CheckStability {mlica}R Documentation

Tests stability of inferred ICA modes.

Description

Performs a correlation test to see which of the inferred ICA modes are reproducible across multiple runs using different random initialisations. Returns a set of consensus ICA modes and stability scores for each following the algorithm of Chiappetta,...et.al (2004).

Usage

CheckStability(a.best.l, corr.th)

Arguments

a.best.l List of a.best objects from mlica runs.
corr.th Correlation threshold to use to decide whether a mode is reproducible.

Value

A list with the following components

consS Consensus source matrix with columns labeling the consensus ICA modes. Has same number of rows as a.best$S.
consA Consensus mixing matrix with rows labeling the consensus ICA modes.
stabM Vector of same length as consM giving the stability measures of each consensus ICA mode. Stability or reproducibility measures are given as fractions, that is, the number of times the ICA mode correlates with one of the other runs at threshold level corr.th divided by the number of runs (length of a.best.l).

Author(s)

Andrew Teschendorff aet21@cam.ac.uk

References

1
Hyvaerinen A., Karhunen J., and Oja E.: Independent Component Analysis, John Wiley and Sons, New York, (2001).
2
Kreil D. and MacKay D. (2003): Reproducibility Assessment of Independent Component Analysis of Expression Ratios from DNA microarrays, Comparative and Functional Genomics 4 (3),300–317.
3
Liebermeister W. (2002): Linear Modes of gene expression determined by independent component analysis, Bioinformatics 18, no.1, 51–60.
4
Chiappetta P., Roubaud MC. and Torresani B.: Blind source separation and the analysis of microarray data, J. Comput. Biol. 2004; 11(6):1090–109.


[Package mlica version 0.6.1 Index]