tests/testthat/test_vcfR2DNAbin.R

In knausb/vcfR: Manipulate and Visualize VCF Data

#
library(testthat)
library(vcfR)
context("vcfR2DNAbin functions")

#data(vcfR_example)
#gene <- dna[1,gff[1,4]:gff[1,5]]

#
#
# Missing data handling ----
#
#


test_that("Works with no variants",{
  data(vcfR_example)
  vcf <- vcf[0,]
  my_DNAbin <- vcfR2DNAbin( vcf, verbose = FALSE )
  expect_true( inherits(my_DNAbin, "DNAbin") )
  expect_equal(dim(my_DNAbin)[1], ncol(vcf@gt) - 1 )
})

test_that("Works with only indels, no SNPs",{
  data(vcfR_example)
  vcf <- extract.indels( vcf, return.indels = TRUE)
  my_DNAbin <- vcfR2DNAbin( vcf, verbose = FALSE )
  expect_true( inherits(my_DNAbin, "DNAbin") )
  expect_equal(dim(my_DNAbin)[1], ncol(vcf@gt) - 1 )
})

test_that("Works with no variants, ref.seq is not NULL",{
  data(vcfR_example)
  gene <- dna[1,gff[1,4]:gff[1,5]]
  vcf <- vcf[0,]
  my_DNAbin <- vcfR2DNAbin( vcf, ref.seq = gene, start.pos = gff[1,4], verbose = FALSE )
  expect_true( inherits(my_DNAbin, "DNAbin") )
  expect_equal( dim(my_DNAbin)[1], ncol(vcf@gt) - 1 )
  expect_equal( dim(my_DNAbin)[2], length(gene) )
  expect_equal( length(ape::seg.sites(my_DNAbin)), 0 )
  expect_equal( sum(ape::base.freq(my_DNAbin) > 0), 4 )
})


test_that("Works with one variant, ref.seq is not NULL",{
  data(vcfR_example)
  gene <- dna[1,gff[1,4]:gff[1,5]]
  vcf <- vcf[1,]
  my_DNAbin <- vcfR2DNAbin( vcf, ref.seq = gene, start.pos = gff[1,4], verbose = FALSE )
  expect_true( inherits(my_DNAbin, "DNAbin") )
  expect_equal( dim(my_DNAbin)[1], (ncol(vcf@gt) - 1) * 2 )
  expect_equal( dim(my_DNAbin)[2], length(gene) )
  expect_equal( length(ape::seg.sites(my_DNAbin)), 0 )
  expect_equal( sum(ape::base.freq(my_DNAbin) > 0), 4 )
})




#
#
# Fabricate data which varies in ploidy ----
#
#

data(vcfR_example)

# Fabricate a vcf of haploid data.
#haps <- extract.haps(vcf)
haps <- extract.gt(vcf, return.alleles=FALSE
#                   , allele.sep="|"
                   )
haps <- apply(haps, MARGIN=2, function(x){unlist(lapply(strsplit(x, split="|"), function(x){x[1]}))})
gt2 <- extract.gt(vcf, extract = FALSE)
gt2 <- matrix( paste( haps, gt2, sep=":"), nrow=nrow(haps), dimnames=dimnames(haps) )
is.na(gt2[is.na(haps)]) <- TRUE
vcf1 <- vcf
vcf1@gt <- cbind(vcf@gt[,'FORMAT'], gt2)
colnames(vcf1@gt)[1] <- 'FORMAT'
rm(haps)
rm(gt2)


# Fabricate a vcf of triploid data.
#haps <- extract.haps(vcf)
haps <- extract.gt(vcf, return.alleles=FALSE
#                   , allele.sep="|"
                   )
haps2 <- apply(haps, MARGIN=2, function(x){unlist(lapply(strsplit(x, split="|"), function(x){x[1]}))})
haps <- matrix( paste( haps, haps2, sep="|"), nrow=nrow(haps), dimnames=dimnames(haps) )
is.na(haps[is.na(haps2)]) <- TRUE
gt2 <- extract.gt(vcf, extract = FALSE)
gt2 <- matrix( paste( haps, gt2, sep=":"), nrow=nrow(haps), dimnames=dimnames(haps) )
is.na(gt2[is.na(haps)]) <- TRUE
vcf3 <- vcf
vcf3@gt <- cbind(vcf@gt[,'FORMAT'], gt2)
colnames(vcf3@gt)[1] <- 'FORMAT'
rm(haps)
rm(gt2)


#
#
# Test haploid data ----
#
#

test_that("vcfR2DNAbin works for haploid data, no ref.seq",{
#  my_DNAbin <- vcfR2DNAbin( vcf1, gt.split = "|" )
  my_DNAbin <- vcfR2DNAbin( vcf1 )
  expect_equal( dim(my_DNAbin)[2], nrow( extract.indels(vcf)@gt ) )
})

test_that("vcfR2DNAbin works for haploid data, no ref.seq",{
  data(vcfR_test)
  
  # Haploidize
  my_non_gt <- extract.gt(vcfR_test, extract = FALSE)
  my_gt <- extract.gt(vcfR_test)
  my_gt <- unlist(lapply(strsplit(my_gt, split = "[/|]"), function(x){x[1]}))
  # https://stackoverflow.com/a/35589023
  my_non_gt <- paste(my_gt, my_non_gt, sep = ":")
  dim(my_non_gt) <- dim(vcfR_test@gt[,-1])
  vcfR_test@gt[,-1] <- my_non_gt
  
  my_dnabin <- vcfR2DNAbin(vcfR_test)
  expect_equal(grep("[acgt]", as.character(my_dnabin), invert = TRUE), integer(0))
})


