Package 'bioDist'

Title: Different distance measures
Description: A collection of software tools for calculating distance measures.
Authors: B. Ding, R. Gentleman and Vincent Carey
Maintainer: Bioconductor Package Maintainer <[email protected]>
License: Artistic-2.0
Version: 1.79.0
Built: 2024-10-30 04:21:56 UTC
Source: https://github.com/bioc/bioDist

Help Index

Find the closest genes.

Description

Find the closest genes to the supplied target gene based on the supplied distances.

Usage

closest.top(x, dist.mat, top)

Arguments

x

the name of the gene (feature) to use.
dist.mat

either a dist object or a matrix of distances.
top

the number of closest genes desired.

Details

The feature named x must be in the supplied distances. If so, then the top closest other features are returned.

Value

A vector of names of the top closest features.

Author(s)

Beiying Ding

See Also

cor.dist, spearman.dist, tau.dist,euc, man,KLdist.matrix,KLD.matrix,mutualInfo

Examples

data(sample.ExpressionSet)
 sE <- sample.ExpressionSet[1:100,]
 d1 <- KLdist.matrix(sE, sample = FALSE)
 closest.top(featureNames(sE)[1], d1, 5)

Pearson correlational distance

Description

Calculate pairwise Pearson correlational distances, i.e. 1-COR or 1-|COR|, and saves as a 'dist' object

Usage

cor.dist(x, ...)

Arguments

x

n by p matrix or ExpressionSet; if x is an ExpressionSet, then the function uses its 'exprs' slot.
...

arguments passed to cor.dist:

  • absif TRUE, then 1-|COR| else 1-COR, default is TRUE.

  • diagif TRUE, then the diagonal of the distance matrix will be displayed, default is FALSE.

  • upperif TRUE, then the upper triangle of the distance matrix will be displayed, default is FALSE.

  • samplefor objects of classes that extend eSet: if TRUE, then distances are computed between samples(columns) , otherwise, they are computed between features(rows).

Details

The cor function is used to compute the pairwise distances between rows of an input matrix, except if the input is an object of a class that extends eSet and sample is TRUE.

Value

Pairwise Pearson correlational distance object

Author(s)

Beiying Ding

See Also

spearman.dist, tau.dist,euc, man, KLdist.matrix, KLD.matrix, mutualInfo

Examples

x <- matrix(rnorm(200), nrow = 5)
 cor.dist(x)

Euclidean distance

Description

Calculate pairwise Euclidean distances and saves the result as a 'dist' object

Usage

euc(x, ...)

Arguments

x

n by p matrix or an object of a class that extends eSet; if x is a matrix, pairwise distances are calculated between the rows of a matrix. If x is an object of a class that extends eSet, the method makes use of the 'exprs' method and pairwise distances are calculated between samples(columns) if sample is TRUE
...

arguments passed to euc:

  • diagif TRUE, then the diagonal of the distance matrix will be displayed; default is FALSE.

  • upperif TRUE, then the upper triangle of the distance matrix will be displayed; default is FALSE.

  • sampleFor objects of classes that extends eSet, pairwise distances are calculated between samples(columns) if sample is TRUE ; default value is TRUE

Details

The method calculates pairwise euclidean distances, assuming that all samples have the same number of observations

Value

An object of class dist with the pairwise Euclidean distance between rows except in case of objects of class that extend eSet when sample is TRUE

Author(s)

Beiying Ding

See Also

spearman.dist, tau.dist, man,KLdist.matrix,KLD.matrix, mutualInfo

Examples

x <- matrix(rnorm(200), nrow = 5)
 euc(x)

Continuous version of Kullback-Leibler Distance (KLD)

Description

Calculate KLD by estimating by smoothing log⁡(f(x)/g(x))∗f(x)\log(f(x)/g(x))*f(x) and then integrating.

Usage

KLD.matrix(x, ...)

