max/ourMELONS
1
0
Fork 0
Code Issues 1 Pull requests Projects Releases Packages Wiki Activity Actions

Added SAM/BAM support (closes #18)

Browse source
This commit is contained in:
Waldir Leoncio 2021-09-03 12:50:11 +02:00
parent 15f2cbcf08
commit 0cfd13b006
4 changed files with 27 additions and 6 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

14
R/greedyMix.R
View file

@ -1,9 +1,11 @@
#' @title Clustering of individuals #' @title Clustering of individuals
#' @param data data file #' @param data data file
#' @param format Format of the data c("FASTA", "VCF" ,"SAM", or "GenePop") #' @param format Format of the data c("FASTA", "VCF" ,"BAM", or "GenePop")
#' @param verbose if \code{TRUE}, prints extra output information #' @param verbose if \code{TRUE}, prints extra output information
#' @importFrom utils read.delim #' @importFrom utils read.delim
#' @importFrom vcfR read.vcfR #' @importFrom vcfR read.vcfR
#' @references Samtools: a suite of programs for interacting
#' with high-throughput sequencing data. <http://www.htslib.org/>
#' @export #' @export
greedyMix <- function(data, format, verbose = TRUE) { greedyMix <- function(data, format, verbose = TRUE) {
format <- tolower(format) format <- tolower(format)
@ -12,8 +14,14 @@ greedyMix <- function(data, format, verbose = TRUE) {
} else if (format == "vcf") { } else if (format == "vcf") {
out <- vcfR::read.vcfR(data, verbose = verbose) out <- vcfR::read.vcfR(data, verbose = verbose)
} else if (format == "sam") { } else if (format == "sam") {
stop("SAM files not yet supported." ) stop(
# TODO #18: implement load_sam() "SAM files not directly supported. ",
"Install the samtools software and execute ",
"'samtools view -b in_file.sam > out_file.bam' to convert to BAM ",
"and try running this function again with 'format=BAM'"
)
} else if (format == "bam") {
out <- Rsamtools::scanBam(data)
} else if (format == "genepop") { } else if (format == "genepop") {
# TODO #19: implement load_genepop() # TODO #19: implement load_genepop()
stop("GenePop files not yet supported." ) stop("GenePop files not yet supported." )

BIN
inst/ext/bam_example.bam Normal file
View file

Binary file not shown.

6
man/greedyMix.Rd
View file

@ -9,10 +9,14 @@ greedyMix(data, format, verbose = TRUE)
\arguments{ \arguments{
\item{data}{data file} \item{data}{data file}
\item{format}{Format of the data c("FASTA", "VCF" ,"SAM", or "GenePop")} \item{format}{Format of the data c("FASTA", "VCF" ,"BAM", or "GenePop")}
\item{verbose}{if \code{TRUE}, prints extra output information} \item{verbose}{if \code{TRUE}, prints extra output information}
} }
\description{ \description{
Clustering of individuals Clustering of individuals
} }
\references{
Samtools: a suite of programs for interacting
with high-throughput sequencing data. <http://www.htslib.org/>
}

13
tests/testthat/test-greedyMix.R
View file

@ -46,12 +46,21 @@ df_vcf <- greedyMix(
format = "VCF", format = "VCF",
verbose = FALSE verbose = FALSE
) )
# TODO #17: add example reading VCF df_bam <- greedyMix(
# TODO #18: add example reading SAM data = file.path(path_inst, "bam_example.bam"),
format = "BAM",
)
# TODO #19: add example reading Genpop # TODO #19: add example reading Genpop
test_that("Files are imported correctly", { test_that("Files are imported correctly", {
expect_equal(dim(df_fasta), c(5, 99)) expect_equal(dim(df_fasta), c(5, 99))
expect_equal(dim(df_vcf), c(variants = 2, fix_cols = 8, gt_cols = 3)) expect_equal(dim(df_vcf), c(variants = 2, fix_cols = 8, gt_cols = 3))
expect_error(
greedyMix(
data = paste(path_inst, "sam_example.sam", sep="/"),
format = "SAM",
)
)
expect_equal(length(df_bam[[1]]), 13)
}) })
context("Linkage") context("Linkage")