Fig 1.
Visual Representation of Our Approach.
First, we summarize gene annotations made to functional terms in the Gene Ontology hierarchy as a gene-term bipartite graph. From these gene-term relationships, we project a term-term network. We partition this network into communities and compare those term communities to branches of terms in the DAG. Finally, we perform functional enrichment analysis on experimentally-defined gene sets using both the term communities and GO branches.
Fig 2.
Visualization of Communities (Circles) of GO Terms Found at the Six Lowest Levels of Resolution (Rows), in Increasing Order (Top to Bottom).
The width of the line connecting two communities is proportional to the percentage of terms in the child community that are also in the parent community. The size of communities is proportional to the log of the number of terms in the community. Color represents the normalized percentage of terms in the community which belong to the BP (yellow), MF (cyan) and CC (magenta) primary domains.
Fig 3.
A Comparison of Branches in the GO DAG and Term Communities Found by Partitioning the Term Network.
(a) Distribution of Jm, the maximum similarity a community or branch with ten or more members has compared to all other branches or communities with ten or more members, respectively. Although a small number of communities and branches have similar memberships, most are highly dissimilar. (b)-(c) Two example comparisons between communities and branches: (b) TC:0003876 compared to GO:0015298, and (c) TC:0011556 compared to GO:0090559. In each panel on the left hand side a community and its inter-community connections in the annotation-driven term network is shown and on the right hand side the branch with which that community has the the highest Jaccard similarity is illustrated. In the right panel edges represent the ontological associations defined by the Gene Ontology term hierarchy. Each term member of the community or branch is colored both by its associated primary domain (inner color—BP:yellow, MF:cyan, CC:magenta) and its community membership (outer color), determined at the same resolution value as the illustrated community. Terms common between each community and branch pair are circled. To read term-labels, please zoom in.
Fig 4.
Biological Information in Term Communities.
(a-d) Term Communities (TC:0000228, TC:0000227) and branches (GO:0050896, GO:0002376) summarized as word clouds. In each case the color of a word represents how often the term description containing that word belongs to each of the primary domains (BP:yellow, MF:cyan, CC:magenta, also see Fig 2 for mixed-domain coloration) and size represents that word’s statistical enrichment in that community/branch.
Fig 5.
Functional Enrichment Analysis in Branches and Communities.
(A) A plot of the percentage of branches and percent of communities found to be enriched at p < 0.01 in each gene signature. Although both communities and GO-branches have enrichment in these signatures, many signatures are only enriched in communities and not branches at the p < 0.01 significance. We chose a subset of signatures to investigate further, and note those with a blue dot. (b) A heat map showing the statistical enrichment of selected cancer signatures (noted in (a) with a blue dot), as measured by AEA, in both GO branches and term communities.