Arguments

x

n by p matrix or list or an object of a class that extends eSet; if x is an an object of a class that extends eSet (eg ExpressionSet), then the function works against its 'exprs' slot.
...

arguments passed to KLD.matrix:

  • methoduse locfit or density to estimate integrand; default is c("locfit", "density")(i.e. both methods).

  • suppupper and lower limits of the integral; default is NULL in which case the limits of the integral are calculated from the range of the data.

  • subdivisionssubdivisions for the integration; default is 1000.

  • diagif TRUE, then the diagonal of the distance matrix will be displayed; default is FALSE.

  • upperif TRUE, then the upper triangle of the distance matrix will be displayed; default is FALSE.

  • samplefor ExpressionSet methods: if TRUE, then distances are computed between samples, otherwise, they are computed between genes.

Details

The distance is computed between rows of the input matrix (except if the input is an object of a class that extends eSet and sample is TRUE.

The presumption is that all samples have the same number of observations. The list method is meant for use when samples sizes are unequal.

Value

An object of class dist with the pairwise, between rows, Kullback-Leibler distances.

Author(s)

Beiying Ding, Vincent Carey

See Also

cor.dist, spearman.dist, tau.dist, dist, KLdist.matrix, mutualInfo

Examples

x <- matrix(rnorm(100), nrow = 5)
 KLD.matrix(x, method = "locfit", supp = range(x))

Discrete version of Kullback-Leibler Distance (KLD)

Description

Calculate the KLD by binning continuous data.

KL distance is calculated using the formula

KLD(f1(x),f2(x))=∑i=1Nf1(xi)∗log⁡f1(xi)f2(xi)KLD(f_1(x),f_2(x)) = \sum_{i=1}^N{ f_1(x_i)*\log\frac{f_1(x_i)}{f_2(x_i)}}

Usage

KLdist.matrix(x, ...)

Arguments

x

n by p matrix or a list or an object of a class that extends eSet. If x is an object of a class derived from eSet (ExpressionSet,SnpSet etc), then the values returned by the exprs function are used.
...

arguments passed to KLdist.matrix:

gridsize

the number of grid points used to select the optimal bin width of the histogram used to estimate density. If no value is supplied, the grid size is calculated internally; default is NULL.

symmetrize

if TRUE, then symmetrize; the default is FALSE.

diag

if TRUE, then the diagonal of the distance matrix will be displayed; the default is FALSE.

upper

if TRUE, then the upper triangle of the distance matrix will be displayed; default is FALSE.

sample

for eSet methods: if TRUE, then the distances are computed between samples, otherwise, between features; the default is TRUE.

Details

The data are binned, and then the KL distance between the two discrete distributions is computed and used. The distance is computed between rows of the input matrix (except if the input is an object of a class that extends eSet and sample is TRUE.

The presumption is that all samples have the same number of observations. The list method is meant for use when samples sizes are unequal.

Value

An object of class dist is returned.

Author(s)

Beiying Ding

See Also

cor.dist, spearman.dist, tau.dist,euc, man,KLD.matrix,mutualInfo

Examples

x <- matrix(rnorm(100), nrow = 5)
 KLdist.matrix(x, symmetrize = TRUE)

Manhattan distance

Description

Calculate pairwise Manhattan distances and saves as a dist object.

Usage

man(x, ...)

Arguments

x

n by p matrix or an object of class that extends eSet. If x is an object of class that extends eSet, (eg ExpressionSet) then the function uses its 'exprs' slot.
...

arguments passed to man:

  • diagif TRUE, then the diagonal of the distance matrix will be displayed; default is FALSE.

  • upperif TRUE, then the upper triangle of the distance matrix will be displayed; default is FALSE.

Details

This is just an interface to dist with the right parameters set.

Value

An instance of the dist class with the pairwise Manhattan distances between the rows of x in case of a matrix or between the features (rows) in case of a class that extends eSet.

Author(s)

Beiying Ding

See Also

cor.dist, spearman.dist, tau.dist,euc, KLdist.matrix, KLD.matrix,mutualInfo

Examples

x <- matrix(rnorm(200), nrow = 5)
 man(x)

Mutual Information

Description

Calculate mutual information via binning

Usage

mutualInfo(x, ...)
MIdist(x, ...)