#
#
# Test diploid data ----
#
#


test_that("vcfR2DNAbin works for diploid data, no ref.seq",{
#  my_DNAbin <- vcfR2DNAbin( vcf, gt.split = "|", verbose = FALSE )
  my_DNAbin <- vcfR2DNAbin( vcf, verbose = FALSE )
  expect_true( inherits(my_DNAbin, "DNAbin") )  
  expect_equal( dim(my_DNAbin)[2], nrow( extract.indels(vcf)@gt ) )
  expect_equal( dim(my_DNAbin)[1], 2 * (ncol(vcf@gt) - 1) )
})


#test_that("vcfR2DNAbin works for diploid data, no ref.seq, wrong gt.split",{
#  
#})


test_that("vcfR2DNAbin with consensus works",{
#  my_DNAbin <- vcfR2DNAbin( vcf, consensus = TRUE, extract.haps = FALSE, gt.split = "|" )
  my_DNAbin <- vcfR2DNAbin( vcf, consensus = TRUE, extract.haps = FALSE )
  expect_true( inherits(my_DNAbin, "DNAbin") )
  expect_equal( dim(my_DNAbin)[1], ncol(vcf@gt) - 1 )
  expect_equal( dim(my_DNAbin)[2], nrow(extract.indels(vcf)@gt) )
})


test_that("vcfR2DNAbin works for diploid data, ref.seq is not NULL",{
#  my_DNAbin <- vcfR2DNAbin( vcf, gt.split = "|", ref.seq = gene, start.pos = gff[1,4], verbose = FALSE )
  data(vcfR_example)
  gene <- dna[1,gff[1,4]:gff[1,5]]
  my_DNAbin <- vcfR2DNAbin( vcf, ref.seq = gene, start.pos = gff[1,4], verbose = FALSE )
  expect_true( inherits(my_DNAbin, "DNAbin") )
  expect_equal( dim(my_DNAbin)[1], 2 * (ncol(vcf@gt) - 1) )
  expect_equal( dim(my_DNAbin)[2], dim(gene)[2] )
  expect_true( length(ape::seg.sites(my_DNAbin)) > 0 )
#  expect_equal( dim(my_DNAbin)[2], nrow( extract.indels(vcf)@gt ) )
})



#
#
# Test triploid data ----
#
#


test_that("vcfR2DNAbin works for triploid data, no ref.seq",{
#  my_DNAbin <- vcfR2DNAbin( vcf3, gt.split = "|", verbose = FALSE )
  my_DNAbin <- vcfR2DNAbin( vcf3, verbose = FALSE )
  expect_true( inherits(my_DNAbin, "DNAbin") )  
  expect_equal( dim(my_DNAbin)[2], nrow( extract.indels(vcf)@gt ) )
  expect_equal( dim(my_DNAbin)[1], 3 * (ncol(vcf@gt) - 1) )
})


#
#
# Variant at end of locus ----
#
#


test_that("vcfR2DNAbin does not include variants at end.pos + 1",{
  data(vcfR_example)
  gene <- dna[1,gff[1,4]:gff[1,5]]
#  vcf@fix[586,1:6]
#  vcf@fix[586,2] <- "24527"
  vcf@fix[586,2] <- "24528"
#  my_DNAbin <- vcfR2DNAbin( vcf, gt.split = "|", ref.seq = gene, start.pos = gff[1,4], verbose = FALSE )
  my_DNAbin <- vcfR2DNAbin( vcf, ref.seq = gene, start.pos = gff[1,4], verbose = FALSE )
  expect_true( inherits(my_DNAbin, "DNAbin") )  
  expect_equal(length(ape::seg.sites(my_DNAbin)), 40)
})


#
#
# Asterisk allele ----
#
#


test_that("vcfR2DNAbin manages the asterisk allele",{
  data(vcfR_test)
  vcfR_test@fix[,'ALT'][3] <- "*,T"
  myDNA <- vcfR2DNAbin(vcfR_test, asterisk_as_del = TRUE, verbose = FALSE)
  myDNA <- as.character(myDNA)
  expect_equal(sum(myDNA[c(1,4),3] == "-"), 2)
  
  data(vcfR_test)
  vcfR_test@fix[,'ALT'][3] <- "*,T"
  myDNA <- vcfR2DNAbin(vcfR_test, asterisk_as_del = FALSE, verbose = FALSE)
  myDNA <- as.character(myDNA)
  expect_equal(sum(myDNA[c(1,4),3] == "n"), 2)
})


#
#
# Indels ----
#
#

test_that("vcfR2DNAbin manages indels, no reference",{
  data(vcfR_test)
  myDNA <- vcfR2DNAbin(vcfR_test, extract.indels = FALSE, unphased_as_NA = FALSE, verbose = FALSE)
  expect_is(myDNA, 'DNAbin')
  expect_equal(dim(ape::as.character.DNAbin(myDNA)), c(6,8))
})

test_that("vcfR2DNAbin manages indels with reference",{
  data(vcfR_test)
  data(vcfR_example)
  dna <- dna[1:12]
  vcfR_test@fix[,'POS'] <- c(2, 4, 6, 8, 10)
  myDNA <- vcfR2DNAbin(vcfR_test, extract.indels = FALSE, 
                       unphased_as_NA = FALSE, ref.seq = dna,
                       start.pos=1, verbose = FALSE)
  expect_is(myDNA, 'DNAbin')
  expect_equal(dim(ape::as.character.DNAbin(myDNA)), c(6,15))
})


####
# EOF. ----