Abstract After publication of this supplement [1], it was brought to our attention that the wrong year was used in the title of the supplement. The 16th Annual UT-KBRIN Bioinformatics Summit was held in 2017.

Memphis, Tennessee

I1 Proceedings of the Fifteenth Annual UT- KBRIN Bioinformatics Summit 2016 Eric C. Rouchka, Julia H. Chariker, Benjamin J. Harrison, Juw Won Park P1 CC-PROMISE: Projection onto the Most Interesting Statistical Evidence (PROMISE) with Canonical Correlation to integrate gene expression and methylation data with multiple pharmacologic and clinical endpoints Xueyuan Cao, Stanley Pounds, Susana Raimondi, James Downing, Raul Ribeiro, Jeffery Rubnitz, Jatinder Lamba P2 Integration of microRNA-mRNA interaction networks with gene expression data to increase experimental power Bernie J Daigle, Jr. P3 Designing and writing software for in silico subtractive hybridization of large eukaryotic genomes Deborah Burgess, Stephanie Gehrlich, John C Carmen P4 Tracking the molecular evolution of Pax gene Nicholas Johnson; Chandrakanth Emani P5 Identifying genetic differences in thermally dimorphic and state specific fungi using in silico genomic comparison Stephanie Gehrlich, Deborah Burgess, John C Carmen P6 Identification of conserved genomic regions and variation therein amongst Cetartiodactyla species using next generation sequencing Kalpani De Silva, Michael P Heaton, Theodore S Kalbfleisch P7 Mining physiological data to identify patients with similar medical events and phenotypes Teeradache Viangteeravat, Rahul Mudunuri, Oluwaseun Ajayi, Fatih Şen, Eunice Y Huang P8 Smart brief for home health monitoring Mohammad Mohebbi, Luaire Florian, Douglas J Jackson, John F Naber P9 Side-effect term matching for computational adverse drug reaction predictions AKM Sabbir, Sally R Ellingson P10 Enrichment vs robustness: A comparison of transcriptomic data clustering metrics Yuping Lu, Charles A Phillips, Michael A Langston P11 Deep neural networks for transcriptome-based cancer classification Rahul K Sevakula, Raghuveer Thirukovalluru, Nishchal K. Verma, Yan Cui P12 Motif discovery using K-means clustering Mohammed Sayed, Juw Won Park P13 Large scale discovery of active enhancers from nascent RNA sequencing Jing Wang, Qi Liu, Yu Shyr P14 Computationally characterizing genomic pipelines and benchmarking results using GATK best practices on the high performance computing cluster at the University of Kentucky Xiaofei Zhang, Sally R Ellingson P15 Development of approaches enabling the identification of abnormal gene expression from RNA-Seq in personalized oncology Naresh Prodduturi, Gavin R Oliver, Diane Grill, Jie Na, Jeanette Eckel-Passow, Eric W Klee P16 Processing RNA-Seq data of plants infected with coffee ringspot virus Michael M Goodin, Mark Farman, Harrison Inocencio, Chanyong Jang, Jerzy W Jaromczyk, Neil Moore, Kelly Sovacool P17 Comparative transcriptomics of three Acinetobacter baumanii clinical isolates with different antibiotic resistance patterns Leon Dent, Mike Izban, Sammed Mandape, Shruti Sakhare, Siddharth Pratap, Dana Marshall P18 Metagenomic assessment of possible microbial contamination in the equine reference genome assembly M Scotty DePriest, James N MacLeod, Theodore S Kalbfleisch P19 Molecular evolution of cancer driver genes Chandrakanth Emani, Hanady Adam, Ethan Blandford, Joel Campbell, Joshua Castlen, Brittany Dixon, Ginger Gilbert, Aaron Hall, Philip Kreisle, Jessica Lasher, Bethany Oakes, Allison Speer, Maximilian Valentine P20 Biorepository Laboratory Information Management System Naga Satya V Rao Nagisetty, Rony Jose, Teeradache Viangteeravat, Robert Rooney, David Hains

The University of Tennessee (UT) and the Kentucky Biomedical Research Infrastructure Network (KBRIN) have collaborated over the past fourteen years to share research and educational expertise in bioinformatics.One result is an

Results Computational analyses of the novel UTR sequences, focusing on RNA-binding protein (RNAbp) interaction motifs revealed strongly over-represented RNAbps with known roles in nervous system pathologies. We consider the implications of 3’UTR transcript extension and protein interaction in the context of axonal plasticity and the consequences of mis-regulation of this process during neurological disease.

