MOTIVATION: To provide high quality, computationally tractable annotation of binding sites for biologically relevant (cognate) ligands in UniProtKB using the chemical ontology ChEBI (Chemical Entities of Biological Interest), to better support efforts to study and predict functionally relevant interactions between protein sequences and structures and small molecule ligands. RESULTS: We structured the data model for cognate ligand binding site annotations in UniProtKB and performed a complete reannotation of all cognate ligand binding sites using stable unique identifiers from ChEBI, which we now use as the reference vocabulary for all such annotations. We developed improved search and query facilities for cognate ligands in the UniProt website, REST API and SPARQL endpoint that leverage the chemical structure data, nomenclature and classification that ChEBI provides. AVAILABILITY AND IMPLEMENTATION: Binding site annotations for cognate ligands described using ChEBI are available for UniProtKB protein sequence records in several formats (text, XML and RDF) and are freely available to query and download through the UniProt website (www.uniprot.org), REST API (www.uniprot.org/help/api), SPARQL endpoint (sparql.uniprot.org/) and FTP site (https://ftp.uniprot.org/pub/databases/uniprot/). SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

A common assumption in multimodal learning is the completeness of training data, i.e., full modalities are available in all training examples. Although there exists research endeavor in developing novel methods to tackle the incompleteness of testing data, e.g., modalities are partially missing in testing examples, few of them can handle incomplete training modalities. The problem becomes even more challenging if considering the case of severely missing, e.g., ninety percent of training examples may have incomplete modalities. For the first time in the literature, this paper formally studies multimodal learning with missing modality in terms of flexibility (missing modalities in training, testing, or both) and efficiency (most training data have incomplete modality). Technically, we propose a new method named SMIL that leverages Bayesian meta-learning in uniformly achieving both objectives. To validate our idea, we conduct a series of experiments on three popular benchmarks: MM-IMDb, CMU-MOSI, and avMNIST. The results prove the state-of-the-art performance of SMIL over existing methods and generative baselines including autoencoders and generative adversarial networks.

In recent years, the US has experienced an opioid epidemic with an unprecedented number of drugs overdose deaths. Research finds such overdose deaths are linked to neighborhood-level traits, thus providing opportunity to identify effective interventions. Typically, techniques such as Ordinary Least Squares (OLS) or Maximum Likelihood Estimation (MLE) are used to document neighborhood-level factors significant in explaining such adverse outcomes. These techniques are, however, less equipped to ascertain non-linear relationships between confounding factors. Hence, in this study we apply machine learning based techniques to identify opioid risks of neighborhoods in Delaware and explore the correlation of these factors using Shapley Additive explanations (SHAP). We discovered that the factors related to neighborhoods’ environment, followed by education and then crime, were highly correlated with higher opioid risk. We also explored the change in these correlations over the years to understand the changing dynamics of the epidemic. Furthermore, we discovered that, as the epidemic has shifted from legal (i.e., prescription opioids) to illegal (e.g., heroin and fentanyl) drugs in recent years, the correlation of environment, crime and health related variables with the opioid risk has increased significantly while the correlation of economic and socio-demographic variables has decreased. The correlation of education related factors has been higher from the start and has increased slightly in recent years suggesting a need for increased awareness about the opioid epidemic.

SUMMARY: The global response to the COVID-19 pandemic has led to a rapid increase of scientific literature on this deadly disease. Extracting knowledge from biomedical literature and integrating it with relevant information from curated biological databases is essential to gain insight into COVID-19 etiology, diagnosis and treatment. We used Semantic Web technology RDF to integrate COVID-19 knowledge mined from literature by iTextMine, PubTator and SemRep with relevant biological databases and formalized the knowledge in a standardized and computable COVID-19 Knowledge Graph (KG). We published the COVID-19 KG via a SPARQL endpoint to support federated queries on the Semantic Web and developed a knowledge portal with browsing and searching interfaces. We also developed a RESTful API to support programmatic access and provided RDF dumps for download. AVAILABILITY AND IMPLEMENTATION: The COVID-19 Knowledge Graph is publicly available under CC-BY 4.0 license at https://research.bioinformatics.udel.edu/covid19kg/. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

BACKGROUND: Protein-protein interaction (PPI) prediction remains a central task in systems biology to achieve a better and holistic understanding of cellular and intracellular processes. Recently, an increasing number of computational methods have shifted from pair-wise prediction to network level prediction. Many of the existing network level methods predict PPIs under the assumption that the training network should be connected. However, this assumption greatly affects the prediction power and limits the application area because the current golden standard PPI networks are usually very sparse and disconnected. Therefore, how to effectively predict PPIs based on a training network that is sparse and disconnected remains a challenge. RESULTS: In this work, we developed a novel PPI prediction method based on deep learning neural network and regularized Laplacian kernel. We use a neural network with an autoencoder-like architecture to implicitly simulate the evolutionary processes of a PPI network. Neurons of the output layer correspond to proteins and are labeled with values (1 for interaction and 0 for otherwise) from the adjacency matrix of a sparse disconnected training PPI network. Unlike autoencoder, neurons at the input layer are given all zero input, reflecting an assumption of no a priori knowledge about PPIs, and hidden layers of smaller sizes mimic ancient interactome at different times during evolution. After the training step, an evolved PPI network whose rows are outputs of the neural network can be obtained. We then predict PPIs by applying the regularized Laplacian kernel to the transition matrix that is built upon the evolved PPI network. The results from cross-validation experiments show that the PPI prediction accuracies for yeast data and human data measured as AUC are increased by up to 8.4 and 14.9% respectively, as compared to the baseline. Moreover, the evolved PPI network can also help us leverage complementary information from the disconnected training network and multiple heterogeneous data sources. Tested by the yeast data with six heterogeneous feature kernels, the results show our method can further improve the prediction performance by up to 2%, which is very close to an upper bound that is obtained by an Approximate Bayesian Computation based sampling method. CONCLUSIONS: The proposed evolution deep neural network, coupled with regularized Laplacian kernel, is an effective tool in completing sparse and disconnected PPI networks and in facilitating integration of heterogeneous data sources.

The human genome contains a large number of protein polymorphisms due to individual genome variation. How many of these polymorphisms lead to altered protein-protein interaction is unknown. We have developed a method to address this question. The intersection of the SKEMPI database (of affinity constants among interacting proteins) and CAPRI 4.0 docking benchmark was docked using HADDOCK, leading to a training set of 166 mutant pairs. A random forest classifier based on the differences in resulting docking scores between the 166 mutant pairs and their wild-types was used, to distinguish between variants that have either completely or partially lost binding ability. Fifty percent of non-binders were correctly predicted with a false discovery rate of only 2 percent. The model was tested on a set of 15 HIV-1 - human, as well as seven human- human glioblastoma-related, mutant protein pairs: 50 percent of combined non-binders were correctly predicted with a false discovery rate of 10 percent. The model was also used to identify 10 protein-protein interactions between human proteins and their HIV-1 partners that are likely to be abolished by rare non-synonymous single-nucleotide polymorphisms (nsSNPs). These nsSNPs may represent novel and potentially therapeutically-valuable targets for anti-viral therapy by disruption of viral binding.

