compareModelStat: Reports a number of statistically significant enhancer~...
In BIMSBbioinfo/reg2gene: Predicting the target genes for regulatory elements

Description Usage Arguments Details Value Author(s) Examples

compareModelStat() reports the number of statistically significant enhancer~ promoter associations for voting or meta-analysis and corresponing individual modelling results. IMORTANT NOTE: since it uses results of previously runned voting procedure it is important that statistics and threshold values used in voting procedure are the same as used in this algorithm. However, meta-analysis statistics thresholding is performed as a part of this function ,thus statistics and threshold [typeSuccess&thresholdSuccess] values are used to filter both: meta-analysis results and individual modelling results.

1
2
3

compareModelStat(indexTable = IndexTable, type = "votingAlgorithm",
  typeSuccess = "pval", thresholdSuccess = 0.1, method = "H3K4me1",
  algorithm = "dcor", cohort = "Roadmap")

`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 or meta-analysis, AND individual modelling (indexTable$type=="ind"). It is necessary that individual modelling paths are present in the index table (indexTable$type=="ind"), because results of these models are compared to the results of voting or meta-analysis procedure.
`type`	"votingAlgorithm" (default; character). Describes which voting procedure is used to produce voting results (common options:"votingAlgoritm", "votingCohorts","votingMethods") or if meta-analysis is used (type="metaA"). This argument corresponds to the type argument in `makeIndexTable`, thus if some other type character object is used with `makeIndexTable` then the same argument should be used
`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"

This function allows an easy assesment of the success of voting or meta-analysis procedure compared to individual modelling procedures, through returning the number of statically significant cases reported in voting or meta-analysis and corresponding individual modelling procedures.

returns a dataframe with information about the number of identified statistically enhancer~promoter associations and info about corresponding input modelling datasets used - cohort,method,algorithm combination. The same statistics is returned for the requested voting method or meta-analysis.

Inga Patarcic

library(stringr)
library(InteractionSet)
## Not run: 

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

compareModelStat(indexTable=IndexTable, 
type="votingAlgorithm",
method="H3K4me1", 
algorithm="dcor",
cohort="Blueprint")


## End(Not run)