The University of Tennessee (UT) and the Kentucky Biomedical Research Infrastructure Network (KBRIN) have collaborated over the past thirteen years to share research and educational expertise in bioinformatics.One result is an annual regional summit for researchers, educators and students interested in bioinformatics.The Thirteenth Annual UT-KBRIN Bioinformatics Summit was held at Lake Barkley State Park in Cadiz, Kentucky on April 11-13, 2014.A total of 172 participants pre-registered, with 104 from Tennessee and 61 from Kentucky.Among the registrants were 69 faculty, 45 students, 38 staff, and 20 postdocs.The conference program consisted of a workshop on Cytoscape and three days of presentations broken into plenary sessions on Genetic Variation and Mutational Analysis, Genomics, and P4 Systems Biology.Nine short talks were selected from 51 submitted poster abstracts. Friday workshopsKeiichiro Ono (UCSD) opened the Summit with the workshop "PART I: Introduction to Cytoscape: an Open Source Platform for Biological Network Data Analysis and Visualization."He provided a description of the basic UI, methods for importing and exporting data, and methods for highlighting and exploring information in a biological network [1-3].Keiichiro discussed methods for importing publicly available databases with a network structure, such as STRING [4], IntAct [5], BioGRID [6], ChEMBL [7], KEGG [8], Reactome [9], WikiPathways [10], and PathwayCommons [11].Keiichiro presented a number of commands for simplifying networks, including selecting first neighbors of a node, creation of a subgraph, and filtering based on node and edge attributes.He also demonstrated tools for highlighting

Background Stress granules (SG) are formed as collections of protein and RNA (ribonucleoprotein structures) and continuously assembled/disassembled in response to stresses such as heat, osmotic, or oxidant stress; representing an attempt to survive the stress through salvage of important proteins and RNA. Recent research suggests diabetic nephropathy (DN) may change or alter SG biology and in conditions that model DN may involve the receptor for activated C-kinases (RACK1). The incorporation of RACK1 into stress granules may down-regulate programmed cell death and further may impart the ability to scaffold to and sequester key signaling proteins to affect cell survival or death. We hypothesized that the inappropriate or dysregulated scaffolding of proteins or RNA into stress granules may be of importance in the development of diabetic nephropathy. The goal of this study is to qualitatively and semi-quantitatively characterize the effects of cell culture conditions modeling diabetes and ER stress in conjunction with over-expression studies of SG stabilizing proteins on RNA transcripts.

The University of Tennessee (UT), the Oak Ridge National Laboratory (ORNL), and the Kentucky Biomedical Research Infrastructure Network (KBRIN), have collaborated over the past twelve years to share research and educational expertise in bioinformatics.One result of this collaboration is the joint sponsorship of an annual regional summit to bring together researchers, educators and students who are interested in bioinformatics from a variety of research and educational institutions.This summit provides unique opportunities for collaboration and forging links between members of the various institutions.This year, the Twelfth Annual UT-ORNL-KBRIN Bioinformatics Summit was held at Paris Landing State Park in Buchanan, TN from March 22-24, 2013.A total of 182 participants pre-registered for the summit, with 116 from various Tennessee institutions and 54 from various Kentucky institutions.A number of additional participants came from universities and research institutions from other states and countries, e.g.