MOTIVATION: The enormous number of redundant sequenced genomes has hindered efforts to analyze and functionally annotate proteins. As the taxonomy of viruses is not uniformly defined, viral proteomes pose special challenges in this regard. Grouping viruses based on the similarity of their proteins at proteome scale can normalize against potential taxonomic nomenclature anomalies. RESULTS: We present Viral Reference Proteomes (Viral RPs), which are computed from complete virus proteomes within UniProtKB. Viral RPs based on 95, 75, 55, 35 and 15% co-membership in proteome similarity based clusters are provided. Comparison of our computational Viral RPs with UniProt's curator-selected Reference Proteomes indicates that the two sets are consistent and complementary. Furthermore, each Viral RP represents a cluster of virus proteomes that was consistent with virus or host taxonomy. We provide BLASTP search and FTP download of Viral RP protein sequences, and a browser to facilitate the visualization of Viral RPs. AVAILABILITY AND IMPLEMENTATION: http://proteininformationresource.org/rps/viruses/ CONTACT: [email protected] SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Reconstruction of PPI networks is a central task in systems biology, and inference from multiple heterogeneous data sources offers a promising computational approach to making de novo PPI prediction by leveraging complementary information and the partial network structure. However, how to quickly learn weights for heterogeneous data sources remains a challenge. In this work, we developed a method to infer de novo PPIs by combining multiple data sources represented in kernel format and obtaining optimal weights based on random walk over the existing partial network. Our proposed method utilizes Baker algorithm and the training data to construct a transition matrix which constrains how a random walk would traverse the partial network. Multiple heterogeneous features for the proteins in the network, including gene expression and Pfam domain profiles, are then combined into the form of a weighted kernel, which provides a new “adjacency matrix” for the whole network but is required to comply with the transition matrix on the part of the training subnetwork. This requirement is met by adjusting the weights to minimize the element-wise difference between the transition matrix and the weighted kernel. The minimization problem is solved by linear programming. The weighted kernel is then transformed to regularized Laplacian (RL) kernel to infer missing or new edges in the PPI network. The results on synthetic data and real data from Yeast show that the accuracy of PPI prediction measured by AUC is increased by up to 19% as compared to a control method without using optimal weights. Moreover, the weights learned by our method Weight Optimization by Linear Programming (WOLP) are very consistent with that learned by sampling, and can provide insights into the relations between PPIs and various feature kernels, thereby improving PPI prediction.

Abstract Motivation: UniRef databases provide full-scale clustering of UniProtKB sequences and are utilized for a broad range of applications, particularly similarity-based functional annotation. Non-redundancy and intra-cluster homogeneity in UniRef were recently improved by adding a sequence length overlap threshold. Our hypothesis is that these improvements would enhance the speed and sensitivity of similarity searches and improve the consistency of annotation within clusters. Results: Intra-cluster molecular function consistency was examined by analysis of Gene Ontology terms. Results show that UniRef clusters bring together proteins of identical molecular function in more than 97% of the clusters, implying that clusters are useful for annotation and can also be used to detect annotation inconsistencies. To examine coverage in similarity results, BLASTP searches against UniRef50 followed by expansion of the hit lists with cluster members demonstrated advantages compared with searches against UniProtKB sequences; the searches are concise (∼7 times shorter hit list before expansion), faster (∼6 times) and more sensitive in detection of remote similarities (>96% recall at e-value <0.0001). Our results support the use of UniRef clusters as a comprehensive and scalable alternative to native sequence databases for similarity searches and reinforces its reliability for use in functional annotation. Availability and implementation: Web access and file download from UniProt website at http://www.uniprot.org/uniref and ftp://ftp.uniprot.org/pub/databases/uniprot/uniref. BLAST searches against UniRef are available at http://www.uniprot.org/blast/ Contact: [email protected]

MicroRNAs (miRNAs) regulate a wide range of cellular and developmental processes through gene expression suppression or mRNA degradation. Experimentally validated miRNA gene targets are often reported in the literature. In this paper, we describe miRTex, a text mining system that extracts miRNA-target relations, as well as miRNA-gene and gene-miRNA regulation relations. The system achieves good precision and recall when evaluated on a literature corpus of 150 abstracts with F-scores close to 0.90 on the three different types of relations. We conducted full-scale text mining using miRTex to process all the Medline abstracts and all the full-length articles in the PubMed Central Open Access Subset. The results for all the Medline abstracts are stored in a database for interactive query and file download via the website at http://proteininformationresource.org/mirtex. Using miRTex, we identified genes potentially regulated by miRNAs in Triple Negative Breast Cancer, as well as miRNA-gene relations that, in conjunction with kinase-substrate relations, regulate the response to abiotic stress in Arabidopsis thaliana. These two use cases demonstrate the usefulness of miRTex text mining in the analysis of miRNA-regulated biological processes.

Revealing the underlying evolutionary mechanism plays an important role in understanding protein interaction networks in the cell. While many evolutionary models have been proposed, the problem about applying these models to real network data, especially for differentiating which model can better describe evolutionary process for the observed network remains a challenge. The traditional way is to use a model with presumed parameters to generate a network, and then evaluate the fitness by summary statistics, which however cannot capture the complete network structures information and estimate parameter distribution. In this work, we developed a novel method based on Approximate Bayesian Computation and modified Differential Evolution algorithm (ABC-DEP) that is capable of conducting model selection and parameter estimation simultaneously and detecting the underlying evolutionary mechanisms for PPI networks more accurately. We tested our method for its power in differentiating models and estimating parameters on simulated data and found significant improvement in performance benchmark, as compared with a previous method. We further applied our method to real data of protein interaction networks in human and yeast. Our results show duplication attachment model as the predominant evolutionary mechanism for human PPI networks and Scale-Free model as the predominant mechanism for yeast PPI networks.

We introduce RLIMS-P version 2.0, an enhanced rule-based information extraction (IE) system for mining kinase, substrate, and phosphorylation site information from scientific literature. Consisting of natural language processing and IE modules, the system has integrated several new features, including the capability of processing full-text articles and generalizability towards different post-translational modifications (PTMs). To evaluate the system, sets of abstracts and full-text articles, containing a variety of textual expressions, were annotated. On the abstract corpus, the system achieved F-scores of 0.91, 0.92, and 0.95 for kinases, substrates, and sites, respectively. The corresponding scores on the full-text corpus were 0.88, 0.91, and 0.92. It was additionally evaluated on the corpus of the 2013 BioNLP-ST GE task, and achieved an F-score of 0.87 for the phosphorylation core task, improving upon the results previously reported on the corpus. Full-scale processing of all abstracts in MEDLINE and all articles in PubMed Central Open Access Subset has demonstrated scalability for mining rich information in literature, enabling its adoption for biocuration and for knowledge discovery. The new system is generalizable and it will be adapted to tackle other major PTM types. RLIMS-P 2.0 online system is available online (http://proteininformationresource.org/rlimsp/) and the developed corpora are available from iProLINK (http://proteininformationresource.org/iprolink/).

