Filter and plot genes to visualize/summarize genes that are differentially translated.

Share:

Description

This function filters the output from the anotaGetSigGenes function based on many user defined thresholds and flags to generate a summary table and optional per gene plots.

Usage

1
anotaPlotSigGenes(anotaSigObj, selIds=NULL, selContr=NULL, minSlope=NULL, maxSlope=NULL, slopeP=NULL, minEff=NULL, maxP=NULL, maxPAdj=NULL, maxRvmP=NULL, maxRvmPAdj=NULL, selDeltaPT=NULL, selDeltaP=NULL, sortBy=NULL, performPlot=TRUE, fileName="ANOTA_selected_significant_genes_plot.pdf", geneNames=NULL)

Arguments

anotaSigObj

The output from the anotaGetSigGenes function.

selIds

The function can consider only a subset of the identifiers from the input data set (which can be further filtered) or used for custom plotting of identifiers of interest (leaving all filters as NULL). For custom selection of identifiers, supply a vector of identifiers (row names from the original data set) to be included. Default is NULL i.e. filtering is performed on all identifiers. Minimum length of selIds is currently 2. However, if only one identifier is of interest this identifier can be at position one and two of the supplied vector which will lead to that the data for the identifier of interested will be plotted twice.

selContr

Which contrast should be evaulated during the filtering, sorting and plotting? Descriptions of the contrasts can be found in the output from the anotaGetSigGenes object in the usedContrasts slot. Indicate the contrast by the column number.

minSlope

The output can be filtered so that genes whose identified slopes are too small can be excluded. Default is NULL i.e. no filtering based on lower boundary of the slope. To exclude genes with e.g. a slope <(-1) assign -1 to minSlope.

maxSlope

The output can be filtered so that genes whose identified slopes are too large can be excluded. Default is NULL i.e. no filtering based on upper boundary of the slope. To exclude genes with e.g. a slope >2 assign 2 to maxSlope.

slopeP

A p-value for the slope being <0 or >1 is calculated if the estimate for the slope is <0 or >1. This p-value can be used to filter the output based on unrealistic slopes. When set low fewer genes will be disqualified. Default is NULL i.e. no filtering based on slope p-value. We recommend setting slopeP between 0.01 and 0.1 depending on data set characteristics.

minEff

The output can be filtered based on minimum effect for inclusion. The value is applied both to negative and positive effects: e.g. a value of 1 will evaluate if the effects are >1 OR <(-1). Default is NULL i.e. no filtering based on effect.

maxP

The output can be filtered based on raw p-values from the anota analysis without RVM (i.e. smaller compared to assigned value). Default is NULL i.e. no filtering.

maxPAdj

The output can be filtered based on adjusted p-values from the anota analysis without RVM (i.e. smaller compared to assigned value). The adjustment method that was used when running anotaGetSigGenes will be evaluated. Default is NULL i.e. no filtering.

maxRvmP

The output can be filtered based on raw p-values from the anota analysis with RVM (i.e. smaller compared to assigned value). Default is NULL i.e. no filtering.

maxRvmPAdj

The output can be filtered based on adjusted p-values from the anota analysis with RVM (i.e. smaller compared to assigned value). The adjustment method that was used when running anotaGetSigGenes will be evaluated. Default is NULL i.e. no filtering.

selDeltaPT

The output can be filtered based on the mean log2(translational activity data / cytosolic mRNA data) between groups difference. The groups are defined by the selected contrast. Default is NULL i.e. no filtering.

selDeltaP

The output can be filtered based on the translational activity data only so that the minimum absolute between groups delta translation is used for gene inclusion. The groups are defined by the selected contrast. Default is NULL i.e. no filtering.

sortBy

The output can be sorted by effect ("Eff"), raw p-value("p") or raw RVM p-value ("apvRvmP"). Default is NULL i.e. no sorting.

performPlot

The function can generate a graphical output per gene. Default is TRUE i.e. generate plots.

fileName

The plots are printed to a file whose file name is given here. Default is "ANOTA_selected_significant_genes_plot.pdf".

geneNames

When anotaPlotSigGenes plots the individual gene plots they will be named by the original row names supplied to the anotaGetSigGenes function. geneNames allows the user to add additional names when plotting to e.g. include gene symbols. Input is a matrix with one column where the original row names match the row names of the input matrix and the desired new names are given in column 1. Default is NULL i.e. no additional names.