Neurons interact with, and are influenced by, tissues that are remote from the cell body. For example, sensory neuron cell bodies are located within peri-spinal ganglia but are connected to both the spinal cord and skin via their axons projecting through dorsal roots and peripheral nerves. Biochemical signals from anatomical compartments (spinal cord / root / ganglion / nerve / skin) modulate the molecular biology of neurons which can respond to signals from any/all of these remote regions. One mechanism by which neurons respond to these signals and interact with their targets is by actively transporting mRNA to that region. There, the mRNA is translated to produce protein at locally-determined positions and times. A growing body of evidence shows that untranslated regions (UTRs) of genes are important for this targeting. For example, 3’-UTRs contain 50nt “zip code” consensus binding sites for cis-acting zip code-binding proteins (ZPBs) that drive axonal targeting of mRNA [ 1 ]. We therefore hypothesized that gene expression during collateral sprouting, an axonal growth process that is highly responsive to target-derived factors, might involve differential regulation of UTR components. Axonal collateral sprouting of sensory neuron axons was induced and progressed for 7 or 14 days. RNA was then harvested from sensory ganglia of experimental and control animals for total transcriptome sequencing using the Illumina platform. Sequence reads were aligned to the currently available reference Rat genome (rn4), and all known exons (coding and untranslated) were tested for differential expression at each time-point compared to those from control ganglia. We also identified the subset of genes for which the sprouting-associated expression change was due to regulation of the UTR while the coding region of that gene was unchanged between sprouting and control. Thus, we identified the genes whose sprouting-associated regulation was due to changes in expression of the UTR and not of the coding sequence (CDS). These genes with differentially-expressed UTRs and non-regulated coding regions were then tested for over-representation of functional classes (Gene Ontology) and for known RNA binding protein binding sites. Ontology enrichment analysis indicates that UTR-specific regulation of expression of numerous genes may be important for neuronal plasticity responses to target-derived factors, and local regulation of axonal plasticity.

BACKGROUND: High-throughput molecular biology techniques yield vast amounts of data, often by detecting small portions of ribonucleotides corresponding to specific identifiers. Existing bioinformatic methodologies categorize and compare these elements using inferred descriptive annotation given this sequence information irrespective of the fact that it may not be representative of the identifier as a whole. RESULTS: All annotations, no matter the granularity, can be aligned to genomic sequences and therefore annotated by genomic intervals. We have developed AbsIDconvert, a methodology for converting between genomic identifiers by first mapping them onto a common universal coordinate system using an interval tree which is subsequently queried for overlapping identifiers. AbsIDconvert has many potential uses, including gene identifier conversion, identification of features within a genomic region, and cross-species comparisons. The utility is demonstrated in three case studies: 1) comparative genomic study mapping plasmodium gene sequences to corresponding human and mosquito transcriptional regions; 2) cross-species study of Incyte clone sequences; and 3) analysis of human Ensembl transcripts mapped by Affymetrix®; and Agilent microarray probes. AbsIDconvert currently supports ID conversion of 53 species for a given list of input identifiers, genomic sequence, or genome intervals. CONCLUSION: AbsIDconvert provides an efficient and reliable mechanism for conversion between identifier domains of interest. The flexibility of this tool allows for custom definition identifier domains contingent upon the availability and determination of a genomic mapping interval. As the genomes and the sequences for genetic elements are further refined, this tool will become increasingly useful and accurate. AbsIDconvert is freely available as a web application or downloadable as a virtual machine at: http://bioinformatics.louisville.edu/abid/.

Background In order to develop treatments or preventive measures for auditory hair cell loss, an understanding of both the process of auditory hair cell regeneration and factors that influence this process, is needed. Our previous microarray analysis showed that growth hormone (GH) was significantly upregulated during zebrafish auditory hair cell regeneration, coupled with cell proliferation [1,2]. We further tested the effects of GH on zebrafish auditory hair cell regeneration by injecting GH after sound exposure and found that GH can efficiently promote post-trauma auditory hair cell regeneration, which may be achieved through stimulating proliferation and suppressing apoptosis [3]. In the current study, we used Next Generation Sequencing (NGS) to examine the possible GH pathways involved in zebrafish auditory hair cell regeneration.

The University of Tennessee (UT), the Oak Ridge National Laboratory (ORNL), and the Kentucky Biomedical Research Infrastructure Network (KBRIN), have collaborated over the past eleven years to share research and educational expertise in bioinformatics.One result of this collaboration is the joint sponsorship of an annual regional summit to bring together researchers, educators and students who are interested in bioinformatics from a variety of research and educational institutions.This summit provides unique opportunities for collaboration and forging links between members of the various institutions.This year, the Eleventh Annual UT-ORNL-KBRIN Bioinformatics Summit was held at the Seelbach Hilton Hotel in Louisville, Kentucky from March 30-April 1, 2012.A total of 232 participants pre-registered for the summit, with 126 from various Kentucky institutions and 80 from various Tennessee institutions.A number of additional participants came from universities and research institutions from other states and countries, e.g.

