PECA: PECA differential gene expression

In PECA: Probe-level Expression Change Averaging

Description

Calculates the PECA ordinary or modified t-statistic to determine differential expression between two groups of samples in Affymetrix gene expression studies or peptide-based proteomic studies.

Usage

## Read AffyBatch object
PECA_AffyBatch(affy=NULL, normalize=FALSE, test="t", type="median", paired=FALSE, progress=FALSE)

## Read CEL-files
PECA_CEL(samplenames1=NULL, samplenames2=NULL, normalize=FALSE, test="t",
   type="median", paired=FALSE, progress=FALSE)

## Read tab separated text file	
PECA_tsv(file=NULL, samplenames1=NULL, samplenames2=NULL, normalize=FALSE,
   test="t", type="median", paired=FALSE, progress=FALSE)

## Read dataframe	
PECA_df(df=NULL, id=NULL, samplenames1=NULL, samplenames2=NULL, normalize=FALSE,
   test="t", type="median", paired=FALSE, progress=FALSE)

Arguments

affy	AffyBatch object.
normalize	A character string indicating if ("quantile") or ("median") normalization is performed.
test	A character string indicating whether the ordinary t-test ("t"), modified t-test ("modt"), or reproducibility-optimized test statistic ("rots") is performed.
type	A character string indicating whether ("median") or ("tukey") is used when calculating gene/protein values.
paired	A logical indicating whether a paired test is performed.
file	Filename of tab separated data.
samplenames1	A character vector containing the names of the .CEL-files/columns in the first group.
samplenames2	A character vector containing the names of the .CEL-files/columns in the second group.
df	Dataframe to be used as an input.
id	Column name of dataframe used for aggregating results.
progress	A logical indicating whether a progress bar is shown.

Details

PECA determines differential gene expression using directly the probe-level measurements from Affymetrix gene expression microarrays or proteomic datasets. An expression change between two groups of samples is first calculated for each probe/peptide on the array. The gene/protein-level expression changes are then defined as medians over the probe-level changes. For more details about the probe-level expression change averaging (PECA) procedure, see Elo et al. (2005), Laajala et al. (2009) and Suomi et al.

PECA calculates the probe-level expression changes using the ordinary or modified t-statistic. The ordinary t-statistic is calculated using the function rowttests in the Bioconductor genefilter package. The modified t-statistic is calculated using the linear modeling approach in the Bioconductor limma package. Both paired and unpaired tests are supported.

The significance of an expression change is determined based on the analytical p-value of the gene-level test statistic. Unadjusted p-values are reported along with the corresponding p-values looked up from beta ditribution. The quality control and filtering of the data (e.g. based on low intensity or probe specificity) is left to the user.

Value

PECADE returns a matrix which rows correspond to the genes under analysis and columns indicate the corresponding signal log-ratio (slr), t-statistic, p-value and FDR adjusted p-value.

References

T. Suomi, G.L. Corthals, O. Nevalainen and L.L. Elo: Using peptide-level proteomics data for detecting differentially expressed proteins. 2015

L.L. Elo, L. Lahti, H. Skottman, M. Kylaniemi, R. Lahesmaa and T. Aittokallio: Integrating probe-level expression changes across generations of Affymetrix arrays. Nucleic Acids Research 33(22), e193, 2005.

E. Laajala, T. Aittokallio, R. Lahesmaa and L.L. Elo: Probe-level estimation improves the detection of differential splicing in Affymetrix exon array studies. Genome Biology 10(7), R77, 2009.

H. Bengtsson, K. Simpson, J. Bullard and K. Hansen: aroma.affymetrix: A generic framework in R for analyzing small to very large Affymetrix data sets in bounded memory. Tech Report \#745, Department of Statistics, University of California, Berkeley, 2008.

Examples

## Generate example data frame
df <- data.frame(id=c(rep("a",10),rep("b",10),rep("c",10)))
df$A1 <- rnorm(30, mean=50, sd=5)
df$A2 <- rnorm(30, mean=48, sd=5)
df$A3 <- rnorm(30, mean=50, sd=5)
df$B1 <- rnorm(30, mean=52, sd=5)
df$B2 <- rnorm(30, mean=53, sd=5)
df$B3 <- rnorm(30, mean=51, sd=5)

## Run the test
group1 <- c("A1","A2","A3")
group2 <- c("B1","B2","B3")
results <- PECA_df(df, group1, group2, id=id)

Example output

Performing log-transformation
Calculating low-level statistics
Aggregating statistics
Done

PECA documentation built on Nov. 8, 2020, 7:24 p.m.