Bioinformatics and Computational Biology
url: http://cbcb.umd.edu/
The Center for Bioinformatics and Computational Biology (CBCB) at the University of Maryland develops computational methods that can be used to explore the genomes of all living things, ranging from viruses to bacteria to humans. Scientists in CBCB are using these methods to discover new genes, to better understand the evolution of species, to help fight infectious diseases, and to understand the workings of the cell. They also design algorithms to take advantage of the latest biotechnology breakthroughs, including new high-throughput sequencing machines and DNA chips. CBCB is a highly interdisciplinary center whose faculty, postdoctoral fellows, and students come from Computer Science, Cell Biology and Molecular Genetics, Mathematics, Physics, Biology and other departments. All of the members of the center share a common interest in gaining a better understanding of how life works. The Center collaborates with numerous research institutes in the region, around the country, and throughout the world.
CBCB has research projects focused in the following areas:
- Genome sequence assembly
- Computational gene finding
- Genome sequencing projects
- Comparative genomics
- Computational proteomics
- Metagenomics and environmental sequencing
- Molecular evolution and phylogenetics
- RNA splicing
These and other projects are described in more detail at http://cbcb.umd.edu/research/. CBCB research is funded by grants from the National Institutes of Health, the National Science Foundation, the Department of Homeland Security, and the Naval Medical Research Center.
Faculty in the Center, including those from the life sciences, are listed at http://cbcb.umd.edu/people/, which also includes links to each faculty member's home pages. CBCB faculty in Computer Science include:
Steven Salzberg is the Director of CBCB. His research interests center on genome assembly, computational gene finding, and comparative genomics, particularly the use of comparisons to better understand the short-term evolution of infectious bacteria and viruses. His recent genomics projects include the Influenza Genome Sequencing Project, an international collaboration to sequence numerous influenza viruses, and projects targeting several bacterial species. His computational projects include the bacterial gene finder Glimmer, the AMOS open source assembly project, and the development of systems for eukaryotic genome annotation. He is a strong proponent of free, open-source software, of open-access publishing, and of the rapid and unrestricted release of genome sequence data, and his editorials on these topics can be found on his home page. Details on his research projects can be found on the CBCB website at http://cbcb.umd.edu/research/.
Art Delcher's research is in the area of designing algorithms to assemble and analyse biological sequence data, and in developing practical software tools that implement those algorithms. He is the original developer of the widely used GLIMMER microbial gene-finding system and of the MUMmer suite of programs for large-scale sequence comparison. He also was the author of several modules used in the Celera Assembler program that was used for the first shotgun sequence assemblies of the Drosophila and human genomes. His current work includes modifications to the assembler to better handle DNA sequences produced by newer technologies with a different error pattern than previous technologies, and in the development of rapid algorithms to compute distinguishing DNA signature sequences that can be used to identify biological threats. He also is interested in the evolutionary development of microbial genes and how that can be exploited to produce more accurate gene predictions within the GLIMMER system.
Carl Kingsford develops computational methods to explore the organization and bacterial genomes, the evolution of viral genomes, and the three-dimensional structure of proteins, among other topics in computational biology. Recently, he has created the TransTermHP program (http://transterm.cbcb.umd.edu/) for predicting genomic signals that cause transcription to stop in hundreds of bacterial species. He has also investigated (with Art Delcher) the phenomenon of overlapping genes in prokaryotic organisms and proposed evolutionary mechanisms by which these overlaps are created and eliminated. He has contributed to recent studies on the evolution of the human and avian influenza viruses, describing the westward movement of the H5N1 strain of influenza. To help with the determination of the shapes of proteins, he has developed methods to predict protein structure based on mathematical programming. He is interested in continuing to apply combinatorial optimization, statistical methods, machine learning, and other techniques to predict protein function and protein-protein interactions, to study the movement and evolution of influenza, and to answer other questions about viral, bacterial and eukaryotic genomes.
Mihaela Pertea's research interests include ab-initio and comparative approaches to gene finding in eukaryotic organisms, motif and regulatory sites prediction, as well as evolutionary and statistical modeling of biologically functional genomic DNA. Gene finding is a key step in understanding the genome of a species once it has been sequenced and assembled. Gene finding is especially difficult in eukaryotes, especially complex organisms like humans, because the genes and regulatory signals that control them are more complex and less well-understood than in prokaryotes. Our group has developed several programs for gene finding in eukaryotes, by taking both ab-initio and comparative approaches to solve this problem. In the ab-initio approach, genomic DNA sequence alone is searched for signs of protein coding genes. The comparative approach is based on the principle that there is evolutionary pressure for the conservation of genes in related species. Therefore it detects genes by incorporating evidence from the genomes of one or more additional organisms. In all approaches, most of the techniques used are statistical in nature, employing sophisticated models to estimate the probabilities of different gene features. Dr. Pertea's current work concentrates in obtaining more accurate gene structure predictions by combining the output of different gene prediction programs with extrinsic evidence such as protein sequence alignments, expressed sequence tag and cDNA alignments, or splice site predictions.
Mihai Pop's research addresses the challenges to genome assembly posed by new sequencing technologies as well as the sequencing of mixtures of organisms (metagenomics). He is the author of the standalone hierarchical scaffolding package Bambus and the leader of the AMOS project - an open-source effort to develop an infrastructure for developing genome assembly software. Dr. Pop was a key participant in a study to characterize the bacteria colonizing the human gut. This study, published in the journal Science, revealed a large level of diversity in this community and highlighted biochemical processes important to the interaction between the bacteria and their human host.
Chau-Wen Tseng conducts research in the area of high-performance computing for bioinformatic applications. Recent advances in molecular biology techniques such as automated DNA sequencing and DNA microarrays allow scientists to quickly gather huge amounts of gene sequence and expression data. Popular algorithms for analyzing bioinformatic data may be oversimplified due to concerns about limited processing power. Dr. Tseng's research investigates methods to exploit the rapidly increasing power of high performance computing architectures to improve the speed and/or quality of bioinformatic algorithms. Areas of current interest include parallel sequence alignment, EST clustering, high-throughput DNA to genome alignment, and protein identification via tandem mass spectrometry.
