Arvind K. Bansal

ASC: an associative-computing paradigm

Todays increased computing speeds allow conventional sequential machines to effectively emulate associative computing techniques. We present a parallel programming paradigm called ASC (ASsociative Computing), designed for a wide range of computing engines. Our paradigm has an efficient associative-based, dynamic memory-allocation mechanism that does not use pointers. It incorporates data parallelism at the base level, so that programmers do not have to specify low-level sequential tasks such as sorting, looping and parallelization. Our paradigm supports all of the standard data-parallel and massively parallel computing algorithms. It combines numerical computation (such as convolution, matrix multiplication, and graphics) with nonnumerical computing (such as compilation, graph algorithms, rule-based systems, and language interpreters). This article focuses on the nonnumerical aspects of ASC.<<ETX>>

Evolutionary Analysis by Whole-Genome Comparisons

A total of 37 complete genome sequences of bacteria, archaea, and eukaryotes were compared. The percentage of orthologous genes of each species contained within any of the other 36 genomes was established. In addition, the mean identity of the orthologs was calculated. Several conclusions result: (i) a greater absolute number of orthologs of a given species is found in larger species than in smaller ones; (ii) a greater percentage of the orthologous genes of smaller genomes is contained in other species than is the case for larger genomes, which corresponds to a larger proportion of essential genes; (iii) before species can be specifically related to one another in terms of gene content, it is first necessary to correct for the size of the genome; (iv) eukaryotes have a significantly smaller percentage of bacterial orthologs after correction for genome size, which is consistent with their placement in a separate domain; (v) the archaebacteria are specifically related to one another but are not significantly different in gene content from the bacteria as a whole; (vi) determination of the mean identity of all orthologs (involving hundreds of gene comparisons per genome pair) reduces the impact of errors in misidentification of orthologs and to misalignments, and thus it is far more reliable than single gene comparisons; (vii) however, there is a maximum amount of change in protein sequences of 37% mean identity, which limits the use of percentage sequence identity to the lower taxa, a result which should also be true for single gene comparisons of both proteins and rRNA; (viii) most of the species that appear to be specifically related based upon gene content also appear to be specifically related based upon the mean identity of orthologs; (ix) the genes of a majority of species considered in this study have diverged too much to allow the construction of all-encompassing evolutionary trees. However, we have shown that eight species of gram-negative bacteria, six species of gram-positive bacteria, and eight species of archaebacteria are specifically related in terms of gene content, mean identity of orthologs, or both.

An automated comparative analysis of 17 complete microbial genomes.

MOTIVATION As sequenced genomes become larger and sequencing becomes faster, there is a need to develop accurate automated genome comparison techniques and databases to facilitate derivation of genome functionality; identification of enzymes, putative operons and metabolic pathways; and to derive phylogenetic classification of microbes. RESULTS This paper extends an automated pair-wise genome comparison technique (Bansal et al., Math. Model. Sci. Comput., 9, 1-23, 1998, Bansal and Bork, in First International Workshop of Declarative Languages, Springer, pp. 275-289, 1999) used to identify orthologs and gene groups to derive orthologous genes in a group of genomes and to identify genes with conserved functionality. Seventeen microbial genomes archived at ftp://ncbi.nlm.nih.gov/genbank/genomes have been compared using the automated technique. Data related to orthologs, gene groups, gene duplication, gene fusion, orthologs with conserved functionality, and genes specifically orthologous to Escherichia coli and pathogens has been presented and analyzed. AVAILABILITY A prototype database is available at ftp://www.mcs.kent.edu/arvind/intellibio / orthos.html. The software is free for academic research under an academic license. The detailed database for every microbial genome in NCBI is commercially available through intellibio software and consultancy corporation (Web site: http://www.mcs.kent.edu/årvind/intellibio . html). CONTACT arvind@mcs.kent.edu.

Bioinformatics in microbial biotechnology--a mini review.

The revolutionary growth in the computation speed and memory storage capability has fueled a new era in the analysis of biological data. Hundreds of microbial genomes and many eukaryotic genomes including a cleaner draft of human genome have been sequenced raising the expectation of better control of microorganisms. The goals are as lofty as the development of rational drugs and antimicrobial agents, development of new enhanced bacterial strains for bioremediation and pollution control, development of better and easy to administer vaccines, the development of protein biomarkers for various bacterial diseases, and better understanding of host-bacteria interaction to prevent bacterial infections. In the last decade the development of many new bioinformatics techniques and integrated databases has facilitated the realization of these goals. Current research in bioinformatics can be classified into: (i) genomics – sequencing and comparative study of genomes to identify gene and genome functionality, (ii) proteomics – identification and characterization of protein related properties and reconstruction of metabolic and regulatory pathways, (iii) cell visualization and simulation to study and model cell behavior, and (iv) application to the development of drugs and anti-microbial agents. In this article, we will focus on the techniques and their limitations in genomics and proteomics. Bioinformatics research can be classified under three major approaches: (1) analysis based upon the available experimental wet-lab data, (2) the use of mathematical modeling to derive new information, and (3) an integrated approach that integrates search techniques with mathematical modeling. The major impact of bioinformatics research has been to automate the genome sequencing, automated development of integrated genomics and proteomics databases, automated genome comparisons to identify the genome function, automated derivation of metabolic pathways, gene expression analysis to derive regulatory pathways, the development of statistical techniques, clustering techniques and data mining techniques to derive protein-protein and protein-DNA interactions, and modeling of 3D structure of proteins and 3D docking between proteins and biochemicals for rational drug design, difference analysis between pathogenic and non-pathogenic strains to identify candidate genes for vaccines and anti-microbial agents, and the whole genome comparison to understand the microbial evolution. The development of bioinformatics techniques has enhanced the pace of biological discovery by automated analysis of large number of microbial genomes. We are on the verge of using all this knowledge to understand cellular mechanisms at the systemic level. The developed bioinformatics techniques have potential to facilitate (i) the discovery of causes of diseases, (ii) vaccine and rational drug design, and (iii) improved cost effective agents for bioremediation by pruning out the dead ends. Despite the fast paced global effort, the current analysis is limited by the lack of available gene-functionality from the wet-lab data, the lack of computer algorithms to explore vast amount of data with unknown functionality, limited availability of protein-protein and protein-DNA interactions, and the lack of knowledge of temporal and transient behavior of genes and pathways.

