Package 'VanillaICE'

Description

A wrapper for the function acf that returns the autocorrelation for the specified lag. Missing values are removed.

Usage

acf2(x, lag = 10, ...)

Arguments

See Also

acf

Description

ArrayViews provides views to the low-level data – log R ratios, B allele frequencies, and genotypes that are stored in parsed files on disk, often scaled and coerced to an integer. Accessors to the low-level data are provided that extract the marker-level summaries from disk, rescaling when appropriate.

Usage

ArrayViews(
  class = "ArrayViews",
  colData,
  rowRanges = GRanges(),
  sourcePaths = character(),
  scale = 1000,
  sample_ids,
  parsedPath = getwd(),
  lrrFiles = character(),
  bafFiles = character(),
  gtFiles = character()
)

## S4 method for signature 'ArrayViews,ANY,ANY,ANY'
x[i, j, ..., drop = FALSE]

colnames(x) <- value

## S4 method for signature 'ArrayViews'
colnames(x, do.NULL = TRUE, prefix = "col")

## S4 method for signature 'ArrayViews'
x$name

## S4 replacement method for signature 'ArrayViews'
x$name <- value

## S4 method for signature 'ArrayViews'
show(object)

## S4 method for signature 'ArrayViews'
sapply(X, FUN, ..., simplify = TRUE, USE.NAMES = TRUE)

## S4 method for signature 'ArrayViews'
ncol(x)

## S4 method for signature 'ArrayViews'
nrow(x)

## S4 method for signature 'ArrayViews'
dim(x)

## S4 method for signature 'ArrayViews'
start(x)

Arguments

Slots

colData

A character string

rowRanges

A DataFrame. WARNING: The accessor for this slot is rowRanges, not rowRanges!

index

A GRanges object

sourcePaths

A character string providing complete path to source files (one file per sample) containing low-level summaries (Log R ratios, B allele frequencies, genotypes)

scale

A length-one numeric vector

parsedPath

A character string providing full path to where parsed files should be saved

lrrFiles

character vector of filenames for log R ratios

bafFiles

character vector of filenames for BAFs

gtFiles

character vector of filenames for genotypes

See Also

CopyNumScanParams parseSourceFile

Examples

ArrayViews()
## From unit test
  require(BSgenome.Hsapiens.UCSC.hg18)
  require(data.table)
  extdir <- system.file("extdata", package="VanillaICE", mustWork=TRUE)
  features <- suppressWarnings(fread(file.path(extdir, "SNP_info.csv")))
  fgr <- GRanges(paste0("chr", features$Chr), IRanges(features$Position, width=1),
                 isSnp=features[["Intensity Only"]]==0)
  fgr <- SnpGRanges(fgr)
  names(fgr) <- features[["Name"]]
  bsgenome <- BSgenome.Hsapiens.UCSC.hg18
  seqlevels(fgr, pruning.mode="coarse") <- seqlevels(bsgenome)[seqlevels(bsgenome) %in% seqlevels(fgr)]
  seqinfo(fgr) <- seqinfo(bsgenome)[seqlevels(fgr),]
  fgr <- sort(fgr)
  files <- list.files(extdir, full.names=TRUE, recursive=TRUE, pattern="FinalReport")
  ids <- gsub(".rds", "", gsub("FinalReport", "", basename(files)))
  views <- ArrayViews(rowRanges=fgr,
                      sourcePaths=files,
                      sample_ids=ids)
  lrrFile(views)
  ## view of first 10 markers and samples 3 and 5
  views <- views[1:10, c(3,5)]

Description

This function is not meant to be called directly by the user. It is exported in the package NAMESPACE for internal use by other BioC packages.

Usage

baumWelchUpdate(param, assay_list)

Arguments

Description

Given the data and an object containing parameters for the HMM, this function computes emission probabilities. This function is not intended to be called by the user and is exported for internal use by other BioC packages.

Usage

calculateEmission(x, param = EmissionParam())

Arguments

Value

A matrix of emission probabilities. Column correspond to the HMM states and rows correspond to markers on the array (SNPs and nonpolymorphic markers)

See Also

baumWelchUpdate

Description

Parameters for computing emission probabilities for a 6-state HMM, including starting values for the mean and standard deviations for log R ratios (assumed to be Gaussian) and B allele frequencies (truncated Gaussian), and initial state probabilities.