Existing analytical tools enable broad-scale experimentation ("-omics") to provide a great deal of information about intracellular processes. However, extraction of information regarding intercellular interactions, particularly from separate datasets, is generally more limited, principally for lack of specialized analytical tools. In turn, few experiments are designed to examine intercellular interactions. Using the large number of previously identified interactions available in databases may provide a useful platform for analyzing these interactions. However, finding all possible interactions is a computationally intensive task and quickly becomes intractable using a naive approach on networks with hundreds of thousands of nodes and edges. A heuristic algorithm similar to the "Backtracking algorithm" is proposed to find all possible protein interactions across any two gene sets. The algorithm starts with an initial set of genes and incrementally adds a candidate to the interaction network and abandons each candidate x as soon as it is determined that x does not lead to a valid solution. An exclusion vector (EV) is used to accomplish this task and is populated at each step, maintaining a list of those nodes that need to be excluded from interactions in the future and thus restricting the size of the network. The EV also allows location awareness by using Gene Ontology (GO) cell component classifications to discard nodes that are not relevant for the network. This algorithm can be readily applied to pathway analysis and the determination of elements underlying intercellular interactions.

A variety of systems exist in which annotations are available at various levels of granularity to a reference coordinate system, such as roads and landmarks on a map, features within a 2-dimensional or 3-dimensional image, or genetic entities (GEs) mapped to a reference genome. As the number of annotations grows, methods to efficiently locate overlapping entities within a specific interval of interest are needed. In this paper, the efficiency of using interval trees for storing, maintaining, and querying large numbers of intervals with special attention to genetic entities is demonstrated. The results suggest a significant speed -- up when compared to relational database approaches. As such, interval trees serve as a suitable alternative for storing and searching annotations to a reference coordinate system.

Many current DNA microarray and other high-throughput data meta-analysis studies concentrate on deriving a concordant list of genes across many experiments to discover the "true" genes responsible for a particular disease process or biological pathway or cellular response (Figure 1A ). However, by concentrating on the genes in common, similarities or differences that exist at a pathway or process level may be missed. A - Usual method of high-throughput experiment meta-analysis comparing gene lists (L1 and L2) directly. B - categoryCompare compares the gene lists on the basis of enriched annotations. We describe a meta-analysis approach that allows comparison and contrast of gene lists at the level of categorical annotation (pathway or Gene Ontology annotations). This categorical evaluation compares enriched annotations between gene lists (Figure 1B ), and displays the results graphically to allow intuitive visualization and exploration of the similarities and differences. False discovery correction via simulation is implemented to control for the effect of different sized gene lists as inputs. The approach was tested using two gene lists, genes involved in the response to denervation in muscle (a literature compendium), and in skin (experimentally determined). Using the categorical comparison highlights known biological processes that are common in the two cases, while also allowing one to easily see areas of difference that are not apparent from examining the gene lists alone.

Given NASA’s recent focus on long-duration space travel, potential adverse effects on astronauts of ionizing radiation need to be minimized. Levels of exposure for astronauts on such flights can be high enough to cause damage to DNA, possibly causing mutation and cancer. By analyzing the mRNA expression levels of genes exposed to different doses of gamma radiation in laboratory experiments, it can be determined which genes act as biomarkers of dosage-specific radiation exposure which can then be used on lab-on-a-chip (LOC) diagnostic platforms for early detection of ionizing radiation exposure. Currently, our group is incorporating three genes known to be involved in double-stranded DNA damage and repair, including: p21, p53, and γH2AX. While these genes should show up in response to ionizing radiation exposure, they are not ideal biomarkers due to their lack of specificity. Therefore, by using meta-analysis, our goal is to gain further insight into additional biomarkers for radiation exposure. More than 40 publicly-available data sets appropriate to our study were obtained from the online repository, Gene Expression Omnibus (GEO) [ 1 ]. These gene expression experiments were subsequently categorized by radiation type and dose. Meta-analysis was performed using a combination of statistical tests using the Differential Expression via Distance Synthesis (DEDS) package [ 2 ]. Preliminary meta-analysis of these publicly available datasets yields potential biomarkers from the P53 signaling pathway (CDKN1A, GADD45A, MDM2, PMAIP1), stress response transcription factors (ATF3, JUN, JUNB, JUND), and cell surface receptors (CD69, CD70, CD83). Additional microarray experiments involving irradiated blood samples are underway. The analysis of both publicly available data and our own datasets will yield a broader picture of genes most sensitive to exposure of ionizing radiation for use as biomarkers on LOC diagnostic platforms for early detection of radiation exposure, leading to subsequent treatment.

