Scoring function problem

As discussed in the introduction, scoring functions still pose problems (see also Chapter 14). Some of these problems arise from insufficient consideration of details of favorable and unfavorable protein-ligand interactions. 

In this section we briefly discuss an approximate formalism that allows incorporation of the experimental error variances in the constrained maximisation of the Bayesian score. The problem addressed here is the derivation of a likelihood function that not only gives the distribution of a structure factor amplitude as computed from the current structural model, but also takes into account the variance due to the experimental error. 

Difficulties in these approaches include that, according to literature [57-59] and the authors own experience, there is no general preference to one superior docking tool and scoring function that adequately works for different problems. Furthermore, it is an open question whether the detection of the best docking mode and the ranking of molecules require different scoring functions. 

Practically all virtual screening procedures rely at least in part on some numerical optimization, be it an optimization of overlap (as in many alignment programs) [81-90], the generation of energetically favorable conformations of a molecule (for example CONCORD ]91] and CORINA [92]), or the relaxation of a compound in complex with the protein (for example [93-97]). The particular virtual screening problem as a whole may be solved this way. Once a decent scoring function is defined, numerical methods... 

Smith, R., Hubbard, R.E., Gschwend, D.A., Leach, A.R., and Good, A.C. Analysis and optimization of stmcture-based virtual screening protocols (3). New methods and old problems in scoring function design. 

---