makeIndexTable: Creates Index table of files produced by reg2gene data...

In BIMSBbioinfo/reg2gene: Predicting the target genes for regulatory elements

Description

The function creates a data frame with the following columns: 1) path to .rds files - results produced by modelling or data integration re2gene f() and corresponding info: 2) methods - which method type is used for this analysis: eg. H3K4me1,... 3) algorithms - which algorithm type is used for this analysis: eg. dcor,... 4) cohorts - which cohort type is used for this analysis: eg. Roadmap,... 5) type - define whether results are produced by voting, meta-analysis,from individual modelling or a result of regActivityAroundTSS()

Usage

makeIndexTable(pathModels, type = "ind", method = "H3K4me1",
  algorithm = "pearson", cohort = "Roadmap")

Arguments

pathModels	(character) a path to folder where .rds files are stored or an object with paths to .rds files. # IMPORTANT! cohort/method/algorithm names should be indicated in the path, or .rds files with this name should be present in the defined directory
type	"ind" (default; character) OPTIONS: ("votingAlgoritm","votingCohorts","votingMethods","metaA","regAct", "ind"). Define whether results are produced by voting by("votingAlgoritm", "votingCohorts","votingMethods"), meta-analysis("metaA"), or from individual modelling ("ind") or a result of regActivityAroundTSS() ("regAct"). Important for analyses when all these info is gathered together
method	"H3K4me1" (default; character). A character vector of all methods used for this analysis. Written for the option: method=c("H3K4me1","H3K27ac","Methylation","DNase")
algorithm	"pearson" (default; character). A character vector of all methods used for this analysis. Written for the option: algorithm=c("pearson","spearman","elasticnet","dcor", "randomForest")
cohort	"Roadmap" (default; character). A character vector of all cohorts used for this analysis. Written for the option: cohort= c("Roadmap","Blueprint","CEMT","McGill")
...	maybe to add in the future

Details

An IndexTable which is a result of makeIndexTable() is used as an input for following functions: selectEP(),votedEP(),plotGEEA() to gathers paths to different reg2gene data integration and modelling results, and allows easier downstream analysis. This f() should be separately runned for voting("votingAlgoritm","votingCohorts","votingMethods"), meta-analysis and individual modelling, and regActivityAroundTSS() results, and individual IndexTables should be pooled together, such that plotGEEA() function can be runned (it requires paths to "regAct" .rds files - results of regActivityAroundTSS() and for example modelling paths.

Value

A data frame with the following columns (n=5): 1) path to .rds files - results produced by modelling or data integration re2gene f() and corresponding info: 2) methods - which method type is used for this analysis: eg. H3K4me1,... 3) algorithms - which algorithm type is used for this analysis: eg. dcor,... 4) cohorts - which cohort type is used for this analysis: eg. Roadmap,... 5) type - define whether results are produced by voting("votingAlgoritm", "votingCohorts","votingMethods"), meta-analysis("metaA"), from individual modelling ("ind") or a result of regActivityAroundTSS() ("regAct")

Author(s)

Inga Patarcic

Examples

require(stringr)

## Not run: 
makeIndexTable(pathModels="~/H3K4me1pearsonRoadmap.rds",
type="ind",method="H3K4me1",algorithm="pearson",cohort="Roadmap")

# cohort/method/algorithm names should be indicated in the path, or .rds 
# files with this name should be present in the defined directory

makeIndexTable(pathModels="~/test.rds",
type="ind",method="H3K4me1",algorithm="pearson",cohort="Roadmap")


# HOW-TO-USE-IT EXAMPLE

IndexTable1 <- makeIndexTable(pathModels="~/H3K4me1pearsonRoadmap.rds",
type="ind",method="H3K4me1",algorithm="pearson",cohort="Roadmap")

IndexTable2 <- makeIndexTable(pathModels="~/ra_H3K4me1pearsonRoadmap.rds",
type="regAct",method="H3K4me1",algorithm="pearson",cohort="Roadmap")

IndexTable <- rbind(IndexTable1,IndexTable2) # IndexTable use downstream  


algorithm <- c("pearson","spearman","elasticnet","dcor","randomForest")
cohort=c("Roadmap","Blueprint","CEMT","McGill")
method=c("H3K4me1","H3K27ac","Methylation","DNase")

votedPath="/data/akalin/Projects/AAkalin_Catalog_RI/Results/Validation/Fishillevic/VoteD/"
metaPath="/data/akalin/Projects/AAkalin_Catalog_RI/Results/Validation/Fishillevic/MetaA/"
pathRegActTSS="/data/akalin/Projects/AAkalin_reg2gene/Results/regActivityAroundTSS/FishilevichChromHMMoneTOone/"
pathModels="/data/akalin/Projects/AAkalin_Catalog_RI/Data/ValidationDataset/Fishillevich/ModellingResults/"
votedAlg="/data/akalin/Projects/AAkalin_Catalog_RI/Results/Validation/Fishillevic/VoteD/VotingAlg/"
votedC="/data/akalin/Projects/AAkalin_Catalog_RI/Results/Validation/Fishillevic/VoteD/VotingCohort/"
votedM="/data/akalin/Projects/AAkalin_Catalog_RI/Results/Validation/Fishillevic/VoteD/VotingMethod/"

IndexTable <- data.frame(rbind(makeIndexTable(metaPath,"metaA"),
                               makeIndexTable(votedAlg,"votingAlgorithm"),
                               makeIndexTable(votedC,"votingCohorts"),
                               makeIndexTable(votedM,"votingMethods"),
                               makeIndexTable(pathRegActTSS,"regAct"),
                               makeIndexTable(pathModels,"ind")),stringsAsFactors = F)
saveRDS(IndexTable,"/data/akalin/Projects/AAkalin_Catalog_RI/Data/ValidationDataset/Fishillevich/IndexTable.rds")


## End(Not run)

BIMSBbioinfo/reg2gene documentation built on May 3, 2019, 6:42 p.m.