This function is exported primarily for internal use by other BioC packages.

Usage

cn_means(object)

cn_sds(object)

baf_means(object)

baf_sds(object)

baf_means(object) <- value

baf_sds(object) <- value

cn_sds(object) <- value

cn_means(object) <- value

EmissionParam(
  cn_means = CN_MEANS(),
  cn_sds = CN_SDS(),
  baf_means = BAF_MEANS(),
  baf_sds = BAF_SDS(),
  initial = rep(1/6, 6),
  EMupdates = 5L,
  CN_range = c(-5, 3),
  temper = 1,
  p_outlier = 1/100,
  modelHomozygousRegions = FALSE
)

EMupdates(object)

## S4 method for signature 'EmissionParam'
show(object)

Arguments

Details

The log R ratios are assumed to be emitted from a normal distribution with a mean and standard deviation that depend on the latent copy number. Similarly, the BAFs are assumed to be emitted from a truncated normal distribution with a mean and standard deviation that depends on the latent number of B alleles relative to the total number of alleles (A+B).

Value

numeric vector

Details

When modelHomozygousRegions is FALSE (the default in versions >= 1.28.0), emission probabilities for B allele frequences are calculated from a mixture of a truncated normal densities and a Unif(0,1) density with the mixture probabilities given by the probability that a SNP is homozygous. In particular, let p denote a 6 dimensional vector of density estimates from a truncated normal distribution for the latent genotypes 'A', 'B', 'AB', 'AAB', 'ABB', 'AAAB', and 'ABBB'. The probability that a genotype is homozygous is estimated as

$prHom=(p["A"] + p["B"])/sum(p)$

and the probability that the genotype is heterozygous (any latent genotype that is not 'A' or 'B') is given by

$prHet = 1-prHom$

Since the density of a Unif(0,1) is 1, the 6-dimensional vector of emission probability at a SNP is given by

$emit = prHet * p + (1-prHet)$

The above has the effect of minimizing the influence of BAFs near 0 and 1 on the state path estimated by the Viterbi algorithm. In particular, the emission probability at homozygous SNPs will be virtually the same for states 3 and 4, but at heterozygous SNPs the emission probability for state 3 will be an order of magnitude greater for state 3 (diploid) compared to state 4 (diploid region of homozygosity). The advantage of this parameterization are fewer false positive hemizygous deletion calls. [ Log R ratios tend to be more sensitive to technical sources of variation than the corresponding BAFs/ genotypes. Regions in which the log R ratios are low due to technical sources of variation will be less likely to be interpreted as evidence of copy number loss if heterozygous genotypes have more 'weight' in the emission estimates than homozgous genotypes. ] The trade-off is that only states estimated by the HMM are those with copy number alterations. In particular, copy-neutral regions of homozygosity will not be called.

By setting modelHomozygousRegions = TRUE, the emission probabilities at a SNP are given simply by the p vector described above and copy-neutral regions of homozygosity will be called.#'

Examples

ep <- EmissionParam()
cn_means(ep)
ep <- EmissionParam()
cn_sds(ep)
ep <- EmissionParam()
baf_means(ep)
ep <- EmissionParam()
baf_sds(ep)
ep <- EmissionParam()
baf_means(ep) <- baf_means(ep)
ep <- EmissionParam()
baf_sds(ep) <- baf_sds(ep)
ep <- EmissionParam()
cn_sds(ep) <- cn_sds(ep)
ep <- EmissionParam()
cn_means(ep) <- cn_means(ep)
ep <- EmissionParam()
show(ep)
cn_means(ep)
cn_sds(ep)
baf_means(ep)
baf_sds(ep)

Description

The HMM-derived genomic ranges are represented as a GRanges-derived object. cnvFilter returns a GRanges object using the filters stipulated in the filters argument.