We present the Weighted Infinite Relational Model (WIRM) that jointly detects biologically sensible ligand groups and protein groups by integrating the clustering of various data types including chemical compound descriptors, protein sequences, ligand-target bindings and pharmaceutical effects. WIRM takes advantage of the Bayesian nonparametric paradigm for integrating multiple data types, for allowing for missing values (e.g. unknown ligand-target interaction) in the data, for automatically inferring the number of clusters without explicit model comparison, and for predicting the ligand-target interactions. Because some of these data types, to varying degrees, may suggest relationships having no implication for ligand-target interactions or for biological sensible ligand and protein groups, WIRM allows different types of data to have different weights based on prior knowledge of their quality or relevance. We apply WIRM to the ion channel proteins and G-protein-coupled receptors. We validate its performance using functional annotation and ligand-target interaction. We also test the relationship among multiple data types by varying the weights which indicate the impact of each data type on the model. The categories and interactions inferred by WIRM both confirm known biology and suggest novel predictions.

The University of Tennessee (UT), the Oak Ridge National Laboratory (ORNL), and the Kentucky Biomedical Research Infrastructure Network (KBRIN), have collaborated over the past decade to share research and educational expertise in bioinformatics.One result of this collaboration is the joint sponsorship of an annual regional summit to bring together researchers, educators and students who are interested in bioinformatics from a variety of research and educational institutions.This summit provides unique opportunities for collaboration and forging links between members of the various institutions.This year, the Tenth Annual UT-ORNL-KBRIN Bioinformatics Summit was held at the University of Memphis in Memphis, Tennessee from April 1-3, 2011.A total of 225 participants pre-registered for the summit, with 146 from various Tennessee institutions and 60 from various Kentucky institutions.A number of additional participants came from universities and research institutions from other states and countries, e. g.University of Arkansas Medical Sciences, University of British Columbia, University of Cincinnati, Iowa State University, etc. Seventy-four registrants were faculty, with an additional 73 students, 50 staff, and 28 postdoctoral participants.The conference program consisted of three days of presentations.The first day included a pre-summit of talks by researchers supported by the Kentucky Biomedical Research Infrastructure Network (KBRIN), and three workshops covering the topics Next-Generation Sequencing, the workflow platform Galaxy, and Gene-MANIA.The next two days were dedicated to scientific presentations divided into three plenary sessions on Pediatric Genetics, Pharmacogenomics and Systems Biology.Each session also included short talks selected from the submitted poster abstracts. Pre-summit KBRIN sessionDr. Eric Rouchka started the pre-summit KBRIN session with an update on the supplement to the primary KBRIN grant awarded in 2009.Thanks to the funding provided by the supplement, five postdoctoral research associates have been hired, along with two masters level bioinformatics staff positions.In addition, a weekly seminar series has been implemented that includes externally invited speakers with strong research programs in bioinformatics.Following was a series of short talks from various University of Louisville researchers: "Developing an Analysis Pipeline for FT-ICR-MS Isotopologue Data from Stable Isotope Resolved Metabolomics (SIRM) Experiments" (Hunter Moseley), "Time Series Classifier Model for miR-mRNA Relationships" (Jovan D. Rebolledo-Mendez), "Assessing Variations in NGS Data" (Alex Kemper), "A Systems Based Approach to Find Protein Interactions Across Tissues" (Fahim Mohammad), "cate-goryCompare: High-Throughput Data Meta-Analysis Using Gene Annotations" (Robert M. Flight), "Collaborations between Biology/Bioinformatics" (Benjamin J. Harrison). Friday workshopsJon Armstrong of Cofactor Genomics (St. Louis, MO; http://www.cofactorgenomics.com)started the official summit program with an excellent overview of the past, present and future of DNA sequencing, beginning with the Maxam-Gilbert [1] and Sanger [2] methods and transitioning to the approaches used by next-generation sequencers.Among the technologies covered were pyrosequencing [3], reversible dye terminator technology [4], sequencing by ligation [5], and single molecule real time

