Terry W. Clark

Gene conversion in the rice genome

BackgroundGene conversion causes a non-reciprocal transfer of genetic information between similar sequences. Gene conversion can both homogenize genes and recruit point mutations thereby shaping the evolution of multigene families. In the rice genome, the large number of duplicated genes increases opportunities for gene conversion.ResultsTo characterize gene conversion in rice, we have defined 626 multigene families in which 377 gene conversions were detected using the GENECONV program. Over 60% of the conversions we detected were between chromosomes. We found that the inter-chromosomal conversions distributed between chromosome 1 and 5, 2 and 6, and 3 and 5 are more frequent than genome average (Z-test, P < 0.05). The frequencies of gene conversion on the same chromosome decreased with the physical distance between gene conversion partners. Ka/Ks analysis indicates that gene conversion is not tightly linked to natural selection in the rice genome. To assess the contribution of segmental duplication on gene conversion statistics, we determined locations of conversion partners with respect to inter-chromosomal segment duplication. The number of conversions associated with segmentation is less than ten percent. Pseudogenes in the rice genome with low similarity to Arabidopsis genes showed greater likelihood for gene conversion than those with high similarity to Arabidopsis genes. Functional annotations suggest that at least 14 multigene families related to disease or bacteria resistance were involved in conversion events.ConclusionThe evolution of gene families in the rice genome may have been accelerated by conversion with pseudogenes. Our analysis suggests a possible role for gene conversion in the evolution of pathogen-response genes.

Parallelizing molecular dynamics using spatial decomposition

Several algorithms have been used for parallel molecular dynamics, including the replicated algorithm and those based on spatial decompositions. The replicated algorithm stores the entire systems coordinates and forces at each processor, and therefore has a low overhead in maintaining the data distribution. Spatial decompositions distribute the data, providing better locality and scalability with respect to memory and computation. We present EULERGROMOS, a parallelization of the GROMOS molecular dynamics program which is based on a spatial decomposition. EULERGROMOS parallelizes all molecular dynamics phases, with most data structures using O(N/P) memory. The paper focuses on the structure of EULERGROMOS and analyses its performance using molecular systems of current interest in the molecular dynamics community. EULERGROMOS achieves performance increases with as few as twenty atoms per processor. We also compare EULERGROMOS with an earlier parallelization of GROMOS, UHGROMOS, which uses the replicated algorithm.<<ETX>>

Identification and characterization of insect-specific proteins by genome data analysis

BackgroundInsects constitute the vast majority of known species with their importance including biodiversity, agricultural, and human health concerns. It is likely that the successful adaptation of the Insecta clade depends on specific components in its proteome that give rise to specialized features. However, proteome determination is an intensive undertaking. Here we present results from a computational method that uses genome analysis to characterize insect and eukaryote proteomes as an approximation complementary to experimental approaches.ResultsHomologs in common to Drosophila melanogaster, Anopheles gambiae, Bombyx mori, Tribolium castaneum, and Apis mellifera were compared to the complete genomes of three non-insect eukaryotes (opisthokonts) Homo sapiens, Caenorhabditis elegans and Saccharomyces cerevisiae. This operation yielded 154 groups of orthologous proteins in Drosophila to be insect-specific homologs; 466 groups were determined to be common to eukaryotes (represented by three opisthokonts). ESTs from the hemimetabolous insect Locust migratoria were also considered in order to approximate their corresponding genes in the insect-specific homologs. Stress and stimulus response proteins were found to constitute a higher fraction in the insect-specific homologs than in the homologs common to eukaryotes.ConclusionThe significant representation of stress response and stimulus response proteins in proteins determined to be insect-specific, along with specific cuticle and pheromone/odorant binding proteins, suggest that communication and adaptation to environments may distinguish insect evolution relative to other eukaryotes. The tendency for low Ka/Ks ratios in the insect-specific protein set suggests purifying selection pressure. The generally larger number of paralogs in the insect-specific proteins may indicate adaptation to environment changes. Instances in our insect-specific protein set have been arrived at through experiments reported in the literature, supporting the accuracy of our approach.