Usage

cnvFilter(object, filters = FilterParam())

cnvSegs(object, filters = FilterParam(state = c("1", "2", "5", "6")))

duplication(object, filters = FilterParam(state = c("5", "6")))

deletion(object, filters = FilterParam(state = c("1", "2")))

hemizygous(object, filters = FilterParam(state = "2"))

homozygous(object, filters = FilterParam(state = "1"))

## S4 method for signature 'HMM'
cnvSegs(object, filters = FilterParam(state = as.character(c(1, 2, 5, 6))))

## S4 method for signature 'HMMList'
segs(object)

## S4 method for signature 'HMMList'
hemizygous(object)

## S4 method for signature 'HMMList'
homozygous(object)

## S4 method for signature 'HMMList'
duplication(object)

## S4 method for signature 'HMMList'
cnvSegs(object, filters = FilterParam(state = as.character(c(1, 2, 5, 6))))

## S4 method for signature 'HMMList'
cnvFilter(object, filters = FilterParam())

## S4 method for signature 'HmmGRanges'
cnvSegs(object, filters = FilterParam(state = as.character(c(1, 2, 5, 6))))

Arguments

See Also

FilterParam

Examples

data(snp_exp)
fit <- hmm2(snp_exp)
segs(fit) ## all intervals
cnvSegs(fit)
filter_param <- FilterParam(probability=0.95, numberFeatures=10, state=c("1", "2"))
cnvSegs(fit, filter_param)
filter_param <- FilterParam(probability=0.5, numberFeatures=2, state=c("1", "2"))
cnvSegs(fit, filter_param)
hemizygous(fit)
homozygous(fit)
duplication(fit)

Description

Raw SNP array processed files have headers and variable labels that may depend the software, how the output files was saved, the software version, and other factors. The purpose of this container is to collect the parameters relevant for reading in the source files for a particular project in a single container. This may require some experimentation as the example illustrates. The function fread in the data.table package greatly simplifies this process.

Usage

CopyNumScanParams(
  cnvar = "Log R Ratio",
  bafvar = "B Allele Freq",
  gtvar = c("Allele1 - AB", "Allele2 - AB"),
  index_genome = integer(),
  select = integer(),
  scale = 1000,
  row.names = 1L
)

## S4 method for signature 'CopyNumScanParams'
show(object)

Arguments

Slots

index_genome

an integer vector

cnvar

the column label for the log R ratios

bafvar

the column label for the B allele frequencies

gtvar

the column label(s) for the genotypes

scale

length-one numeric vector indicating how the low-level data should be scaled prior to saving on disk

select

numeric vector indicating which columns to read

row.names

length-one numeric vector indicating which column the SNP names are in

See Also

ArrayViews parseSourceFile

Examples

CopyNumScanParams() ## empty container

Description

This function is not intended to be called directly by the user, and is exported only for internal use by other BioC packages.

Usage

doUpdate(param)

Arguments

See Also

HmmParam

Description

If there are multiple markers on the same chromosome with the same annotated position, only the first is kept.

Usage

dropDuplicatedMapLocs(object)

Arguments

Value

an object of the same class with duplicated genomic positions removed

Examples

data(snp_exp)
g <- rowRanges(snp_exp)
## duplicate the first row
g[length(g)] <- g[1]
 rowRanges(snp_exp) <- g
 snp_exp2 <- dropDuplicatedMapLocs(snp_exp)

Description

Removes markers on chromosomes X and Y.

Usage

dropSexChrom(object)

Arguments

Value

an object of the same class as the input

Description

Get or set a matrix of emission probabilities. This function is exported primarily for internal use by other BioC packages.

Usage

emission(object)

emission(object) <- value

Arguments

Value

matrix

Description

Parameters for computing emission probabilities include the starting values for the Baum Welch update and initial state probabilities.

Usage

emissionParam(object)

emissionParam(object) <- value

Arguments

Value

EmissionParam instance

Examples

hparam <- HmmParam()
emissionParam(hparam)
ep <- EmissionParam()
cn_means(ep) <- log2(c(.1/2, 1/2, 2/2, 2/2, 3/2, 4/2))
emissionParam(hparam) <- ep

Description

The maximum a posteriori estimate of the trio copy number state for each genomic range is represented in a GRanges-derived class. Ultimately, these ranges will be filtered based on the trio copy number state (e.g., denovo deletions), size, number of features (SNPs), or chromosome. FilterParam is a container for the parameters commmonly used to filter the genomic ranges.

