http://precedings.nature.com/tags/literature%20search Recently posted documents tagged with 'literature search' Nature Publishing Group en Nature Precedings http://precedings.nature.com/images/header_logo.gif http://precedings.nature.com http://dx.doi.org/10.1038/npre.2009.3775.1 Key assertions are extracted from “conclusions” sections of PubMed abstracts andconverted into Semantic Web / Linked Data format. The results are made accessible viafiles, a SPARQL endpoint, and a faceted search interface. Conclusion sections areidentified as valuable resources for machine-augmented key assertion identification andintegration in the biomedical domain. Results are discussed and opportunities for futurework and cooperation are highlighted. http://dx.doi.org/10.1038/npre.2009.3775.1 Mon, 21 Sep 2009 08:29:29 UTC Extracting conclusion sections from PubMed abstracts for rapid key assertion integration in biomedical research doi:10.1038/npre.2009.3775.1 2009-09-21 Nature Precedings 2009-09-21T08:29:29Z Manuscript Bioinformatics http://creativecommons.org/licenses/by/3.0/ http://dx.doi.org/10.1038/npre.2009.3122.1 Manual curation of biomedical data is highly accurate but time consuming, and does not scale with the ever increasing growth of biomedical literature. Text mining as a high-throughput computational technique scales well but requires human expertise to produce highly accurate results. Ontologies can help organizing large quantities of unstructured information. Here we present three systems, namely GoGene, GoPubMed and GoWeb, employing biomedical ontologies and show how they can assist manual curation of biomedical data.GoGene associates all genes from different model organisms to concepts of the Gene Ontology (GO) and the Medical Subject Headings (MeSH). The hierarchical structures of both terminologies support clustering and summarizing long lists of genes. Through the integration of known gene annotations from UniProt and EntrezGene with text-mined annotations from all abstracts in PubMed, GoGene currently contains up to 4,000,000 associations between genes and concepts from GO and MeSH for ten model organisms. The quality of all associations can be verified by following the links to their origin, that is, literature or database entries.GoPubMed aims at reducing the limitations of classical keyword search. It handles inconsistent vocabulary such as synonyms and specialized terminology. It shows the most relevant concepts in GO and MeSH for a search and thus reveals information which otherwise remains buried in the masses of text. This feature as well as the entire bibliography of all authors in PubMed facilitate comprehensive literature search. GoWeb translates these ideas to the World Wide Web and is thus not only limited to PubMed abstracts. GoWeb uses a standard web-search service and organizes search results based on GO, MeSH, and other concepts such as companies and institutions. http://dx.doi.org/10.1038/npre.2009.3122.1 Wed, 22 Apr 2009 21:14:18 UTC Ontology-based Assisted Curation of Biomedical Data doi:10.1038/npre.2009.3122.1 2009-04-22 Conrad Plake Nature Precedings 2009-04-22T21:14:18Z Presentation Bioinformatics http://creativecommons.org/licenses/by/3.0/ http://dx.doi.org/10.1038/npre.2009.3148.1 EU Directive 86/609/EEC for the protection of laboratory animals obliges scientists to consider whether a planned animal experiment can be replaced, reduced or refined (3Rs principle). To meet this regulatory obligation, scientists must consult the relevant scientific literature prior to any experimental study using laboratory animals. More than 50 million potentially 3Rs relevant documents are spread over the World Wide Web, biomedical literature and patent databases. In April 2008, the beta version of Go3R (www.Go3R.org), the first knowledge-based semantic search engine for alternative methods to animal experiments, was released. Go3R is free of charge and enables scientists and regulatory authorities involved in the planning, authorisation and performance of animal experiments to determine the availability of alternative methods in a fast and comprehensive manner. The technical basis of this search engine is specific 3Rs expert knowledge captured within the Go3R Ontology containing 87,218 labels and synonyms. A total of 16,620 concepts were structured in 28 branches, where 1,227 concepts were newly defined to specifically describe directly 3Rs relevant knowledge. Additionally relevant headings from MeSH where referenced to reflect the topics associated with the definition of Animal Testing Alternatives. Therefore it is distinguished between thematic-defining and directly 3Rs relevant branches. In addition to the assignment of direct parent-child relationships, further relationship types were introduced to allow to model 3Rs relevant domain knowledge. Examples for such knowledge are e.g. (1) the characteristics of cell culture tests methods, which usually utilize “specific cell types” or “cell lines” and are associated with a specific “endpoint” and “endpoint detection method” or (2) named test methods