Optimized per-ligand active states of receptors.

[primaryodors]

Based on the findings in #200, it was not the right approach to conform receptors to G proteins and then measure their ligand binding potentials.

Consequently, issues #270, #272, #281, #294 are now obsolete, as are predict/couple.pepd and predict/method_coupled.php. #218 was technically obsolete already, however the changes introduced in #310 supersede that requirement.

We are trying the opposite approach, namely to begin by finding the receptor's optimal active state conformer for each of its ligands, then to calculate its binding energy to the various G protein α subunits, and use both ligand binding energies and G-protein binding energies to arrive at a prediction.

This issue may serve as a hub from which to create manageable subtasks. These include:

1. Develop a variation of the "soft dock" feature that translates, pivots, and flexes transmembrane helices in limited ways to arrive at optimal ligand contact. #314
2. Add the option to output the best resulting pose of a dock as a whole PDB of the entire modified protein. #314
3. Add the ability to dock a Gαolf, Gαs, Gαq, or Gαo subunit into the cytoplasmic cavity of the PDB result from step 2 and give binding energy metrics. #315
4. Create a new prediction method using the features from steps 1-3 and delete the obsolete coupled method and couple pepd. #316