Introducing meta-services for biomedical information extraction

Leitner, Florian; Krallinger, Martin; Rodríguez-Penagos, Carlos; Hakenberg, Jörg; Plake, Conrad; Kuo, Chung‐Hao; Hsu, Chun‐Nan; Tsai, Richard Tzong-Han; Hung, Hsi-Chuan; Lau, William; Johnson, Calvin A.; Sætre, Rune; Yoshida, Kazuhiro; Chen, Yanhua; Kim, Sun Chang; Shin, Soo-Yong; Zhang, Byoung-Tak; Baumgartner, William A.; Hunter, Lawrence; Haddow, Barry; Matthews, Michael; Wang, Xinglong; Ruch, Patrick; Ehrler, Frédéric; Özgür, Arzucan; Erkan, Güneş; Radev, Dragomir; Krauthammer, Michael; Luong, ThaiBinh; Hoffmann, Robert; Sander, Chris; Valencia, Alfonso

doi:10.1186/gb-2008-9-s2-s6

Cited by 61 publications

(56 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…1). To this end, we implemented a specialized version of the BioCreative Meta-Server [14]. Through this platform, participants could register their classification servers ("annotation servers") via a Web interface (up to five annotation servers per participating group), instruct the BCMS to send them annotation requests together with the next article in the queue, and review the current state of the process, including error and technical problem reports.…”

Section: The Biocreative Meta-servermentioning

confidence: 99%

See 1 more Smart Citation

An Overview of BioCreative II.5

Leitner

Mardis

Krallinger

et al. 2010

IEEE/ACM Trans. Comput. Biol. and Bioinf.

View full text Add to dashboard Cite

Abstract-We present the results of the BioCreative II.5 evaluation in association with the FEBS Letters experiment, where authors created Structured Digital Abstracts to capture information about protein-protein interactions. The BioCreative II.5 challenge evaluated automatic annotations from 15 text mining teams based on a gold standard created by reconciling annotations from curators, authors, and automated systems. The tasks were to rank articles for curation based on curatable protein-protein interactions; to identify the interacting proteins (using UniProt identifiers) in the positive articles (61); and to identify interacting protein pairs. There were 595 fulltext articles in the evaluation test set, including those both with and without curatable protein interactions. The principal evaluation metrics were the interpolated area under the precision/recall curve (AUC iP/R), and (balanced) F-measure. For article classification, the best AUC iP/R was 0.70; for interacting proteins, the best system achieved good macroaveraged recall (0.73) and interpolated area under the precision/recall curve (0.58), after filtering incorrect species and mapping homonymous orthologs; for interacting protein pairs, the top (filtered, mapped) recall was 0.42 and AUC iP/R was 0.29. Ensemble systems improved performance for the interacting protein task.

show abstract

Section: The Biocreative Meta-servermentioning

confidence: 99%

“…This motivated the implementation of the BioCreative MetaServer platform (BCMS) [14], which is able to display and integrate multiple predictions from various annotation servers based on Web services and accessible to end users via a Web interface.…”

Section: Introductionmentioning

confidence: 99%

An Overview of BioCreative II.5

Leitner

Mardis

Krallinger

et al. 2010

IEEE/ACM Trans. Comput. Biol. and Bioinf.

View full text Add to dashboard Cite

show abstract

“…The BioCreative MetaServer (BCMS) [12] is a textmining platform that was built following the blueprint of similar and highly successful result aggregation platforms in Bioinformatics, such as structure prediction services [5]. The common principle is unifying results from various algorithms running on the same task but in a distributed manner, thereby providing a collated view on the data that enables the service's user to directly compare and access different results.…”

Section: Bcms the Biocreative Metaservermentioning

confidence: 99%

“…For requests made to the BCMS, these arrays are grouped by Annotation Server and forwarded to the client. Optionally, the raw results can also be viewed in a web browser [12].…”

Section: Bcmsmentioning

confidence: 99%

SOA-Based Integration of Text Mining Services

Starlinger

Leitner

Valencia

et al. 2009

2009 Congress on Services - I

Self Cite

View full text Add to dashboard Cite

show abstract

“…For BioCreative II.5, fifteen participating research groups returned 134 submissions in online and offline scenarios: the online scenario required each team to set up CASP-like annotations servers [9], [18], which would take a previously unseen query document and return predictions within ten minutes per document; the offline scenario allowed for bulk download and annotation of a set of texts, with a one week time constraint for returning predictions. In this paper, we describe our approaches to extract protein-protein interactions from fulltext articles (IPT task) and to map individual proteins to their respective UniProt identifiers (INT).…”

Section: Introductionmentioning

confidence: 99%

Efficient Extraction of Protein-Protein Interactions from Full-Text Articles

Hakenberg¹,

Leaman²,

Vo³

et al. 2010

IEEE/ACM Trans. Comput. Biol. and Bioinf.

View full text Add to dashboard Cite

Abstract-Proteins and their interactions govern virtually all cellular processes, such as regulation, signaling, metabolism, and structure. Most experimental findings pertaining to such interactions are discussed in research articles, which in turn get curated by protein interaction databases. Authors, editors, and publishers benefit from efforts to alleviate the tasks of searching for relevant articles, evidence for physical interactions, and proper identifiers for each protein involved. The BioCreative II.5 community challenge addressed these tasks in a competition-style assessment, to evaluate and compare different methodologies, to make aware of the increasing accuracy of automated methods, and to guide future implementations. In this paper, we present our approaches for protein named entity recognition including normalization, and for extraction of proteinprotein interactions from full text. Our overall goal is to identify efficient individual components, and we compare various compositions to handle a single full-text article in between ten seconds and two minutes. We propose strategies to transfer document-level annotations to the sentence-level, which allows for the creation of a more fine-grained training corpus; we use this corpus to automatically derive around 5000 patterns. We rank sentences by relevance to the task of finding novel interactions with physical evidence, using a sentence classifier built from this training corpus. Heuristics for paraphrasing sentences help to further remove unnecessary information that might interfere with patterns, such as additional adjectives, clauses, or bracketed expressions. In BioCreative II.5, we achieved an f-score of 22% for finding protein interactions, and 43% for mapping proteins to UniProt IDs; disregarding species, f-scores are 30% and 55%, respectively. On average, our best-performing setup required around two minutes per full text. All data and pattern sets as well as Java classes that extend third-party software are available as supplementary information.

show abstract

Introducing meta-services for biomedical information extraction

Cited by 61 publications

References 29 publications

An Overview of BioCreative II.5

An Overview of BioCreative II.5

SOA-Based Integration of Text Mining Services

Efficient Extraction of Protein-Protein Interactions from Full-Text Articles

Contact Info

Product

Resources

About