|
- Influenza Virus Database, Animesh Ray
- Evolution Experiments with Digital Organisms, Christoph Adami
- Structure and Origin of Functional Modules in Cellular Biology, Christoph Adami, Alpan Raval
- Neutral Protein Evolution, Alpan Raval
- Analysis and Modeling of Integrated Functional Networks, Alpan Raval, Animesh Ray
- Understanding the Basis of Robustness of Gene Regulatory Networks, Animesh Ray
- Analysis of Networks that Regulate Complex Developmental Pathways, Animesh Ray
- Epigenetic Processes, Animesh Ray
- Novel Representations of Protein Structure, Alpan Raval
- Topological Algorithms for Structural Genomics, Greg Dewey
- Inferring Gene Regulatory Networks in Cancer, Greg Dewey
Influenza Virus Database
Animesh Ray Professor Ray's laboratory has launched the first prototype of a Flu Virus integrative database (currently available at: http://aqua.sdsc.edu:8080/flu/). The database has been populated with over 2,800 complete flu virus genome sequences, isolated from over 13,000 sources representing over 1,000 geographical locations from 108 different countries. These complete genomes and additional gene fragment sequences were isolated from over 228 host organisms including humans. While the database contents represent fairly comprehensive coverage of publicly available virus sequences, the uniqueness of the database is defined by the search/query capabilities over the warehoused information. Specifically, there are now available phylogenetic trees that map the evolutionary relationships among these sequences. The database can be searched for evolutionary distances between any two specific sequences or, more importantly, a specified range of distance from a chosen sequence and all sequences that fall within this range of evolutionary distance can be recovered. Since evolution of new pandemic strains requires novel mutations to arise (by de novo mutation or natural recombination) in evolutionarily related viral sequences-and some of these mutations are now known or anticipated-good information about which sequences have the most opportunity to evolve to a potential pandemic variety can be obtained. The second advantage of the database is that it can be queried using the parameters of three-dimensional structures of specific viral proteins, such as the major viral antigens. This is very important because several structural parameters presumed to be important for lethality in the infamous 1918 flu virus have recently been described. The lab is working on further improvements to the query functionality. Evolution Experiments with Digital Organisms
Christoph Adami Digital organisms are self-replicating computer programs that compete with each other and evolve and adapt to a user-specified environment. These organisms (a form of non-terrestrial life) have been used extensively to study fundamental problems in evolutionary biology, from the evolution of complex features, to the problem of speciation and the evolution of sex. Many other problems can be addressed with digitals. Among the most interesting are the evolution of sexual recombination and its relation to parasitism, the origin of genome complexity via genome doubling events, and the origin of information. The Adami lab is studying how information can be generated spontaneously in a transition from a thermodynamical regime of random sequences to an information-dominated regime. Structure and Origin of Functional Modules in Cellular Biology
Christoph Adami
Alpan Raval Biological function is the complex consequence of the action of a large number of molecules that interact in many different ways. Elucidating the contribution of each molecule to a particular function would seem hopeless, had evolution not shaped the interaction of molecules in such a way that they form functional units, or building blocks, of the organisms function. These building blocks can be called modules, whose interactions, interconnections, and fault-tolerance can be investigated from a higher-level point of view, allowing for a synthetic view of biological systems. Professors Adami and Raval are taking an integrated computational and experimental approach to determine the extent of modularity in the gene-interaction network, the properties of modules, their contribution to robustness and fault-tolerance, their origin and evolution, and how existing and emerging criteria of module definition and function affect our ability to predict biological function. Neutral Protein Evolution
Alpan Raval Mutations that are neutral to selection dominate the substitution process over the course of evolution of genomes. Nevertheless, neutrally evolving populations display unexpected properties, including the emergence of mutational robustness and overdispersion of the molecular clock, or non-adherance to the predicted rate of occurrence. Dr. Raval's research involves studying the nature of the substitution process in neutral evolution with particular emphasis on the conditions that could lead to significant overdispersion of the molecular clock. He also works on the development of a biophysical framework for neutral protein evolution in which properties of evolving protein populations can be characterized in terms of biophysical measurements on the wild-type protein. Analysis and Modeling of Integrated Functional Networks
Alpan Raval
Animesh Ray Are there organizational principles in the network of interactions among a few thousand RNA molecules, proteins, and their binding sites on DNA in a cell? How do such complex networks of interaction function reliably in the noisy environment of a cell, and how did these networks evolve? These questions are at the heart of some of the research work being conducted by Professors Raval and Ray. Genome-scale biological networks of different types (for example, protein-protein interaction, DNA-protein interaction, genetic interaction, and metabolic networks) can be integrated and analyzed for the presence of novel topological motifs and for possible correlation of motif occurrence with sequence features, expression levels and functional significance. Professors Raval and Ray, in collaboration with colleagues at the University of California San Diego, are developing mathematical methods and software tools to mine and parse networks for different types of motifs. Preliminary results reveal the inter-dependence of various functional networks in yeast and relationships between gene function and network properties. For example, the physical organization of protein-protein interaction networks seems to reflect information about certain genetic interactions among the genes that encode these proteins. Predicted genetic interactions derived from computational models are being tested by laboratory experiments. The computational methods should provide new ways to identify targets of combination drug therapy in which each drug acting alone has little effect, but in combination have a significant effect on a targeted disease pathway. A public computational resource, available at: http://biologicalnetworks.net, has been developed by Professor Ray's laboratory in collaboration with colleagues at the University of California San Diego. Understanding the Basis of Robustness of Gene Regulatory Networks
