Special Issue "Algorithms and Molecular Sciences"

Assistant Editor
Ms. Laura Simon
MDPI, Kandererstrasse 25, CH-4057 Basel, Switzerland
E-mail:

Related Special Issue: Algorithms and Molecular Sciences in the International Journal of Molecular Sciences

Submission

All papers should be submitted to algorithms@mdpi.org. To be published continuously until the deadline and papers will be listed together at the special issue website.

Submitted papers should not have been published nor be under consideration for publication elsewhere. All papers are refereed through a peer-review process. A guide for authors is available on the Instructions for Authors page. Algorithms is an international peer-reviewed quarterly journal published by Molecular Diversity Preservation International.

Open Access publication fees are 300 CHF per paper. English correction fees and/or formatting fees (250 CHF) will be added in certain cases (550 CHF per paper for those papers that require extensive additional formatting and/or English corrections.). For the first two issues to be published in 2008, the publishing fees will be waived for well-prepared manuscripts.

Algorithms and Molecular Sciences

Title: Association Rules and Biological Data: a Huge Amount of Extracted Knowledge
Author: Simona Rossi

Title: Pathway Analysis and Feature Selection with Genetic Algorithms in Metabolomics
Authors: Wei ZOU¹ and Vladimir V. TOLSTIKOV²*
Affiliations: ¹UC Davis Genome Center, University of California, Davis, CA, USA
²Director of Metabolomics Core Lab, Genome Center, University of California, Davis, 451 East Health Science Drive, GBSF 1313 Davis, CA 95616, USA
* corresponding Author: Vladimir V. Tolstikov, PhD
Abstract: A robust and complete workflow for metabolomics measurement and data mining is described in detail. Three independent and complementary analytical techniques for metabolic profiling were applied: hydrophilic interaction chromatography (HILIC–LC–ESI–MS), reversed-phase liquid chromatography (RP–LC–ESI–MS), and gas chromatography (GC–TOF–MS) all coupled to mass spectrometry. Unsupervised methods, such as principle component analysis (PCA) and clustering, and supervised methods, such as classification were used for data mining. Genetic Algorithms (GA), a multivariate approach, was probed for selection of the smallest subsets of potentially discriminative classifiers. From thousands of peaks found in total, small subsets selected by GA are considered as highly potential classifiers allowing discrimination. Annotated GC–TOF–MS data allowed generating small subset of identified metabolites. LC–ESI–MS data and small subsets require further annotation. Annotated GC–TOF–MS data generate small subset of identified metabolites. LC–ESI–MS data and small subsets require further annotation using Fourier-transform ion cyclotron resonance (FT-ICR) MS accurate mass measurement and nuclear magnetic resonance spectroscopy (NMR) structure elucidation. GA were also used to generate correlated networks for pathway analysis. Several case studies demonstrated that such a workflow combining comprehensive metabolic profiling and advanced data mining techniques provides a powerful metabolomic approach for small molecules biomarker discovery.
Key words: pathway analysis, HILIC, metabolic profiling, LC–ESI–MS, GC–TOF–MS, feature selection, genetic algorithms, network construction

Title: A novel algorithm for macromolecular similarity search
Authors: Stanislav Jakuschev and Daniel Hoffmann
Abstract: Many macromolecules, namely proteins, show functional substructures or epitopes defined by characteristic spatial arrangements of groups of specific atoms or residues. The identification of such substructures in a set of macromolecular 3D-structures solves an important problem in molecular biology as it allows the assignment of functions to molecular moieties and thus opens the possibility of a mechanistic understanding of molecular function. We have devised an algorithm that models a functional epitope formed by a group of atoms or residues as set of points in cartesian space with associated functional properties. The algorithm searches for similar epitopes in a database of structures by an efficient multi-stage comparison of distance sets in the epitope and in the structures from the database. The search results in a scored list of matches and corresponding optimal superpositions of query
epitope and matching epitopes from the database. The algorithm is discussed against the background of related approaches, and it is successfully tested on pairs of ligand binding pockets of apo- and holo-enzymes, and on heme-binding pockets in hemoglobin. Finally, it is applied to the prediction of a binding pocket of the ligand 2'O-acetyl-ADP-ribose on the Sir2 enzyme.

Title: Bayesian Maximum Entropy Approach To Digital X-ray Mammogram Processing: A Robust Algorithm for Early Breast Cancer Detection
Author: Radu Mutihac
Affiliation: University of Bucharest, Romania
Abstract: Basics of Bayesian statistics in inverse problems using the maximum entropy principle are summarized in connection with the reconstruction of positive, additive images from various types of data like X-ray digital mammograms. An efficient iterative algorithm for image reconstruction from large data sets based on the conjugate gradient method and Lagrange multipliers in nonlinear optimization of a specific potential function was developed. The point spread function was defined by numerical simulations of homogeneous mammalian tissue with microcalcification inclusions of various opacities such as to match the characteristics of the present solid-state X-ray detectors. Incorporating a priori knowledge in blurred images increased the overall quality and interpretability of most X-ray mammograms. The processed images proved to be superior compared with their raw counterparts in terms of contrast, resolution, noise, and visibility of details.

