compEpiTools (version 1.6.3)

findLncRNA: Identify putative long non coding RNAs (lncRNA)

Description

Identify putative long non coding RNAs (lncRNA) based on ChIP-seq chromatin features and RNAseq data

Usage

findLncRNA(k4me3gr, k4me3bam, k4me1bam, k79bam, k36bam, RNAseqbam,
sizeLNC=10000, extDB= NULL, txdb, org=NULL, Qthr=0.95)

Arguments

k4me3gr

GRanges; the set of H3K4me3 peaks from a ChIP-seq experiment

k4me3bam

character; a path to BAM file containing H3K4me3 ChIP-seq aligned reads

k4me1bam

character; a path to BAM file containing H3K4me1 ChIP-seq aligned reads

k79bam

either NA or a path to BAM file containing H3K79me2 ChIP-seq aligned reads

k36bam

either NA or a path to BAM file containing H3K36me3 ChIP-seq aligned reads

RNAseqbam

either NA or a path to BAM file containing RNA-seq aligned reads

sizeLNC

numeric; the size of the putative lncRNA

extDB

GRanges; a set lncRNAs to be included in the analysis

txdb

an object of class TxDb

org

either NULL or an object of class BSgenome

Qthr

numeric in [0,1]; the percentile of the signal in random genomic regions to be considered as minumum cutoff

Value

coordinates are reported as row names) are reported on the rows and the columns indicate the library-size normalized reads density of the following marks: H3K4me3, H3K4me1, H3K79me2, H3K36me3 and RNAseq reads. Reads density is reported for both up- and down-stream regions of width sizeLNC, see details, while for extDB regions the reads density is reported only for the regions as they are defined.

Details

Putative long non coding RNAs (lncRNAs) are identified based on the associated chromatin features and, possibly, RNAseq signal. Only putative lncRNAs distal from genebodies are identified. Briefly, H3K4me3 peaks are used as the main mark indicating transcriptional activity. Only H3K4me3 peaks outside genebodies +/- 10Kb are considered. Only peaks where the signal of H3K4me1 is lower than H3K4me3 are kept, to discard possible enhancer sites. Regions of interest of length sizeLNC (default 10Kb) are considered from the mid point of remaining H3K4me3 peaks, either in the forward or reverse direction (ROIs). The rational is that distal H3K4me3 marks could indicate the TSS of distal lncRNAs.

Optionally, to increase the likelihood of having identified a bona-fide transcriptional unit, downstream regions (ROIs on either the forward or reverse strand) are evaluated for the existence of significant transcriptional signal. This is achieved based on the optional data (optional tracks) to be provided as BAM files: ChIP-seq for H3K79me2 or H3K36me3, or RNA-seq. Density of H3K79me2, H3K36me3 and RNAseq reads in the remaining H3K4me3 regions and in 100K random regions of 10Kb each is determined (non overlapping with ROIs), normalized by the respective library size. ROIs where the signal of any of these is higher than the Qthr percentile of the random regions (profiled for the same mark) are considered as putative lncRNAs.

An optional GRanges containing regions to be considered in any case (for example based on lists of a priori known lncRNAs) can be provided as extDB. These regions will be evaluated as they are, and subject to the same filtering procedure based on H3K79me2, H3K36me3 and RNAseq data, if provided. Passing at least one of H3K79me2, H3K36me3 and RNAseq while setting Qthr to 0 correspond to profile the ROIs (and the extDB regions if provided) for those optional tracks and avoid filtering based on the signal of random regions. All bam files have to be associated to the corresponding index .bai files. Please refer to the documentation of samtools on how to create them.

References

http://genomics.iit.it/groups/computational-epigenomics.html

Examples

  require(TxDb.Mmusculus.UCSC.mm9.knownGene)
  txdb <- TxDb.Mmusculus.UCSC.mm9.knownGene
  # loading H3K4me3 peaks as a GRanges object
  # built based on the BED file from the GEO GSM1234483 sample
  # limited to chr19:3200000-4000000
  H3K4me3GR <- system.file("extdata", "H3K4me3GR.Rda", package="compEpiTools")
  load(H3K4me3GR)
  # pointing to Pol2 BAM file (it could be used as a replacement of the K79bam or K36bam ..)
  # BAM file from the GEO GSM1234478 sample, limited to chr19:3200000-4000000
  Pol2bam <- system.file("extdata", "Pol2.bam", package="compEpiTools")
  # pointing to H3K4me3 BAM file
  # BAM file from the GEO GSM1234483 sample, limited to chr19:3200000-4000000
  H3K4me3bam <- system.file("extdata", "H3K4me3.bam", package="compEpiTools")
  # pointing to H3K4me1 BAM file
  # BAM file from the GEO GSM1234488 sample, limited to chr19:3200000-4000000
  H3K4me1bam <- system.file("extdata", "H3K4me1.bam", package="compEpiTools")
  res <- findLncRNA(k4me3gr=H3K4me3GR, k4me3bam=H3K4me3bam, k4me1bam=H3K4me1bam, 
    k79bam=Pol2bam, k36bam=NA, RNAseqbam=NA, 
    sizeLNC=10000, txdb=txdb, org=NULL, Qthr=0)