Hyper basis function (HyperBF) networks are generalized radial basis function neural networks (where the activation function is a radial function of a weighted distance). Such generalization provides HyperBF networks with high capacity to learn complex functions, which in turn make them susceptible to overfitting and poor generalization. Moreover, training a HyperBF network demands the weights, centers, and local scaling factors to be optimized simultaneously. In the case of a relatively large dataset with a large network structure, such optimization becomes computationally challenging. In this paper, a new regularization method that performs soft local dimension reduction in addition to weight decay is proposed. The regularized HyperBF network is shown to provide classification accuracy competitive to a support vector machine while requiring a significantly smaller network structure. Furthermore, a practical training to construct HyperBF networks is presented. Hierarchal clustering is used to initialize neurons followed by a gradient optimization using a scaled version of the Rprop algorithm with a localized partial backtracking step. Experimental results on seven datasets show that the proposed training provides faster and smoother convergence than the regular Rprop algorithm.

Relational databases provide unprecedented opportunities for knowledge discovery. Various approaches have been proposed to infer structures over entity types and predict relationships among elements of these types. However, discovering structures beyond the entity type level, e.g. clustering over relation concepts, remains a challenging task. We present a Bayesian nonparametric model for joint relation and domain clustering. The model can automatically infer the number of relation clusters, which is particularly important in novel cases where little prior knowledge is known about the number of relation clusters. The approach is applied to clustering various relations in a gene database.

DOAJ is a unique and extensive index of diverse open access journals from around the world, driven by a growing community, committed to ensuring quality content is freely available online for everyone.

the Oak

BACKGROUND: In humans, copies of the Long Interspersed Nuclear Element 1 (LINE-1) retrotransposon comprise 21% of the reference genome, and have been shown to modulate expression and produce novel splice isoforms of transcripts from genes that span or neighbor the LINE-1 insertion site. RESULTS: In this work, newly released pilot data from the 1000 Genomes Project is analyzed to detect previously unreported full length insertions of the retrotransposon LINE-1. By direct analysis of the sequence data, we have identified 22 previously unreported LINE-1 insertion sites within the sequence data reported for a mother/father/daughter trio. CONCLUSIONS: It is demonstrated here that next generation sequencing data, as well as emerging high quality datasets from individual genome projects allow us to assess the amount of heterogeneity with respect to the LINE-1 retrotransposon amongst humans, and provide us with a wealth of testable hypotheses as to the impact that this diversity may have on the health of individuals and populations.

In current proteomics research, a big challenge is to differentiate the correlative proteins for a given biological function from all existing proteins, and if it does, how strong is the relationship between the proteins and function. Statistical significance testing can be used to address this question. However, every traditional statistical test method may suffer from the inability to identify important differentially expressed proteins if the biological samples do not completely meet the assumptions of each test method [ 1 ]. To detect the regulated proteins for differential proteomics, we analyze multiple significance test methods and discover some significance proteins. We use the four statistical methods, i.e., Kolmogorov-Smirnov test (KS-test), Baumgartner-Weib-Schindler test (BWS-test), T-test, Brunner-Munzel test (BM-test) to measure the difference of the expression level of individual protein under two experimental conditions, respectively. The results had been successfully used for the discovery of protein biomarkers in breast cancer.

In response to the ICMLA 2009 "Functional Clustering of Gene Expression Profiles in Human Cancers Challenge", we present a new dimension reduction approach that ranks features based on their localized discriminative power. The proposed method is based on a localized dimension reduction penalty added to the objective function for training a hyper basis function (hyper BF or generalized RBF) neural network. The localized dimension reduction in a hyper BF network is motivated by the fact that many patterns are recognized to belong to one class due to the co-occurrence of specific values along a specific set of dimensions. Furthermore, one class is very likely to be composed of multiple sub-models that are different in their characteristics. The proposed method is applied at multiple iterations. Features are ranked each iteration and a smaller subset is passed to the next iteration. A support vector machine (SVM) is used to evaluate the remaining dimensions in a ten-fold cross validation setting. Experimental results show the proposed method effectively reduces the number of dimensions from 54,613 to 65 while increasing the cross-validation classification accuracy from 92.3% to 96.4%. The classification accuracy of the final model on unseen data was reported by the challenge coordinators to be 94% accurate.

