Table 1.
Evolutionary properties of communities and DAVID groups in AML 2.3.
Gene evolutionary rates (ERs) take real values from 0 (most conserved) to approximately 6.9 (most variable), and ages take integer values from 0 (oldest) to 12 (youngest). The table is organized as follows. "Comm. Index" is the index of the ten largest communities. "Num. genes" is the number of genes in the community. "Comm. ER" indicates whether the community is significantly hotter (i.e. has a higher ER) or colder (i.e. has a lower ER) than the mean of 300 equally-sized sets of genes randomly selected from the network, with a significance threshold of p = 10−3. "Diff. in mean" is the difference between the mean ER of the community and the mean ER of the 300 randomly selected sets. "p-value" is the significance of the difference. "Comm. age", "Diff. in mean", and "p-value" are the same as previously stated, but for age rather than ER. "DAVID group name" is the name of the DAVID group that DAVID identified as enriched in each community. “Group type” states whether the DAVID group is a protein type (P), location of final gene product (L), biological process (B), or cellular component (C). "Num. genes" is the number of genes in the DAVID group. "DAVID Benjamini" is the significance of the enrichment of the DAVID group, as reported by DAVID. The remaining DAVID group columns are computed in the same manner as the community columns.
Table 2.
Evolutionary properties of communities and DAVID groups in HumanNet.
See Table 1 for explanation of column headers.
Fig 1.
Ages and evolutionary rates for enriched DAVID groups in AML 2.3.
(A) Distribution of evolutionary rates (ERs), measured in units of the number of nonsynonymous substitutions per amino acid site per billion years, for all genes (purple) and for genes in the translational elongation DAVID group (green). This DAVID group has a very low ER compared to the background distribution. (B) Distribution of ages for all genes (purple) and genes in the transmembrane DAVID group (green), where age = 0 is the oldest and age = 12 is the youngest. Transmembrane genes are much younger than average. (C) Summary of mean ER and mean age for DAVID groups in Table 1. The relative ERs on the x-axis are computed from ERrelative = ERfunc. group mean − ERnetwork mean, and likewise for relative age on the y-axis. The DAVID groups from (A) and (B) have bold labels in (C). Each marker type corresponds to one of communities 0 through 9. As expected, old DAVID groups tend to have a low average ER (i.e. are “cold”), and young DAVID groups tend to evolve frequently (i.e. are “hot”). Unabbreviated DAVID group names are listed in Table 1.
Table 3.
Centrality and evolutionary measures in AML 2.3.
Single-node centrality measures exhibit a small but significant correlation with evolutionary rate and age. The DAVID groups’ average centrality measures show stronger correlation with evolutionary properties, particularly between PageRank and age.
Fig 2.
PageRank and DAVID groups for AML 2.3.
Mean PageRank versus mean age of each DAVID group from Table 1 (age = 0 is the oldest and age = 12 is the youngest). Old DAVID groups tend to have high PageRank. Unabbreviated DAVID group names are listed in Table 1.
Fig 3.
Set efficiency and evolutionary rate for AML 2.3.
The cumulative set efficiency (SE) of all genes below a given evolutionary rate (ER) rank (lowest to highest ER, i.e. “coldest” to “hottest”). The SE of the 500 coldest genes is significantly higher than the control, and including hotter genes monotonically decreases the SE. This indicates that the coldest genes exchange information efficiently, while the hottest genes are more dispersed and thus communicate less efficiently.
Fig 4.
Interset efficiency and age for AML 2.3.
Interset efficiency from DAVID group in column j to DAVID group in row i. The list of DAVID groups was sorted by average age from oldest (transcriptional elongation) to youngest (hemoglobin complex). Old DAVID groups exchange information efficiently, as indicated by the high interset efficiency values in the lower-left corner. Younger DAVID groups, particularly the blood cell-specific DAVID groups of lymphocyte activation and hemoglobin complex, are remote from most other DAVID groups. Note that the above matrix is asymmetric because the network is directed, and that the colors are log-scaled.
Fig 5.
Interset efficiency and age for HumanNet.
See Fig 4 for explanation.