Graphs of Human Disease and Disease Genes
http://www.pnas.org/content/104/21/8685.full
This paper demonstrates and analyzes the relationship between human diseases and the genes that control those diseases. The authors created network graphs based on correlations between diseases/disorders, and the gene mutations that are known to be tied to them. The creation of such a figure provides a “genome-wide roadmap” that can be used in future work as a visual reference to the interconnectedness of genes and their diseases. As is seen in this paper, these networks are fairly dense, so such a reference will prove to be very useful. This work is limited, however, to only those disorders that have been researched so completely that there is great certainty of the link between itself and specific genes. Further expansion of these networks is dependent on great efforts of study and resources. Analysis of the specific genes and diseases led the researchers to find that disease genes are “nonessential” and are largely localized to the periphery of the network, whereas essential genes typically encode “hub proteins” that are expressed widely through the body. The authors state this suggests that disorders caused by mutations in genes that effect the physical body should have peripheral symptoms and implications.
In class, we’ve discussed how graphs of this format can represent all sorts of relationships, from diplomatic ties between countries to the food chain, but here we see the familiar format being applied to a notably less “social” subject. In the Human Disease Network (HDN), each node represents a specific disease, edges represent a gene mutation exists in common between the two nodes, the size of the node correlates to the number of genes that participate in the disease, and the color of the node indicates what disorder class it belongs to. By comparison, in the Disease Gene Network (DGN), each node represents a specific gene, an edge represents the connected genes participate in a common disorder, the size of the node is proportional to the number of diseases the gene is present in, and the color again indicates the class of the disorder associated with the node. It is interesting to observe the amount of triadic closures that exist within this network. Looking at the DGN in Figure 1, there are very few exceptions to the rule of Triadic Closure. The HDN from the same Figure 1 demonstrates a larger occurrence of exceptions to the rule. This makes me wonder about the strength of the edges, and if there is any way to define a “strong” versus a “weak” bond in such a network where edges are based on definitions and not subjective perceptions. What is it about genes versus diseases that makes them satisfy the Triadic Closure Property so closely? It is notable that the DGN only displays nodes that have at least one link, so we are only looking at the population of genes that is the most closely tied.