Assembly scoring ties, how to treat them?

[josemduarte]

In some cases (e.g. 5b1l) we can have assembly scoring ties. At the moment they are solved by assigning bio to the smallest assembly (see #166). Should it be the largest?

Notice that in the 5b1l example there's another independent issue, whereby protein-DNA interfaces have 0 score (this is issue #161), which then results in 0 scores for all assemblies