BACKGROUND: Text mining is increasingly used in the biomedical domain because of its ability to automatically gather information from large amount of scientific articles. One important task in biomedical text mining is relation extraction, which aims to identify designated relations among biological entities reported in literature. A relation extraction system achieving high performance is expensive to develop because of the substantial time and effort required for its design and implementation. Here, we report a novel framework to facilitate the development of a pattern-based biomedical relation extraction system. It has several unique design features: (1) leveraging syntactic variations possible in a language and automatically generating extraction patterns in a systematic manner, (2) applying sentence simplification to improve the coverage of extraction patterns, and (3) identifying referential relations between a syntactic argument of a predicate and the actual target expected in the relation extraction task. RESULTS: A relation extraction system derived using the proposed framework achieved overall F-scores of 72.66% for the Simple events and 55.57% for the Binding events on the BioNLP-ST 2011 GE test set, comparing favorably with the top performing systems that participated in the BioNLP-ST 2011 GE task. We obtained similar results on the BioNLP-ST 2013 GE test set (80.07% and 60.58%, respectively). We conducted additional experiments on the training and development sets to provide a more detailed analysis of the system and its individual modules. This analysis indicates that without increasing the number of patterns, simplification and referential relation linking play a key role in the effective extraction of biomedical relations. CONCLUSIONS: In this paper, we present a novel framework for fast development of relation extraction systems. The framework requires only a list of triggers as input, and does not need information from an annotated corpus. Thus, we reduce the involvement of domain experts, who would otherwise have to provide manual annotations and help with the design of hand crafted patterns. We demonstrate how our framework is used to develop a system which achieves state-of-the-art performance on a public benchmark corpus.

Vast research efforts have been devoted to providing clinical diagnostic markers of myocardial infarction (MI), leading to over one million abstracts associated with "MI" and "Cardiovascular Diseases" in PubMed. Accumulation of the research results imposed a challenge to integrate and interpret these results. To address this problem and better understand how the left ventricle (LV) remodels post-MI at both the molecular and cellular levels, we propose here an integrative framework that couples computational methods and experimental data. We selected an initial set of MI-related proteins from published human studies and constructed an MI-specific protein-protein-interaction network (MIPIN). Structural and functional analysis of the MIPIN showed that the post-MI LV exhibited increased representation of proteins involved in transcriptional activity, inflammatory response, and extracellular matrix (ECM) remodeling. Known plasma or serum expression changes of the MIPIN proteins in patients with MI were acquired by data mining of the PubMed and UniProt knowledgebase, and served as a training set to predict unlabeled MIPIN protein changes post-MI. The predictions were validated with published results in PubMed, suggesting prognosticative capability of the MIPIN. Further, we established the first knowledge map related to the post-MI response, providing a major step towards enhancing our understanding of molecular interactions specific to MI and linking the molecular interaction, cellular responses, and biological processes to quantify LV remodeling.

SUMMARY: We have developed a new web application for peptide matching using Apache Lucene-based search engine. The Peptide Match service is designed to quickly retrieve all occurrences of a given query peptide from UniProt Knowledgebase (UniProtKB) with isoforms. The matched proteins are shown in summary tables with rich annotations, including matched sequence region(s) and links to corresponding proteins in a number of proteomic/peptide spectral databases. The results are grouped by taxonomy and can be browsed by organism, taxonomic group or taxonomy tree. The service supports queries where isobaric leucine and isoleucine are treated equivalent, and an option for searching UniRef100 representative sequences, as well as dynamic queries to major proteomic databases. In addition to the web interface, we also provide RESTful web services. The underlying data are updated every 4 weeks in accordance with the UniProt releases. AVAILABILITY: http://proteininformationresource.org/peptide.shtml. CONTACT: [email protected]. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

MOTIVATION: Prediction of protein-protein interaction has become an important part of systems biology in reverse engineering the biological networks for better understanding the molecular biology of the cell. Although significant progress has been made in terms of prediction accuracy, most computational methods only predict whether two proteins interact but not their interacting residues-the information that can be very valuable for understanding the interaction mechanisms and designing modulation of the interaction. In this work, we developed a computational method to predict the interacting residue pairs-contact matrix for interacting protein domains, whose rows and columns correspond to the residues in the two interacting domains respectively and whose values (1 or 0) indicate whether the corresponding residues (do or do not) interact. RESULTS: Our method is based on supervised learning using support vector machines. For each domain involved in a given domain-domain interaction (DDI), an interaction profile hidden Markov model (ipHMM) is first built for the domain family, and then each residue position for a member domain sequence is represented as a 20-dimension vector of Fisher scores, characterizing how similar it is as compared with the family profile at that position. Each element of the contact matrix for a sequence pair is now represented by a feature vector from concatenating the vectors of the two corresponding residues, and the task is to predict the element value (1 or 0) from the feature vector. A support vector machine is trained for a given DDI, using either a consensus contact matrix or contact matrices for individual sequence pairs, and is tested by leave-one-out cross validation. The performance averaged over a set of 115 DDIs collected from the 3 DID database shows significant improvement (sensitivity up to 85%, and specificity up to 85%), as compared with a multiple sequence alignment-based method (sensitivity 57%, and specificity 78%) previously reported in the literature. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

In this paper, we attempt to build a pipeline that identifies and extracts disease-drug relationships via sentence classification, and demonstrate the feasibility and utility of our approach using tardive dyskinesia as a case study. We manually developed and annotated a biomedicai training corpus for tardive dyskinesia. Using 10-fold cross validation, we tested and trained a naïve Bayes classifier to identify sentences pertaining to disease-drug relationships. Our precision, recall, and F-measure were all approximately 66%, and area under the ROC curve was over 80%. Our method helps to elucidate various drug effects on tardive dyskinesia and constitutes an initial effort toward the task of disease-drug relationship extraction.

Figures play an important role in illustrating concepts, methodology and results in biomedicai literature. However, figures in biomedicai literature are often composed of multiple subfigures (panels), which may illustrate diverse methodologies or results. Robust and accurate panel partitioning is crucial to support article categorization based on methods or experimental results and to provide the evidence source for derived assertions. But, it is a challenging task. In this paper, we present a comprehensive framework for harvesting multimodal panels in biomedicai literature, and demonstrate its application to protein-protein interaction (PPI)-related literature as a use case. A unique feature of our solution is that we combine pixel-level representations of images with figure captions. Our approach first analyzes figure captions to identify the label style used to mark panels. We then use pixel-level representations to partition a figure into a set of bounding boxes of connected components. We also perform a lexical analysis on the text within the figure to locate panel labels that match the caption analysis results. Finally, we estimate the optimal panel layout and use the layout to partition the figure. We tested our system on a dataset provided by the Molecular INTeraction database (MINT), and show that our approach surpasses pure caption-based and pure image-based approaches, achieving a 96.64% precision.

