|
|
BCB 690 Spring 2005
 |
Friday 1:10-2:00 PM 0296 Town Engineering
BCB 690 Student Seminar Jing Ding, Dept. of Electrical and Computer Engineering; (co-major professors: Berleant and Schnable)
Abstract: "Extracting Biochemical Interactions from MEDLINE Using a Link Grammar Parser" Many natural language processing approaches at various complexity levels have been reported for extracting biochemical interactions from MEDLINE. While some algorithms using simple template matching are unable to deal with the complex syntactic structures, others exploiting sophisticated parsing techniques are hindered by greater computational cost. This study investigates link grammar parsing for extracting biochemical interactions. Link grammar parsing can handle many syntactic structures and is computationally relatively efficient. We experimented on a sample MEDLINE corpus. Although the parser was originally developed for conversational English and made many mistakes in parsing sentences from the biochemical domain, it nevertheless achieved better overall performance than a co-occurrence-only method. Customizing the parser for the biomedical domain is expected to improve its performance further. Reference: Ding J, Berleant D, Xu J, and Fulmer AW (2003) Extract biochemical interactions from MEDLINE abstracts using a link-grammar parser. Proceedings of the Fifteenth IEEE Conference on Tools with Artificial Intelligence (ICTAI 2003), Nov. 3-5, Sacramento, pp. 467-471.
Aimin Yan, Dept. of Biochemistry, Biophysics and Molecular Biology (co-major professors: Robert Jernigan and Zhijun Wu) Abstract: "How do sidechains orient globally in protein structures?" An angle is defined to serve as a metric for global sidechain orientations, which reflects the orientation of the sidechain relative to the radial vector from the center of the protein to this amino acid. The sidechain orientations of buried residues exhibit characteristically different orientations than do exposed residues, in both monomeric and dimeric structures. Overall, buried sidechains point mostly inward; whereas surface sidechains tend to point outward from the surface. This difference in behavior also correlates well with a residue’shydrophobicity; so a global sidechain orientation can be viewed as a direct structural manifestation of hydrophobicity. In the case of interfacial residues between subunits, there are statistically significant differences between exposed residues and interface residues for ALA, ARG, ASN, ASP, GLU, HIS, LYS, THR, VAL, MET, PRO, and overall the interface residues have an increased tendency to point inward. Presumably these substantial differences in orientations of sidechains could be a manifestation of intermolecular interactions. Also there are some variations of the angle for different residue types in different solvent accessible layers, which could be due to the local environments of residues. Raul Piaggio, Dept. of Computer Science (major professor, Oliver Eulenstein, co-major professor, Drena Dobbs) Abstract: "Phylogenetic Compression" I will present a recently developed method to compress DNA sequences: by using phylogenetic trees. This approach provides two advantages. On the one hand, it achieves very high compression rates on homologous sequences. On the other hand, it provides a framework in which to decide when more than one phylogenetic tree may be a better explanation of the unrelying evolutionary process than just one total-evidence tree.
