Motivation: Although the amount of data in biology is rapidly increasing, critical information for understanding biological events like phosphorylation or gene expression remains locked in the biomedical literature. Most current text mining (TM) approaches to extract information about biological events are focused on either limited-scale studies and/or abstracts, with data extracted lacking context and rarely available to support further research.

Results: Here we present BioContext, an integrated TM system which extracts, extends and integrates results from a number of tools performing entity recognition, biomolecular event extraction and contextualization. Application of our system to 10.9 million MEDLINE abstracts and 234 000 open-access full-text articles from PubMed Central yielded over 36 million mentions representing 11.4 million distinct events. Event participants included over 290 000 distinct genes/proteins that are mentioned more than 80 million times and linked where possible to Entrez Gene identifiers. Over a third of events contain contextual information such as the anatomical location of the event occurrence or whether the event is reported as negated or speculative.

Availability: The BioContext pipeline is available for download (under the BSD license) at http://www.biocontext.org, along with the extracted data which is also available for online browsing.

Contact: martin.gerner@gmail.com

Supplementary information: Supplementary data are available at Bioinformatics online.

Manual curation has long been used for extracting key information found within the primary literature for input into biological databases. The human immunodeficiency virus type 1 (HIV-1), human protein interaction database (HHPID), for example, contains 2589 manually extracted interactions, linked to 14 312 mentions in 3090 articles. The advancement of text-mining (TM) techniques has offered a possibility to rapidly retrieve such data from large volumes of text to a high degree of accuracy. Here, we present a recreation of the HHPID using the current state of the art in TM. To retrieve interactions, we performed gene/protein named entity recognition (NER) and applied two molecular event extraction tools on all abstracts and titles cited in the HHPID. Our best NER scores for precision, recall and F-score were 87.5%, 90.0% and 88.6%, respectively, while event extraction achieved 76.4%, 84.2% and 80.1%, respectively. We demonstrate that over 50% of the HHPID interactions can be recreated from abstracts and titles. Furthermore, from 49 available open-access full-text articles, we extracted a total of 237 unique HIV-1–human interactions, as opposed to 187 interactions recorded in the HHPID from the same articles. On average, we extracted 23 times more mentions of interactions and events from a full-text article than from an abstract and title, with a 6-fold increase in the number of unique interactions. We further demonstrated that more frequently occurring interactions extracted by TM are more likely to be true positives. Overall, the results demonstrate that TM was able to recover a large proportion of interactions, many of which were found within the HHPID, making TM a useful assistant in the manual curation process. Finally, we also retrieved other types of interactions in the context of HIV-1 that are not currently present in the HHPID, thus, expanding the scope of this data set. All data is available at http://gnode1.mib.man.ac.uk/HIV1-text-mining.

We describe and evaluate an automated approach used as part of the i2b2 2011 challenge to identify and categorise statements in suicide notes into one of 15 topics, including Love, Guilt, Thankfulness, Hopelessness and Instructions. The approach combines a set of lexico-syntactic rules with a set of models derived by machine learning from a training dataset. The machine learning models rely on named entities, lexical, lexico-semantic and presentation features, as well as the rules that are applicable to a given statement. On a testing set of 300 suicide notes, the approach showed the overall best micro F-measure of up to 53.36%. The best precision achieved was 67.17% when only rules are used, whereas best recall of 50.57% was with integrated rules and machine learning. While some topics (eg, Sorrow, Anger, Blame) prove challenging, the performance for relatively frequent (eg, Love) and well-scoped categories (eg, Thankfulness) was comparatively higher (precision between 68% and 79%), suggesting that automated text mining approaches can be effective in topic categorisation of suicide notes.

Background

The last two decades have witnessed a dramatic acceleration in the production of genomic sequence information and publication of biomedical articles. Despite the fact that genome sequence data and publications are two of the most heavily relied-upon sources of information for many biologists, very little effort has been made to systematically integrate data from genomic sequences directly with the biological literature. For a limited number of model organisms dedicated teams manually curate publications about genes; however for species with no such dedicated staff many thousands of articles are never mapped to genes or genomic regions.

Methodology/Principal Findings

To overcome the lack of integration between genomic data and biological literature, we have developed pubmed2ensembl (http://www.pubmed2ensembl.org), an extension to the BioMart system that links over 2,000,000 articles in PubMed to nearly 150,000 genes in Ensembl from 50 species. We use several sources of curated (e.g., Entrez Gene) and automatically generated (e.g., gene names extracted through text-mining on MEDLINE records) sources of gene-publication links, allowing users to filter and combine different data sources to suit their individual needs for information extraction and biological discovery. In addition to extending the Ensembl BioMart database to include published information on genes, we also implemented a scripting language for automated BioMart construction and a novel BioMart interface that allows text-based queries to be performed against PubMed and PubMed Central documents in conjunction with constraints on genomic features. Finally, we illustrate the potential of pubmed2ensembl through typical use cases that involve integrated queries across the biomedical literature and genomic data.

