SiteSifter finds highly conserved DNA motifs embedded within coding
 regions. Each instance of a motif is scored based on the chance that
 its constituent codons are conserved over and above that required for
 amino acid conservation.
Software  Paper

COALESCE uses large collections of genomic data and Bayesian
 integration to predict coregulated gene modules, the conditions of
 regulation, and the consensus binding motifs for regulation. It uses
 a synthesis of gene expression biclustering, motif prediction, and
 data integration (including expression, sequence, nucleosome
 positioning, and evolutionary conservation). It is available as part
 of the Sleipnir library.
Software  Paper

Nearest Neighbor Networks
Nearest Neighbor Networks (NNN) is a graph-based algorithm used to
 cluster genes with similar microarray expression profiles. The NNN
 clustering method is an alternative to classical techniques such as
 hierarchical and K-means clustering. NNN generates clusters of
 functionally related genes with high precision, and the clusters
 generally represent a broader selection of biological processes than
 those produced by other methods; NNN performs best on data sets with
 many conditions and on datasets that are modular (i.e. contain several
 grouped subsets of conditions).
Software  Paper