Details

This function allows the user to filter the output generated from the anotaGetSigGenes function to derive a reduced selection of genes that are considered for further evaluation. This is done by setting one or several of the filtering parameters described above. The function also generates a graphical output which helps when evaluating a single gene's regulation. In the graphical output, the results for each gene is displayed on separate rows. The first graph shows all samples and per sample class regression lines using the common slope with different colors for each sample class. The magnitude of the common slope is indicated. The second graph shows key statistics for the gene without the RVM model for all contrasts analyzed when running anotaGetSigGenes but any ordering and selection of genes is performed on the contrast given by the selContr argument as described above. The third graph is similar to the second but with RVM statistics instead (if RVM was used in the anotaGetSigGenes analysis).

Value

anotaPlotSigGenes generates a graphical output as described above and a list object containing summary data for those genes that passed the selected set of filters. The output list object contains the following slots:

selectedData

A matrix containing non-RVM data for the filtered identifiers. Columns are "apvSlope" (the common slope used in APV); "apvSlopeP" (if the slope is <0 or >1 a p-value for the slope being <0 or >1 is calculated; if the slope is >=0 & <=1 this value is set to 1); "unadjustedResidError" (the residual error before calculating the effective residual error); "apvEff" (the group effect); "apvMSerror" (the effective mean square error); "apvF" (the F-value); "residDf" (the residual degrees of freedom); "apvP" (the p-value); "apvPAdj" (the adjusted p-value).

selectedRvmData

A matrix containing RVM data for the filtered identifiers. Columns are "apvSlope" (the common slope used in APV); "apvSlopeP" (if the slope is <0 or >1 a p-value for the slope being <0 or >1 is calculated; if the slope is >=0 & <=1 this value is set to 1); "apvEff" (the group effect); "apvRvmMSerror" (the effective mean square error after RVM); "apvRvmF" (the RVM F-value); "residRvmDf" (the residual degrees of freedom after RVM); "apvRvmP" (the RVM p-value); "apvRvmPAdj" (the adjusted RVM p-value).

groupIntercepts

A matrix with the group intercepts, i.e. the translational activity for each group independent of cytosolic mRNA level. Can be used for e.g. clustering of translational activity. Data for all groups defined when using the anotaGetSigGenes function are supplied although the filtering is based on the contrast defined under the selContr argument.

deltaData

Mean delta translational activity data ("deltaP"), mean delta cytosolic mRNA data ("deltaT") and mean delta log ratio data ("deltaPT") comparing the sample classes specified by the selected contrast.

usedThresholds

A list object with the user set values for the filtering.

Author(s)

Ola Larsson ola.larsson@ki.se, Nahum Sonenberg nahum.sonenberg@mcgill.ca, Robert Nadon robert.nadon@mcgill.ca

See Also

anotaPerformQc, anotaResidOutlierTestanotaGetSigGenes

Examples

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
 ## Load the library and dataset (two phenotypes)
  library(anota)
  data(anotaDataSet)
 ## Quality control of the data set.
  anotaQcOut <- anotaPerformQc(dataT= anotaDataT[1:200,], dataP=anotaDataP[1:200,],
phenoVec=anotaPhenoVec, nDfbSimData=500)
 ##Test normality of residuals
  anotaResidOut <- anotaResidOutlierTest(anotaQcObj=anotaQcOut)
 ##Identify differentially translated genes.
  anotaSigGeneOut <- anotaGetSigGenes(dataT= anotaDataT[1:200,],
dataP=anotaDataP[1:200,], phenoVec=anotaPhenoVec, anotaQcObj=anotaQcOut)
 ##Plot some of the differentially expressed mRNAs
  anotSigGeneOutFiltered <-
anotaPlotSigGenes(anotaSigObj=anotaSigGeneOut, selContr=1,
maxP=0.05,slopeP=0.05, maxSlope=1.5, minSlope=(-0.5), selDeltaPT=0.5)