Friday, February 4 at 1:10 p.m. Jae-Hyung Lee, Dept. of Genetics, Development and Cell Biology (Major Professor: Drena Dobbs; co-major professor: Kai-Ming Ho) Abstract: Designing the model of the EIAV Rev protein and gaining the experimental evidences of the model The Rev protein of EIAV is an essential regulatory protein that facilitates export of incompletely spliced viral RNAs from the nucleus to the cytoplasm. Discrete functional domains of Rev mediate protein-RNA and protein-protein interactions that are required for nuclear import, RNA binding, multimerization, and nuclear export. Because the Rev protein has an aggregation property, there is no high-dimensional structural information of EIAV Rev. To get computational model of the EIAV Rev protein, we used a structural threading approach. The predicted exon 2 of EIAV Rev is an alpha-helical protein which has 6 helix bundles. Verifying the computational model, various experimental methods were performed. The result of one of biophysical experiments, circular dichroism shows that the EIAV Rev protein is an alpha-helical protein. Also we confirmed agreements with computational model in the residues and motifs of EIAV Rev which are important for function and structure using mutational and functional assays. Together with the results, the computational prediction of protein structure is a powerful tool for gaining insights the relationship between the protein function and structure. Also more detail verifications of the model may help us to feedback on the existing model and develop better procedures for getting accurate predictions. Also February 4 Michael Terribilini, Dept. of Genetics, Development and Cell Biology (Major Professor: Drena Dobbs; co-major professor Vasant Honavar) Abstract: Computational prediction of RNA-binding sites in proteins based on amino acid sequence Protein-RNA interactions are vitally important in a wide range of biological processes, including regulation of gene expression, protein synthesis, and replication and assembly of many viruses. The ability to reliably predict which residues of a protein directly contribute to RNA binding - without the requirement for structural information regarding either the protein or the protein-RNA complex - would significantly enhance our understanding of how proteins recognize RNA and potentially generate new strategies for clinical intervention in both genetic and infectious diseases. We have developed a machine learning approach for predicting which amino acids of an RNA-binding protein are involved in protein-RNA interactions, using only the protein sequence as input. Interfaces from known protein-RNA complexes in the PDB were extracted to generate a non-redundant set of 109 protein chains from which a total of 3581 "interface" and 21,537 "non-interface" residues were obtained. Using this dataset, a Naïve Bayes classifier was trained to predict which residues in a given RNA binding protein are located at the protein-RNA interface. The classifier identifies interface residues with < 86% overall accuracy and correlation coefficient of > 0.35. Classifiers can be calibrated to increase the specificity of interface residue prediction for specific functional classes of RNA-binding proteins. To our knowledge, this simple approach provides the best available sequence-based prediction of protein-RNA interaction sites. It should be valuable both for hypothesis-driven investigations of specific RNA-binding proteins and for discovery-based large scale functional genomics efforts.
Friday, February 11 at 1:10 p.m. Haitao Cheng, Dept. of Computer Science (Major Professor: Dr. Robert Jernigan; co-major professor: Dr. Dimitris Margaritis) Title: Identifying structure fragments by sequence alignments to predict protein secondary structures Abstract: A new method of predicting protein secondary structure from amino acid sequence has been developed. The method is based on multiple sequence alignment of the query sequence with all other sequences with known structure from the Protein Data Bank (PDB) by using BLAST. The fragments of the alignments belonging to proteins from the PBD are then used for further analysis. We have studied various schemes of assigning weights for matching segments and calculated normalized scores to predict one of the three secondary structures: α-helix, β-sheet, or coil. We applied several artificial intelligence techniques: decision trees (DT), neural networks (NN) and support vector machines (SVM) to improve the accuracy of predictions and found that SVM gave the best performance. Preliminary data show that combining the fragment mining approach with GOR V (Kloczkowski et al, 2002 Proteins 49, 154- 166) for regions of low sequence similarity improves the prediction accuracy. Also Friday, February 11 LaRon M. Hughes, Dept. of Electrical and Computer Engineering (Major Professor: Dr. Daniel Berleant)
Title: PathBinderH: a Tool for Linnaean Taxonomy-Aware Literature Searches Abstract: PathBinderH is a Web-served text mining tool that allows users to search PubMed (including MEDLINE) to identify abstracts that contain user-specified terms co- occurring in the same sentence. Unlike standard tools that allow users to identify scientific abstracts containing one or more query terms, PathBinderH allows abstracts to be included or excluded from a search based on given plant taxa. This enables (1) filtering out abstracts dealing with species of less interest while retrieving sentences from abstracts about any of the potentially many species within the specified taxa, and (2) identifying abstracts that are more likely to prove relevant to a user than abstracts that contain the query terms but in different sentences, because the query terms are more likely to be coordinated in their use. Preliminary results have shown that for a given query the results of a PubMed and PathBinderH query render a perhaps surprisingly small overlap in returned items. We present an empirical investigation of this important problem. We show that there is a large and significant difference in the query results of the two tools. Furthermore, the results show that PathBinderH returns results that are just as relevant to researchers due to its taxonomic and sentence- based retrieval methods.
