| Computational Biology at the University of Virginia encompasses a variety of research techniques and perspectives that use computational and mathematical approaches to understand biological, biophysical, and physiological problems. Today, rapid genome sequencing methods, large-scale gene-expression analysis, and high-throughput structural genomics projects have dramatically increased the amounts of data available to build, and to test, new biological hypotheses. Advances in data acquisition and imaging technologies and computer power now make it possible to extract detailed structural and physiological details that were inaccessible a decade ago. Computational Biology Faculty at the University of Virginia are developing new approaches to analysing large datasets, exploring fundamental questions in protein evolution and structural biology, and developing better models for changes in human disease.
Computational Genome Biology
Every time a genome is fully sequenced, scientists get a complete list of every gene, and thus every protein, an organism can make. The announcement of the completed human genome sequence, together with fully completed genome sequences for the fruit-fly, yeast, and more than one hundred bacteria, allow us to to pursue, for the first time, experimental strategies that look at physiological processes at the "genome-scale", with the assurance that the role of every gene in an organism can be examined during critical physiological changes, without imposing an investigator's bias.
While the explosion of genome information can change fundamentally our research strategies, it also presents the potential for information overload. Most investigators find it challenging just to remember the functions of the 1% of the 60,000 human proteins that they recognize. Genome-scale technologies not only tell us the structure of these proteins, they can also determine when and where each protein is expressed. Understanding a physiological whole from the sums of thousands of parts, each with distinct spacial and temporal patterns of existence, can be overwhelming.
The promise of bioinformatics and computational genomics is both to develop tools that allow "classically" trained investigators to sift through masses of genome data to understand better their favorite problems, and to develop fundamentally new research strategies to address fundamental problems in biology and medicine. Investigators at the University of Virginia are building improved algorithms, statistical methods, and databases to better exploit genome-scale sequence and expression data. UVa, as part of the Virginia Bioinformatics Consortium, has enhanced the Genex RNA expression database, which is currently being used to store large scale Affymetrix GeneChip(tm) results. Reseachers are also developing a detailed Cell Migration knowledgebase. ( Kretsinger, Pearson, Skalak )
Computational Structural Biology
The success of the human genome project has spawned interest structural genomics - the determination of hundreds of truly novel protein structures each year, with the goal of determining by experimental methods a representative set of macromolecular structures, including medically important human proteins and proteins from important pathogens and model organisms. Currently, protein structure determination is limited by both experimental bottlenecks (protein purification and crystallization) and computational barriers - data acquisition, transformation, and actual polypeptide chain tracing are computationally demanding. Researchers at UVa are developing automated strategies for protein structure determination, with the goal of dramatically reducing the time and effort required to produce a finished protein structure once appropriate crystals are
available. Other investigators are exploring computational approaches for predicting protein structure, exploiting the rapidly grown protein structure data bases.
While X-ray structure determination has provided the most detailed models of biological macromolecules, many biological structures are too large for X-ray crystallography. In many cases, however, structural models of large macromolecular assemblies can be computed from hundreds of electron microscope images. Researchers at UVa examine the function of these assemblies using new computational approaches. ( Egelman, Minor )
Mathematical Models of Physiological Change
Simple mathematical descriptions provide the foundations for many physical, chemical, and biological phenomena. While every student has heard at least one fundamental formula of physics (E=mc2, F=ma), it is widely thought that the complexity and variety of living systems preclude accurate mathematical models. However, mathematics has provided accurate models of critical physiological functions, from oxygen binding to hemoglobin to the electrical processes underlying a heart beat. As data acquistion advances and genome projects allow researchers to monitor physiological processes in greater and greater detail, it has become practical to develop accurate predictive models for more complex physiological processes, and computational biologists are developing more sophisticated data analysis methods, and mathematical models, that seek to specifically model physiological changes associated with diabetes and cardiovascular disease. (Johnson, Lee, Skalak )
For a list of faculty currently conducting research in this area, click here
|
