On computer simulation methods used to study models of two-component lipid bilayers.

We show that the use of a computer simulation method introduced to calculate the equilibrium thermodynamic properties of a model of a two-component lipid bilayer membrane [Freire, E., & Snyder, B. (1980) Biochemistry 19, 88-94] is incorrect. This is done by comparing the method to that of Metropolis, which has been proven to generate equilibrium distribution of that model, and by showing that back-processes have been omitted in the implicit master equation of Freire and Snyder. We have illustrated this explicitly by first generating distributions according to the method of Freire and Snyder and then allowing the system to relax via the Kawasaki method, which uses the technique of Metropolis. We show that relaxation to a different distribution occurs. We also remark that the cluster distributions generated by the Freire-Snyder method are substantially different from those occurring in equilibrium distributions. Thus, conclusions about equilibrium thermodynamic properties such as specific heats and transition enthalpies or about transport properties or cluster properties at equilibrium cannot be drawn from the results obtained by using this method. Finally, we point out that the method of Freire and Snyder is appropriate to so-called aggregation models, which have been used to study irreversible growth; and we suggest biological systems that might be simulated by their method.