Skip to contents

Filter and normalise gene expression

Source: R/data_geNormalisationFiltering.R
normaliseGeneExpression.Rd

Gene expression is filtered and normalised in the following steps:

  • Filter gene expression;

  • Normalise gene expression with calcNormFactors;

  • If performVoom = FALSE, compute counts per million (CPM) using cpm and log2-transform values if log2transform = TRUE;

  • If performVoom = TRUE, use voom to compute log2-CPM, quantile-normalise (if method = "quantile") and estimate mean-variance relationship to calculate observation-level weights.

Usage

normaliseGeneExpression(
  geneExpr,
  geneFilter = NULL,
  method = "TMM",
  p = 0.75,
  log2transform = TRUE,
  priorCount = 0.25,
  performVoom = FALSE
)

normalizeGeneExpression(
  geneExpr,
  geneFilter = NULL,
  method = "TMM",
  p = 0.75,
  log2transform = TRUE,
  priorCount = 0.25,
  performVoom = FALSE
)

Arguments

geneExpr

Matrix or data frame: gene expression

geneFilter

Boolean: filtered genes (if NULL, skip filtering)

method

Character: normalisation method, including TMM, RLE, upperquartile, none or quantile (see Details)

p

numeric value between 0 and 1 specifying which quantile of the counts should be used by method="upperquartile".

log2transform

Boolean: perform log2-transformation?

priorCount

Average count to add to each observation to avoid zeroes after log-transformation

performVoom

Boolean: perform mean-variance modelling (using voom)?

Value

Filtered and normalised gene expression

Details

edgeR::calcNormFactors will be used to normalise gene expression if method is TMM, RLE, upperquartile or none. If performVoom = TRUE, voom will only normalise if method = "quantile".

Available normalisation methods:

  • TMM is recommended for most RNA-seq data where more than half of the genes are believed not differentially expressed between any pair of samples;

  • RLE calculates the median library from the geometric mean of all columns and the median ratio of each sample to the median library is taken as the scale factor;

  • upperquartile calculates the scale factors from a given quantile of the counts for each library, after removing genes with zero counts in all libraries;

  • quantile forces the entire empirical distribution of each column to be identical (only performed if performVoom = TRUE).

See also

Other functions for gene expression pre-processing: convertGeneIdentifiers(), filterGeneExpr(), plotGeneExprPerSample(), plotLibrarySize(), plotRowStats()

Examples

geneExpr <- readFile("ex_gene_expression.RDS")
normaliseGeneExpression(geneExpr)
#>                Normal 1 Normal 2 Normal 3 Normal 4 Normal 5 Cancer 1 Cancer 2
#> ACTN1|87       16.44383 16.78304 16.48223 17.24207 17.07503 17.10327 16.49269
#> ADAM15|8751    16.70099 16.09395 16.28898 15.96189 15.31259 15.63044 15.97064
#> AKAP8L|26993   14.14424 14.80911 14.24193 14.41006 15.30310 14.58999 14.11916
#> AKR1A1|10327   15.14212 15.06161 14.85207 15.74210 15.51387 15.57613 15.64140
#> ALDOA|226      18.58900 20.00823 18.57857 18.85194 18.54547 18.33072 18.78489
#> ANXA6|309      16.80155 16.85176 16.67479 16.76024 16.67841 16.79276 16.81781
#> ARAP1|116985   16.01864 15.57970 15.45777 15.44823 15.82390 15.95734 15.89010
#> ATP5C1|509     14.74478 15.47092 16.10871 15.23351 15.62633 15.44610 15.49723
#> BOLA2|552900   14.03579 15.05401 14.67957 13.53149 14.32091 14.37295 14.08933
#> C16orf13|84326 13.75582 14.21114 14.03332 13.69816 14.23023 13.78593 13.60679
#>                Cancer 3 Cancer 4 Cancer 5
#> ACTN1|87       16.81314 18.48316 17.61675
#> ADAM15|8751    16.39172 15.97958 16.13043
#> AKAP8L|26993   15.14118 14.98893 13.49192
#> AKR1A1|10327   15.32114 15.90415 14.74053
#> ALDOA|226      18.40211 18.50496 18.54164
#> ANXA6|309      17.12530 16.53170 17.37841
#> ARAP1|116985   16.17556 16.06011 15.56131
#> ATP5C1|509     16.52061 16.49485 15.97965
#> BOLA2|552900   14.96354 14.09810 14.11716
#> C16orf13|84326 13.85646 13.88293 13.58011