Text mining applications using natural language processing are often confronted with long and complicated sentences. This is observed particularly in the abstracts of scientific articles where authors summarize, in few sentences, the various facts described throughout the manuscript. Being rich in novel and important information, the abstract has been the primary target of biomedicai text mining applications. In this work, we aim to simplify complex sentences in abstracts of biomedicai text so that they can be readily processed by text mining applications. We focus on syntactic constructs that are frequently encountered in the biomedicai literature, such as coordinations, relative clauses, and appositions, with emphasis on their boundary detection. Our approach yielded good detection performance (average F-measure between 86.5% and 92.7%), and aided in improving biomedicai text mining applications, RLIMS-P and RankPref.

We present a new computational method for predicting ligand binding residues and functional sites in protein sequences. These residues and sites tend to be not only conserved, but also exhibit strong correlation due to the selection pressure during evolution in order to maintain the required structure and/or function. To explore the effect of correlations among multiple positions in the sequences, the method uses graph theoretic clustering and kernel-based canonical correlation analysis (kCCA) to identify binding and functional sites in protein sequences as the residues that exhibit strong correlation between the residues’ evolutionary characterization at the sites and the structure-based functional classification of the proteins in the context of a functional family. The results of testing the method on two well-curated data sets show that the prediction accuracy as measured by Receiver Operating Characteristic (ROC) scores improves significantly when multipositional correlations are accounted for.

Figures in biomedical articles often constitute direct evidence of experimental results. Image analysis methods can be coupled with text-based methods to improve knowledge discovery. However, automatically harvesting figures along with their associated captions from full-text articles remains challenging. In this paper, we present an automatic system for robustly harvesting figures from biomedical literature. Our approach relies on the idea that the PDF specification of the document layout can be used to identify encoded figures and figure boundaries within the PDF and enforce constraints among figure-regions. This allows us to harvest fragments of figures (subflgures), from the PDF, correctly identify subfigures that belong to the same figure, and identify the captions associated with each figure. Our method simultaneously recovers figures and captions and applies additional filtering process to remove irrelevant figures such as logos, to eliminate text passages that were incorrectly identified as captions, and to re-group subflgures to generate a putative figure. Finally, we associate figures with captions. Our preliminary experiments suggest that our method achieves an accuracy of 95% in harvesting figures-caption pairs from a set of 2,035 full-text biomedical documents from BioCreative III, containing 12,574 figures.

MOTIVATION: Identifier (ID) mapping establishes links between various biological databases and is an essential first step for molecular data integration and functional annotation. ID mapping allows diverse molecular data on genes and proteins to be combined and mapped to functional pathways and ontologies. We have developed comprehensive protein-centric ID mapping services providing mappings for 90 IDs derived from databases on genes, proteins, pathways, diseases, structures, protein families, protein interaction, literature, ontologies, etc. The services are widely used and have been regularly updated since 2006. AVAILABILITY: www.uniprot.org/mappingandproteininformation-resource.org/pirwww/search/idmapping.shtml CONTACT: [email protected].

BACKGROUND: The overall goal of the BioCreative Workshops is to promote the development of text mining and text processing tools which are useful to the communities of researchers and database curators in the biological sciences. To this end BioCreative I was held in 2004, BioCreative II in 2007, and BioCreative II.5 in 2009. Each of these workshops involved humanly annotated test data for several basic tasks in text mining applied to the biomedical literature. Participants in the workshops were invited to compete in the tasks by constructing software systems to perform the tasks automatically and were given scores based on their performance. The results of these workshops have benefited the community in several ways. They have 1) provided evidence for the most effective methods currently available to solve specific problems; 2) revealed the current state of the art for performance on those problems; 3) and provided gold standard data and results on that data by which future advances can be gauged. This special issue contains overview papers for the three tasks of BioCreative III. RESULTS: The BioCreative III Workshop was held in September of 2010 and continued the tradition of a challenge evaluation on several tasks judged basic to effective text mining in biology, including a gene normalization (GN) task and two protein-protein interaction (PPI) tasks. In total the Workshop involved the work of twenty-three teams. Thirteen teams participated in the GN task which required the assignment of EntrezGene IDs to all named genes in full text papers without any species information being provided to a system. Ten teams participated in the PPI article classification task (ACT) requiring a system to classify and rank a PubMed® record as belonging to an article either having or not having "PPI relevant" information. Eight teams participated in the PPI interaction method task (IMT) where systems were given full text documents and were required to extract the experimental methods used to establish PPIs and a text segment supporting each such method. Gold standard data was compiled for each of these tasks and participants competed in developing systems to perform the tasks automatically.BioCreative III also introduced a new interactive task (IAT), run as a demonstration task. The goal was to develop an interactive system to facilitate a user's annotation of the unique database identifiers for all the genes appearing in an article. This task included ranking genes by importance (based preferably on the amount of described experimental information regarding genes). There was also an optional task to assist the user in finding the most relevant articles about a given gene. For BioCreative III, a user advisory group (UAG) was assembled and played an important role 1) in producing some of the gold standard annotations for the GN task, 2) in critiquing IAT systems, and 3) in providing guidance for a future more rigorous evaluation of IAT systems. Six teams participated in the IAT demonstration task and received feedback on their systems from the UAG group. Besides innovations in the GN and PPI tasks making them more realistic and practical and the introduction of the IAT task, discussions were begun on community data standards to promote interoperability and on user requirements and evaluation metrics to address utility and usability of systems. CONCLUSIONS: In this paper we give a brief history of the BioCreative Workshops and how they relate to other text mining competitions in biology. This is followed by a synopsis of the three tasks GN, PPI, and IAT in BioCreative III with figures for best participant performance on the GN and PPI tasks. These results are discussed and compared with results from previous BioCreative Workshops and we conclude that the best performing systems for GN, PPI-ACT and PPI-IMT in realistic settings are not sufficient for fully automatic use. This provides evidence for the importance of interactive systems and we present our vision of how best to construct an interactive system for a GN or PPI like task in the remainder of the paper.

