Biology

Computational Biology research in CSMD at ORNL encompasses many important aspects including molecular biophysics for bio-energy, genetic level pathogen identification for bio-security and development of data analytics for human health. We develop and apply theoretical methods and computational algorithms for investigating processes that occur at the biomolecular level to entire organisms and microbial communities. At the fundamental level, we are interested in biomolecular structure, dynamics and function particularly in relation to the biophysical mechanism of enzyme catalysis. For applications in the area of bio-energy, we focus on the discovery of pathways and regulatory networks in plant-microbial communities (in association with the Plant Microbial Interfaces, PMI, initiative) and bio-energy related crops (in association with the Bio-energy Science Center, BESC). In particular, we utilize novel algorithms to assign functions to the so-called hypothetical genes that could not be annotated by existing techniques and advance genome-wide association approaches to link genotypes and phenotypes in plants. These investigations take advantage of analytical approaches in mass-spectrometry and high-throughput transcriptomics algorithms developed in-house. We also utilize molecular simulations to dissect the mechanism for enzyme catalyzed biochemical processes that allow conversion of biomass into easily fermentable sugars. For the area of bio-security, we develop and apply high-performance algorithms to analyze metagenomics, transcriptomics, proteomics, and metabolomics data for genetic level pathogen identification. Another major emphasis of the Computational Biology effort is to obtain fundamental knowledge that cannot be obtained by experimental techniques, but we also work closely with the experimental facilities including the Spallation Neutron Source (SNS) at ORNL for improving the quality of information that is obtained. Another aspect of the research involves most effective utilization of heterogeneous and energy efficient computer architectures for modeling and simulations related to biology. In the area of human health, we are developing hardware and software infrastructure that will enable real-time and near real-time analysis and extraction of knowledge from the vast amount of data that is being generated.

Computational Biophysics

(this represents work of Chongle, Pratul, Xiaolin, Bobby Sumpter, Miguel, but not bioinformatics):

The power of computational biology lies in obtaining knowledge that is beyond the reach of experimental techniques. In particular, the use of computational simulations and modeling provides insights into the structure, dynamics, folding and function of biomolecular assemblies. The long term goal of the group is to develop and utilize novel computational algorithms to obtain detailed information about the bio-molecular assemblies. In addition, the interface of biology with other domains including chemistry and materials provides a unique opportunity.

The areas of expertise and interest include:

  1. Characterization of microbial systems using genomics and proteomics: Genomics and proteomics are providing systems biology insights into isolated microbes and natural microbial communities of interest. A protein family database for accurate functional annotation of sequenced microbial genomes is under construction. This will enable large-scale comparative genomics analysis for evolution study, pathway discovery and phenotype prediction. Complementary to genomics, proteomics measures globalprotein expression patterns. We are developing new algorithms for quantifying protein abundance changes, tracking stable isotope incorporation in proteins, and identifying post-translation modifications and amino acid mutations. Results from these approaches will be integrated into a knowledgebase to facilitatecommunity collaboration and information dissemination.
  2. Biomolecular characterization based on detailed atomistic simulations: Biological systems are dynamic in nature; characterization of their dynamics at the atomistic level is therefore essential to understanding many biological phenomena. Computer simulations at the atomistic level provide an excellenttool for capturing dynamics of biological systems. When closely coupled with experiments, it opens the opportunity to address many fundamental mechanistic problems in bio-energy and environmental sciences. Current efforts include the characterization of cellulose and ribulose-1,5-bisphosphate carboxylase/oxygenase enzymes, and gating mechanisms in a bacterial ion channels.
  3. Catalysis including enzyme catalysis: Enzymes are important for industrial applications, human health related uses and more importantly for mission of DOE in renewable energy and carbon sequestration strategies for climate change. Engineered enzymes have been sought with improved efficiency to lower the cost of applications and improve the efficiency of various biochemical processes. Computational modeling and simulations allow a detailed understanding of the factors that provide high catalytic efficiency and detailed understanding the mechanism of catalysis. In collaboration with experiments, computational techniques could break the next break-through in enzyme engineering.
  4. "Nano-bio" interface/materials: The interface between biology, chemistry and materials offers a unique opportunity for design of new bio-inspired material, new probes and nano-devices. Computational methods allow new opportunities to explore various properties of interest for the materials during the exploration phase and help in selecting and the development of right "Nano-bio" materials.
  5. Joint computational-neutron sciences: Dynamical of biological molecules has gained significant interest, particularly in relation to the biological function. Use of joint computational and experimental (neutron sciences) techniques can provide vital information about dynamics at different time-scales, as well as information about structure and other properties of biomolecules.
  6. Optimizing software of emerging architectures and developing novel algorithms: Methods are being developed and implemented to enable more scalable fast multi-pole method for treating long-range electrostatic interactions in large-scale moleculardynamics simulations. In particular, emphasis is on the emerging heterogeneous multi- and many-core architectures with graphical processing units (GPUs). Ensemble molecular dynamics algorithms that use multiple, cooperative trajectories to not only improve the exploration of the phase space but also expand the scalability on emerging extreme-scale platforms are also being developed.