Parallelization of a molecular dynamics non-bonded force algorithm for MIMD architecture

Abstract A method for parallelizing the non-bonded pair list generation and non-bonded force calculation algorithm for molecular dynamics is presented. Using the parallelism inherent to existing algorithms, it is possible, with minor modifications, to adapt the non-bonded routines to multiple-instruction, multiple-data (MIMD) computer architectures. This methodology has been applied to the molecular dynamics program GROMOS for the Stellar GS1000 Graphics Supercomputer. Aspects of the Stellar GS1000 architecture and programming environment are presented with attention to the performance of the molecular dynamics program. A speed enhancement factor of about 3 has been obtained relative to the serial execution of the program, which is close to the theoretical maximum factor of 4 for this machine. The overall speed enhancement factor increases to about 6 with the additional use of vectorization for a version that has been extensively rewritten to be more highly vectorizable than the standard code where gains from vectorization are slight. In the former case, the program executes at about 35% of the speed obtained on a single-processor Cray X-MP.

Pfortran: a parallel dialect of Fortran

IPfortran is a language designed to facilitate the programming of multi-process, data-parallel applications [BCS91]. Based on a message-passing paradigm, IPfortran extends Fortran with a small set of extensions. The duality of the send and receive operations is encapsulated in an infix operator (Section 2) or reducing functions (Section 3). With system-dependent message-passing pushed out of sight, code is streamlined and development time reduced. Errors in writing message-passing code are reduced by leaving the generation of the message-passing logic to the translator.

Evaluating parallel languages for molecular dynamics computations

The paper describes the practicalities of porting a basic molecular dynamics computation to a distributed-memory machine. In the process, it shows how program annotations can aid in parallelizing a moderately complex code. It also argues that algorithm replacement may be necessary in parallelization, a task which cannot be performed automatically. The paper closes with some results from a parallel GROMOS implementation.<<ETX>>

Bioinformatics process management: information flow via a computational journal

This paper presents the Bioinformatics Computational Journal (BCJ), a framework for conducting and managing computational experiments in bioinformatics and computational biology. These experiments often involve series of computations, data searches, filters, and annotations which can benefit from a structured environment. Systems to manage computational experiments exist, ranging from libraries with standard data models to elaborate schemes to chain together input and output between applications. Yet, although such frameworks are available, their use is not widespread–ad hoc scripts are often required to bind applications together. The BCJ explores another solution to this problem through a computer based environment suitable for on-site use, which builds on the traditional laboratory notebook paradigm. It provides an intuitive, extensible paradigm designed for expressive composition of applications. Extensive features facilitate sharing data, computational methods, and entire experiments. By focusing on the bioinformatics and computational biology domain, the scope of the computational framework was narrowed, permitting us to implement a capable set of features for this domain. This report discusses the features determined critical by our system and other projects, along with design issues. We illustrate the use of our implementation of the BCJ on two domain-specific examples.

The Performance of Different Communication Mechanisms and Algorithms Used for Parallelization of Molecular Dynamics Code

Communication performance appears to have the most important influence on parallelization efficiency of large scientific applications. Different communication algorithms and communication mechanisms were used in parallelization of molecular dynamics code. In is shown that in the case of fast communication hardware well scaling algorithm must be used. Presented data shows that MD code can be also run efficiently on the pentium cluster but low latency communication mechanism must be used.

Application of Pfortran and Co-Array Fortran in the parallelization of the GROMOS96 molecular dynamics module

After at least a decade of parallel tool development, parallelization of scientific applications remains a significant undertaking. Typically parallelization is a specialized activity supported only partially by the programming tool set, with the programmer involved with parallel issues in addition to sequential ones. The details of concern range from algorithm design down to low-level data movement details. The aim of parallel programming tools is to automate the latter without sacrificing performance and portability, allowing the programmer to focus on algorithm specification and development. We present our use of two similar parallelization tools, Pfortran and Crays Co-Array Fortran, in the parallelization of the GROMOS96 molecular dynamics module. Our parallelization started from the GROMOS96 distributions shared-memory implementation of the replicated algorithm, but used little of that existing parallel structure. Consequently, our parallelization was close to starting with the sequential version. We found the intuitive extensions to Pfortran and Co-Array Fortran helpful in the rapid parallelization of the project. We present performance figures for both the Pfortran and Co-Array Fortran parallelizations showing linear speedup within the range expected by these parallelization methods.

Education and research challenges in parallel computing

Over three decades of parallel computing, new computational requirements and systems have steadily evolved, yet parallel software remains notably more difficult relative to its sequential counterpart, especially for fine-grained parallel applications. We discuss the role of education to address challenges posed by applications such as informatics, scientific modeling, enterprise processing, and numerical computation. We outline new curricula both in computational science and in computer science. There appear to be new directions in which graduate education in parallel computing could be directed toward fulfilling needs in science and industry.