Arguments

x

an n by p matrix or ExpressionSet; if x is an ExpressionSet, then the function uses its 'exprs' slot.
...

arguments passed to mutualInfo and MIdist:

  • nbinnumber of bins to calculate discrete probabilities; default is 10.

  • diagif TRUE, then the diagonal of the distance matrix will be displayed; default is FALSE.

  • upperif TRUE, then the upper triangle of the distance matrix will be displayed; default is FALSE.

  • samplefor ExpressionSet methods, if TRUE, then distances are computed between samples, otherwise, between genes.

Details

For mutualInfo each row of x is divided into nbin groups and then the mutual information is computed, treating the data as if they were discrete.

For MIdist we use the transformation proposed by Joe (1989), δ∗=(1−exp⁡(−2δ))1/2\delta^* = (1 - \exp(-2 \delta))^{1/2} where δ\delta is the mutual information. The MIdist is then 1=δ∗1 = \delta^*. Joe argues that this measure is then similar to Kendall's tau, tau.dist.

Value

An object of class dist which contains the pairwise distances.

Author(s)

Robert Gentleman

References

H. Joe, Relative Entropy Measures of Multivariate Dependence, JASA, 1989, 157-164.

See Also

dist, KLdist.matrix, cor.dist, KLD.matrix

Examples

x <- matrix(rnorm(100), nrow = 5)
 mutualInfo(x, nbin = 3)

Spearman correlational distance

Description

Calculate pairwise Spearman correlational distances, i.e. 1-SPEAR or 1-|SPEAR|, for all rows of a matrix and return a dist object.

Usage

spearman.dist(x, ...)

Arguments

x

n by p matrix or ExpressionSet; if x is an ExpressionSet, then the function uses its 'exprs' slot.
...

arguments passed to spearman.dist:

  • absif TRUE, then 1-|SPEAR| else 1-SPEAR; default is TRUE.

  • diagif TRUE, then the diagonal of the distance matrix will be displayed; default is FALSE.

  • upperif TRUE, then the upper triangle of the distance matrix will be displayed; default is FALSE.

  • samplefor the ExpressionSet method: if TRUE (the default), then distances are computed between samples.

Details

We call cor with the appropriate arguments to compute the row-wise correlations.

Value

One minus the Spearman correlation, between rows of x, are returned, as an instance of the dist class.

Author(s)

Beiying Ding

See Also

cor.dist, tau.dist, euc, man, KLdist.matrix, KLD.matrix, mutualInfo, dist

Examples

x <- matrix(rnorm(200), nrow = 5)
 spearman.dist(x)

Kendall's tau correlational distance

Description

Calculate pairwise Kendall's tau correlational distances, i.e. 1-TAU or 1-|TAU|, for all rows of the input matrix and return an instance of the dist class.

Usage

tau.dist(x, ...)

Arguments

x

n by p matrix or ExpressionSet; if x is an ExpressionSet, then the function uses its 'exprs' slot.
...

arguments passed to tau.dist:

  • absif TRUE, then 1-|TAU| else 1-TAU; default is TRUE.

  • diagif TRUE, then the diagonal of the distance matrix will be displayed; default is FALSE.

  • upperif TRUE, then the upper triangle of the distance matrix will be displayed; default is FALSE.

  • samplefor the ExpressionSet method: if TRUE (the default), then distances are computed between samples.

Details

Row-wise correlations are computed by calling the cor function with the appropriate arguments.

Value

One minus the row-wise Kendall's tau correlations are returned as an instance of the dist class. Note that this can be extremely slow for large data sets.

Author(s)

Beiying Ding

See Also

cor.dist, spearman.dist, euc, man, KLdist.matrix, KLD.matrix, mutualInfo

Examples

x <- matrix(rnorm(200), nrow = 5)
 tau.dist(x)