like “PREDISAFE™”, which replaces an animal test namely the “eye irritation test” in rabbits and uses specific cells namely “SIRC Cells” or (3) the “Haemagglutinin-Neuraminidase Protein Assay”, which detects a protein of the “Newcastle disease virus”. Thereby, an article in which e.g. a specific 3Rs method is not explicitly mentioned could still be recognized as relevant for the specific topic searched for in an indirect manner, for example if it mentions specific cells, endpoints or endpoint detection methods, which are relevant for the respective application. The search engine Go3R with its novel ontology is already well recognized by the 3Rs community and will be further maintained and developed. http://dx.doi.org/10.1038/npre.2009.3148.1 Wed, 22 Apr 2009 20:06:24 UTC An Ontology To Represent Knowledge On Animal Testing Alternatives doi:10.1038/npre.2009.3148.1 2009-04-23 Thomas Wächter Nature Precedings 2009-04-22T20:06:24Z Poster Bioinformatics http://creativecommons.org/licenses/by/3.0/ http://precedings.nature.com/documents/2101/version/2 With increasing number of bio-literature search engines, scientists and health professionals either make a subjective choice of tool(s) or face a challenge of analyzing multiple features of a plethora of bibliographic software. There is an urgent need for a thorough comparative analysis of the available literature scanning tools, from the user’s perspective. We report results of the first time semi-quantitative comparison of 21 search programs, which can search published (partial or full text) documents in life science areas. The observations can assist life science researchers and medical professionals to make an informed selection among the programs, depending on their search objectives. Some of the important findings are: 1. Most of the hits obtained from Scopus, ReleMed, EBImed, CiteXplore, and HighWire Press were usually relevant (i.e., these tools showed a better precision than other tools). 2. But a very high number of relevant citations were retrieved by HighWire Press, Google Scholar, CiteXplore and Pubmed Central (they had better recall). 3. HWP and CiteXplore seemed to have a good balance of precision and recall efficiencies. 4. PubMed Central, PubMed and Scopus provided the most useful query systems. 5. GoPubMed, BioAsk, EBIMed, ClusterMed could be more useful among the tools that can automatically process the retrieved citations for further scanning of bio-entities such as proteins, diseases, tissues, molecular interactions etc). The authors suggest the use of PubMed, Scopus, Google Scholar and HighWire Press – for better coverage, and GoPubMed – to view the hits categorized based on the MeSH and gene ontology terms. http://precedings.nature.com/documents/2101/version/2 Thu, 24 Jul 2008 15:38:26 UTC How do you choose your literature search tool(s)? hdl:10101/npre.2008.2101.2 2008-07-24 Kshitish K. Acharya Nature Precedings 2008-07-24T15:38:26Z Manuscript Bioinformatics http://creativecommons.org/licenses/by/3.0/ http://precedings.nature.com/documents/2101/version/1 With increasing number of bibliographic software, scientists and health professionals either make a subjective choice of tool(s) that could suit their needs or face a challenge of analyzing multiple features of a plethora of search programs. There is an urgent need for a thorough comparative analysis of the available bio-literature scanning tools, from the user’s perspective. We report results of the first time semi-quantitative comparison of 21 programs, which can search published (partial or full text) documents in life science areas. The observations can assist life science researchers and medical professionals to make an informed selection among the programs, depending on their search objectives. Some of the important findings are: 1. Most of the hits obtained from Scopus, ReleMed, EBImed, CiteXplore, and HighWire Press were usually relevant (i.e. these tools show a better precision than other tools). 2. But a very high number of relevant citations were retrieved by HighWire Press, Google Scholar, CiteXplore and Pubmed Central (they had better recall). 3. HWP and CiteXplore seemed to have a good balance of precision and recall efficiencies. 4. PubMed Central, PubMed and Scopus provided the most useful query systems. 5. GoPubMed, BioAsk, EBIMed, ClusterMed could be more useful among the tools that can automatically process the retrieved citations for further scanning of bio-entities such as proteins, diseases, tissues, molecular interactions, etc. The authors suggest the use of PubMed, Scopus, Google Scholar and HighWire Press – for better coverage, and GoPubMed – to view the hits categorized based on the MeSH and gene ontology terms. The article is relavant to all life science subjects. http://precedings.nature.com/documents/2101/version/1 Tue, 22 Jul 2008 10:50:58 UTC A comparative analysis of 21 literature search engines hdl:10101/npre.2008.2101.1 2008-07-22 Kshitish K. Acharya Nature Precedings 2008-07-22T10:50:58Z Manuscript Bioinformatics http://creativecommons.org/licenses/by/3.0/