", gsub(":", "", data$V1), sep="")
         id_snps <- paste(snpID, 0:(ncol(genotypes) -1), sep="")
-        format_files = NULL
+        format_files <- NULL
     } else if (format == "delim") {
-        data <- data.table::fread(filename, data.table=FALSE,
+        data <- data.table::fread(filename, data.table=FALSE, header=TRUE,
                                   sep=delimiter)
-        id_snps <- data$V1
+        id_snps <- data[,1]
         genotypes <- t(data[,-1])
         colnames(genotypes) <- id_snps
         if (N > nrow(genotypes)) {

README.Rmd

@@ -51,7 +51,7 @@ be customised.
 Full documentation of **PhenotypeSimulator** is available at
 http://HannahVMeyer.github.io/PhenotypeSimulator/.
 
-The current github version of *PhenotypeSimulator* is 0.3.3 and can be
+The current github version of *PhenotypeSimulator* is 0.3.4 and can be
 installed via
 
 ```{r gh-installation, eval = FALSE}

README.md

@@ -44,7 +44,7 @@ customised.
 Full documentation of **PhenotypeSimulator** is available at
 <http://HannahVMeyer.github.io/PhenotypeSimulator/>.
 
-The current github version of *PhenotypeSimulator* is 0.3.3 and can be
+The current github version of *PhenotypeSimulator* is 0.3.4 and can be
 installed via
 
 ``` r