Title: Exhaustive enumeration of kinetic model topologies for the analysis of time-resolved RNA folding
Authors: Joshua Martin¹, Katrina Simmons¹, and Alain Laederach^1,2,*
Affiliations: ¹Computational and Structural Biology Department, Wadsworth Center, Albany, NY 12208
²Biomedical Sciences Program, School of Public Health, SUNY, Albany, NY 12208
*corresponding Author, e-mail: alain@wadsworth.org
Abstract: Unlike protein folding, the process by which a large RNA molecule adopts a functionally active conformation remains poorly understood. Chemical mapping techniques, such as Hydroxyl Radical (·OH) footprinting, report on local changes in an RNA as it folds. The analysis and interpretation of this kinetic data requires the identification and subsequent optimization of a kinetic model and its parameters. We detail our approach to this problem, specifically focusing on a novel strategy to enumerate kinetic model topologies. These topologies relate the time evolution of the different species in solution to the experimentally observed local changes in the structure. In our initial implementation of the enumeration strategy, only systems with one or two intermediates were tractable using a distributed computing approach. However, as the number of intermediates in the system increased beyond two the problem became intractable due to a factorial increase in the number of models to be tested. The number of model topologies increases exponentially with our new enumeration strategy allowing the analysis of molecules with more than three intermediates. We demonstrate that we still test all necessary models and still identify the correct model. This enables us to analyze data sets from larger molecules such as the Ribosome, which were previously computationally intractable.

Title: Mutual Information to Investigate Specificity-Determining Positions: The Case of the Intrinsically Fluorescent Proteins
Authors: Walter Rocchia, Sara Bonella, Pietro Amat, Riccardo Nifosi', Valentina Tozzini

Title: Recent Advances in the Computational Discovery of Functional Transcription Factor Binding Sites
Authors: N.T. Tung and I.P. Androulakis (Rutgers University)
Abstract: The discovery of gene regulatory elements e.g. transcriptional factor binding sites (TFBSs) - upstream regions where specialized regulatory proteins bind - requires the synergism between computational and experimental techniques tin order o reveal the underlying regulatory mechanisms that drive gene expression in response to external cues and signals. Utilizing the large amount of high-throughput experimental data constantly growing in recent years, researchers have attempted to decipher the coded patterns which are hidden in the genomic sequences. These patterns called motifs are potential binding sites to transcription factors which are hypothesized as the main regulators of the transcriptional process.
Consequently, 'precise' detection of these elements is required and thus a large number of computational approaches have been developed to support the de novo identification of TFBSs. Early approaches simply addressed the problem by concentrating on a single motif in a promoter sequence set of either 'coregulated' genes or a single gene of orthologous species. However, recent advances have established that considering a set of neighbor TFBSs, called cis-regulatory modules or TF-modules, is more relevant to functionally-activated binding site region in a specific condition and/or tissue. Even though novel approaches are continuously proposed and almost all have reported some success in yeast and other lower organisms, regulatory organization in higher organism remains a challenger. Besides, due to the variety of underlying algorithms as well as complex strategies, the strength and weakness of each approach is usually due to our lack of understanding, leading to difficulty in choosing the method as well as designing the strategy for a specific problem.
We will present a brief review on the history of TFBS discovery along with approaches for binding site representation through position weight metrices (PWM) and promoter identification. We then review the techniques for scanning PWM to locate physical TFBSs, using orthologous information to identify functional BSs, and determining cis-regulatory modules to infer functionally-activated TFBSs. Finally, we extend these findings to explore the possibility of identifying condition-specific regulatry mechanism, thus evaluating alternative cellular response mechanisms depending on the nature of the various extracellular signals.

Title: Algorithm for Nanotubes Computer Generation with Different Configurations
Authors: M. Leonor Contreras^1,*, Eliseo Benítez², José Alvarez² and Roberto Rozas¹
Affiliations: ¹Environmental Sciences Department, Faculty of Chemistry and Biology, University of Santiago de Chile, Usach, Avda. L. B. O’Higgins 3363, Santiago, Chile
²Informatics Department, Faculty of Engineering, University of Santiago de Chile, Usach, Avda. Ecuador 3659, Santiago, Chile
E-mails: leonor.contreras@usach.cl; eliseo.benitez@gmail.com; jalvarez@diinf.usach.cl; roberto.rozas@usach.cl
* Author to whom correspondence should be addressed.
Abstract: The algorithm here described refer to generation, visualization and manipulation of molecular nanostructures of single wall nanotubes (NTs) of particular configurations (armchair, zipper, multi-zipper, zigzag, chiral) by means of a Graphical User Interface (GUI). NTs are constructed based on a carbon graphene sheet according to certain parameters defining required nanostructures. Generated NTs can easily be modified changing for instance carbon atoms for nitrogen or boron, visualized and exported into a standard format useful as input to be analyzed and studied through other applications in order to get optimized geometries and to carry out further calculations of molecular and electronic properties.