Stratification of co-evolving genomic groups using ranked phylogenetic profiles

BMC Bioinformatics

Table 3 Intra- and inter-cluster correlation in the four clusters dominated by S. pyogenes proteins.

	Mean correlation between all protein-pair combinations†				Mean correlation with the cluster-profile††				Mean correlation with the reduced cluster-profile††
	cl 1	cl 4	cl 14	cl 19	Cl 1	cl 4	cl 14	cl 19	cl 1	cl 4	cl 14	cl 19
Cluster 1	0.54	0.30‡	0.40‡	0.14‡	0.56	0.57	0.41‡	034‡	0.56	0.52	0.40‡	034‡
Cluster 4	0.30‡	0.51	0.17‡	0.56	0.43‡	0.53	0.27‡	0.16‡	0.00	0.04	0.00	0.00
Cluster 14	0.40	0.17‡	0.39	0.18‡	0.44	0.17‡	0.48	0.27‡	0.47	0.18‡	0.49	0.28‡
Cluster 19	0.14‡	0.56	0.18‡	0.32	0.22‡	0.43	0.29‡	0.47	0.25‡	0.43	0.31‡	0.47

Intra-cluster combinations are shown in bold.
† Each cell shows the mean correlation for all possible protein-pair combination between the protein-members of the row-cluster and the protein-members of the column-cluster. Diagonal values show the intra-cluster correlation.
†† Each cell shows the mean correlation between cluster-profile of the row-cluster and all members of the column-cluster.
ℑ Reduced cluster profile is composed solely of data points defined as highly-agreeable. Agreeability was calculated for each data point (species) in a cluster as the inverse coefficient of variation weighted by the fraction of appearances of a species in a cluster:
Agr = (mean*app)/(sd*mmb)
Where mean is the mean position of a species in a cluster; app is the number of appearances of a species in a cluster; sd is the standard deviation calculated for the species; and mmb is number of proteins in a cluster.
Data points for which Agr > 0.5 are considered highly agreeable.
‡ The inter-cluster correlation is significantly lower than the intra-cluster correlation observed for the row-cluster. Significance is defined as P value < 0.05 in the Wilcoxon rank sum test (equivalent to the Mann-Whitney test).

ISSN: 1471-2105