OVERCOMING STABILITY LIMITATIONS IN BIOMOLECULAR DYNAMICS - I - COMBINING FORCE SPLITTING VIA EXTRAPOLATION WITH LANGEVIN DYNAMICS IN LN

We present an efficient new method termed LN for propagating biomolecu lar dynamics according to the Langevin equation that arose fortuitousl y upon analysis of the range of harmonic validity of our normal-mode s cheme LIN. LN combines force linearization with force splitting techni ques and disposes of LIN'S computationally intensive minimization (anh armonic correction) component. Unlike the competitive multiple-timeste pping (MTS) schemes today-formulated to be symplectic and time-reversi ble-LN merges the slow and fast forces via extrapolation rather than ' 'impulses;'' the Langevin heat bath prevents systematic energy drifts. This combination succeeds in achieving more significant speedups than these MTS methods which are Limited by resonance artifacts to an oute r timestep less than some integer multiple of half the period of the f astest motion (around 4-5 fs for biomolecules). We show that LN achiev es very good agreement with small-timestep solutions of the Langevin e quation in terms of thermodynamics (energy means and variances), geome try, and dynamics (spectral densities) for two proteins in vacuum and a large water system. Significantly, the frequency of updating the slo w forces extends to 48 fs or more, resulting in speedup factors exceed ing 10. The implementation of LN in any program that employs force-spl itting computations is straightforward, with only partial second-deriv ative information required, as well as sparse Hessian/vector multiplic ation routines. The linearization part of LN could even be replaced by direct evaluation of the fast components. The application of LN to bi omolecular dynamics is well suited for configurational sampling, therm odynamic, and structural questions. (C) 1998 American Institute of Phy sics.