Minor paper suggestions

[j-andrews7]

The paper is pretty solid, just a few comments. Given the usage of networkx for the network generation, probably worth throwing a mention/reference to it into the text somewhere.

Other minor edits (recommended deletions or additions):

For example, in the GO data, GO:0090630 defines activation of GTPase activity and is a child of GO:0043547, which is a positive regulation of GTPase activity which in turn is a child of GO:0051345 representing a positive regulation of hydrolase activity.

These two instances of "a" feel out of place.

The complete human genome GAF representation contains 288,575 associations of 19,606 gene symbols with over 18,680 GO terms.

Seems like "with" is missing here.

Since the resulting graphs may also be large, with thousands of nodes, the main interface provides input widgets that allow users to interactively reduce the subgraph to nodes for which:

1. The function definitions match certain patterns**.**
2. A minimum number of genes share a function**.**
3. Nodes belong to a specific GO subtree: Biological Process (BP), Molecular Function (MF), or Cellular Component (CC).

Just unify punctuation of bullet points here.

[ialbert]

Word edits applied, added networkx as a citation.