Kenton Weber, Dept. of Computer Science (Major Professor H. H. Chou) Abstract: Biological Data Extraction Using VECT VECT is a package designed to allow biological researchers to create programs that are able to extract specific types of data from information rich documents from sources such as (although not limited to) GenBank. VECT allows the creation of such a program without requiring the biologist to have knowledge of programming. This is accomplished by providing a graphical interface that allows the biologist, using point and click methods, to give directions on what type of information is required and how to recognize it; what operations to perform on it and how the results are to be formatted. A Perl program is then created that will perform the required tasks. An example use of VECT might be to create a program that will parse the GenBank report of the an entire microbial genome and create an output file that gives the gene id number, the predicted function, and the protein sequence of each predicted gene, with the results presented in FASTA format. Using VECT, such a program could be created in a few minutes and then used on any genome for which a GenBank file existed. My demonstration will give an introduction to VECT and show how it can be used for biological data extraction. VECT is freely downloadable and available at: http://www.complex.iastate.edu/download/Vect/index.html VECT was created in the laboratory of Dr. Hui-Hsien Chou Complex Computation Laboratory, Iowa State University Also, February 18, 2005 Michael E. Sparks, Department of Genetics, Development and Cell Biology (Major Prof: Volker Brendel Co-Major Prof: Thomas Peterson) Abstract: We have investigated the compositional features of noncanonical spliceosomal introns--specifically those with GC donor and AG acceptor termini--in a diverse range of multicellular eukaryotic taxa. We present results indicating that these are distinctive from canonical U12-type spliceosomal introns in that, while signals at acceptor termini are similar, GC donor sites are generally more conserved than their GT couterparts. This observation was consistent across all species studied. As well, we describe the implementation of gene prediction software that specifically addresses scoring of non-canonical GC-AG splice sites that occur in gene structures. The performance of these models will be presented.
Friday, Feb 25 Jennifer Quammen, First year rotation student Title: Digestive enzyme activities in anuran tadpoles under varying food conditions. Abstract: Aquatic tadpoles encounter foods that vary in terms of nutritional value and quality. Studying digestive enzymes of tadpoles gives insight into their capabilities to consume and digest differing foods available at unpredictable intervals. This study compared the digestive enzymatic activities in two tadpole species, the Wood frog (Rana sylvatica) and the American toad (Bufo americanus) exposed to changing diets throughout a 108 hour experimental period after a 12 hour acclimation period. The diets fluctuated from algae (carbohydrate), to shrimp (protein), and back to algae (carbohydrate) with equally divided feeding periods on each food. This investigation is the first to report evidence of digestive enzyme plasticity in larval anurans. We determined that the levels of amylase and trypsin activities varied between species and among time periods, and that both enzymes were more active in B. americanus than R. sylvatica. Pepsin had the lowest specific activities and was not significantly different between species. Lipase did not vary across time periods or between these species throughout the experiment. Trends in the levels of carbohydrase and protease activity varied with different foods consumed across the experimental periods, indicating that enzymatic plasticity is present for these tadpoles as diet shifts. Knowledge of tadpole digestive enzymes can help us decipher the physiological changes that must take place for these animals to reach metamorphosis, when encountering these unpredictable environmental food resources. Also Friday, Feb 25 Wuyan Zhang, Dept. of Statistics, (Co-major professors: Alicia Carriquiry & Jack Dekkers) "pQTL transcriptome mapping: an efficient method to integrate QTL mapping and gene expression analysis to discover the genetic basis of complex traits" Abstract: High-through-put microarray gene expression analysis has enabled genome-wide identification of genes whose expression is affected by specified factors or treatments. Of particular interest is the use of this technology to study the genetic architecture of complex traits, which includes many traits of interest. This requires combining the power of quantitative trait locus (QTL) mapping with gene expression analysis through what is called genetical genomics or transcriptome mapping. Transcriptome mapping enables genome-wide identification of positional candidate genes for QTL and genes involved in metabolic path-ways for the phenotypic trait. Current methods for transcriptome mapping require conducting individual microarray assays on a large (>500) number of animals for sufficient statistical power, which is prohibitively expensive for most labs. We propose a more efficient and directed approach for combining the power of QTL mapping and expression analysis, which can reduce the number of microassays required by 100-fold or more and targets the generation of expression data that is relevant to the phenotypic traits of interest.