Conclusion/Significance

By allowing biologists to find the relevant literature on specific genomic regions or sets of functionally related genes more easily, pubmed2ensembl offers a much-needed genome informatics inspired solution to accessing the ever-increasing biomedical literature.

Summary: Identifying mentions of named entities, such as genes or diseases, and normalizing them to database identifiers have become an important step in many text and data mining pipelines. Despite this need, very few entity normalization systems are publicly available as source code or web services for biomedical text mining. Here we present the Gnat Java library for text retrieval, named entity recognition, and normalization of gene and protein mentions in biomedical text. The library can be used as a component to be integrated with other text-mining systems, as a framework to add user-specific extensions, and as an efficient stand-alone application for the identification of gene and protein names for data analysis. On the BioCreative III test data, the current version of Gnat achieves a Tap-20 score of 0.1987.

Availability: The library and web services are implemented in Java and the sources are available from http://gnat.sourceforge.net.

Contact: jorg.hakenberg@roche.com

In the future, we hope to see an open and thriving data market in which users can find and select data from a wide range of data providers. In such an open access market, data are products that must be packaged accordingly. Increasingly, eCommerce sellers present heterogeneous product lines to buyers using faceted browsing. Using this approach we have developed the Ontogrator platform, which allows for rapid retrieval of data in a way that would be familiar to any online shopper. Using Knowledge Organization Systems (KOS), especially ontologies, Ontogrator uses text mining to mark up data and faceted browsing to help users navigate, query and retrieve data. Ontogrator offers the potential to impact scientific research in two major ways: 1) by significantly improving the retrieval of relevant information; and 2) by significantly reducing the time required to compose standard database queries and assemble information for further research. Here we present a pilot implementation developed in collaboration with the Genomic Standards Consortium (GSC) that includes content from the StrainInfo, GOLD, CAMERA, Silva and Pubmed databases. This implementation demonstrates the power of ontogration and highlights that the usefulness of this approach is fully dependent on both the quality of data and the KOS (ontologies) used. Ideally, the use and further expansion of this collaborative system will help to surface issues associated with the underlying quality of annotation and could lead to a systematic means for accessing integrated data resources.

Background

There have been a number of recent efforts (e.g. BioCatalogue, BioMoby) to systematically catalogue bioinformatics tools, services and datasets. These efforts rely on manual curation, making it difficult to cope with the huge influx of various electronic resources that have been provided by the bioinformatics community. We present a text mining approach that utilises the literature to automatically extract descriptions and semantically profile bioinformatics resources to make them available for resource discovery and exploration through semantic networks that contain related resources.

Results

The method identifies the mentions of resources in the literature and assigns a set of co-occurring terminological entities (descriptors) to represent them. We have processed 2,691 full-text bioinformatics articles and extracted profiles of 12,452 resources containing associated descriptors with binary and tf*idf weights. Since such representations are typically sparse (on average 13.77 features per resource), we used lexical kernel metrics to identify semantically related resources via descriptor smoothing. Resources are then clustered or linked into semantic networks, providing the users (bioinformaticians, curators and service/tool crawlers) with a possibility to explore algorithms, tools, services and datasets based on their relatedness. Manual exploration of links between a set of 18 well-known bioinformatics resources suggests that the method was able to identify and group semantically related entities.

Conclusions

The results have shown that the method can reconstruct interesting functional links between resources (e.g. linking data types and algorithms), in particular when tf*idf-like weights are used for profiling. This demonstrates the potential of combining literature mining and simple lexical kernel methods to model relatedness between resource descriptors in particular when there are few features, thus potentially improving the resource description, discovery and exploration process. The resource profiles are available at http://gnode1.mib.man.ac.uk/bioinf/semnets.html

Background

The task of recognizing and identifying species names in biomedical literature has recently been regarded as critical for a number of applications in text and data mining, including gene name recognition, species-specific document retrieval, and semantic enrichment of biomedical articles.

Results

In this paper we describe an open-source species name recognition and normalization software system, LINNAEUS, and evaluate its performance relative to several automatically generated biomedical corpora, as well as a novel corpus of full-text documents manually annotated for species mentions. LINNAEUS uses a dictionary-based approach (implemented as an efficient deterministic finite-state automaton) to identify species names and a set of heuristics to resolve ambiguous mentions. When compared against our manually annotated corpus, LINNAEUS performs with 94% recall and 97% precision at the mention level, and 98% recall and 90% precision at the document level. Our system successfully solves the problem of disambiguating uncertain species mentions, with 97% of all mentions in PubMed Central full-text documents resolved to unambiguous NCBI taxonomy identifiers.

Conclusions

LINNAEUS is an open source, stand-alone software system capable of recognizing and normalizing species name mentions with speed and accuracy, and can therefore be integrated into a range of bioinformatics and text-mining applications. The software and manually annotated corpus can be downloaded freely at http://linnaeus.sourceforge.net/.

