Auflistung nach:
Auflistung German Conference on Bioinformatics nach Erscheinungsdatum
1 - 10 von 156
Treffer pro Seite
Sortieroptionen
- KonferenzbeitragGoPubMed: ontology-based literature search applied to gene ontology and pubmed(German Conference on Bioinformatics 2004, GCB 2004, 2004) Delfs, Ralph; Doms, Andreas; Kozlenkov, Alexander; Schroeder, MichaelThe biomedical literature grows at a tremendous rate, so that finding the relevant literature is becoming more and more difficult. To address this problem we introduce ontology-based literature search, which structures search results thorugh the categories of an ontology. We develop and implement GoP- ubMed, which submits keywords to PubMed, extracts GeneOntology-terms from the retrieved abstracts, and presents the relevant sub-ontology for browsing. For GoPubMed we develop a novel term extraction algorithm and evaluate its performance. GoPubMed is available at www.gopubmed.org
- KonferenzbeitragIntegrating public databases into an existing protein visualization and modeling program – BRAGI(German Conference on Bioinformatics 2004, GCB 2004, 2004) Dieterich, Guido; Kvesic, Marsel; Schomburg, Dietmar; Heinz, Dirk W.; Reichelt, JoachimBRAGI offers an efficient visual access to sequence alignment information, 3D alignments and annotated protein structure-function correlations. As a new feature we have mapped information from SWISS-PROT and InterPro to individual entries of the PDB. 3D structural alignments from DALI database were converted to XML files for easy access in BRAGI. BRAGI provides interactive access to NCBI-Blast and the DALI server. Linking and visualizing different types of information hopefully allow the structure function of proteins to be appreciated more intuitively.
- KonferenzbeitragInherent size and complexity constraints in prokaryote gene regulatory networks(German Conference on Bioinformatics 2004, GCB 2004, 2004) Gagen, MichaelThere are no fully scalable technologies, including in biology. Whatever the technology – oxygen uptake by lungs rather than through the skin, bones to support weight on land, regulatory technologies within cells or between cells in multicellular organisms – the adoption of any one technology to solve one set of limitations both potentiates novel behaviours and at the same time imposes additional constraints and limitations. The gene regulatory networks of prokaryotes (single-cell life) will also be size and complexity constrained by the adopted regulatory technology. Here, we generalize network growth models in response to recent comparative genomics results showing that the number of prokaryote gene regulators scales quadratically with gene number. Our generalized models then predict an upper limit to the size and complexity of prokaryote genomes which is very close to that observed in the evolutionary record. An immediate implication is that the protein-only regulatory networks of prokaryotes cannot explain the much larger and more complex genomes of multicellular eukaryotes. This conclusion is consistent with recent exciting findings of novel RNA regulatory mechanisms in eukaryotes. We compare these results with two alternate models proposing that regulator numbers scale linearly with gene number (though no plausible physical mechanism underlies such models). To make this comparison as accessible as possible, all three models have been encoded into a JAVA applet showing actual comparative genomics results, simulated statistics on the numbers of inbound and outbound regulatory links, the sizes of the largest fully connected island, the predicted maximum gene count for prokaryotes, and so on.
- KonferenzbeitragComparison of centralities for biological networks(German Conference on Bioinformatics 2004, GCB 2004, 2004) Koschützki, Dirk; Schreiber, FalkThe analysis of biological networks involves the evaluation of the vertices within the connection structure of the network. To support this analysis we discuss five centrality measures and demonstrate their applicability on two example networks, a protein-protein-interaction network and a transcriptional regulation network. We show that all five centrality measures result in different valuations of the vertices and that for the analysis of biological networks all five measures are of interest.
- KonferenzbeitragInformation flux through proteins(German Conference on Bioinformatics 2004, GCB 2004, 2004) Vriend, Gert
- KonferenzbeitragSuperhelically destabilized sites in the E. Coli genome: Implications for promoter prediction in prokaryotes(German Conference on Bioinformatics 2004, GCB 2004, 2004) Wang, Huiquan; Benham, Craig J.Stress-induced DNA duplex destabilization (SIDD) analysis exploits the known structural and energetic properties of DNA to predict which sites become susceptible to strand separation under superhelical stress. Experimental results show that this analysis is quantitatively accurate in predicting destabilized sites that occur in transcriptional regulatory regions, matrix/scaffold attachment sites and replication origins. Here we report the results of a SIDD analysis of the complete E. coli genome, performed using a new algorithm specific for long genomic DNA sequences. Our results demonstrate that less than 7% of the E. coli genome has the propensity to become highly destabilized at physiological superhelical densities. Those SIDD sites with high destabilization potential are statistically significantly associated with divergent and tandem intergenic regions, but not with convergent intergenic or coding regions. More than 80% of the intergenic regions containing experimentally characterized promoters are found to overlap these SIDD sites. Strong SIDD sites are highly enriched in the 5' upstream regions of genes regulating stress responses in E. coli, suggesting a possible functional role in their regulation. We discuss the possibility of using SIDD properties for promoter prediction, and potential roles of predicted SIDD sites in transcriptional regulation.
