extractGenomeWideAnalysis: Genome wide Analysis of Consequences due to isoform switching

In kvittingseerup/IsoformSwitchAnalyzeR: Identify, Annotate and Visualize Isoform Switches with Functional Consequences from both short- and long-read RNA-seq data

Genome wide Analysis of Consequences due to isoform switching

Description

This function enables a genome wide analysis of changes in isoform usage of isoforms with a common annotation.

Specifically this function extract isoforms of interest and for each category of annotation (such as signal peptides) the global distribution of IF (measuring isoform usage) are plotted for each subset of features in that category (e.g with and without signal peptides). This enables a global analysis of isoforms with a common annotation. The annotation considered are (if added to the switchAnalyzeRlist) coding potential, intron retentions, isoform class code (Cufflinks/Cuffdiff data only), NMD status, ORFs, protein domains, signal peptide and whether switch consequences were identified.

The isoforms of interest can either be defined by isoforms form gene differentially expressed, isoform that are differential expressed or isoforms from genes with isoform switching - as controlled by featureToExtract. Please note that the extractConsequenceEnrichment function probably more relevant than using featureToExtract='isoformUsage' since it directly uses the paired information from switches.

This function offers both visualization of the result as well as analysis via summary statistics of the comparisons.

Usage

extractConsequenceGenomeWide(
    switchAnalyzeRlist,
    featureToExtract = 'isoformUsage',
    annotationToAnalyze = 'all',
    alpha=0.05,
    dIFcutoff = 0.1,
    log2FCcutoff = 1,
    violinPlot=TRUE,
    alphas=c(0.05, 0.001),
    localTheme=theme_bw(),
    plot=TRUE,
    returnResult=TRUE
)

extractGenomeWideAnalysis(
    switchAnalyzeRlist,
    featureToExtract = 'isoformUsage',
    annotationToAnalyze = 'all',
    alpha=0.05,
    dIFcutoff = 0.1,
    log2FCcutoff = 1,
    violinPlot=TRUE,
    alphas=c(0.05, 0.001),
    localTheme=theme_bw(),
    plot=TRUE,
    returnResult=TRUE
)

Arguments

switchAnalyzeRlist	A switchAnalyzeRlist object containing the result of an isoform switch analysis (such as the one provided by isoformSwitchTestDEXSeq()) as well as additional annotation data for the isoforms.
featureToExtract	This argument, given as a string, defines the set isoforms which should be analyzed. The available options are: 'isoformUsage' (Default): Analyze a subset of isoforms defined by change in isoform usage (controlled by dIFcutoff) and the significance of the change in isoform expression (controlled by alpha). Please note that the extractConsequenceEnrichment function probably more relevant than using featureToExtract='isoformUsage' since it directly uses the paired information from switches. 'isoformExp' :Analyze a subset of isoforms defined by change in isoform expression (controlled by log2FCcutoff) and the significance of the change in isoform expression (controlled by alpha) 'geneExp' :Analyze all isoforms from a subset of genes defined by by change in gene expression (controlled by log2FCcutoff) and the significance of the change in gene expression (controlled by alpha) 'all' : Analyze all isoforms stored in the switchAnalyzeRlist (note that this is highly depending on the parameter reduceToSwitchingGenes in isoformSwitchTestDEXSeq - which should be set to FALSE (default is TRUE) if the 'all' option should be used here).
annotationToAnalyze	A vector of strings indicating what categories of annotation to analyze. Annotation types given here but not (yet) analyzed in the switchAnalyzeRlist will not be plotted. See details for full list of usable strings, their meaning and dependencies. Default is 'All'.
alpha	The cutoff which the FDR correct p-values (q-values) must be smaller than for calling significant switches. Default is 0.05.
dIFcutoff	The cutoff which the changes in (absolute) isoform usage must be larger than before an isoform is considered switching. This cutoff can remove cases where isoforms with (very) low dIF values are deemed significant and thereby included in the downstream analysis. This cutoff is analogous to having a cutoff on log2 fold change in a normal differential expression analysis of genes to ensure the genes have a certain effect size. Default is 0.1 (10%).
log2FCcutoff	The cutoff which the changes in (absolute) isoform or gene expression must be larger than before an isoform is considered for inclusion.
violinPlot	A logical indicating whether to make a violin plots (if TRUE) or boxplots (if FALSE). Violin plots will always have added 3 black dots, one of each of the 25th, 50th (median) and 75th percentile of the data. Default is TRUE.
alphas	A numeric vector of length two giving the significance levels represented in plots. The numbers indicate the q-value cutoff for significant (*) and highly significant (***) respectively. Default 0.05 and 0.001 which should be interpret as q<0.05 and q<0.001 respectively). If q-values are higher than this they will be annotated as 'ns' (not significant).
localTheme	General ggplo2 theme with which the plot is made, see ?ggplot2::theme for more info. Default is theme_bw().
plot	A logic indicting whether the analysis should be plotted. If TRUE and returnResult = FALSE the ggplot2 object will be returned instead. Default is TRUE.
returnResult	A logical indicating whether to return a data.frame with summary statistics of the comparisons (if TRUE) or not (if FALSE). If FALSE (and plot=TRUE) the ggplot2 object will be returned instead. Default is TRUE.

