windowValues: Bins sequencing data for plotting

Description

Given a variable with corresponding genomic positions, the function bins the values in windows of a specified size and calculates weighted mean and 25th and 75th percentile for each window. The resulting object are visualized by the function plotWindows.

Usage

windowValues(x, positions, chromosomes, window = 1e6, overlap = 0,
        weight = rep.int( x = 1, times = length(x)), start.coord = 1)
    windowBf(Af, Bf, good.reads, positions, chromosomes, window = 1e6,
        overlap = 0, start.coord = 1, conf = 0.95)

Arguments

variable to be windowed.

positions

base-pair positions.

chromosomes

names or numbers of the chromosomes.

window

size of windows used for binning data. Smaller windows will take more time to compute.

overlap

integer defining the number of overlapping windows. Default is 0, no overlap.

weight

weights to be assigned to each value of x, usually related to the read depth.

start.coord

coordinate at which to start computing the windows. If NULL, will start at the first position available.

A-allele frequency for the Bf calculation.

B-allele frequency for the Bf calculation.

good.reads

number of reads passing filter for the Bf calculation.

conf

confidence intervals of the binned Bf value.

Value

a list of data.frame, one per chromosome. Each data.frame contains base-pair windows covering the chromosome. Each row of the data.frame correspond to a window and its weighted mean, 25th and 75th percentiles of the input values, and the number of data points within each window.

Details

DNA sequencing produces an amount of data too large to be handled by standard graphical devices. In addition, for samples analyzed with older machines and with low or middle coverage (20x to 50x), measures such as read depth are subject to big variations due to technical noise. Using windowValues prior to plotting reduces the noise and the amount of data to be plotted.

The binning of the B-allele frequency requires a separate function, windowBf, as the B-allele frequency calculation uses multiple values: Af, Bf and good.reads.

The output of windowValues and windowBf can be used as input for plotWindows.

Examples

Run this code

# NOT RUN {
# }
# NOT RUN {
    data.file <-  system.file("extdata", "example.seqz.txt.gz",
        package = "sequenza")
    seqz.data <- read.seqz(data.file)
    # 1Mb windows, each window is overlapping with 1 other
    # adjacent window: depth ratio
    seqz.ratio <- windowValues(x = seqz.data$depth.ratio,
        positions = seqz.data$position, chromosomes = seqz.data$chromosome,
        window = 1e6, weight = seqz.data$depth.normal, start.coord = 1,
        overlap = 1)

    seqz.hom  <- seqz.data$zygosity.normal == 'hom'
    seqz.het  <- seqz.data[!seqz.hom, ]
    # 1Mb windows, each window is overlapping with 1 other adjacent window:
    # B-allele frequency
    seqz.bafs  <- windowValues(x = seqz.het$Bf, positions = seqz.het$position,
        chromosomes = seqz.het$chromosome, window = 1e6,
        weight = seqz.het$depth.tumor, start.coord = 1, overlap = 1)
    # Repeat the same operation using windowBf
    seqz.bafs  <- windowBf(Af = seqz.het$Bf, Bf = seqz.het$Bf,
        good.reads = seqz.het$good.reads, positions = seqz.het$position,
        chromosomes = seqz.het$chromosome, window = 1e6,
        start.coord = 1, overlap = 1, conf = 0.95)
# }

Run the code above in your browser using DataLab

Data Engineering and BI courses are free this week!