reg2gene: Predicting the target genes for regulatory elements

Description Usage Arguments Details Value Author(s) Examples

The function returns GInteractions object with selected N TOP/BOTTOM/RANDOM associations per cohort/method/algorithm/voting/metaA. It can return N random statistically not significant associations as well.

1
2
3

selectEP(indexTable, topN = "all", select = "top", success = TRUE,
  typeSuccess = "pval", thresholdSuccess = 0.05, method = "H3K4me1",
  algorithm = "pearson", cohort = "Roadmap", metaA = FALSE)

`indexTable`	an output of the makeIndexTable() for given combination of methods/algorithms/cohorts; a table should contain the following columns: path - paths to .rds files; methods - a method type used for the analysis: eg. H3K4me1; algorithms - algorithm used for the analysis: eg. dcor, cohorts - cohort used for the analysis: eg. Roadmap; type - define whether results are produced by voting, meta-analysis,from individual modelling or a result of regActivityAroundTSS()[OPTIONAL]
`topN`	"all" (default) or integer. Defines how many EP associations to report. If topN=="all" then all associations with predefined statistics [typeSuccess<thresholdSuccess], eg pval<0.05 are reported. If topN is an integer then corresponding number of EP associations (as defined by select and success arguments) is reported.
`select`	"top" (default). Other options "bottom" and "random". Indicates whether to select top N genes [topN], bottom N genes [topN] or randomly select N genes [topN]
`success`	TRUE (default). Whether to select statistically significant EP associations (succes==TRUE), or those that were found to be statistically not significant EP associations (succes==FALSE).
`typeSuccess`	"pval" (default) Other options "qval". Which statistics to threshold to assess a statistical significance.
`thresholdSuccess`	0.05 (default, numeric). A threshold useD to assess a statistical significance.
`method`	"H3K4me1" (default; character). For which method to extract Enh~Promoter pairs. Common options: "H3K4me1","H3K27ac", "Methylation","DNase"
`algorithm`	"pearson" (default; character). For which algorithm to extract Enh~Promoter pairs. Common options:"pearson","spearman", "elasticnet","dcor","randomForest"
`cohort`	"Roadmap" (default; character). For which cohort to extract Enh~Promoter pairs. Common options:"Roadmap","Blueprint", "CEMT","McGill"
`metaA`	FALSE (default). Whether to perform this analysis on results of meta-analysis or not. This argument overwrites cohort argument.

This function allows easy export of TOPN gene~enhancer pairs per method, cohort and algorithm combination of modelling results or meta-analysis. It can return all statistically significant/unsignificant genes or select top, bottom or random N of genes. Which statistics to filter can be choosen by user (typeSuccess), as well as a threshold value for it (thresholdSuccess).

returns a GInteractions object filtered for N (or retained all) enhancer-promoter interactions. Corresponding statistics from either modelling analysis associateReg2Gene or meta-analysis metaInteractions is retained.

Inga Patarcic

## Not run: 

require(stringr)
library(InteractionSet)

# ONE .rds file will need to be added into package!!!

IndexTable <- readRDS("/data/akalin/Projects/AAkalin_Catalog_RI/Results/Validation/Fishillevic/VoteD//CohortVoting_McGillH3K4me1.rds")


TOPgenes <- selectEP(IndexTable,
topN=16, #change argument "all" not all
select="top", #"bottom","random",
success=TRUE,
typeSuccess="pval",
thresholdSuccess=0.05,
method="H3K4me1",
algorithm="dcor",
cohort="Roadmap",
metaA=FALSE)

saveRDS(TOPgenes,"/data/akalin/Projects/AAkalin_Catalog_RI/Results/Validation/Fishillevich/TOPgenesH3K4me1dcorRoadmap.rds")

## End(Not run)