Details

extractGenomeWideAnalysis is just a wrapper for extractGenomeWideConsequenceAnalysis included for backward comparability.

Changes in isoform usage are measure as the difference in isoform fraction (dIF) values, where isoform fraction (IF) values are calculated as <isoform_exp> / <gene_exp>.

The significance test is performed with R's build in wilcox.test() (aka 'Mann-Whitney-U') with default parameters and resulting p-values are corrected via p.adjust() using FDR (Benjamini-Hochberg).

The arguments passed to annotationToAnalyze must be a combination of:

• isoform_class_code : Divide transcripts based on differences in the transcript classification provide by cufflinks (only available for data imported from Cufflinks/Cuffdiff). For a updated list of class codes see http://cole-trapnell-lab.github.io/cufflinks/cuffcompare/#transfrag-class-codes.

• coding_potential : Divide transcripts based on differences in coding potential, as indicated by the CPAT analysis. Requires that importCPATanalysis have been used to add external CPAT analysis to the switchAnalyzeRlist.

• intron_retention : Divide transcripts based on presence intron retentions (and their genomic positions). Require that analyzeIntronRetention have been run.

• ORF : Divide transcripts based on whether an ORF is annotated or not. Requires that both the isoforms have been annotated with ORF either via identifyORF or by supplying a GTF file and setting addAnnotatedORFs=TRUE when creating the switchAnalyzeRlist.

• NMD_status : Divide transcripts based on differences in sensitivity to Nonsense Mediated Decay (NMD). Requires that both the isoforms have been annotated with PTC either via identifyORF or by supplying a GTF file and setting addAnnotatedORFs=TRUE when creating the switchAnalyzeRlist.

• domains_identified : Divide transcripts based on differences in the name and order of which domains are identified by the Pfam in the transcripts. Requires that importPFAManalysis have been used to add external Pfam analysis to the switchAnalyzeRlist. Requires that both the isoforms are annotated with a ORF either via identifyORF or by supplying a GTF file and setting addAnnotatedORFs=TRUE when creating the switchAnalyzeRlist.

• signal_peptide_identified : Divide transcripts based on differences in whether a signal peptide was identified or not by the SignalP analysis. Requires that analyzeSignalP have been used to add external SignalP analysis to the switchAnalyzeRlist. Requires that both the isoforms are annotated with a ORF either via analyzeORF or by supplying a GTF file and setting addAnnotatedORFs=TRUE when creating the switchAnalyzeRlist (and are thereby also affected by removeNoncodinORFs=TRUE in analyzeCPAT).

• switch_consequences : Whether the gene is involved in isoform switches with predicted consequences. Requires that analyzeSwitchConsequences have been used).

Value

If plot=TRUE: A plot of the distribution of IF values as a function of the annotation and condition compared. If returnResult=TRUE: A data.frame with the summary statistics from the comparison of the two conditions with a Wilcox.test.

Author(s)

Kristoffer Vitting-Seerup

References

Vitting-Seerup et al. The Landscape of Isoform Switches in Human Cancers. Mol. Cancer Res. (2017).

See Also

analyzeAlternativeSplicing
analyzeSwitchConsequences
extractConsequenceEnrichment
extractConsequenceEnrichmentComparison

Examples

### Load example data
data("exampleSwitchListAnalyzed")

### make the genome wide analysis
symmaryStatistics <- extractConsequenceGenomeWide(
    switchAnalyzeRlist = exampleSwitchListAnalyzed,
    featureToExtract = 'isoformUsage', # alternatives are 'isoformExp' and 'geneExp'
    plot=TRUE,
    returnResult = TRUE
)

kvittingseerup/IsoformSwitchAnalyzeR documentation built on June 28, 2024, 5:41 p.m.