Usage

FilterParam(
  probability = 0.99,
  numberFeatures = 10,
  seqnames = paste0("chr", c(1:22, "X", "Y")),
  state = as.character(1:6),
  width = 1L
)

## S4 method for signature 'FilterParam'
probability(object)

## S4 method for signature 'FilterParam'
state(object)

## S4 method for signature 'FilterParam'
show(object)

Arguments

Slots

probability

a length-one numeric vector indicating the minimum posterior probability for the called state. Genomic intervals with posterior probabilities below probability will be filtered.

numberFeatures

a positive integer indicating the minimum number of features in a segment

seqnames

a character vector of seqnames to select (i.e., 'chr1' for only those intervals on chromosome 1)

width

positive integer indicating the minimal width of genomic intervals

state

character string indicating which hidden Markov model states to select

See Also

cnvFilter cnvSegs hmm2

Examples

fp <- FilterParam()
width(fp)
numberFeatures(fp)
seqnames(fp)
## To select CNV segments for which
## - the CNV call has a 'posterior' probability of at least 0.95
## - the number of features is at least 10
## - the HMM states are 1 (homozygous deletion) or 2 (hemizygous deletion)
FilterParam(probability=0.95, numberFeatures=10, state=c("1", "2"))

Description

Accessor for HMM filter parameters

Usage

filters(object)

Arguments

Description

Extract SNP genotypes. Genotypes are assumed to be represented as integers: 1=AA, 2=AB, 3=BB.

Usage

genotypes(object)

## S4 method for signature 'ArrayViews'
lrr(object)

## S4 method for signature 'ArrayViews'
baf(object)

## S4 method for signature 'ArrayViews'
genotypes(object)

## S4 method for signature 'SnpArrayExperiment'
baf(object)

## S4 method for signature 'SnpArrayExperiment'
copyNumber(object)

## S4 method for signature 'SnpArrayExperiment'
lrr(object)

## S4 method for signature 'SnpArrayExperiment'
genotypes(object)

Arguments

See Also

copyNumber

Description

Create an example SnpArrayExperiment from source files containing marker-level genomic data that are provided in this package

Usage

getExampleSnpExperiment(bsgenome)

Arguments

Value

A SnpArrayExperiment

Examples

## Not run: 
   if(require("BSgenome.Hsapiens.UCSC.hg18")){
     genome <- BSgenome.Hsapiens.UCSC.hg18
     snp_exp <- getExampleSnpExperiment(genome)
   }

## End(Not run)

Description

Accessor for HMM model parameters

Usage

getHmmParams(object)

Arguments

Examples

hmm_object <- HMM()
getHmmParams(hmm_object)

Description

The contructor HMM creates and object of class HMM. Not typically called directly by the user.

Usage

HMM(
  granges = GRanges(),
  param = HmmParam(),
  posterior = matrix(),
  filters = FilterParam()
)

## S4 method for signature 'HMM'
state(object)

## S4 method for signature 'HMM'
show(object)

Arguments

Slots

granges

a GRanges object

param

a HmmParam object

posterior

a matrix of posterior probabilities

filters

a FilterParam object

See Also

hmm2

Examples

data(snp_exp)
hmm_list <- hmm2(snp_exp[,1])
resultsFirstSample <- hmm_list[[1]]
resultsFirstSample
HMM()

Description

This function is intended for estimating the integer copy number from germline or DNA of clonal origin using a 6-state HMM. The states are homozygous deletion, hemizygous deletion, diploid copy number, diploid region of homozygosity, single copy gain, and two+ copy gain. Because heterozygous markers are more informative for copy number than homozygous markers and regions of homozgosity are common in normal genomes, we currently computed a weighted average of the BAF emission matrix with a uniform 0,1 distribution by the probability that the marker is heterozygous, thereby downweighting the contribution of homozygous SNPs to the likelihood. In addition to making the detection of copy-neutral regions of homozgosity less likely, it also helps prevent confusing hemizygous deletions with copy neutral regions of homozygosity – the former would be driven mostly by the log R ratios. This is experimental and subject to change.