BACKGROUND: The BioCreative challenge evaluation is a community-wide effort for evaluating text mining and information extraction systems applied to the biological domain. The biocurator community, as an active user of biomedical literature, provides a diverse and engaged end user group for text mining tools. Earlier BioCreative challenges involved many text mining teams in developing basic capabilities relevant to biological curation, but they did not address the issues of system usage, insertion into the workflow and adoption by curators. Thus in BioCreative III (BC-III), the InterActive Task (IAT) was introduced to address the utility and usability of text mining tools for real-life biocuration tasks. To support the aims of the IAT in BC-III, involvement of both developers and end users was solicited, and the development of a user interface to address the tasks interactively was requested. RESULTS: A User Advisory Group (UAG) actively participated in the IAT design and assessment. The task focused on gene normalization (identifying gene mentions in the article and linking these genes to standard database identifiers), gene ranking based on the overall importance of each gene mentioned in the article, and gene-oriented document retrieval (identifying full text papers relevant to a selected gene). Six systems participated and all processed and displayed the same set of articles. The articles were selected based on content known to be problematic for curation, such as ambiguity of gene names, coverage of multiple genes and species, or introduction of a new gene name. Members of the UAG curated three articles for training and assessment purposes, and each member was assigned a system to review. A questionnaire related to the interface usability and task performance (as measured by precision and recall) was answered after systems were used to curate articles. Although the limited number of articles analyzed and users involved in the IAT experiment precluded rigorous quantitative analysis of the results, a qualitative analysis provided valuable insight into some of the problems encountered by users when using the systems. The overall assessment indicates that the system usability features appealed to most users, but the system performance was suboptimal (mainly due to low accuracy in gene normalization). Some of the issues included failure of species identification and gene name ambiguity in the gene normalization task leading to an extensive list of gene identifiers to review, which, in some cases, did not contain the relevant genes. The document retrieval suffered from the same shortfalls. The UAG favored achieving high performance (measured by precision and recall), but strongly recommended the addition of features that facilitate the identification of correct gene and its identifier, such as contextual information to assist in disambiguation. DISCUSSION: The IAT was an informative exercise that advanced the dialog between curators and developers and increased the appreciation of challenges faced by each group. A major conclusion was that the intended users should be actively involved in every phase of software development, and this will be strongly encouraged in future tasks. The IAT Task provides the first steps toward the definition of metrics and functional requirements that are necessary for designing a formal evaluation of interactive curation systems in the BioCreative IV challenge.

BACKGROUND: Representing species-specific proteins and protein complexes in ontologies that are both human- and machine-readable facilitates the retrieval, analysis, and interpretation of genome-scale data sets. Although existing protin-centric informatics resources provide the biomedical research community with well-curated compendia of protein sequence and structure, these resources lack formal ontological representations of the relationships among the proteins themselves. The Protein Ontology (PRO) Consortium is filling this informatics resource gap by developing ontological representations and relationships among proteins and their variants and modified forms. Because proteins are often functional only as members of stable protein complexes, the PRO Consortium, in collaboration with existing protein and pathway databases, has launched a new initiative to implement logical and consistent representation of protein complexes. DESCRIPTION: We describe here how the PRO Consortium is meeting the challenge of representing species-specific protein complexes, how protein complex representation in PRO supports annotation of protein complexes and comparative biology, and how PRO is being integrated into existing community bioinformatics resources. The PRO resource is accessible at http://pir.georgetown.edu/pro/. CONCLUSION: PRO is a unique database resource for species-specific protein complexes. PRO facilitates robust annotation of variations in composition and function contexts for protein complexes within and between species.

One of the aims of the Protein Structure Initiative (PSI) in the post genome-sequencing era is to elucidate biochemical and biophysical functions of each protein structure. Thus, the development of new methods for a large-scale analysis/annotation of protein functional residues is inevitable. Currently existing methods are not capable to do so due to the lack of automation, availability, and/or poor performance. In our previous work we were able to improve the accuracy of the prediction to ~86%, although the number of false-positives remained high. In this paper we present a fully-automated method for the prediction of catalytic residues in proteins that improves accuracy by reduction of false-positives, and is applicable for a large-scale analysis. Here, catalytic residues are predicted by machine learning approach followed by hierarchical analysis of the predicted residues. The capability of the method was tested on diverse family of hydrolytic enzymes with a/b hydrolase fold with widely differing phylogenetic origins and catalytic functions. The method was executed manually and then fully reproduces automatically. In the manual analysis, in 17 enzymes, the method correctly predicted all 3 residues of the catalytic triad with 3 false-positives out of 282 residues on average. Our method successfully eliminates the number of false-positives, while being applicable for a large-scale analysis of the protein function.

We present a new computational method for predicting ligand binding sites in protein sequences. The method uses kernelbased canonical correlation analysis and linear regression to identify binding sites in protein sequences as the residues that exhibit strong correlation between the residues' evolutionary characterization at the sites and the structure based functional classification of the proteins in the context of a functional family. We explore the effect of correlations among multiple positions in the sequences and show that their inclusion enhances the prediction accuracy significantly.

BACKGROUND: The Protein Ontology (PRO) is designed as a formal and principled Open Biomedical Ontologies (OBO) Foundry ontology for proteins. The components of PRO extend from a classification of proteins on the basis of evolutionary relationships at the homeomorphic level to the representation of the multiple protein forms of a gene, including those resulting from alternative splicing, cleavage and/or post-translational modifications. Focusing specifically on the TGF-beta signaling proteins, we describe the building, curation, usage and dissemination of PRO. RESULTS: PRO is manually curated on the basis of PrePRO, an automatically generated file with content derived from standard protein data sources. Manual curation ensures that the treatment of the protein classes and the internal and external relationships conform to the PRO framework. The current release of PRO is based upon experimental data from mouse and human proteins wherein equivalent protein forms are represented by single terms. In addition to the PRO ontology, the annotation of PRO terms is released as a separate PRO association file, which contains, for each given PRO term, an annotation from the experimentally characterized sub-types as well as the corresponding database identifiers and sequence coordinates. The annotations are added in the form of relationship to other ontologies. Whenever possible, equivalent forms in other species are listed to facilitate cross-species comparison. Splice and allelic variants, gene fusion products and modified protein forms are all represented as entities in the ontology. Therefore, PRO provides for the representation of protein entities and a resource for describing the associated data. This makes PRO useful both for proteomics studies where isoforms and modified forms must be differentiated, and for studies of biological pathways, where representations need to take account of the different ways in which the cascade of events may depend on specific protein modifications. CONCLUSION: PRO provides a framework for the formal representation of protein classes and protein forms in the OBO Foundry. It is designed to enable data retrieval and integration and machine reasoning at the molecular level of proteins, thereby facilitating cross-species comparisons, pathway analysis, disease modeling and the generation of new hypotheses.

OBJECTIVES: Biomedical named entity recognition (BNER) is a critical component in automated systems that mine biomedical knowledge in free text. Among different types of entities in the domain, gene/protein would be the most studied one for BNER. Our goal is to develop a gene/protein name recognition system BioTagger-GM that exploits rich information in terminology sources using powerful machine learning frameworks and system combination. DESIGN: BioTagger-GM consists of four main components: (1) dictionary lookup-gene/protein names in BioThesaurus and biomedical terms in UMLS Metathesaurus are tagged in text, (2) machine learning-machine learning systems are trained using dictionary lookup results as one type of feature, (3) post-processing-heuristic rules are used to correct recognition errors, and (4) system combination-a voting scheme is used to combine recognition results from multiple systems. MEASUREMENTS: The BioCreAtIvE II Gene Mention (GM) corpus was used to evaluate the proposed method. To test its general applicability, the method was also evaluated on the JNLPBA corpus modified for gene/protein name recognition. The performance of the systems was evaluated through cross-validation tests and measured using precision, recall, and F-Measure. RESULTS: BioTagger-GM achieved an F-Measure of 0.8887 on the BioCreAtIvE II GM corpus, which is higher than that of the first-place system in the BioCreAtIvE II challenge. The applicability of the method was also confirmed on the modified JNLPBA corpus. CONCLUSION: The results suggest that terminology sources, powerful machine learning frameworks, and system combination can be integrated to build an effective BNER system.

