Biology Reference
In-Depth Information
species, the available expression data indicated that relatively few of these
relationships have been conserved among the six diverse species.
35
Global properties of expression data can also be studied by con-
structing “expression networks,” where genes with similar expression
profiles are connected.
52
Despite the small proportion of conserved rela-
tionships, we found that the topological properties of such networks are
conserved in all six organisms. This includes power-law connectivity dis-
tributions (with similar exponents), increased likelihood of connecting
genes of similar connectivity, and a high degree of clustering. In addition,
highly connected genes were significantly more likely to be essential and
conserved.
35
2.3. Differential Clustering Algorithm
We developed the differential clustering algorithm (DCA) to study sys-
tematically whether the coexpression of genes in one organism has been
conserved fully, partially, or not at all in another organism.
53
The DCA is
applied to a set of orthologous genes that are present in both organisms.
Such a set can be defined according to a functional category, a common
motif in the upstream sequence, or a transcription module in either
dataset. As a first step, the pairwise correlations between these genes are
measured in each organism separately, defining two pairwise correlation
matrices (PCMs) of the same size. Next, the PCM of the primary
(“reference”) organism is clustered, assigning genes into subsets that are
coexpressed in this organism, but not necessarily in the second (“target”)
organism. Finally, the genes within each coexpressed subgroup are
reordered by clustering according to the PCM of the target organism.
This procedure is performed twice, reciprocally, such that each PCM is
used once for the primary and once for the secondary clustering, yield-
ing two distinct orderings of the genes.
The results of the DCA are presented in terms of the rearranged
PCMs (Fig. 5). Since these matrices are symmetric and refer to the same
set of orthologous genes, they can be combined into a single matrix
without losing information. Specifically, we join the two PCMs into one
composite matrix such that the lower-left triangle depicts the pairwise
correlations in the reference organism, while the upper-right triangle
