Well, there are two ways to think about grouping. First, grouping a set of nodes into a compound node can help visualize some distinct aspect of those nodes. For example, a group of proteins might form a complex. Second, in very dense networks, you may want to abstract a set of nodes and collapse them -- in sequence similarity networks, for example, it's not uncommon to collapse a set of very similar sequences into a single node. Here are two papers you can look at that have suggestions for when you might want to use grouping:

1. Ten simple rules to create biological network figures for communication
2. Visualization of omics data for systems biology

-- scooter