The ever-increasing scientific literature and the exponential growth of large-scale molecular data have prompted active research in biological text mining to facilitate literature-based curation of molecular databases. Meanwhile, systems biology and bio-ontologies are emerging as critical tools in biological research where complex data in disparate resources are generated, integrated and analyzed. Both rely on literature for data annotation and analysis. The challenges facing us are to develop broadly utilized text mining tools and systems, and to bring together developer and user communities for system development and evaluation. We describe a framework for linking text mining tools with ontology and systems biology, extending from a previously developed text mining resource, iProLINK. We focus on molecular and ontological resources, including genes/proteins, protein-protein interaction (PPI), and Protein Ontology. The framework consists of two major components: a user interface for text mining of PPI from an integrated tool server and software modules to allow text mining outputs to be created, ranked, and used by the community. Use cases are presented for assessing the gaps and making recommendations for future development.

Due to the heightened concern about bioterrorism and emerging/reemerging infectious diseases, a flood of molecular data about human pathogens has been generated and maintained in disparate databases. However, scientific findings regarding these pathogens and their host responses are buried in the growing volume of biomedical literature and there is an urgent need to mine information pertaining to pathogenesis-related proteins especially host-pathogen protein-protein interactions from literature. In this paper, we report our exploration of developing an automated system to identify MEDLINE abstracts referring to host-pathogen protein-protein interactions. An annotated corpus consisting of 1,360 MEDLINE abstracts was generated. With this corpus, we developed and evaluated document classification systems using support vector machines (SVMs). We also investigated the effects of feature selection using the information gain (IG) measure. Document classification systems were designed at two levels, abstract-level and sentence-level. We observed that feature selection was effective not only in reducing the dimensionality of features to build a compact system, but also in improving document classification performance. We also observed abstract-level systems and sentence-level systems yielded different classification of MEDLINE abstracts, and the combination of these systems could improve the overall document classification.

MOTIVATION: Our purpose is to develop a statistical modeling approach for cancer biomarker discovery and provide new insights into early cancer detection. We propose the concept of dependence network, apply it for identifying cancer biomarkers, and study the difference between the protein or gene samples from cancer and non-cancer subjects based on mass-spectrometry (MS) and microarray data. RESULTS: Three MS and two gene microarray datasets are studied. Clear differences are observed in the dependence networks for cancer and non-cancer samples. Protein/gene features are examined three at one time through an exhaustive search. Dependence networks are constructed by binding triples identified by the eigenvalue pattern of the dependence model, and are further compared to identify cancer biomarkers. Such dependence-network-based biomarkers show much greater consistency under 10-fold cross-validation than the classification-performance-based biomarkers. Furthermore, the biological relevance of the dependence-network-based biomarkers using microarray data is discussed. The proposed scheme is shown promising for cancer diagnosis and prediction. AVAILABILITY: See supplements: http://dsplab.eng.umd.edu/~genomics/dependencenetwork/

MOTIVATION: Redundant protein sequences in biological databases hinder sequence similarity searches and make interpretation of search results difficult. Clustering of protein sequence space based on sequence similarity helps organize all sequences into manageable datasets and reduces sampling bias and overrepresentation of sequences. RESULTS: The UniRef (UniProt Reference Clusters) provide clustered sets of sequences from the UniProt Knowledgebase (UniProtKB) and selected UniProt Archive records to obtain complete coverage of sequence space at several resolutions while hiding redundant sequences. Currently covering >4 million source sequences, the UniRef100 database combines identical sequences and subfragments from any source organism into a single UniRef entry. UniRef90 and UniRef50 are built by clustering UniRef100 sequences at the 90 or 50% sequence identity levels. UniRef100, UniRef90 and UniRef50 yield a database size reduction of approximately 10, 40 and 70%, respectively, from the source sequence set. The reduced redundancy increases the speed of similarity searches and improves detection of distant relationships. UniRef entries contain summary cluster and membership information, including the sequence of a representative protein, member count and common taxonomy of the cluster, the accession numbers of all the merged entries and links to rich functional annotation in UniProtKB to facilitate biological discovery. UniRef has already been applied to broad research areas ranging from genome annotation to proteomics data analysis. AVAILABILITY: UniRef is updated biweekly and is available for online search and retrieval at http://www.uniprot.org, as well as for download at ftp://ftp.uniprot.org/pub/databases/uniprot/uniref. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Biomedical ontologies are emerging as critical tools in genomic and proteomic research, where complex data in disparate resources need to be integrated. A number of ontologies describe properties that can be attributed to proteins. For example, protein functions are described by the Gene Ontology (GO) and human diseases by SNOMED CT or ICD10. There is, however, a gap in the current set of ontologies - one that describes the protein entities themselves and their relationships. We have designed the PRotein Ontology (PRO) to facilitate protein annotation and to guide new experiments. The components of PRO extend from the classification of proteins on the basis of evolutionary relationships to the representation of the multiple protein forms of a gene (products generated by genetic variation, alternative splicing, proteolytic cleavage, and other post-translational modifications). PRO will allow the specification of relationships between PRO, GO and other ontologies in the OBO Foundry. Here we describe the initial development of PRO, illustrated using human and mouse proteins involved in the transforming growth factor-beta and bone morphogenetic protein signaling pathways.

MOTIVATION: With more and more research dedicated to literature mining in the biomedical domain, more and more systems are available for people to choose from when building literature mining applications. In this study, we focus on one specific kind of literature mining task, i.e., detecting definitions of acronyms, abbreviations, and symbols in biomedical text. We denote acronyms, abbreviations, and symbols as short forms (SFs) and their corresponding definitions as long forms (LFs). The study was designed to answer the following questions; i) how well a system performs in detecting LFs from novel text, ii) what the coverage is for various terminological knowledge bases in including SFs as synonyms of their LFs, and iii) how to combine results from various SF knowledge bases. METHOD: We evaluated the following three publicly available detection systems in detecting LFs for SFs: i) a handcrafted pattern/rule based system by Ao and Takagi, ALICE, ii) a machine learning system by Chang et al., and iii) a simple alignment-based program by Schwartz and Hearst. In addition, we investigated the conceptual coverage of two terminological knowledge bases: i) the UMLS (the Unified Medical Language System), and ii) the BioThesaurus (a thesaurus of names for all UniProt protein records). We also implemented a web interface that provides a virtual integration of various SF knowledge bases. RESULTS: We found that detection systems agree with each other on most cases, and the existing terminological knowledge bases have a good coverage of synonymous relationship for frequently defined LFs. The web interface allows people to detect SF definitions from text and to search several SF knowledge bases. AVAILABILITY: The web site is http://gauss.dbb.georgetown.edu/liblab/SFThesaurus.

