Normalize counts in a DGEList object — normalize

This function wraps additional normalization methods (qsmooth and cyclic loess) for a given DGEList object in addition to the standard normalization methods implemented by edgeR::normlibSizes().

Usage

normalize_counts(
  x,
  method = c("TMM", "TMMwsp", "RLE", "upperquartile", "none", "qsmooth", "loess"),
  refColumn = NULL,
  logratioTrim = 0.3,
  sumTrim = 0.05,
  doWeighting = TRUE,
  Acutoff = -1e+10,
  p = 0.75,
  group_factor = NULL,
  batch = NULL,
  norm_factors = NULL,
  window = 0.05,
  prior_count = 0.1,
  weights = NULL,
  span = 0.3,
  iterations = 4L,
  min.weight = 1e-05,
  max.weight = 1e+05,
  equal.weights.as.null = TRUE,
  loess_method = c("weightedLowess", "loess", "locfit")
)

Arguments

x: DGEList object
method: Normalization method. One of "TMM", "TMMwsp", "RLE", "upperquartile", "none", "qsmooth", or "loess".
refColumn: column to use as reference for method="TMM". Can be a column number or a numeric vector of length nrow(object).
logratioTrim: the fraction (0 to 0.5) of observations to be trimmed from each tail of the distribution of log-ratios (M-values) before computing the mean. Used by method="TMM" for each pair of samples.
doWeighting: logical, whether to use (asymptotic binomial precision) weights when computing the mean M-values. Used by method="TMM" for each pair of samples.
Acutoff: minimum cutoff applied to A-values. Count pairs with lower A-values are ignored. Used by method="TMM" for each pair of samples.
p: numeric value between 0 and 1 specifying which quantile of the counts should be used by method="upperquartile".
group_factor: a group level continuous or categorial covariate associated with each sample or column in the object. The order of the group_factor must match the order of the columns in object. Used in qsmooth normalization
batch: (Optional) batch covariate (multiple batches are not allowed). If batch covariate is provided, Combat() from sva is used prior to qsmooth normalization to remove batch effects. See Combat() for more details. Used in qsmooth.
norm_factors: optional normalization scaling factors. Used in qsmooth.
window: window size for running median which is a fraction of the number of rows in object. Default is 0.05. Used in qsmooth.
prior_count: prior count to add to 0 counts when computing offsets. Default = 0.1. Used in qsmooth.
weights: numeric vector of non-negative prior weights. Missing values are treated as zero. Default = NULL, Used in loess.
span: positive numeric value between 0 and 1 specifying proportion of data to be used in the local regression moving window. Larger numbers give smoother fits. Default = 0.3. Used in loess.
iterations: number of local regression fits. Values greater than 1 produce robust fits. Default = 4L. Used in loess.
min.weight: minimum weight. Any lower weights will be reset. Default = 1e-5. Used in loess.
max.weight: maximum weight. Any higher weights will be reset. Default = 1e5. Used in loess.
equal.weights.as.null: should equal weights be treated as if weights were NULL, so that lowess is called? Applies even if all weights are all zero. Default = TRUE. Used in loess.
loess_method: method used for weighted lowess. Possibilities are "weightedLowess", "loess" or "locfit". Used in loess.
sumTrimthe: fraction (0 to 0.5) of observations to be trimmed from each tail of the distribution of A-values before computing the mean. Used by method="TMM" for each pair of samples.

Value

DGEList with norm.factors or offset matrix

Details

If "TMM", "TMMwsp", "RLE", "upperquartile", or "none" is selected, then edgeR::normLibSizes(x, ...) is used. If method = "qsmooth" then qsmooth::qsmooth(x, group_factor, ...) is used. If method = "loess" then csaw::normOffsets(x, ...) is used.

Examples

counts <- matrix(rnbinom(1e3, mu = 10, size = 20), ncol = 20)

y <- edgeR::DGEList(
  counts = counts,
  group = gl(n = 2, k = 10, labels = c("control", "treatment"))
)

# TMM normalization
y.tmm <- normalize_counts(y, method = "TMM")

# Qsmooth normalization
y.qs <- normalize_counts(y, method = "qsmooth", group_factor = y$samples$group)

# Cyclic loess
y.loess <- normalize_counts(y, method = "loess")