mapFeaturesToCRMs: R interface to bed_to_matrix REST in server

In LedPred: Learning from DNA to Predict Enhancers

Description

The mapFeaturesToCRMs function allows the user to create a training set matrix to build a predictive model. The training set is composed of positive regions (known to be involved in the pathway of interest) and negative regions (randomly picked or known to not be involved in the pathway of interest) that will be described (scored) by features. Three types of features file format are accepted: Position specific scoring matrices modeling motifs recognised by transcription factors, bed files containing region coordinates for any discrete feature (NGS peaks, conservation blocks) and wig/bigWig files containing signal data. This script has been tested with version 0.99 of the online server. Go here to see current version of the server http://ifbprod.aitorgonzalezlab.org/map_features_to_crms.php

Usage

  mapFeaturesToCRMs(URL = "http://ifbprod.aitorgonzalezlab.org/map_features_to_crms.php",
  positive.bed = NULL, genome = NULL, negative.bed = NULL,
  shuffling = NULL, background.seqs = NULL, genome.info = NULL,
  pssm = NULL, background.freqs = NULL, ngs = NULL, bed.overlap = NULL,
  my.values = NULL, feature.ranking = NULL, feature.nb = NULL,
  crm.feature.file = NULL, stderr.log.file = NULL, stdout.log.file = NULL)

Arguments

URL	URL of the server REST target
positive.bed	Positive bed file path. Compulsory
genome	Genome code, eg. dm3 for Drosophila Melanogaster. Compulsory
negative.bed	Negative bed file path.
shuffling	Integer with number of time shuffle background sequences (background.seqs). If negative.bed is NULL and shuffling is set at 0, the feature matrix does not contain negative sequences. It is useful to produce a test set matrix.
background.seqs	Background sequences used for shuffling. If shuffling = 0, set this parameter at 0.
genome.info	File require for shuffling bed. If shuffling = 0, set this parameter at 0.
pssm	Position specific scoring matrices
background.freqs	Background frequencies of nucleotides in genome
ngs	NGS (bed and wig) files
bed.overlap	Minimal overlap as a fraction of query sequence with NGS bed peak. Equivalent with intersectBed -f argument. Default 1bp.
my.values	Bed file where fourth column are values to append to the SVM matrix
feature.ranking	File with ranked features (Output of rankFeatures). It is used for scoring a query bed file
feature.nb	Integer with feature.nb
crm.feature.file	Path to feature matrix file
stderr.log.file	Path to error log
stdout.log.file	Path to standard output log

Value

A list

feature.matrix	a data frame where each row is a region and each column a feature, each cell carry a score, the first column is the response vector
stdout.log	Standard output log of mapFeaturesToCRMs script in server
stderr.log	Standard error log of mapFeaturesToCRMs script in server

Examples

## Not run: 
 dirPath <- system.file("extdata", package="LedPred")
 file.list <-   list.files(dirPath, full.names=TRUE)
 background.freqs <- file.list[grep("freq", file.list)]
 positive.regions <-  file.list[grep("positive", file.list)]
 negative.regions <-  file.list[grep("negative", file.list)]
 TF.matrices <-  file.list[grep("tf", file.list)]
 ngs.path <- system.file("extdata/ngs", package="LedPred")
 ngs.files=list.files(ngs.path, full.names=TRUE)
 crm.features.list <- mapFeaturesToCRMs(positive.bed=positive.regions,
     negative.bed=negative.regions,  background.freqs=background.freqs,
     pssm=TF.matrices, genome="dm3", ngs=ngs.files,
     crm.feature.file = "crm.features.tab",
     stderr.log.file = "stderr.log", stdout.log.file = "stdout.log")
 names(crm.features.list)
 class(crm.features.list$crm.features)
 crm.features.list$stdout.log
 crm.features.list$stderr.log

## End(Not run)

LedPred documentation built on Nov. 8, 2020, 8 p.m.