MEGA-SQ: A method using the SQuEAL function to find the optimal superposition of several quasi-flexible molecules

One of the most common operations in molecular modeling is to superimpose s ets of active molecules in order to determine what features they have in co mmon and what the three-dimensional disposition of these features might be. In this paper we describe MEGA-SQ, a procedure based on our previously des cribed method SQ (Miller et al. J. Med. Chem. 1999, 42, 1505-1514), which g enerates and scores superpositions of pairs of rigid molecules. Given a set of explicitly generated conformations for each of several molecules, MEGA- SQ generates pairwise comparisons of the conformations and then builds the pairwise comparisons into high-scoring "ensembles" with a genetic algorithm . An ensemble contains one conformation of each molecule with the conformat ions oriented so that they have the best mutual superposition. We demonstra te the utility of MEGA-SQ with two examples: angiotensin-II antagonists and neurokinin-1 antagonists. We can show that, without any user bias as to wh at groups are common to the molecules, MEGA-SQ can produce superpositions t hat closely resemble published pharmacophore models.