Welcome to Clemson Computational Biophysics Lab

We are a theoretical and computational biophysics group in the Department of Physics and Astronomy at Clemson University, College of Science. We apply concepts and methods in Physics, especially Statistical Mechanics and Thermodynamics, to study biological systems, and try to learn new physics emerging from the complex systems. The major theme throughout our research is to integrate dynamics into the study of structure-function relationship of biomolecules and molecular complexes. We believe that uncovering the interrelationship between structure, dynamics, and function of biomolecules can help better understand biology and human diseases, and accelerate biomedical research.

Our research focus is multiscale modeling of biomolecules and molecular complexes. Experimental characterization of biological systems is often hindered by the limited ability to cover a wide range of time and length scales, associated with many biological processes. Multiscale modeling, which innovatively combines atomistic and coarse-grained simulations, provides a unique opportunity to bridge the gaps of time and length scales between experimental observation and underlying molecular systems. We apply the multiscale modeling approach to study structure, dynamics, and function of large biomolecules, formation of molecular complexes, and also interactions between nanomaterials and biological systems.

We are always interested in enthusiastic colleagues to join our lab.

Inhibition of islet amyloid polypeptide aggregation in type-II diabetes

Accumulating evidence suggests that the aggregation of islet amyloid polypeptide (IAPP, a.k.a. amylin) is associated with β-cell death in type 2 diabetes (T2D). IAPP is co-secreted with insulin by pancreatic beta-cells, and also works together with insulin to control the serum glucose level. In vitro studies suggest that IAPP is one of the most amyloidogenic peptide, which forms amyloid fibrils within hours at micromolar concentrations. However, no apparent IAPP amyloid aggregates are observed in healthy individuals where IAPP is stored in β-cell granules at milimolar concentrations for hours before being secreted to the blood stream. Therefore, physiological conditions of β-cell granules natively inhibit the amyloid aggregation of IAPP.

The cellular environment of β-cell granules is unique in its high concentrations of Zn2+, insulin and proinsulin c-peptide in addition to IAPP. C-peptide is the co-product of insulin synthesis, which connects insulin A- and B-chains in the precursor proinsulin and is co-secreted with insulin in equal molar. A high concentration of Zn2+, maintained by a β-cell specific zinc transporter – ZnT8, is believed to be important for the efficient storage of insulin: zinc coordinates the formation of insulin hexamers, which form crystals in the dense core of β-cell granules. ). We hypothesize that intermolecular interactions with insulin, zinc and c-peptide are important for the native inhibition of IAPP aggregation inside β-cell granules. Using state-of-the-art discrete molecular dynamics (DMD) simulations, we showed that both insulin monomers and dimers could bind IAPP monomer and inhibit IAPP self-association by competing with same amyloidogenic regions, subsequently preventing aggregation. Continue reading

Interactions between Nanoparticles and Biomolecules at the Nano-Bio interface

The advancement of nanomedicine and increasing applications of nanoparticles (NPs) in consumer products have led to administered biological exposure and unintentional environmental accumulation of NPs, causing concerns over the safety and sustainability of nanotechnology. Other sources of NP pollutions include combustion processes from human activities, such as petroleum-fueled vehicles. Therefore, there is a crucial need to understand the molecular mechanism of nanotoxicity to ensure safe nanotechnology and enable the vast applications of nanomedicine.