An associative model to minimize matching and backtracking overhead in logic programs with large knowledge bases

Abstract A model is presented which exploits data level massive parallelism present in associative computers for the efficient execution of logic programs with large knowledge bases. The exploitation of data parallelism in goal reduction efficiently prunes non-unifiable clauses resulting in effective reduction of shallow backtracking, and marks the potential bindings for the variables with single occurrence in a manner which is independent of the number of clauses. During deep backtracking, bindings are released simultaneously using associative search resulting in a significant reduction in execution time overhead of backtracking and garbage collection. A scheme for a logical data structure representation incorporating direct interface between lists and vectors is described. This allows the efficient integration of symbolic computation and a large class of vectorizable numerical computation on associative supercomputers.

Distributed Storage of Replicated Beliefs to Facilitate Recovery of Distributed Intelligent Agents

We address the problem of recovering the state of an agent after a hardware/software failure of the system. We address the replication and reincarnation sub-problems of agent recovery under certain assumptions. An algorithm for distributed storage of replicated beliefs is provided and its correctness is proved formally. This algorithm allows the reincarnation of multiple crashed agents in a system of distributed autonomous intelligent agents. The scheme uses replication and distributed storage in the immediate neighboring agents, and uses distributed logical clocks to preserve the causality and to terminate retransmission.

Integrating co-regulated gene-groups and pair-wise genome comparisons to automate reconstruction of microbial pathways

This paper extends previously described automated techniques by automatically integrating the information about automatically derived co-transcribed gene-groups, functionally similar gene-groups derived using automated pair-wise genome comparisons and automatically derived orthologs (functionally equivalent genes) to derive microbial metabolic pathways. The method integrates automatically derived co-transcribed gene-groups with orthologous and homologous gene-groups (http://www.mcs.kent.edu//spl sim/arvind/orthos.html), the biochemical pathway template available at the KEGG database. (http://www.genome.ad.jp), the enzyme information derived from the SwissProt enzyme database (http://expasy.hcuge.ch/) and Ligand database (http://www.genome.ad.jp). The technique refines existing pathways (based upon network of reactions of enzymes) by associating corresponding non-enzymatic, regulatory, and cotranscribed proteins to enzymes. The technique has been illustrated by deriving a major pathway of M. tuberculosis by comparison with seven microbial genomes including E coli and B. subtilis - two microbes well explored in wet laboratories.

An integrated approach for efficient analysis of facial expressions

This paper describes a new automated facial expression analysis system that integrates Locality Sensitive Hashing (LSH) with Principal Component Analysis (PCA) and Linear Discriminant Analysis (LDA) to improve execution efficiency of emotion classification and continuous identification of unidentified facial expressions. Images are classified using feature-vectors on two most significant segments of face: eye segments and mouth-segment. LSH uses a family of hashing functions to map similar images in a set of collision-buckets. Taking a representative image from each cluster reduces the image space by pruning redundant similar images in the collision-buckets. The application of PCA and LDA reduces the dimension of the data-space. We describe the overall architecture and the implementation. The performance results show that the integration of LSH with PCA and LDA significantly improves computational efficiency, and improves the accuracy by reducing the frequency-bias of similar images during PCA and SVM stage. After the classification of image on database, we tag the collision-buckets with basic emotions, and apply LSH on new unidentified facial expressions to identify the emotions. This LSH based identification is suitable for fast continuous recognition of unidentified facial expressions.

Establishing a framework for comparative analysis of genome sequences

This paper describes a framework and a high level language tool for comparative analysis of genome sequence alignment. The framework integrates the information derived from multiple sequence alignment and phylogeny to derive new properties about homologous sequences. Multiple sequence alignments are treated as an abstract data type. Abstract operations have been described to manipulate a multiple sequence alignment, and to derive mutation related information from a phylogenetic tree by superimposing parsimony. The framework has been applied to derive constrained columns which exhibit evolutionary pressure to preserve a common property in a column despite mutation. A Prolog tool based on the the framework has been implemented and demonstrated on alignments containing 3000 sequences and 3904 columns.<<ETX>>

An abstract interpretation scheme for identifying inherent parallelism in logic programs

We describe a new scheme for the abstract interpretation of logic programs. The scheme was developed to identify and integrate different forms of inherent parallelism in logic programs at compile time. The scheme has four components: generalization, abstract unification, summarization and concretization, algorithms for which are discussed. The abstract domain for interpretation consists of type expressions which are used as program modes. The generated mode information has been applied to identify different classes of procedures exhibiting different forms of inherent parallelism. The mode information has also been applied for detection of guards and producer-consumer relationship. The advantages and limitations of our resulting scheme are discussed.