Objective

The authors present a system developed for the Challenge in Natural Language Processing for Clinical Data—the i2b2 obesity challenge, whose aim was to automatically identify the status of obesity and 15 related co-morbidities in patients using their clinical discharge summaries. The challenge consisted of two tasks, textual and intuitive. The textual task was to identify explicit references to the diseases, whereas the intuitive task focused on the prediction of the disease status when the evidence was not explicitly asserted.

Design

The authors assembled a set of resources to lexically and semantically profile the diseases and their associated symptoms, treatments, etc. These features were explored in a hybrid text mining approach, which combined dictionary look-up, rule-based, and machine-learning methods.

Measurements

The methods were applied on a set of 507 previously unseen discharge summaries, and the predictions were evaluated against a manually prepared gold standard. The overall ranking of the participating teams was primarily based on the macro-averaged F-measure.

Results

The implemented method achieved the macro-averaged F-measure of 81% for the textual task (which was the highest achieved in the challenge) and 63% for the intuitive task (ranked 7th out of 28 teams—the highest was 66%). The micro-averaged F-measure showed an average accuracy of 97% for textual and 96% for intuitive annotations.

Conclusions

The performance achieved was in line with the agreement between human annotators, indicating the potential of text mining for accurate and efficient prediction of disease statuses from clinical discharge summaries.

Background

Availability of information about transcription factors (TFs) is crucial for genome biology, as TFs play a central role in the regulation of gene expression. While manual literature curation is expensive and labour intensive, the development of semi-automated text mining support is hindered by unavailability of training data. There have been no studies on how existing data sources (e.g. TF-related data from the MeSH thesaurus and GO ontology) or potentially noisy example data (e.g. protein-protein interaction, PPI) could be used to provide training data for identification of TF-contexts in literature.

Results

In this paper we describe a text-classification system designed to automatically recognise contexts related to transcription factors in literature. A learning model is based on a set of biological features (e.g. protein and gene names, interaction words, other biological terms) that are deemed relevant for the task. We have exploited background knowledge from existing biological resources (MeSH and GO) to engineer such features. Weak and noisy training datasets have been collected from descriptions of TF-related concepts in MeSH and GO, PPI data and data representing non-protein-function descriptions. Three machine-learning methods are investigated, along with a vote-based merging of individual approaches and/or different training datasets. The system achieved highly encouraging results, with most classifiers achieving an F-measure above 90%.

Conclusions

The experimental results have shown that the proposed model can be used for identification of TF-related contexts (i.e. sentences) with high accuracy, with a significantly reduced set of features when compared to traditional bag-of-words approach. The results of considering existing PPI data suggest that there is not as high similarity between TF and PPI contexts as we have expected. We have also shown that existing knowledge sources are useful both for feature engineering and for obtaining noisy positive training data.

Linking gene and protein names mentioned in the literature to unique identifiers in referent genomic databases is an essential step in accessing and integrating knowledge in the biomedical domain. However, it remains a challenging task due to lexical and terminological variation, and ambiguity of gene name mentions in documents. We present a generic and effective rule-based approach to link gene mentions in the literature to referent genomic databases, where pre-processing of both gene synonyms in the databases and gene mentions in text are first applied. The mapping method employs a cascaded approach, which combines exact, exact-like and token-based approximate matching by using flexible representations of a gene synonym dictionary and gene mentions generated during the pre-processing phase. We also consider multi-gene name mentions and permutation of components in gene names. A systematic evaluation of the suggested methods has identified steps that are beneficial for improving either precision or recall in gene name identification. The results of the experiments on the BioCreAtIvE2 data sets (identification of human gene names) demonstrated that our methods achieved highly encouraging results with F-measure of up to 81.20%.

Background

Text mining has spurred huge interest in the domain of biology. The goal of the BioCreAtIvE exercise was to evaluate the performance of current text mining systems. We participated in Task 2, which addressed assigning Gene Ontology terms to human proteins and selecting relevant evidence from full-text documents. We approached it as a modified form of the document classification task. We used a supervised machine-learning approach (based on support vector machines) to assign protein function and select passages that support the assignments. As classification features, we used a protein's co-occurring terms that were automatically extracted from documents.

Results

The results evaluated by curators were modest, and quite variable for different problems: in many cases we have relatively good assignment of GO terms to proteins, but the selected supporting text was typically non-relevant (precision spanning from 3% to 50%). The method appears to work best when a substantial set of relevant documents is obtained, while it works poorly on single documents and/or short passages. The initial results suggest that our approach can also mine annotations from text even when an explicit statement relating a protein to a GO term is absent.

Conclusion

A machine learning approach to mining protein function predictions from text can yield good performance only if sufficient training data is available, and significant amount of supporting data is used for prediction. The most promising results are for combined document retrieval and GO term assignment, which calls for the integration of methods developed in BioCreAtIvE Task 1 and Task 2.