Högskolan i Skövde

In this study, the use of most frequent spatial prepositions in English, Polish, Swedish, and Russian is analyzed. The prepositions and their contexts are extracted from corpora by means of concordance tools. The collostructional strength between the prepositions and the most frequent nouns in the PPs (Gries et al. 2005) is then computed in order to get a more detailed picture of the contexts in which a given preposition is most likely to appear. The results of the investigation are then analysed within the framework of cognitive semantics, especially Croft and Cruse's (2004) taxonomy of construal operations, and Talmy’s (2005) classification of spatial images

Motivation: Automated text analysis is an important tool for facilitating the extraction of knowledge from biomedical abstracts, thereby enabling researchers to build pathway models that integrate and summarize information from a large number of sources. Advanced methods of in-depth analysis of texts using grammar-based approaches developed within the field of computational linguistics must be adapted to the special requirements and challenges posed by biomedical texts, so that these methods can be made available to the bioinformatics and computational biology communities. Results: Our system for automated text analysis and extraction of pathway information is here applied to a set of PubMed abstracts concerning the CBF signaling pathway, which is a key pathway involved in the cold-adaptation response of plants subjected to cold non-freezing temperatures. The system successfully and accurately re-discovers the main features of this pathway, while also pointing to interesting and plausible new hypotheses. The evaluation also reveals a number of issues which will be important targets in the continued development of the system, e.g. the need for an extended lexicon of taxonomic terms and an improved procedure for recognition of sentence boundaries.

We demonstrate how automated text analysis can be used to support the large-scale analysis of metabolic and regulatory pathways by deriving pathway maps from textual descriptions found in the scientific literature. The main assumption is that correct syntactic analysis combined with domain-specific heuristics provides a good basis for relation extraction. Our method uses an algorithm that searches through the syntactic trees produced by a parser based on a Referent Grammar formalism, identifies relations mentioned in the sentence, and classifies them with respect to their semantic class and epistemic status (facts, counterfactuals, hypotheses). The semantic categories used in the classification are based on the relation set used in KEGG (Kyoto Encyclopedia of Genes and Genomes), so that pathway maps using KEGG notation can be automatically generated. We present the current version of the relation extraction algorithm and an evaluation based on a corpus of abstracts obtained from PubMed. The results indicate that the method is able to combine a reasonable coverage with high accuracy. We found that 61% of all sentences were parsed, and 97% of the parse trees were judged to be correct. The extraction algorithm was tested on a sample of 300 parse trees and was found to produce correct extractions in 90.5% of the cases.

An essential part of bioinformatic research concerns the iterative process of validating hypotheses by analyzing facts stored in databases and in published literature. This process can be enhanced by language technology methods, in particular by automatic text understanding. Since it is becoming increasingly difficult to keep up with the vast number of scientific articles being published, there is a need for more easily accessible representations of the current knowledge. The goal of the research described in this paper is to develop a system aimed to support the large-scale research on metabolic and regulatory pathways by extracting relations between biological objects from descriptions found in literature. We present and evaluate the procedures for semantico-syntactic tagging, dividing the text into parts concerning previous research and current research, syntactic parsing, and transformation of syntactic trees into logical representations similar to the pathway graphs utilized in the Kyoto Encyclopaedia of Genes and Genomes.

The long-term goal of the research presented in this paper is to incorporate linguistic text analysis into a system for evaluation of biological pathways. In this system, relations extracted from biomedical texts will be compared with pathways encoded in existing specialized databases. In this way, the biologist's conclusions regarding the plausibility and/or novelty of a certain relation between genes, proteins, etc., can be supported by fused information from biological databases and biological literature. We aim at overcoming the shortcomings of existing systems for information retrieval by proposing a method based on thorough linguistic analysis of a large text corpus. In this paper, we present a comparative analysis of two corpora: one consisting of biomedical texts from PubMed, the other one of general English prose. The results stress the importance of taking multiword entries into account when constructing a system for extracting biological relations from texts

An essential part of bioinformatic research concerns the iterative process of validating hypotheses by analyzing facts stored in databases and in published literature. This process can be enhanced by automatic in-depth text understanding. A prerequisite for this is an adequate syntactic and semantic analysis. The paper presents the results of syntactic, semantic, and textual analysis of a corpus of biomedical abstracts. It focuses on the ways in which relevant molecular interactions are referred to in the abstracts, and proposes a strategy for linking natural language expressions to the standard notation used in Kyoto Encyclopedia of Genes and Genomes. The syntactic and semantic regularities observed in the language of biomedicine are also discussed from the cognitive point of view.

In this paper, we describe a method for extracting biological relations (in the first place, relations used in the KEGG ontology of biological pathways) from scientific texts. The main assumption is that correct syntactic analysis combined with domain-specific heuristics provides a good basis for relation extraction. We propose an algorithm that searches through the syntactic trees produced by a parser based on a formalism called Referent Grammar (inspired by Categorial Grammar), identifies relations mentioned in the sentence, and classifies them with respect to their semantic class and epistemic status (facts, counterfactuals, hypotheses).