Release 2

Contains mayor refactoring of the whole code base, but most important differences are these:

• Population information is now managed and stored in a Pandas DataFrame instead of a dedicated EvoDesign class. This should make integrating EvoDesign into other code bases easier.
• Instead of using Python's global RNG, a dedicated instance of NumPy's default RNG is now used. This should improve experiment reproducibility.

Finally, this version of the code was used to design alternative sequences for the IDR-2009 peptide.