support for multiple gene family definitions

[adf-ncgr]

at the level of actually representing proteins as belonging to multiple trees, there is really no problem. There is a slight problem in terms of MSAs due to uniqueness constraints in chado, but we can either workaround this or modify our chado constraints slightly and propose that this be accepted into the core. The main problem as far as this module is concerned is how we'd want to expose the different sets of families through the search interface. An auxiliary problem is how to show relationships between different gene families. There may be some other things that need to get resolved, such as how link-ins on the basis of gene name would work if there are multiple possible answers (the current redirection mechanism would not seem to fit that scenario).