Friday, March 4 Lixia Jin, Department of Biochemistry, Biophysics and Molecular Biology (Major professors: Dr. Robert Jernigan and Zhijun Wu) Title: GNM/ANM model and its applications Abstract: GNM (Gaussian Network Model) and ANM(Anisotropic Network Model) were proposed to simulate the interactions between residues for a given protein structure. In these two models, a protein structure is described using an elastic network, in which the nodes are the residues and the links are the interactions between residues. It is assumed that the fluctuations of the residues obey Gaussian distribution. GNM is non-directional (isotropic), it can reproduce the B-factors curves of crystal structure quite precisely. ANM is an improvement of GNM by considering the difference of three different directions. It can simulation the motions of the molecules. They can be applied to study the dynamics and functions of large-scale molecules. References: Ivet Bahar, Ali Rana Atigan and Burak Erman, Folding & Design, 1997, 2:173-181 A.R. Atilgan, S.R. Durell, R.L.Jernigan, M.N.Demirel, O.Keskin, and I. Bahar, Biophysical Journal, 2001, 80:505-515 Also Friday, March 4 Sachet Shukla, Dept. of Plant Pathology (Major professor: W. Allen Miller) Title: Identifying long-distance interactions in RNA molecules Abstract: Long distance tertiary interactions in viral RNA genomes have been known to play an important role in various stages of the viral life cycle. Current methods for the prediction of such interactions are typically very computationally intensive and tend to compromise the accuracy of stronger secondary interactions. We have tried to address this problem by post-processing results from standard RNA secondary structure tools. We believe that such consensus predictions supplemented by phylogentic information, sorted by reliability score(s) would provide good starting points for further analysis of tertiary interactions. The eventual goal is to apply these programs on viral genomes/other RNA molecules and identify candidates for experimental validation. Initial results for the BYDV genome and future directions will be discussed.
Friday, March 11 Timothy Alcon, First year rotation student Title: Proteomic Study of Retinal Development Abstract: There is currently a great deal of interest in stem cells and their possible theraputic uses. Retinal progenitor cells are multipotent, rather than pluripotent, but it is hoped that ways may be discovered for exercising sufficient control over their differentiation that they can be used to treat certain diseases that cause blindness, such as macular degeneration. We would be better able to do so if we had better information about the signaling and gene regulatory networks that govern the behavior of retinal progenitor cells. One of the current questions in proteomics is how best to infer the structure of such networks from gene expression data. I will not be presenting any new results, but merely giving an overview of the status quo.
