1: \begin{abstract}

2: Assessing the convergence of a biomolecular simulation is an essential 

3: part of any careful computational investigation, because many 

4: fundamental aspects of molecular behavior 

5: %important 

6: %quantities (e.g., free energy differences) 

7: depend on the relative populations 

8: of different conformers.

9: %; insufficient convergence translates into 

10: %systematic errors. 

11: Here we present a physically intuitive method to self-consistently 

12: assess the convergence of trajectories generated by molecular dynamics 

13: and related methods. 

14: Our approach reports directly and systematically on the 

15: structural diversity of a simulation trajectory. Straightforward 

16: clustering and classification steps are the key ingredients, 

17: allowing the approach to be trivially applied to systems of any size. 

18: Our initial study on met-enkephalin strongly suggests that even fairly 

19: long trajectories ($\sim 50$ nsec) may not be converged for this 

20: small---but highly flexible---system.

21: %Standard clustering methods first 

22: %generate a set of reference structures to any desired precision. The 

23: %trajectory is then classified by proximity to the reference structures, 

24: %yielding a one-dimensional histogram of structurally distinct populations. 

25: %Comparing ensembles 

26: %of different trajectories (or different parts of the same trajectory) built 

27: %with the same reference structures provides a sensitive, quantitative measure

28: %of convergence. Please note: this is a preliminary manuscript, and should 

29: %be read as such. Comments are most welcome, especially regarding pertinent 

30: %prior work. 

31: \end{abstract}

32: