Motifs in Biology: Analysis of Ambiguous Data Hawaii International Conference on System Sciences - 27 Biotechnology Computing Track Maui, Hawaii, January 4 - 7, 1994 Co-chairs: Lloyd Allison, Monash University, Australia Toni Kazic, Washington University, USA David States, Washington University, USA The Hawaii International Conference on System Sciences (HICSS) - 27 announces a call for papers for the Biotechnology Computing Minitrack on the topic of Motifs in Biology. This minitrack will focus on finding, classifying, and testing motifs in the organization, structure and dynamics of biological systems. The use, reuse, and recombination of motifs in biology is fundamental to genetics and evolution. By ``motif'' we mean any nonrandom, recurrent organizational pattern, at any level of biological complexity. Classification in biology depends on the recognition of motifs, whether these are structural or ecological features of an organism in evolutionary taxonomy, spatial or temporal motifs in embryology and development, or similarities in catalytic mechanisms in enzymology. Motif induction, representation, and recognition in biological databases is a challenging area of computational biology. Often the high level representation of pattern information is as important as the ability to recognize a motif. In other cases the problem is to assemble scattered information of many types into a credible pattern. Many techniques have been used to induce patterns from anonymous data, ranging from stochastic-search to a variety of classification schemes. A second task is the recognition of instances of hypothesized patterns in apparently related data. Both activities take place against a background of inherent biological variation and diversity, experimental error and various sources of "noise". Variation is present both in the biological material and the data. It produces multiple "versions" of data items. Information from different sources may be contradictory. All this confounds the induction and recognition of motifs. Success is strongly determined by how variation in biological data is treated computationally. Outputs are of little use unless accompanied by indications of their biological and statistical significance. The identification of novel motifs in complex data is the essence of scientific discovery. The purpose of this minitrack is to bring pattern recognition, classification, variation handling and information extraction technologies together with motif discovery applications in biological domains. Emphasis will be given to biological areas which are still unfamiliar to computer scientists, such as developmental biology, and which might benefit from advanced pattern induction techniques. Papers should describe applications or algorithms for real biological systems, not abstractions of biological problems which conveniently remove variation to simplify the computation or increase the elegance of the abstraction. Thus, papers describing algorithms or heuristics, statistical or other measures of biological merit which treat either inherent biological variation or gaps or uncertainties in data are particularly welcome. Algorithms should have been run on real biological data, or on simulated data with explicitly stated assumptions, or on both. Computational, statistical and other methods should be described in sufficient detail for another worker to reproduce them. We believe the time is ripe to begin considering the problem of motif recognition in more general terms. Though the literature has been preoccupied with the recognition of linear motifs in DNA and proteins, there is a growing appreciation of the importance of spatial, temporal and functional motifs. Representation of motifs and the underlying data structures can have profound consequences for our ability to compute with them. Appropriate representation can reveal motifs which would be obscured in other data structures. Alert biologists are beginning to wonder what computer science can offer to aid recognition of complex, multifaceted motifs. The imperfections of biological data mean that the straightforward application of computational techniques, which usually assume perfect data or complete matching, is inappropriate. Biological application include: o Identification and characterization of signalling or structural motifs in linear sequences or chromosomes o Analysis of folding patterns in DNA and RNA. o Analysis of packing and folding patterns in protein tertiary structure o Recognition of motifs in nonlinear data such as physiological networks (topological, dynamic and regulatory) o Analysis of temporal and spatial patterns in development o Molecular approaches to systematic biology o Origin and evolution of motifs. Computational techniques include: o Bayesian and other statistical classification methods o Statistical and non-logical retrieval methods for databases o Hidden Markov models o Information theory and algorithmic complexity o Genetic algorithms o Artificial neural nets and techniques to extract information from them o Novel algorithms and data structures o Graph topology approaches o Database dredging o Case-based reasoning o Formal grammars of biological structure and function o Validation techniques o Construction and use of artificial data sets o Data representation o Parallel computation INSTRUCTIONS FOR AUTHORS A miniabstract consisting of a title, list of authors and two or three sentences describing the manuscript's application area and techniques is required. Manuscripts should be 22-26 typewritten, double-spaced pages in 10 or 12 point type. The final form for the printed papers will be 10-12 pages, 10 point, single spaced, in two columns. Authors are encouraged to check their manuscript prior to submission to be sure it falls within these final boundaries. Papers must not have been previously presented or published, nor currently submitted for journal publication. Once accepted to the conference, a paper may be submitted for journal publication. Each manuscript will be refereed by five reviewers. Manuscripts should include a title page that identifies the title of the paper, the full name(s) of the author(s), affiliation(s), complete mailing and electronic address(es), telephone number(s) and a 300 word abstract of the paper. Deadlines: * Miniabstract is due by May 1, 1993. * Six copies of the manuscript are due by June 7, 1993 * Notification of accepted papers by August 31, 1993 * Accepted camera ready manuscripts are due by October 1, 1993 Note: The miniabstract should be sent by electronic mail. Fax or electronic submission of manuscripts is not acceptable. Send submissions and questions regarding this minitrack to any of the cochairs: Lloyd Allison, Dept. Computer Science, Monash University, AUSTRALIA 3168. lloyd@cs.monash.edu.au 61 3 565 5205 (tel) 61 3 565 5146 (fax) Toni Kazic, Institute for Biomedical Computing, Box 8036, Washington University School of Medicine, 700 South Euclid Ave., St. Louis MO 63110, USA. toni@athe.wustl.edu 314-362-3121 (tel) 314-362-0234 (fax) David States, Institute for Biomedical Computing, Box 8068, Washington University School of Medicine, 700 South Euclid Ave., St. Louis MO 63110, USA. states@ibc.wustl.edu 314-362-2138 (tel) 314-362-0234 (fax)