Usage

hmm2(
  object,
  emission_param = EmissionParam(),
  transition_param = TransitionParam(),
  ...
)

## S4 method for signature 'SnpArrayExperiment'
hmm2(
  object,
  emission_param = EmissionParam(),
  transition_param = TransitionParam(),
  ...
)

## S4 method for signature 'oligoSnpSet'
hmm2(
  object,
  emission_param = EmissionParam(),
  transition_param = TransitionParam(),
  ...
)

## S4 method for signature 'ArrayViews'
hmm2(
  object,
  emission_param = EmissionParam(),
  transition_param = TransitionParam(),
  tolerance = 2,
  verbose = FALSE,
  ...
)

Arguments

Details

The hmm2 method allows parallelization across samples using the foreach paradigm. Parallelization is automatic when enabled via packages such as snow/doSNOW.

Examples

tp <- TransitionParam()
TransitionParam(taup=1e12)
data(snp_exp)
emission_param <- EmissionParam(temper=1/2)
fit <- hmm2(snp_exp, emission_param)
unlist(fit)
cnvSegs(fit)
## There is too little data to infer cnv reliably in this trivial example.
## To illustrate filtering options on the results, we select
## CNVs for which
## - the CNV call has a posterior probability of at least 0.5
## - the number of features is 2 or more
## - the HMM states are 1 (homozygous deletion) or 2 (hemizygous deletion)
fp <- FilterParam(probability=0.5, numberFeatures=2, state=c("1", "2"))
cnvSegs(fit, fp)
## for parallelization
## Not run: 
   library(snow)
   library(doSNOW)
   cl <- makeCluster(2, type = "SOCK")
   registerDoSNOW(cl)
   fit <- hmm2(snp_exp, emission_param)

## End(Not run)

Description

The constructor function for the HMMList class. The constructor is useful for representing a list of HMM objects.

Usage

HMMList(object)

Arguments

See Also

HMMList HMM hmm2

Description

Each element of the HMMList contains the genomic intervals of the HMM segmentation (GRanges-derived object), parameters from the Baum-Welch, and a FilterParam object.

Usage

## S4 method for signature 'HMMList'
show(object)

## S4 method for signature 'HMMList'
unlist(x, recursive = TRUE, use.names = TRUE)

Arguments

Slots

.Data

a list. Each element of the list should be a HMM object.

See Also

HMM

Examples

data(snp_exp)
fit <- hmm2(snp_exp)
class(fit)
identical(length(fit), ncol(snp_exp))
unlist(fit)

Description

Contains emission probabilities, parameters for emission probabilities, and transition probabilities required for computing the most likely state path via the Viterbi algorithm

Usage

HmmParam(
  emission = matrix(0, 0, 0),
  emission_param = EmissionParam(),
  transition = rep(0.99, nrow(emission)),
  chromosome = character(nrow(emission)),
  loglik = LogLik(),
  viterbi = Viterbi(),
  compute_posteriors = TRUE,
  verbose = FALSE
)

## S4 method for signature 'HmmParam'
show(object)

## S4 method for signature 'HmmParam'
nrow(x)

## S4 method for signature 'HmmParam'
ncol(x)

Arguments

Examples

HmmParam()

Description

The results of a 6-state HMM fit to simulated copy number and genotype data.

Format

a GRanges object

Description

Constructor for HmmTrellisParam class

Usage

HmmTrellisParam(
  ylimits = list(c(0, 1), c(-3, 1)),
  expandfun = function(g) {     width(g) * 50 }
)

Arguments

Description

Parameters for plotting idiograms

Usage

IdiogramParams(
  seqnames = character(),
  seqlengths = numeric(),
  unit = "kb",
  genome = "hg19",
  box = list(color = "blue", lwd = 1)
)

## S4 method for signature 'IdiogramParams,ANY'
plot(x, y, ...)

Arguments

Value

IdiogramParam object

Description

Paramater class for plotting idiograms

Usage

## S4 method for signature 'IdiogramParams'
show(object)

Arguments

Slots

seqnames

length-one character vector providing chromosome name

seqlengths

length-one numeric vector indicating size of chromosome

unit

character string indicating unit for genomic position (default is 'kb')

genome

character string indicating genome build

box

a list of parameters for plotting the box around the part of the idiogram that is plotted.