Continue reading

Protein folding, misfolding, and aggregation

Most proteins fold into specific three-dimensional structures, which determine their functions. The folding process can be described by the free energy landscape as in a first order phase transition. The native state features the lowest free energy and correspond to the most stable and most populated species in physiological conditions. However, due to either environmental changes or mutations, the native states are destabilized. The intermediate(s) and unfolded states are promoted, where protein exposes their hydrophobic core and un-satisfied hydrogen bond donors and acceptors. These non-native species are sticky in nature and tends to aggregate under high concentrations. The aggregation is a nucleation process, and the final aggregates can adopt a fibrillar shape depending on the structural and dynamical properties of the aggregation precursor species. Continue reading

Molecular Recognition

A major challenge in modeling molecular recognition is the conformational flexibility. The structures of the receptors in the bound and un-bound states are often different, known as the induced-fit problem. In these cases, a rigid docking approach will fail to achieve accurate predictions. Existing flexible methods for modeling molecular recognition or docking often adopt the ensemble docking approach, where an ensemble of receptor and/or ligand conformations are pre-constructed. Rigid or semi-rigid docking is performed onto these structures in the ensemble. The predicted binding poses are then selected from the grand ensemble of poses. In another words, the conformational flexibility of ligands and receptors are modeled in a decoupled or loose-coupled manner. Once the bound-conformation is not included in the pre-constructed ensemble, accurate prediction will be difficult. Alternatively, we fully integrate the conformational flexibility into the modeling of molecular recognition.

Continue reading

Post-translational modifications

The structural and energetic determinants of TPST enzyme specificity

Many PTM enzymes have strong sequence preferences in the targeted substrate proteins/peptides while others do not. What are the molecular mechanism of such drastic differences in PTM enzyme specificities? It is expected that the binding affinity between substrate and enzyme plays an important role in defining the specificities. Interestingly, in the case of tyrosylprotein sulfotransferase proteins (TPST) affinity alone cannot explain the differences in sulfated and non-sulfated sequences. We found that the structural properties of the peptide in the host protein also play an important role in determining the TPST specificity. We are trying to extend this idea to other PTM enzymes.

1. P. Nedumpully-Govindan, L. Li, E.G. Alexov, M.A. Blenner, and F. Ding, “Structural and energetic determinants of tyrosylprotein sulfotransferase sulfation specificity”, Bioinformatics, in press (2014)

Protein folding


a) Protein folding thermodynamics and kinetics

b) Protein folding transition states

c) Protein folding pathways

d) Protein conformational dynamics

Continue reading

Protein Design and Protein Engineering

The ability to accurately manipulate a protein’s structure, dynamics and function, and eventually to create de novo proteins with specific functions is the holy grail of modern biology. The success in protein design & engineering requires the accurate knowledge of inter-residue interactions, protein folding, protein-ligand and protein-protein recognitions, and also the molecular mechanism of catalysis. For example, the design of an enzyme needs the sub-angstrom precision in strategically positioning residues important for catalysis. Therefore, computational modeling plays a crucial role in protein design. Despite some recent successes in protein design, we are still far from perfect, requiring better understanding of inter-molecular interactions and faster sampling approaches.


We designed a molecular modeling suite, named Medusa. Medusa features physical-based force field to model inter-residue interactions, and a Monte-Carlo based simulated annealing approach to sample the sequence and sidechain structural spaces. The orientation of amino acid is usually modeled by rotamer library, a set of discrete sidechain conformations tabulated from the protein databank.  Continue reading

Modeling RNA 3D structure using experimental constraints

RNA structure determination is one of the major challenges in structural biology. Many RNAs are not amenable to high-resolution structure characterization by either x-ray or NMR methods because of their conformational flexibility or large size. Recently, novel computational methods to determine RNA structures have begun to emerge, but have often been limited to small RNAs with simple topologies due to either sampling problems or inaccuracy in force field. We developed a discrete molecular dynamics (DMD)-based RNA modeling approach, which allowed robust recapitulation of 3D structure of small RNA structures (< 50 nts). To fold large RNAs with complex 3D structures, we proposed to incorporate experimentally-derived structural information into modeling. Using various types experimentally-derived structural information to bias DMD simulations, we were able to recapitulate 3D structure of RNAs with complex topologies and lengths up to 230 nts. We expect a broad application of our experimentally-driven RNA modeling approach for generating robust structural hypotheses that are useful for guiding explorations of structure-function relationships in RNA.
Continue reading