Many data sets contain an abundance of background data or samples belonging to classes not currently under consideration. We present a new unsupervised learning method based on Fuzzy C-Means to learn sub models of a class using background samples to guide cluster split and merge operations. The proposed method demonstrates how background samples can be used to guide and improve the clustering process. The proposed method results in more accurate clusters and helps to escape locally minimum solutions. In addition, the number of clusters is determined for the class under consideration. The method demonstrates remarkable performance on both synthetic 2D and real world data from the MNIST dataset of hand written digits.

Biological sequence usually contains yet to find knowledge, and mining biological sequences usually involves a huge dataset and long computation time. Common tasks for biological sequence mining are pattern discovery, classification and clustering. The newly developed model, plausible neural network (PNN), provides an intuitive and unified architecture for such a large dataset analysis. This paper introduces the basic concepts of the PNN, and explains how it is applied to biological sequence mining. The specific task of biological sequence mining, exon/intron prediction, is implemented by using PNN. The experimental results show the capability of solving biological sequence mining tasks using PNN

BACKGROUND: Detection of short, subtle conserved motif regions within a set of related DNA or amino acid sequences can lead to discoveries about important regulatory domains such as transcription factor and DNA binding sites as well as conserved protein domains. In order to help assess motif detection algorithms on motifs with varying properties and levels of conservation, we have developed a computational tool, rMotifGen, with the sole purpose of generating a number of random DNA or protein sequences containing short sequence motifs. Each motif consensus can be user-defined, randomly generated, or created from a position-specific scoring matrix (PSSM). Insertions and mutations within these motifs are created according to user-defined parameters and substitution matrices. The resulting sequences can be helpful in mutational simulations and in testing the limits of motif detection algorithms. RESULTS: Two implementations of rMotifGen have been created, one providing a graphical user interface (GUI) for random motif construction, and the other serving as a command line interface. The second implementation has the added advantages of platform independence and being able to be called in a batch mode. rMotifGen was used to construct sample sets of sequences containing DNA motifs and amino acid motifs that were then tested against the Gibbs sampler and MEME packages. CONCLUSION: rMotifGen provides an efficient and convenient method for creating random DNA or amino acid sequences with a variable number of motifs, where the instance of each motif can be incorporated using a position-specific scoring matrix (PSSM) or by creating an instance mutated from its corresponding consensus using an evolutionary model based on substitution matrices. rMotifGen is freely available at: http://bioinformatics.louisville.edu/brg/rMotifGen/.

Motif discovery, the process of discovering a meaningful pattern of nucleotides or amino acids that is shared by two or more molecules, is an important part of the study of gene function. In this paper, we propose a hybrid motif discovery approach based upon a combination of Particle Swarm Optimization (PSO) and the Expectation-Maximization (EM) algorithm. In the proposed algorithm, we use PSO to generate a seed for the EM algorithm.

BACKGROUND: Enhancements in sequencing technology have recently yielded assemblies of large genomes including rat, mouse, human, fruit fly, and zebrafish. The availability of large-scale genomic and genic sequence data coupled with advances in microarray technology have made it possible to study the expression of large numbers of sequence products under several different conditions in days where traditional molecular biology techniques might have taken months, or even years. Therefore, to efficiently study a number of gene products associated with a disease, pathway, or other biological process, it is necessary to be able to design primer pairs or oligonucleotides en masse rather than using a time consuming and laborious gene-by-gene method. RESULTS: We have developed an integrated system, MPrime, in order to efficiently calculate primer pairs or specific oligonucleotides for multiple genic regions based on a keyword, gene name, accession number, or sequence fasta format within the rat, mouse, human, fruit fly, and zebrafish genomes. A set of products created for mouse housekeeping genes from MPrime-designed primer pairs has been validated using both PCR-amplification and DNA sequencing. CONCLUSION: These results indicate MPrime accurately incorporates standard PCR primer design characteristics to produce high scoring primer pairs for genes of interest. In addition, sequence similarity for a set of oligonucleotides constructed for the same set of genes indicates high specificity in oligo design.