Animesh Ray It has been argued that biological interaction networks are robust to many kinds of perturbations. The evolutionary mechanisms of such robustness properties are not easy to appreciate. For example, one mechanism of robustness (one that is incorporated by systems engineers) is redundancy. However, since evolution is thought to act upon the function of single genes, it is unclear how redundancy can be stable over evolution since there should be no selection pressure acting upon purely redundant functional units. To answer such questions Professor Ray's laboratory, in collaboration with colleagues at the University of Rochester Medical School and the University of Toronto, are mutating each essential gene of Saccharomyces cerevisiae. They then select for all possible genes in the genome which, if over-expressed, could potentially suppress the lethal effects of the original mutation. The idea is that each essential gene if mutated leaves 'holes' in the network of genes and proteins, but by over-expressing the communication traffic through other pathways, 'shunt pathways' could potentially be revealed that make life possible even in the presence of these lethal 'holes'. If a dense network of 'holes' and 'shunt pathways' can be obtained, the rules that underlie the readjustment of communication networks under conditions of potentially lethal perturbations can be studied. These studies have the potential to reduce the deleterious effects of drugs on normal cells while maximizing the lethal effects of the same drugs on diseased cells. Analysis of Networks that Regulate Complex Developmental Pathways
Animesh Ray Simple developmental decisions, such as sporulation in yeast triggered by a change in the food quality or the differentiation of a stem cell upon receiving a morphogen signal, appear to involve differential regulation of hundreds of genes. Reverse engineering of the network of regulation among these hundreds of genes and their products requires a combination of diverse genomic scale experimental techniques and mathematical modeling. Professor Ray's laboratory is investigating sporulation in yeast as one such model system. In collaboration with colleagues at the University of California San Diego and the Institute for Systems Biology, Professor Ray's lab has uncovered genome-wide oscillatory patterns of gene expression during the switch from normal growth to sporulation. Using microarray profiling of thousands of mRNAs and mathematical analysis normally used in signal processing in electrical engineering, they are probing the networks of genes and proteins, and their modification states, which underlie these dynamical properties of gene expression. These studies will ultimately help provide better understanding of the workings of normal and pathological human cells, and should enable insights for therapeutic intervention. Epigenetic Processes
Animesh Ray Although cells respond to environmental stimuli in many different ways, it is likely, nonetheless, that there are conserved themes in these processes. One aim of Professor Ray's laboratory is to understand how exposures to different nutritional conditions alter the profile of epigenetic signatures (proteins and their modification states that bind to DNA) acting upon the genome in yeast. Using genome wide chromatin immunoprecipitation techniques (ChIP-Chip), the lab is investigating the histone code associated with different chromosomal regions under different metabolic conditions. A further project in Professor Ray's laboratory investigates the epigenetic signatures associated with chromosome break repair and genetic recombination. Since it is increasingly apparent that many complex human diseases, such as most cancers (many of which involve faulty DNA damage repair), diabetes, obesity and numerous degenerative diseases, are due to abnormal epigenetic processes, understanding the epigenetic mechanisms behind adaptation to nutritional conditions and DNA damage could provide insights into human diseases as well. Professor Ray and his collaborator, Dr. Ranjan Perera, are also extending the analysis of epigenetic signatures to the study of melanoma with the goal of identifying specific biomarkers for this malignant condition. Novel Representations of Protein Structure
Alpan Raval Different representations of the three-dimensional structure of a protein have been traditionally used in computational biology in order to (a) compare and align structures of different proteins, and (b) discover correlations between sequence and structure in anticipation of predicting structural properties from sequence information alone. These representations include global ones like contact maps and distance matrices, and local ones like solvent accessibility and secondary structure. Professor Raval is using recent developments in network biology to create novel representations of protein structure. These representations will be used in structure alignment methods as well as structure prediction methods to define new classes of position-specific substitution matrices that capture global aspects of protein structure. These matrices could be integrated into a protein structure prediction approach. A related direction is using these global properties for structure-structure comparison. Topological Algorithms for Structural Genomics
Greg Dewey As the number of protein structures continues to grow, structure comparison techniques have become increasingly crucial bioinformatics tools. Because protein structures evolve more slowly than protein sequences, structure comparisons can be used to assess distant evolutionary relationships and common functions for pairs that do not have high sequence similarity. Structure alignment is also a central tool for protein classification and structural genomic initiatives. Despite the importance of structure comparison, a number of fundamental issues remain unresolved. Professor Dewey's group is developing new algorithms and scoring systems for structural genomics. These algorithms are based on using topological measures of protein structures. By encoding topological measures into an alphabet, it is possible to convert structure alignment problems into sequence alignment problems. The advantage of this approach is that mature and efficient algorithms previously developed for sequence bioinformatics can be used. This method can be used for rapid searches of structure databases and for multiple structure alignments. Inferring Gene Regulatory Networks in Cancer
Greg Dewey In collaboration with researchers at the Karmonos Cancer Institute in Detroit, Professor Dewey's lab is analyzing microarray time series data and inferring network structure from data on normal and breast tumor cell lines. Time series data for the response of normal and cancer cells to an environmental perturbation are analyzed using a simple linear response theory. Using this simple analysis, hierarchical networks are seen for both cancer and normal cells. These networks both show hub-like structures centered on the expression of just a few genes. The identity of the hubs differs between the tumor lines and normal cells, and the physiological significance of this difference remains to be elucidated. Comparative information using data from a number of different tumor cell lines will be invaluable in identifying genes that are differentially expressed. These genes are potentially new therapeutic targets.
|