Friday, March 25 Jeffry Sander, First year rotation student Title: Protein identification in mycoplasma hyopneumoniae Abstract: Protein sequences are being identified at an explosively increasing rate. Sequence identification alone provides little benefit. Early steps of understanding proteinís biological roles lie in structure/function prediction. Employing existing programs and understanding differences between their methods provide more interesting results than the over used and often only used BLAST. Matthew Wilkerson, First year rotation student Title: Community Gene Annotation Abstract: Investigating increasingly more complex genetic and biological questions requires a solid foundation of the topic’s base elements. Genome scale studies require gene sets, an organism’s collection of genes, as accurate and as complete as possible. The process of specifying the gene set for an organism, referred to as nucleotide level genome annotation, consists of defining the locations and structures of genes and is initially completed in genome projects by automatic gene prediction programs. While this is a good first step in genome annotation, these programs, when used individually or collectively, are not completely accurate and ignore some classes and aspects of genes. Manual annotation has had documented success in a group jamboree format. This method of annotation is not limited by the pre-defined rules of the automatic prediction programs, and has the potential to provide a higher quality of annotations. Here, I will present the User Contributed Annotation System, a web based gene annotation tool that facilitates an online community. It is available for Rice, Arabidopsis, and Maize.
OsGDB
AtGDB
PlantGDB maize GSS
Gene Annotation Tutorial
Community Annotation
Friday, April 1 Yong Huang, First year rotation student Title: A phylogenetic analysis of miRNA families Abstract: MiRNAs are short (~ 22nt) non-coding RNAs that have important regulatory roles in development and growth. In this study, we used a bootstrap value based method to classify miRNAs in to families. We studied the genomic arrangement of these miRNA families. We found that members of an miRNA family did not cluster in general. However, in multiple cases, miRNAs from different families formed similar clusters in different parts of the genome. These finding may gave us some insight into the evolution and function of miRNAs. Also Friday, April 1 Lei Yang, Dept. of Biochemistry, Biophysics and Molecular Biology (Major professor: Dr. Robert Jernigan) Title: Simulation of Conformational Transitions in Proteins by Elastic Network Model A computationally efficient method is used to simulate the transitions of a protein between its two conformations. The method is based on a coarse-grained elastic network model in which contact interactions between spatially proximal residues of the protein are modelled with Gaussian potentials. The computational speed depends on the cutoff distance to define the interactions between the neighborhood residues. An application to large-scale motion is conducted using Calmodulin as an example.
Friday, April 8 Jennifer Deitloff, Dept. of Ecology, Evolution and Organismal Biology (Major Professor: Dean Adams) Title: Morphological variation in populations of Plethodon cinereus and P. electromorphus Abstract: Theoretical evolutionary ecology generates testable predictions concerning the effects of species interactions on ecological, behavioral, and morphological traits. These predictions can then be examined by comparing the characteristics of interacting species in regions where they co-occur (sympatry) to regions where each occurs separately (allopatry). When patterns of morphology, resource use, or behavioral traits covary in a manner consistent with theoretical expectations, one can treat this as primary evidence that a particular interaction is an important force in that ecological community. For example, a pattern of sympatric morphological divergence (character displacement) correlated with sympatric divergence of resource use is often treated as evidence of exploitative competition. In the genus Plethodon, morphological changes associated with exploitative food use, and behavioral interference, have been identified in systems with narrow contact zones. Less understood however is what occurs in larger sympatric regions between species. Here we examined morphological variation among allopatric and sympatric populations of Plethodon cinereus and P. electromorphus in Ohio. The sympatric region is significantly larger than those previously studied in other Plethodon communities (4X larger). We used landmark-based geometric morphometric methods to quantify head morphology, and examined patterns of morphological variation across allopatric and sympatric zones. Implications for competition and possible character displacement are discussed. Erin Myers, Department of Ecology, Evolution and Organismal Biology (Major professors: Drs. Fredric Janzen and Dean Adams) Title: Quantitative Genetics Of Plastron Shape In Slider Turtles (Trachemys Scripta) Abstract: Shape variation is widespread in nature and embodies both a response to and a source for evolution and natural selection. To detect patterns of shape evolution, one must assess the quantitative genetic underpinnings of shape variation as well as the selective environment that the organisms have experienced. Here we used geometric morphometrics to assess variation in plastron shell shape in 1314 neonatal slider turtles (Trachemys scripta) from 162 clutches of lab-incubated eggs from two nesting areas. Multivariate analysis of variance indicated that nesting area has a limited role in describing plastron shape variation among clutches, while differences between individual clutches were highly significant, suggesting a prominent clutch effect. The covariation between plastron shape and several possible maternal effect variables (yolk hormone levels and egg dimensions) was assessed for a subset of clutches, and found to be negligible. We subsequently employed two recently proposed methods for estimating heritability from shape variables, and generalized a univariate approach to accommodate unequal sample sizes. Univariate estimates of shape heritability yielded large values for both nesting populations (h2 ≈ 0.86), and multivariate estimates of maximal additive heritability were also large for both nesting populations (h2 max ≈ 0.57). We also estimated the dominant trend in heritable shape change for each nesting population and found that the direction of shape evolution was not the same for the two sites. Therefore, while the magnitude of shape evolution was similar between nesting populations, the manner in which plastron shape is evolving is not. We conclude that the univariate approach for assessing quantitative genetic parameters from geometric morphometric data has limited utility, as it is unable to accurately describe how shape is evolving.
