sigora: Sigora's main function.

In sigora: Signature Overrepresentation Analysis

Sigora's main function.

Description

This function determines which Signatures (GPS) from a collection of GPS data (GPSrepo argument) for the specified pathway repository are present in the specified list of genes of interest (queryList argument)). It then uses the distribution function of hypergeometric probabilities to identify the pathways whose GPS are over-represented among the present GPS and saves the results to the file specified in the saveFile argument.

Usage

sigora(
  GPSrepo,
  level,
  markers = FALSE,
  queryList = NULL,
  saveFile = NULL,
  weighting.method = "invhm",
  idmap = load_data("idmap")
)

Arguments

GPSrepo	An object created by makeGPS or one of the precompiled GPS data collections that are provided with this package (currently for KEGG and Reactome). e.g. reaH for human Reactome GPS, kegH for human KEGG GPS, and reaM and kegM for corresponding mouse GPS. See the examples section for creating and using your own GPS.
level	In hierarchical repositories (e.g. Reactome) number of levels to consider. Recommended value for KEGG: 2, for Reactome: 4.
markers	Whether to take single genes that are uniquely associated with only one pathway into account (i.e. should pathway unique genes/PUGs be considered GPS?). Recommended value: TRUE (1).
queryList	A user specified list of genes of interest ('query list'), as a vector of ENSEMBL/ ENTREZ IDs or gene symbols (HGNC/MGI).
saveFile	If provided, the results are saved here as a tab delimited File (including , for each pathway, a list of genes ordered by their contribution to the statistical significance of the pathway).
weighting.method	The weighting method or GPS. The default weighting scheme for the GPS is the reciproc of the harmonic mean of the degrees of the two component genes of a GPS. A wide range of alternative weighting schemes are pre-implemented (see below). Additional user defined weighting schemes are also supported. Currently, the following alternatives are pre-implemented: 'noweights','cosine','topov','reciprod','jac','justPUGs'and'invhm'. Additional user defined weighting schemes are also supported (see section examples). 'noweights': assigns a constant of 1 to all GPS. 'cosine': all GPS are weighted by the cosine of the degrees of their consituent genes. 'topov': all GPS are weighted by topological overlap of their consituent genes. 'reciprod': all GPS are weighted by reciproc of product of the number of pathway annotations of their consituent genes. 'jac':all GPS are weighted by the jaccard similarity of the pathway annotations consituent genes. 'justPUGs': Analysis is performed using PUGs only. 'invhm': all GPS are weighted by the reciproc of the harmonic mean of the degrees of their consituent genes (default).
idmap	A dataframe for converting between different gene-identifier types (e.g. ENSEMBL, ENTREZ and HGNC-Symbols of genes). Most users do not need to set this argument, as there is a built-in conversion table.

Value

summary_results	A dataframe listing the analysis results.
detailed_results	A dataframe describing the detailed evidence (present Gene-Pair Signatures) for each pathway.

References

Foroushani AB, Brinkman FS and Lynn DJ (2013).“Pathway-GPS and SIGORA: identifying relevant pathways based on the over-representation of their gene-pair signatures.”PeerJ, 1

See Also

sigora-package , makeGPS

Examples

##query list
ils <- grep("^IL",load_data('idmap')[["Symbol"]],value=TRUE)
## using precompiled GPS repositories:
sigRes.ilreact <- sigora(queryList=ils,GPSrepo=load_data('reaH'),level=4)

sigRes.ilkeg <- sigora(queryList=ils,GPSrepo=load_data('kegH'),level=2)
## user created GPS repository:
nciH<-makeGPS(pathwayTable=load_data('nciTable'))
sigRes.ilnci<-sigora(queryList=ils,GPSrepo=nciH,level=2)
## user defined weighting schemes :
myfunc<-function(a,b){1/log(a+b)}
sigora(queryList=ils,GPSrepo=nciH,level=2, weighting.method = myfunc)

sigora documentation built on March 18, 2022, 8:05 p.m.