Examples

if(require(BSgenome.Hsapiens.UCSC.hg18) && require(grid)){
   si <- seqinfo(BSgenome.Hsapiens.UCSC.hg18)
   iparam <- IdiogramParams(seqnames="chr1",
                            genome="hg18",
                            seqlengths=seqlengths(si)["chr1"],
                            box=list(xlim=c(20e6L, 25e6L), color="blue", lwd=2))
   iparam
   idiogram <- plot(iparam)
   vp <- viewport(x=0.05, y=0.8, width=unit(0.9, "npc"), height=unit(0.2, "npc"),
                  name="vp1", just=c("left", "bottom"))
   grid.newpage()
   pushViewport(vp)
   print(idiogram, vp=vp, newpage=FALSE)
}

Description

Assess whether genotype is heterozygous based on BAFs

Usage

isHeterozygous(object, cutoff)

## S4 method for signature 'ArrayViews'
isHeterozygous(object, cutoff)

## S4 method for signature 'SnpArrayExperiment'
isHeterozygous(object, cutoff)

## S4 method for signature 'numeric'
isHeterozygous(object, cutoff)

## S4 method for signature 'matrix'
isHeterozygous(object, cutoff)

Arguments

Examples

if(require("BSgenome.Hsapiens.UCSC.hg18")){
  bsgenome <- BSgenome.Hsapiens.UCSC.hg18
  snp_exp <- getExampleSnpExperiment(bsgenome)
  is_het <- isHeterozygous(snp_exp[, 1], c(0.4, 0.6))
  table(is_het)
}

Description

A container for the log likelihood of the Viterbi state path. Stores the log likelihood from succesive updates of model parameters. When the difference between the log likelihoods at iteration i and i-1 is below the tolerance, no additional updates are performed.

Usage

LogLik(loglik = numeric(), tolerance = 1L)

Arguments

See Also

LogLik

Description

Exported for internal use by other BioC packages

Usage

## S4 method for signature 'LogLik'
length(x)

## S4 method for signature 'LogLik'
show(object)

Arguments

Slots

loglik

a numeric vector

tolerance

a numeric vector

See Also

LogLik

Description

Accessors for objects of class ArrayViews

Usage

lrrFile(object)

lrrFile(object) <- value

bafFile(object)

gtFile(object)

## S4 method for signature 'ArrayViews'
lrrFile(object)

## S4 replacement method for signature 'ArrayViews'
lrrFile(object) <- value

## S4 method for signature 'ArrayViews'
bafFile(object)

## S4 method for signature 'ArrayViews'
gtFile(object)

Arguments

Examples

views <- ArrayViews(parsedPath=tempdir())
sourcePaths(views)
lrrFile(views)
bafFile(views)
gtFile(views)

Description

Exported for internal use by other BioC packages

Description

Remove SNPs with NAs in any of the low-level estimates

Usage

NA_filter(x, i)

Arguments

Value

An object of the same class

Description

The number of SNP/nonpolymorphic probes contained in a genomic interval

Usage

numberFeatures(object)

Arguments

Description

A character string indicating the complete path for storing parsed files.

Usage

parsedPath(object)

## S4 method for signature 'ArrayViews'
parsedPath(object)

Arguments

See Also

parseSourceFile ArrayViews

ArrayViews

Description

This function parses genome studio files, writing the low-level data for log R ratios, B allele frequencies, and genotypes to disk as integers (1 file per subject per data type).

Usage

parseSourceFile(object, param)

## S4 method for signature 'ArrayViews,CopyNumScanParams'
parseSourceFile(object, param)

Arguments

See Also

ArrayViews ArrayViews CopyNumScanParams

Examples

require(BSgenome.Hsapiens.UCSC.hg18)
  bsgenome <- BSgenome.Hsapiens.UCSC.hg18
  require(data.table)
  extdir <- system.file("extdata", package="VanillaICE", mustWork=TRUE)
  features <- suppressWarnings(fread(file.path(extdir, "SNP_info.csv")))
  fgr <- GRanges(paste0("chr", features$Chr), IRanges(features$Position, width=1),
                 isSnp=features[["Intensity Only"]]==0)
  fgr <- SnpGRanges(fgr)
  names(fgr) <- features[["Name"]]
  seqlevels(fgr) <- seqlevels(bsgenome)[seqlevels(bsgenome) %in% seqlevels(fgr)]
  seqinfo(fgr) <- seqinfo(bsgenome)[seqlevels(fgr),]
  fgr <- sort(fgr)
  files <- list.files(extdir, full.names=TRUE, recursive=TRUE, pattern="FinalReport")
  views <- ArrayViews(rowRanges=fgr, sourcePaths=files, parsedPath=tempdir())
  show(views)

