LFC: Select DCLs based on 'Limit Fold Change' model
Description
The limit fold change (LFC) model is a robust statistical method modeling the relationship between maximum coexpression and log coexpression ratio of genes.
(Mutch, et al., 2002). The algorithm starts with a set of gene coexpression value pairs each
comprising two coexpression values of a gene pair calculated under two different conditions respectively.
a two-column data matrix, with column one the coexpression values for one condition and column two those for another.
nbins
number of x bins for fitting y=a+(b/x).
p
the fraction at y axis for determining boundary points; must be within [0,1].
sign
specifies the sign type of exprs. Exprs is either 'same-sign', with coexpression pairs of the same sign, or 'different-sign', with coexpression pairs of opposite signs.
figname
names of figures of the LFC fitting results.
Value
the identified DCLs will be returned, as a subset of coexpression pairs.
Details
According to how the signs of coexpression values are paired, gene links are divided into two parts: the 'same-signed' set and the 'differently-signed' set. From the 'differently-signed' set, the 'correlation-switched' gene links that have two differently-signed coexpression values both surpassing a cutoff value are subtracted, who make the first part of DCLs. The remaining differently-signed gene links in aggregate inherit the title of 'differently-signed' set.
For the 'same-signed' set, gene links are binned with respect to their maximum coexpression values, and those links ranked the top p of highest fold changes in each bin are fitted with a simple equation of the form y = a + (b/x); for the 'differently-signed set', the horizontal and vertical axes are exchanged and similar binning and fitting procedures are applied. Links lie above the fitted curves are considered DCLs.
References
Mutch, D.M., Berger, A., Mansourian, R., Rytz, A. and Roberts, M.A. (2002) The limit fold change model: a practical approach for selecting differentially expressed genes from microarray data, BMC Bioinformatics, 3, 17.