Can Finite Size Effects in the Poland-Scheraga Model Explain Simulations of a Simple Model for DNA Denaturation ?

Abstract

We compare results of previous simulations of a simple model of DNA denaturation to the predictions of the Poland-Scheraga model. Concentrating on the critical region of the latter model we calculate both thermodynamic quantities and the distribution functions measured in the simulations. We find that the Poland-Scheraga model yields an excellent fit to the data, provided (i) we include a (singular) factor weighting the open ends of the doubly stranded chain, and (ii) we keep the leading corrections to the finite size scaling limit. The exponent c_1, which governs the end-weighting factor, is fairly well determined: 0.1 <~ c_1 <~ 0.15. The exponent c, which governs the length distribution of large loops, is determined only poorly. The data are compatible with values of c in at least the range 1.9 <~ c <~ 2.2. From the data it therefore cannot be decided whether the denaturation transition asymptotically is of first or of second order. We suggest that simulations of doubly stranded chains closed at both ends might allow for a more precise determination of c.