## read the first file
dat <- fread(files[1], skip="[Data]")
## information to store on the markers
select <- match(c("SNP Name", "Allele1 - AB", "Allele2 - AB",
                  "Log R Ratio", "B Allele Freq"), names(dat))
##
## which rows to keep in the MAP file. By matching on the sorted GRanges object
## containing the feature annotation, the low-level data for the log R ratios/
## B allele frequencies will also be sorted
##
index_genome <- match(names(fgr), dat[["SNP Name"]])
scan_params <- CopyNumScanParams(index_genome=index_genome, select=select)
##
## parse the source files
##
parseSourceFile(views, scan_params)
list.files(parsedPath(views))
##
##  Inspecting source data through accessors defined on the views object
##
require(oligoClasses)
## log R ratios
r <- head(lrr(views))
## B allele frequencies
b <- head(baf(views))
g <- head(genotypes(views))

Description

Accessor for probability filter

Usage

probability(object)

Arguments

Description

Rescale a numeric vector

Usage

rescale(x, l, u)

Arguments

Description

Compute the column-wide or row-wise mode of numeric matrices

Usage

rowModes(x)

colModes(x)

rowMAD(x, ...)

Arguments

Value

numeric vector

See Also

mad

mad rowMedians

Examples

X <- matrix(rnorm(100), 10, 10)
rowMAD(X)

Description

Accessor to obtain all segments from the HMM.

Usage

segs(object)

Arguments

Value

a GRanges-derived object

Description

Show method for objects of class Viterbi

Usage

## S4 method for signature 'Viterbi'
show(object)

Arguments

Description

A container for low-level summaries used for downstream copy number estimation, including log R ratios, B allele frequencies, and genotypes

Format

a SnpArrayExperiment object

Description

This function is exported primarily for internal use by other BioC packages.

Usage

snpArrayAssays(cn = new("matrix"), baf = new("matrix"), ...)

Arguments

Examples

data(snp_exp, package="VanillaICE")
r <- lrr(snp_exp)
b <- baf(snp_exp)
sl <- snpArrayAssays(cn=r, baf=b)

Description

Constructor for SnpArrayExperiment

Usage

SnpArrayExperiment(
  cn,
  baf,
  rowRanges = GRanges(),
  colData = DataFrame(),
  isSnp = logical(),
  ...
)

## S4 method for signature 'missing'
SnpArrayExperiment(
  cn,
  baf,
  rowRanges = GRanges(),
  colData = DataFrame(),
  isSnp = logical(),
  ...
)

## S4 method for signature 'matrix'
SnpArrayExperiment(
  cn,
  baf,
  rowRanges = GRanges(),
  colData = DataFrame(row.names = colnames(cn)),
  isSnp = logical(),
  ...
)

Arguments

Examples

## empty container
library(VanillaICE)
data(snp_exp, package="VanillaICE") # example

se <- SnpArrayExperiment(cn=lrr(snp_exp), baf=baf(snp_exp),
                   rowRanges=rowRanges(snp_exp))

Description

A single-argument generic function to construct a SnpArrayExperiment.

Usage

SnpExperiment(object)

## S4 method for signature 'ArrayViews'
SnpExperiment(object)

Arguments

Examples

view <- ArrayViews()
SnpExperiment(view)

Description

An extension to GRanges for representing SNPs

Constructor for SnpGRanges class

Usage

SnpGRanges(object = GRanges(), isSnp, ...)

## S4 method for signature 'missing'
SnpGRanges(object, isSnp)

## S4 method for signature 'GRanges'
SnpGRanges(object, isSnp)

Arguments

Slots

elementMetadata

a SnpDataFrame

Examples

SnpGRanges()
g <- GRanges("chr1", IRanges(15L, 15L))
SnpGRanges(g, isSnp=TRUE)

Description

Files containing SNP-level summaries for log R ratios, B allele frequencies, and genotypes – one sample per subject – are required.

