Generate simulated samples with copy number profiles derived from a specific sample

Source: R/CNVMetricsSimulations.R

The function uses the input sample to simulate new samples. The simulated samples will possess similar sizes of events, proportional to the original chromosome. To generate realistic simulations, the specified sample must contain segments covering the majority of the genome. Most importantly, the NEUTRAL segments should be present.

processSim(curSample, nbSim)

Arguments

curSample

a GRanges that contains a collection of genomic ranges representing copy number events, including amplified/deleted status, from one sample. The sample must have a metadata column called 'state' with a state, in an character string format, specified for each region (ex: DELETION, LOH, AMPLIFICATION, NEUTRAL, etc.) and a metadata column called 'CN' that contains the log2 copy number ratios.

nbSim

a single positive integer which is corresponding to the number of simulations that will be generated.

Value

a data.frame containing the segments for each simulated sample. The data.frame has 6 columns:

  • ID a character string, the name of the simulated sample

  • chr a character string, the name fo the chromosome

  • start a integer, the starting position of the segment

  • end a integer, the ending position of the segment

  • log2ratio a numerical, the log2 copy number ratio assigned to the segment

  • state a character string, the state of the segment (ex: DELETION, AMPLIFICATION, NEUTRAL, etc.)

Details

TODO

Author

Astrid Deschênes, Pascal Belleau

Examples


## Load required package to generate the sample
require(GenomicRanges)

## Create one 'demo' genome with 2 chromosomes and few segments
## The stand of the regions doesn't affect the calculation of the metric
sample01 <- GRanges(seqnames=c(rep("chr1", 4), rep("chr2", 3)),
    ranges=IRanges(start=c(1905048, 4554832, 31686841, 32686222,
        1, 120331, 725531),
    end=c(2004603, 4577608, 31695808, 32689222, 117121,
        325555, 1225582)),
    strand="*",
    state=c("AMPLIFICATION", "NEUTRAL", "DELETION", "LOH",
        "DELETION", "NEUTRAL", "NEUTRAL"),
    log2ratio=c(0.5849625, 0, -1, -1, -0.87777, 0, 0))

## Generates 10 simulated genomes based on the 'demo' genome
simRes <- processSim(curSample=sample01, nbSim=10)