simMat: Similarity Matrices

In npcooley/SynExtend: Tools for Working With Synteny Objects

Similarity Matrices

Description

The simMat object is an internally utilized class that provides similar functionality to the dist object, but with matrix-like accessors.

Like dist, this object stores values as a vector, reducing memory by making use of assumed symmetry. simMat currently only supports numeric data types.

Usage

## Create a blank sym object
simMat(VALUE, nelem, NAMES=NULL, DIAG=FALSE)

## S3 method for class 'vector'
as.simMat(x, NAMES=NULL, DIAG=TRUE, ...)

## S3 method for class 'matrix'
as.simMat(x, ...)

## S3 method for class 'simMat'
print(x, ...)

## S3 method for class 'simMat'
as.matrix(x, ...)

## S3 method for class 'simMat'
as.data.frame(x, ...)

## S3 method for class 'simMat'
Diag(x, ...)

## S3 replacement method for class 'simMat'
Diag(x) <- value

Arguments

VALUE	Numeric (or NA_real_) indicating placeholder values. A vector of values can be provided for this function if desired.
nelem	Integer; number of elements represented in the matrix. This corresponds to the number of rows and columns of the object, so setting nelem=10 would produce a 10x10 matrix.
NAMES	Character (Optional); names for each row/column. If provided, this should be a character vector of length equal to nelem.
DIAG	Logical; Is the diagonal included in the data? If FALSE, the constructor generates 1s for the diagonal.
x	Various; for print and Diag, the "simMat" object to print. For as.vector or as.matrix, the vector or matrix (respectively). Note that as.matrix expects a symmetric matrix–providing a non-symmetric matrix will take only the upper triangle and produce a warning.
value	Numeric; value(s) to replace diagonal with.
...	Additional parameters provided for consistency with generic.

Details

The simMat object has a very similar format to dist objects, but with a few notable changes:

• simMat objects have streamlined print and show methods to make displaying large matrices better. print accepts an additional argument n corresponding to the maximum number of rows/columns to print before truncating.

• simMat objects support matrix-style get/set operations like s[1,] or s[1,3:5]

• simMat objects allow any values on the diagonal, rather than just zeros as in dist objects.

• simMat objects support conversion to matrices and data.frame objects

• simMat objects implement get/set Diag() methods. Note usage of capitalized Diag; this is to avoid conflicts and weirdness with using base diag.

See the examples for details on using these features.

The number of elements printed when calling print or show on a simMat object is determined by the "SynExtend.simMat" option.

Value

simMat and as.simMat return an object of class "simMat". Internally, the object stores the upper triangle of the matrix similar to how dist stores objects.

The object has the following attributes (besides "class" equal to "simMat"):

nrow	the number of rows in the matrix implied by the vector
NAMES	the names of the rows/columns

as.matrix(s) returns the equivalent matrix to a "simMat" object.

as.data.frame(s) returns a data.frame object corresponding to pairwise similarities.

Author(s)

Aidan Lakshman [e-mail]

Examples

## Creating a blank simMat object initialized to zeros
s <- simMat(0, nelem=20)
s

## Print out 5 rows instead of 10
print(s, n=5)

## Create a simMat object with 5 entries from a vector
dimn <- 5
vec <- 1:(dimn*(dimn-1) / 2)
s1 <- as.simMat(vec, DIAG=FALSE)
s1

## Here we include the diagonal
vec <- 1:(dimn*(dimn+1) / 2)
s2 <- as.simMat(vec, DIAG=TRUE)
s2

## Subsetting
s2[1,]
s2[1,3:4]
# all entries except first row
s2[-1,]
# all combos not including 1
s2[-1,-1]

## Replace values (automatically recycled)
s2[1,] <- 10
s2

## Get/set diagonal
Diag(s1)
Diag(s1) <- 5
s1

npcooley/SynExtend documentation built on Nov. 15, 2024, 3:02 p.m.