Usage

sourcePaths(object)

Arguments

Examples

sourcePaths(ArrayViews())

Description

Retrieve genomic location of SNPs

Usage

## S4 method for signature 'oligoSnpSet'
start(x)

Arguments

Description

The states are represented as integers: 1=homozygous deletion, 2=hemizygous deletion, 3=diploid normal heterozygosity, 4=diploid region of homozygosity, 5=single copy gain, 6=two or more copy gain.

Usage

## S4 method for signature 'Viterbi'
state(object)

Arguments

Description

Extract the copy number state for each genomic interval.

Usage

## S4 method for signature 'HmmGRanges'
state(object)

Arguments

Description

This function simplifies the process of sweeping the modal log R ratio from the rows or columns of a SnpArrayExperiment object. It is most useful when a large number of samples (more than 10) are available and the dataset is a collection of germline samples. We assume that the samples are from a single batch and that the modal value will be a robust estimate of the mean log R ratio for diploid copy number. Variation in the modal estimates between markers is presumed to be attributable to probe effects (e.g., differences hybridization efficiency/PCR do to sequence composition). For sex chromosomes, one should apply this function separately to men and women and then recenter the resulting matrix according to the expected copy number.

Usage

sweepMode(x, MARGIN)

## S4 method for signature 'SnpArrayExperiment'
sweepMode(x, MARGIN)

Arguments

Value

an object of the same class as x

Examples

data(snp_exp)
snp_exp_rowcentered <- sweepMode(snp_exp, 1)
snp_exp_colcentered <- sweepMode(snp_exp, 2)
x <- lrr(snp_exp)
x_rowcentered <- sweep(x, 1, rowModes(x))
all.equal(lrr(snp_exp_rowcentered), x_rowcentered)

Description

Threshold numeric values according to user-specific limits. The thresholded values can also be jittered near the limits.

Usage

threshold(x, lim = c(-Inf, Inf), amount = 0)

Arguments

See Also

jitter

Examples

x <- rnorm(1000, 0, 3)
y <- threshold(x, c(-5,5))
range(y)

Description

Contains parameters for computing transition probabilities

Usage

TransitionParam(taup = 1e+10, taumax = 1 - 5e+06)

## S4 method for signature 'TransitionParam'
show(object)

Arguments

Details

Diagonal elements of the transition probability matrix are computed as e^-2*d/taup, where d is the distance between markers i and i-1 and taup is typically in the range of 1xe10. This probability is constrained to be no larger than taumax. The probabilities on the off-diagonal elements are the same and are subject to the constraint that the rows of the transition probability matrix sum to 1.

Examples

TransitionParam()
## higher values of taup make transitions between states less likely
TransitionParam(taup=1e12)

Description

This function is not intended to be called directly by the user. It is exported in the package NAMESPACE for internal use by other BioC packages.

Usage

updateHmmParams(
  object,
  emission_param = EmissionParam(),
  transition_param = TransitionParam()
)

Arguments

Description

Default viewports for plotting CNV data with lattice-style graphics

Usage

viewports()

Value

list

See Also

xyplotList xygrid

Examples

vps <- viewports()

Description

Data for the graphic is generated by a call to grangesData.

Usage

xyplotList(granges, se, param = HmmTrellisParam())

## S4 method for signature 'HmmGRanges,SnpArrayExperiment'
xyplotList(granges, se, param = HmmTrellisParam())

## S4 method for signature 'GRangesList,SnpArrayExperiment'
xyplotList(granges, se, param = HmmTrellisParam())

xygrid(trellis_plot, viewports, granges)

Arguments

See Also

viewports

Examples

if(require("BSgenome.Hsapiens.UCSC.hg18")){
  bsgenome <- BSgenome.Hsapiens.UCSC.hg18
  snp_exp <- getExampleSnpExperiment(bsgenome)
  seqlevels(snp_exp, pruning.mode="coarse") <- "chr22"
  fit <- hmm2(snp_exp)
  g <- reduce(hemizygous(fit), min.gapwidth=500e3)
  trellis_param <- HmmTrellisParam()
  fig <- xyplotList(g, snp_exp, trellis_param)
  vps <- viewports()
  xygrid(fig[[1]], vps, g)
}