MOTIVATION: Attribute selection is a critical step in development of document classification systems. As a standard practice, words are stemmed and the most informative ones are used as attributes in classification. Owing to high complexity of biomedical terminology, general-purpose stemming algorithms are often conservative and could also remove informative stems. This can lead to accuracy reduction, especially when the number of labeled documents is small. To address this issue, we propose an algorithm that omits stemming and, instead, uses the most discriminative substrings as attributes. RESULTS: The approach was tested on five annotated sets of abstracts from iProLINK that report on the experimental evidence about five types of protein post-translational modifications. The experiments showed that Naive Bayes and support vector machine classifiers perform consistently better [with area under the ROC curve (AUC) accuracy in range 0.92-0.97] when using the proposed attribute selection than when using attributes obtained by the Porter stemmer algorithm (AUC in 0.86-0.93 range). The proposed approach is particularly useful when labeled datasets are small.

UNLABELLED: BioThesaurus is a web-based system designed to map a comprehensive collection of protein and gene names to protein entries in the UniProt Knowledgebase. Currently covering more than two million proteins, BioThesaurus consists of over 2.8 million names extracted from multiple molecular biological databases according to the database cross-references in iProClass. The BioThesaurus web site allows the retrieval of synonymous names of given protein entries and the identification of protein entries sharing the same names. AVAILABILITY: BioThesaurus is accessible for online searching at http://pir.georgetown.edu/iprolink/biothesaurus

A web-based version of the RLIMS-P literature mining system was developed for online mining of protein phosphorylation information from MEDLINE abstracts. The online tool presents extracted phosphorylation objects (phosphorylated proteins, phosphorylation sites and protein kinases) in summary tables and full reports with evidence-tagged abstracts. The tool further allows mapping of phosphorylated proteins to protein entries in the UniProt Knowledgebase based on PubMed ID and/or protein name. The literature mining, coupled with database association, allows retrieval of rich biological information for the phosphorylated proteins and facilitates database annotation of phosphorylation features.

BACKGROUND: The number of protein sequences deriving from genome sequencing projects is outpacing our knowledge about the function of these proteins. With the gap between experimentally characterized and uncharacterized proteins continuing to widen, it is necessary to develop new computational methods and tools for functional prediction. Knowledge of catalytic sites provides a valuable insight into protein function. Although many computational methods have been developed to predict catalytic residues and active sites, their accuracy remains low, with a significant number of false positives. In this paper, we present a novel method for the prediction of catalytic sites, using a carefully selected, supervised machine learning algorithm coupled with an optimal discriminative set of protein sequence conservation and structural properties. RESULTS: To determine the best machine learning algorithm, 26 classifiers in the WEKA software package were compared using a benchmarking dataset of 79 enzymes with 254 catalytic residues in a 10-fold cross-validation analysis. Each residue of the dataset was represented by a set of 24 residue properties previously shown to be of functional relevance, as well as a label {+1/-1} to indicate catalytic/non-catalytic residue. The best-performing algorithm was the Sequential Minimal Optimization (SMO) algorithm, which is a Support Vector Machine (SVM). The Wrapper Subset Selection algorithm further selected seven of the 24 attributes as an optimal subset of residue properties, with sequence conservation, catalytic propensities of amino acids, and relative position on protein surface being the most important features. CONCLUSION: The SMO algorithm with 7 selected attributes correctly predicted 228 of the 254 catalytic residues, with an overall predictive accuracy of more than 86%. Missing only 10.2% of the catalytic residues, the method captures the fundamental features of catalytic residues and can be used as a "catalytic residue filter" to facilitate experimental identification of catalytic residues for proteins with known structure but unknown function.

OBJECTIVE: Natural language processing (NLP) approaches have been explored to manage and mine information recorded in biological literature. A critical step for biological literature mining is biological named entity tagging (BNET) that identifies names mentioned in text and normalizes them with entries in biological databases. The aim of this study was to provide quantitative assessment of the complexity of BNET on protein entities through BioThesaurus, a thesaurus of gene/protein names for UniProt knowledgebase (UniProtKB) entries that was acquired using online resources. METHODS: We evaluated the complexity through several perspectives: ambiguity (i.e., the number of genes/proteins represented by one name), synonymy (i.e., the number of names associated with the same gene/protein), and coverage (i.e., the percentage of gene/protein names in text included in the thesaurus). We also normalized names in BioThesaurus and measures were obtained twice, once before normalization and once after. RESULTS: The current version of BioThesaurus has over 2.6 million names or 2.1 million normalized names covering more than 1.8 million UniProtKB entries. The average synonymy is 3.53 (2.86 after normalization), ambiguity is 2.31 before normalization and 2.32 after, while the coverage is 94.0% based on the BioCreAtive data set comprising MEDLINE abstracts containing genes/proteins. CONCLUSION: The study indicated that names for genes/proteins are highly ambiguous and there are usually multiple names for the same gene or protein. It also demonstrated that most gene/protein names appearing in text can be found in BioThesaurus.

Motivation: A large volume of experimental data on protein phosphorylation is buried in the fast-growing PubMed literature. While of great value, such information is limited in databases owing to the laborious process of literature-based curation. Computational literature mining holds promise to facilitate database curation. Results: A rule-based system, RLIMS-P (Rule-based LIterature Mining System for Protein Phosphorylation), was used to extract protein phosphorylation information from MEDLINE abstracts. An annotation-tagged literature corpus developed at PIR was used to evaluate the system for finding phosphorylation papers and extracting phosphorylation objects (kinases, substrates and sites) from abstracts. RLIMS-P achieved a precision and recall of 91.4 and 96.4% for paper retrieval, and of 97.9 and 88.0% for extraction of substrates and sites. Coupling the high recall for paper retrieval and high precision for information extraction, RLIMS-P facilitates literature mining and database annotation of protein phosphorylation. Availability: The program is available on request from the authors. The phosphorylation patterns and datasets used in this study are available at http://pir.georgetown.edu/iprolink/ Contact: [email protected]

BACKGROUND: A large volume of data and information about genes and gene products has been stored in various molecular biology databases. A major challenge for knowledge discovery using these databases is to identify related genes and gene products in disparate databases. The development of Gene Ontology (GO) as a common vocabulary for annotation allows integrated queries across multiple databases and identification of semantically related genes and gene products (i.e., genes and gene products that have similar GO annotations). Meanwhile, dozens of tools have been developed for browsing, mining or editing GO terms, their hierarchical relationships, or their "associated" genes and gene products (i.e., genes and gene products annotated with GO terms). Tools that allow users to directly search and inspect relations among all GO terms and their associated genes and gene products from multiple databases are needed. RESULTS: We present a standalone package called DynGO, which provides several advanced functionalities in addition to the standard browsing capability of the official GO browsing tool (AmiGO). DynGO allows users to conduct batch retrieval of GO annotations for a list of genes and gene products, and semantic retrieval of genes and gene products sharing similar GO annotations. The result are shown in an association tree organized according to GO hierarchies and supported with many dynamic display options such as sorting tree nodes or changing orientation of the tree. For GO curators and frequent GO users, DynGO provides fast and convenient access to GO annotation data. DynGO is generally applicable to any data set where the records are annotated with GO terms, as illustrated by two examples. CONCLUSION: We have presented a standalone package DynGO that provides functionalities to search and browse GO and its association databases as well as several additional functions such as batch retrieval and semantic retrieval. The complete documentation and software are freely available for download from the website http://biocreative.ifsm.umbc.edu/dyngo.

