isa.iterate {isa2}

The Iterative Signature Algorithm

Description

Perform ISA on the (normalized) input matrix.

Usage

## S4 method for signature 'list'
isa.iterate(normed.data, ...)

Arguments

normed.data

The normalized data. A list of two matrices, usually coming from isa.normalize.

...

Additional arguments, see details below.

Details

isa.iterate performs the ISA iteration on the specified input seeds. It can be called as

    isa.iterate(normed.data, row.seeds, col.seeds,
                thr.row, thr.col = thr.row,
		direction = c("updown", "updown"),
		convergence = c("corx", "cor", "eps"),
		cor.limit = 0.99, eps = 1e-04, corx=3,
		oscillation = FALSE, maxiter = 100)
	      

where the arguments are:

normed.data

The normalized data. A list of two matrices, usually coming from isa.normalize.

row.seeds

The row seed vectors to start the ISA runs from. Every column is a seed vector. (If this argument and col.seeds are both present, then both of them are used.)

col.seeds

The column seed vectors to start the ISA runs from, every column is a seed vector. (If this argument and row.seeds are both present, then both of them are used.)

thr.row

Numeric scalar or vector giving the threshold parameter for the rows. Higher values indicate a more stringent threshold and the result biclusters will contain less rows on average. The threshold is measured by the number of standard deviations from the mean, over the values of the row vector. If it is a vector then it must contain an entry for each seed.

thr.col

Numeric scalar or vector giving the threshold parameter for the columns. The analogue of thr.row.

direction

Character vector of length two, one for the rows, one for the columns. It specifies whether we are interested in rows/columns that are higher (‘up’) than average, lower than average (‘down’), or both (‘updown’).

convergence

Character scalar, the convergence criteria for the ISA iteration. If it is ‘cor’, then convergence is measured based on high Pearson correlation (see the cor.limit argument below) of the subsequent row and column vectors. If it is ‘eps’, then all entries of the subsequent row and column vectors must be close to each other, see the eps argument below.

‘corx’ is similar to ‘cor’, but the current row/column vectors are compared to the ones corx steps ago, instead of the ones in the previous step. See the corx argument below, that gives the size of the shift.

cor.limit

The correlation limit for convergence if the ‘cor’ method is used.

eps

Limit for convergence if the ‘eps’ method is used.

corx

The number of iterations to use as a shift, for checking convergence with the ‘corx’ method.

oscillation

Logical scalar, whether to look for oscillating seeds. Usually there are not too many oscillating seeds, so it is safe to leave this on FALSE.

maxiter

The maximum number of iterations allowed.

Value

A named list with many components. Please see the manual page of isa for a complete description.

Author(s)

Gabor Csardi [e-mail]

References

Bergmann S, Ihmels J, Barkai N: Iterative signature algorithm for the analysis of large-scale gene expression data Phys Rev E Stat Nonlin Soft Matter Phys. 2003 Mar;67(3 Pt 1):031902. Epub 2003 Mar 11.

Ihmels J, Friedlander G, Bergmann S, Sarig O, Ziv Y, Barkai N: Revealing modular organization in the yeast transcriptional network Nat Genet. 2002 Aug;31(4):370-7. Epub 2002 Jul 22

Ihmels J, Bergmann S, Barkai N: Defining transcription modules using large-scale gene expression data Bioinformatics 2004 Sep 1;20(13):1993-2003. Epub 2004 Mar 25.

See Also

isa2-package for a short introduction on the Iterative Signature Algorithm. See isa for an easy way of running ISA.

Examples

## A basic ISA work flow for a single threshold combination
## In-silico data
set.seed(1)
insili <- isa.in.silico()

## Random seeds
seeds <- generate.seeds(length=nrow(insili[[1]]), count=100)

## Normalize input matrix
nm <- isa.normalize(insili[[1]])

## Do ISA
isares <- isa.iterate(nm, row.seeds=seeds, thr.row=2, thr.col=1)

## Eliminate duplicates
isares <- isa.unique(nm, isares)

## Filter out not robust ones
isares <- isa.filter.robust(insili[[1]], nm, isares)

## Print the sizes of the modules
cbind( colSums(isares$rows!=0), colSums(isares$columns!=0) )

## Plot the original data and the modules found
if (interactive()) {
  layout(rbind(1:2))
  image(insili[[1]], main="In silico data")
  image(outer(isares$rows[,1],isares$columns[,1])+
        outer(isares$rows[,2],isares$columns[,2])+
        outer(isares$rows[,3],isares$columns[,3]), main="ISA modules")
}