Friday, April 15 Fengli Fu, First-year Rotation student Title: Comparative analysis between sorghum and maize exon/intron strctures Abstract: The knowledge of gene model especially the comparison of gene models between different species will help the value-setting of parameters for gene finding program, such as twinscan. Therefore, this project made comparative analysis of sorghum and maize exon/intron structures. Gene identification program GeneSeqer is used in the prediction of the exons and introns. The input data are SAMIs (Sorghum Assembly genoMic Islands) release 1.0 and SbGI (Sorghum biocolor Gene Index) release 8.0 downloaded from TIGR. The comparative analysis between sorghum and maize exon and intron lengths showed that there is no significant difference between sorghum and maize gene model. Du Pan, Department of Electrical and Computer Engineering (Major Professors: Drs. Julie Dickerson and Eve Wurtele) Title: Genetic Network Inference Based on Time Series Expression Profiles Abstract: Time series expression profiles provide dynamic information for inferring gene regulatory relationships. One of the major challenges of genetic network inference is to tell direct interactions from indirect interactions. Time correlation can estimate the time delay and edge direction. A new algorithm combines time correlation with partial correlation and d-separation theory to tell the direct and indirect interactions. The proposed algorithm was evaluated using simulated data and real yeast cell cycle data. The simulation evaluated the performance of the algorithm under different parameter settings, including different profile lengths, different noise levels, without using time delay and edge direction information and etc. Using the yeast cell cycle data, the algorithm successfully identified the yeast cell cycle development stages, cell cycle loop, negative feedback loops and some interesting uncharacterized ORFs and provided hypothesis for further biological studies.
Friday, April 22 Rajakumar Sankula, First-year Rotation student Title: Evolutionary Information : Protein Context Abstract: Most proteins have been formed by gene duplication, recombination, and divergence. In the process, they resulted in a wealth of evolutionary information. The extent of utilization of this evolutionary information has been increased manifold in the post-genomic era. In the protein context, its application is quite ubiquitous. Besides the phylogeny reconstruction, various forms of evolutionary information such as conserved residues, protein co- evolution, heterotachy etc. are being applied in protein structure prediction, functional assignment, identification of binding sites, predicting interaction etc. This is an overview of some of the pertaining applications and potential questions. Mgavi E.Brathwaite, Department of Computer Science (Major Professors: Drs. Vasant Honavar and Heather Greenlee) Title: Snap25 & Early Development in the Retina Abstract: SNAP25 expression is significant during the early development of photoreceptors. Early data suggest that animals with a decrease in SNAP25 have an abnormal number of photoreceptors during early development. Five genes have been found that have a high correlation to Snap25 expression. This is an initial step to finding out what is going on during photoreceptor development.
URL:
Copyright© 2005, Iowa State University, all rights reserved. Please direct corrections, suggestions, and comments to bcb@iastate.edu. Last Modified: |