A graph kernel for protein-protein interaction extraction

Airola, Antti; Pyysalo, Sampo; Björne, Jari; Pahikkala, Tapio; Ginter, Filip; Salakoski, Tapio

doi:10.3115/1572306.1572308

Cited by 63 publications

(102 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Fundel et al (2007) found that, in extraction of relations between genes and proteins, a system based on the SD scheme greatly outperformed the previous best system on the LLL challenge dataset (by an 18% absolute improvement in F-measure). Airola et al (2008) provide more systematic results on a number of proteinprotein interaction datasets. Their graph kernel approach uses an all-dependency-paths kernel which allows their system to consider full dependency graphs.…”

Section: The Formalism and The Toolmentioning

confidence: 99%

The Stanford typed dependencies representation

Marneffe

Manning

2008

Coling 2008: Proceedings of the Workshop on Cross-Framework and Cross-Domain Parser Evaluation - CrossParser '08

726

614

View full text Add to dashboard Cite

This paper examines the Stanford typed dependencies representation, which was designed to provide a straightforward description of grammatical relations for any user who could benefit from automatic text understanding. For such purposes, we argue that dependency schemes must follow a simple design and provide semantically contentful information, as well as offer an automatic procedure to extract the relations. We consider the underlying design principles of the Stanford scheme from this perspective, and compare it to the GR and PARC representations. Finally, we address the question of the suitability of the Stanford scheme for parser evaluation.

show abstract

Section: The Formalism and The Toolmentioning

confidence: 99%

The Stanford typed dependencies representation

Marneffe

Manning

2008

Coling 2008: Proceedings of the Workshop on Cross-Framework and Cross-Domain Parser Evaluation - CrossParser '08

726

614

View full text Add to dashboard Cite

show abstract

“…To obtain AUC iP/R curves, the highest precision at each recall point is calculated. The interested reader can find a detailed discussion regarding f-measure versus AUC scores (for PPI extraction systems) in [36].…”

Section: Ranking Of Extracted Pairs and Proteinsmentioning

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

“…Features used for classifier training are normally syntactic and lexical patterns derived from dependency relations between individual words in sentences which are revealed automatically by syntactic parsers. Various kernels have been proposed to calculate similarity between syntactic structures, including subsequence kernel [10], tree kernels [11], shortest path kernel [12], graph kernel [13], or a combination of them [14]. Under this kind of problem setting, one sentence in the dataset yields C 2 n distinct instances, where n is the number of proteins in the sentence and each instance represents a pairwise combination of proteins.…”

Section: Work On Protein-protein Interactions Extractionmentioning

confidence: 99%

Biomedical events extraction using the hidden vector state model

Zhou

2011

Artificial Intelligence in Medicine

View full text Add to dashboard Cite

Objective: Biomedical events extraction concerns about extracting events describing changes on the state of bio-molecules from literature. Comparing to the protein-protein interactions (PPIs) extraction task which often only involves the extraction of binary relations between two proteins, biomedical events extraction is much harder since it needs to deal with complex events consisting of embedded or hierarchical relations among proteins, events, and their textual triggers. In this paper, we propose an information extraction system based on the hidden vector state (HVS) model, called HVS-BioEvent, for biomedical events extraction, and investigate its capability in extracting complex events.Methods and Material: HVS has been previously employed for the extractions of PPIs. In HVS-BioEvent, we propose an automated way to generate abstract annotations for HVS training and further propose novel machine learning approaches for event trigger word identification, and for biomedical events extraction from the HVS parse results.Results Our proposed system achieves an F-score of 49.57% on the corpus used in the BioNLP'09 shared task, which is only 2.38% lower than the best performing system by UTurku in the BioNLP'09 share task. Nevertheless, HVS-BioEvent outperforms UTurku's system on complex events extraction with * Corresponding author. Tel:+86-255209876; Fax: +86-255209876.Email addresses: d.zhou@seu.edu.cn (Deyu Zhou), y.he@open.ac.uk (Yulan He) March 11, 2011 36.57% vs 30.52% being achieved for extracting regulation events, and 40.61% vs 38.99% for negative regulation events. Preprint submitted to Elsevier ConclusionsThe results suggest that the HVS model with the hierarchical hidden state structure is indeed more suitable for complex event extraction since it could naturally model embedded structural context in sentences.

show abstract

A graph kernel for protein-protein interaction extraction

Cited by 63 publications

References 18 publications

The Stanford typed dependencies representation

The Stanford typed dependencies representation

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

Biomedical events extraction using the hidden vector state model

Contact Info

Product

Resources

About