BACKGROUND: The identification of unique proteins at different taxonomic levels has both scientific and practical value. Strain-, species- and genus-specific proteins can provide insight into the criteria that define an organism and its relationship with close relatives. Such proteins can also serve as taxon-specific diagnostic targets. DESCRIPTION: A pipeline using a combination of computational and manual analyses of BLAST results was developed to identify strain-, species-, and genus-specific proteins and to catalog the closest sequenced relative for each protein in a proteome. Proteins encoded by a given strain are preliminarily considered to be unique if BLAST, using a comprehensive protein database, fails to retrieve (with an e-value better than 0.001) any protein not encoded by the query strain, species or genus (for strain-, species- and genus-specific proteins respectively), or if BLAST, using the best hit as the query (reverse BLAST), does not retrieve the initial query protein. Results are manually inspected for homology if the initial query is retrieved in the reverse BLAST but is not the best hit. Sequences unlikely to retrieve homologs using the default BLOSUM62 matrix (usually short sequences) are re-tested using the PAM30 matrix, thereby increasing the number of retrieved homologs and increasing the stringency of the search for unique proteins. The above protocol was used to examine several food- and water-borne pathogens. We find that the reverse BLAST step filters out about 22% of proteins with homologs that would otherwise be considered unique at the genus and species levels. Analysis of the annotations of unique proteins reveals that many are remnants of prophage proteins, or may be involved in virulence. The data generated from this study can be accessed and further evaluated from the CUPID (Core and Unique Protein Identification) system web site (updated semi-annually) at http://pir.georgetown.edu/cupid. CONCLUSION: CUPID provides a set of proteins specific to a genus, species or a strain, and identifies the most closely related organism.

Abstract We review recent results in literature data mining for biology and discuss the need and the steps for a challenge evaluation for this field. Literature data mining has progressed from simple recognition of terms to extraction of interaction relationships from complex sentences, and has broadened from recognition of protein interactions to a range of problems such as improving homology search, identifying cellular location, and so on. To encourage participation and accelerate progress in this expanding field, we propose creating challenge evaluations, and we describe two specific applications in this context. Contact: [email protected]@[email protected]@[email protected] * To whom correspondence should be addressed.

Presents new techniques for biosequence classification, with a focus on recognizing E. Coli promoters in DNA. Specifically, given an unlabeled DNA sequence S, we want to determine whether or not S is an E. Coli promoter. We use an expectation maximization (EM) algorithm to locate the -35 and -10 binding sites in an E. Coli promoter sequence. The EM algorithm differs from previously published EM algorithms in that, instead of assuming a uniform distribution for the lengths of the spacer between the -35 binding site and the -10 binding site as well as between the -10 binding site and the transcriptional start site, our algorithm deduces the probability distribution for these lengths. Based on the located binding sites, we select features in each E. Coli promoter sequence according to their information contents and represent the features using an orthogonal encoding method. We then feed the features to a neural network for promoter recognition. Empirical studies show that the proposed approach achieves good performance on different data sets.

Biological data mining aims to extract signi cant information from DNA, RNA and proteins.The signi cant information may refer to motifs, functional sites, clustering and classi cation rules.This paper presents an example of biological data mining: the classi cation of protein sequences using neural netw orks.We proposenew tec hniques to extract features from protein data and use them in combination with the Ba yesianneural network to classify protein sequences obtained from the PIR protein database maintained at the National Biomedical Research F oundation.T o evaluate the performance of the proposed approach, we c o mpare it with other protein classi ers built based on sequence alignment and machine learning methods.Experimental results sho w the high precision of the proposed classi er and the complementarity of the tools studied in the paper.

UNLABELLED: The Protein Information Resource (PIR) has greatly expanded its Web site and developed a set of interactive search and analysis tools to facilitate the analysis, annotation, and functional identification of proteins. New search engines have been implemented to combine sequence similarity search results with database annotation information. The new PIR search systems have proved very useful in providing enriched functional annotation of protein sequences, determining protein superfamily-domain relationships, and detecting annotation errors in genomic database archives. AVAILABILITY: http://pir.georgetown.edu/. CONTACT: [email protected]

UNLABELLED: An integrated database and search system has been developed for protein family identification and information retrieval, as an approach to undertake the highly complex, genomic-scale problem of molecular sequence database search and organization. AVAILABILITY: http://diana.uthct.edu CONTACT: [email protected]

A new method, the motif identification neural design (MOTIFIND), has been developed for rapid and sensitive protein family identification. The method is an extension of our previous gene classification artificial neural system and employs new designs to enhance the detection of distant relationships. The new designs include an n-gram term weighting algorithm for extracting local motif patterns, an enhanced n-gram method for extracting residues of long-range correlation, and integrated neural networks for combining global and motif sequence information. The system has been tested and compared with several existing methods using three protein families, the cytochrome c, cytochrome b and flavodoxin. Overall it achieves 100% sensitivity and > 99.6% specificity, an accuracy comparable to BLAST, but at a speed of approximately 20 times faster. The system is much more robust than the PROSITE search which is based on simple signature patterns. MOTIFIND also compares favorably with BLIMPS, the Hidden Markov Model and PROFILESEARCH in detecting fragmentary sequences lacking complete motif regions and in detecting distant relationships, especially for members of under-represented subgroups within a family. MOTIFIND may be generally applicable to other proteins and has the potential to become a full-scale database search and sequence analysis tool.

A neural network classification method has been developed as an alternative approach to the search/organization problem of large modular databases. Two artificial neural systems have been implemented on a Cray for rapid protein/nucleic acid classification of unknown sequences. The system employs a n-gram hashing function for sequence encoding and modular backpropagation networks for classification. The protein system has achieved a 82 to 100% sensitivity at a speed that is about an order of magnitude faster than other search methods. With the rapid accumulation of sequences, the saving in time will become increasingly significant. The pilot nucleic acid system showed a 96% classification accuracy. The software tool would be valuable for the organization of molecular sequence databases and is generally applicable to any databases that are organized according to family relationships.>

An approach to the prediction of protein structure/function from primary sequence that is a direct protein/domain classification using a neural network is presented. The protein classification system, PROCANS, is able to classify unknown proteins to protein/domain classes by imbedding a domain database in a back-propagation network. The initial results of using PROCANS for 30 training records are discussed. The system converged in 89 Cray CPU seconds during training, and the classification was rapid and accurate for all unknown cases tested. The algorithm differs from many other methods in that it does not require an exhaustive database search or a primary sequence alignment for sequence comparisons. These features make it possible to scale-up the system to embed the entire domain database and classify at least 1000 protein/domain classes