- KonferenzbeitragGenlight: An interactive system for high-throughput sequence analysis and comparative genomics(German Conference on Bioinformatics 2004, GCB 2004, 2004) Beckstette, Michael; Sczyrba, Alexander; Selzer, Paul M.
- KonferenzbeitragPhylogenetic networks with constrained and unconstrained recombination(German Conference on Bioinformatics 2004, GCB 2004, 2004) Gusfield, DanA phylogenetic network is a generalization of a phylogenetic tree, allowing structural properties that are not tree-like. With the growth of genomic data, much of which does not fit ideal tree models, and the increasing appreciation of the genomic role of such phenomena as recombination, recurrent and back mutation, horizontal gene transfer, gene conversion, and mobile genetic elements, there is greater need to understand the algorithmics and combinatorics of phylogenetic networks. Wang et al. studied the problem of constructing a phylogenetic network for a set of nn binary sequences derived from a known ancestral sequence, when each site in the sequence can change state at most once in the network, and recombination between sequences is allowed. They showed that the problem of finding a phylogenetic network that minimizes the number of recombination events is NP-hard, but gave a polynomial-time algorithm (O(nm + n4)-time, for n sequences of length m each) that was intended to determine whether the sequences could be derived on a phylogenetic network in which the recombination cycles are node disjoint. We call such a network a “galled-tree”. The paper by Wang et al. is seminal in defining this problem and asserting that it has an efficient solution. Unfortunately, the algorithm given there is incomplete and only constitutes a sufficient, but not a necessary, test for the existence of a galled-tree for the data. In this talk we do several things. By more deeply analyzing the combinatorial constraints on galled-trees, we obtain an algorithm that runs in O(nm + n3)-time and is guaranteed to be both a necessary and sufficient test for the existence of a galledtree for the data. We show how to relax the assumption that we know the ancestral sequence. We show that when there is a galled-tree, the algorithm constructs a “reduced” galled-tree that minimizes the number of mutations occurring on recombination cycles. We prove that the algorithm produces a reduced galled-tree that minimizes the number of recombinations needed for the data, over all possible phylogenetic networks for the data, even if multiple crossovers, or different ancestral sequences are allowed. Hence, when a galled-tree exists, the problem of minimizing the number of recombinations can be solved efficiently. The effect is that the galled-tree is a phylogenetic network that explains the input sequences with the “littlest deviation” from a true tree model. We show that the reduced galled-tree is “nearly-unique”, but when it is not unique, the algorithm also obtains a count of the number of galled-trees that exist for the input data, and can create these in linear time for each one, starting from the canonical galled-tree. Finally, we consider phylogenetic networks where the recombination networks are not constrained in any way. We discuss new, efficiently computed lower bounds on the number of recombination events needed. Joint work with Satish Eddhu, Chuck Langley, and Dean Hickerson
- Editiertes Buch
- KonferenzbeitragMultiple sequence alignment with user-defined constraints(German Conference on Bioinformatics 2004, GCB 2004, 2004) Morgenstern, Burkhard; Prohaska, Sonja J.; Werner, Nadine; Weyer-Menkhoff, Jan; Schneider, Isabelle; Subramanian, Amarendran R.; Stadler, Peter F.In many situations, automated multi-alignment programs are not able to correctly align families of nucleic acid or protein sequences. Distantly related sequences are generally hard to align, and sequence duplications may present additional challenges to standard alignment algorithms. In the present paper, we describe a semiautomatic approach to multiple sequence alignment. The user can specify parts of the sequences that are thought to be related to each other; our software program will use these sites as anchor points and create a multiple alignment respecting these userdefined constraints. By using functionally, structurally or evolutionarily related positions of the input sequences as anchor points, the proposed method can produce alignments that are biologically more meaningful than alignments produced by fully automated procedures. We apply our approach to genomic sequences around the Hox gene cluster. As a by-product, we obtain useful insights for the further development of alignment algorithms. The described alignment approach has been integrated into the tracker software system.