nucleosomeDynamics: Run nucleosomeDynamics to compare two sets.

In nucleosome-dynamics/NucDyn: Nucleosome Dynamics

Description

This is the main function in NucDyn. It allows to compare the reads of two NGS experiments of nucleosome coverage.

Usage

nucleosomeDynamics(setA, setB, maxLen = 170, equalSize = FALSE,
  readSize = 140, maxDiff = 74, minDiff = 10, mc.cores = 1)

## S4 method for signature 'IRanges,IRanges'
nucleosomeDynamics(setA, setB, maxLen = 170,
  equalSize = FALSE, readSize = 140, maxDiff = 74, minDiff = 10)

## S4 method for signature 'RangedData,RangedData'
nucleosomeDynamics(setA, setB,
  maxLen = 170, equalSize = FALSE, readSize = 140, maxDiff = 74,
  minDiff = 10, mc.cores = 1)

## S4 method for signature 'GRanges,GRanges'
nucleosomeDynamics(setA, setB, maxLen = 170,
  equalSize = FALSE, readSize = 140, maxDiff = 74, minDiff = 10,
  mc.cores = 1)

Arguments

setA	Reads of the first experiment in IRanges, RangedData, or GRanges. The format should the same of the one in setB.
setB	Reads of the second experiment in IRanges, RangedData, or 'GRanges. The format should the the same of the one in setA.
maxLen	Reads longer than this number will be filtered out.
equalSize	If set to TRUE, all sets will be set to the same length, conserving their original dyad position. Use it if the reads in your sets have differences in length (ie, due to differences in the digestion) that you are not interested in.
readSize	Length to which all reads will be set in case equalSize is TRUE. It is ignored when equalSize is set to FALSE.
maxDiff	Maximum distance between the dyads of two reads that allows them to still be considered a "shift".
minDiff	Minimum distance between the dyads of two reads that allows them to still be considered a "shift".
mc.cores	If parallel support, the number of cores available. This option is only used if the provided sets are from more than one chromosome.

Details

The aim of nucleosomeDynamics is to infer "movement" (with direction and magnitude) of the reads between two reference nucleosome maps. In contrast with a simple coverage difference, nucleosomeDynamics can tell how the reads change between two different experiments. This is useful to analyze regions where fine regulatory role of the nucleosomes is suspected to happen.

This method is based on the idea that reads in a reference state (ref1) should match those in another reference state (ref2) after applying a few shifts and/or indels. Both ref1 and ref2 should be experimental nucleosome maps obtained from two experimental conditions.

The function looks for a match to each read in ref1 with a read in ref2 to discard those that correspond to unchanged nucleosomes between the two experimental conditions (equal start and/or end). After all possible matches have been found, those reads are discarded and nucleosomeDynamics looks then for matches in the remaining reads, corresponding to relevant changes (shifts), using a dynamic programming algorithm that minimizes distances between matched reads. Then, to account for the possibility that one sample has a higher coverage than the other, randomly picked reads are removed from the dataset with more reads to obtain sets of the same size. After this, all the remaining reads are considered indels.

The different types of matches, in the order in which they are tried are:

• Coinciding: Reads that start and end in the exact same position in both sets.

• Same start: Reads that start in the same position in both sets but end at a different one.

• Same end: Reads that end in the same position in both sets but start at a different one.

• Contained: Reads from one set that are contained or contain reads from the other set. For a read to be contained by another, it has to start at a more upstream position but end in a more downstream position than the second read.

• Shifts: Reads whose dyads are separated by less than a maximum distance of maxDist.

• Indels: Reads present in only one experiment.

If equalSize=TRUE, all reads are forced to the same size, and the match types "Same start", "Same end" and "Contained" do not apply.

In an attempt to find the optimum pairing for the shifts, we use dynamic programming approach. It is done in such a way that the score is inversely proportional to the dyad distance (to favor shortest possible distance), but with a very high gap penalty (to favor more pairs at longer distance rather than less closer pairs) and a penalty of -Inf for distances higher than maxDist so that those don't happen at all.

Value

An object of class NucDyn.

Author(s)

Ricard Illa, Diana Buitrago [e-mail]

Examples

    data(readsG2_chrII)
    data(readsM_chrII)
    dyn <- nucleosomeDynamics(setA=readsG2_chrII, setB=readsM_chrII)

nucleosome-dynamics/NucDyn documentation built on July 10, 2019, 10:54 a.m.