Seems like it is the wrong time to be in the Devdas mode. I must keep up with the current state of reality because it is totally indifferent to my struggles. So instead of being mushy, and seeking the help of imaginary friends in the sky I must move on.
I must look into the intersection of Constraint Satisfaction Problems and Cellular Automata. CSPs can be easily be parallelized on CAs and Self Timed Cellular Automatas (STCAs) with a zillion times speed ups over conventional Turing Machines. If a CA based Constraint Solver which has the ability to model every constraint (a.k.a universal) can be found, that may be the holy grail of Computational Biology.
I ask myself to develop a Tangrams Solver first. It will give me an idea of how -ary constraints get implemented in spatial problems. Then general constraints experienced by each of the 20 amino acids in various pH values must be deduced by using automatic methods from the large amounts of experimentally verified protein structure data.
In both the Tangrams solver and in proteins, the constituent components are not related to a central sign post. Instead they are related to each other. Hence if there are components, they will be related to each other through binary constraints. These constraints would range from certain spatial ones that enforce steric hindrances to coloumb forces and non-coloumb forces like hydrogen bonds.
Once the solver is implemented on a CA, Hashlife like algorithms can take over and accelerate computation to astronomical speeds.