Apply log2 Transformation to Expression Data

Checks whether the input data is already log-transformed; if not, applies a log2 transformation. This helps ensure comparability of expression values across datasets.

Usage

log_trans(data, pseudo = 1, ...)

# S3 method for class 'matrix'
log_trans(data, pseudo = 1, ...)

# S3 method for class 'ExpressionSet'
log_trans(data, pseudo = 1, ...)

Arguments

data: A matrix-like data object.
pseudo: A numeric value added before transformation to avoid taking log of zero. For example, log2(exprs + pseudo).
...: Additional arguments passed to methods.

Value

A matrix or an ExpressionSet with log2-transformed expression values.

Details

The function heuristically determines whether data has been log-transformed, following the methodology used in GEO2R. If not, it applies log2() with the specified pseudo offset.

References

NCBI GEO2R: https://www.ncbi.nlm.nih.gov/geo/geo2r/?acc=GSE1122

Examples

sample_means <- 2^runif(2, 2, 10)
sample_disp <- 100 / sample_means + 0.5
data <- matrix(
    rnbinom(4, mu = sample_means, size = 1 / sample_disp),
    nrow = 2
)
log_trans(data)
#> Applying log2 transformation with a pseudo count of 1 to the data.
#>          [,1]     [,2]
#> [1,] 0.000000 3.169925
#> [2,] 7.924813 6.149747
log_trans(log2(data))
#> The data is already log-transformed. No transformation applied.
#>          [,1]     [,2]
#> [1,]     -Inf 3.000000
#> [2,] 7.918863 6.129283