Sandeep Somani

News: Presenting a poster and talk based on Publication # 1 at Algorithms in MacroMolecular Modeling Conference at Austin TX from Nov 11-15, 2009.

Research

Overview. My main interests are in developing computational methods for understanding behaviour of molecules in a thermal environment using atomistic simulations. Statistical mechanics theory provides the link between microscopic properties (such as 3-D structure and dynamics) to thermodynamic properties (such as free energy). Accurate prediction of thermodynamic properties from molecular simulations has applications in many fields involving molecular design. My focus is on the biological application of drug design.

Current Research. Molecules are flexible structures and constantly fluctuate in a thermal environment. Molecular dynamics (MD) provides a tool for capturing conformational fluctuations on nanosecond timescale. Conformation of an N-atom molecule is specified by 3N-6 internal coordinates. As a result, even for small (N< 50) drug-like molecules, the conformational space is high dimensional. We developed a method to sample conformations in such high dimensions using low-order distribution functions (or pdfs) among the internal coordinates. The pdfs are obtained as normalized histograms of the time series of internal coordinates from the MD simulation. We find that much of the conformational fluctuations of a molecule are captured by low order distributions (or correlations).

Representative conformations of united-atom cyclohexane from Molecular Dynamics simulation (in green) and from sampling (in black). There are 6 atoms in the molecule corresponding to 12 internal coordinates. Conformational fluctionals are thus described by a 12-dim pdf. The sampled conformations were obtained as follows. MD simulation was used to extract all 1-,2- and 3-dim pdf among the 12 Bond-Angle-Torsion (BAT) coordinates. The using the superposition approximation based sampling algorithm , conformations were sampled from only 1-D or singlet pdfs (BAT-1), singlet and doublet pdfs (BAT-2), and singlet, doublet and triplet pdfs (BAT-3). Each conformation is oriented such that atom 1 is at the origin, atom 2 is along the x axis, and atom 3 is in the x-y plane. Each set contains 100 conformations. The viewing direction is along the x-y place, so that atoms 1, 2 and 3 are on the solid line to the left. Each set contains 100 conformations.

Publications

Computational Structural Biology

1. Somani S, Killian BJ and Gilson MK, Sampling Conformations in High Dimensions Using Low Dimensional Distribution Functions. Journal of Chemical Physics, 130, 134102 (2009) pdf
2. Killian BJ, Kravitz JY, Somani S, Dasgupta P, Pang Y-P, Gilson MK, Configurational Entropy in Protein-Peptide Binding.Computational Study of Tsg101 UEV Domain with an HIV-derived PTAP Nonapeptide. Journal of Molecular Biology, Volume 389, Issue 2, 2009
3. Somani S, Chng Choon-Peng, Verma CS, Hydration of a hydrophobic cavity and its functional role: a simulation study of human interleukin-1beta. Proteins 67(4):868-85 (2007) pdf

Computational Systems Biology

1. Dhar P, Meng TC, Somani S, Ye L, Sairam A, Chitre M, Hao Z, Sakharkar K. Cellware--a multi-algorithmic software for computational systems biology. Bioinformatics 2004 May 22;20(8):1319-2
2. Dhar P, Meng TC, Somani S, Ye L, Sakharkar K, Krishnan A, Ridwan AB, Chitre M, Hao Z. Grid Cellware: the first grid-enabled tool for modeling and simulating cellular processes  Bioinformatics 2005 Apr 1;21(7):1284-7
3. Meng TC, Somani S, Dhar P, Modeling and simulation of biological systems with stochasticity. In Silico Biology 2004 Apr 16;4(2)