Gerold Schneider

The Protein-Protein Interaction tasks of BioCreative III: classification/ranking of articles and linking bio-ontology concepts to full text

BackgroundDetermining usefulness of biomedical text mining systems requires realistic task definition and data selection criteria without artificial constraints, measuring performance aspects that go beyond traditional metrics. The BioCreative III Protein-Protein Interaction (PPI) tasks were motivated by such considerations, trying to address aspects including how the end user would oversee the generated output, for instance by providing ranked results, textual evidence for human interpretation or measuring time savings by using automated systems. Detecting articles describing complex biological events like PPIs was addressed in the Article Classification Task (ACT), where participants were asked to implement tools for detecting PPI-describing abstracts. Therefore the BCIII-ACT corpus was provided, which includes a training, development and test set of over 12,000 PPI relevant and non-relevant PubMed abstracts labeled manually by domain experts and recording also the human classification times. The Interaction Method Task (IMT) went beyond abstracts and required mining for associations between more than 3,500 full text articles and interaction detection method ontology concepts that had been applied to detect the PPIs reported in them.ResultsA total of 11 teams participated in at least one of the two PPI tasks (10 in ACT and 8 in the IMT) and a total of 62 persons were involved either as participants or in preparing data sets/evaluating these tasks. Per task, each team was allowed to submit five runs offline and another five online via the BioCreative Meta-Server. From the 52 runs submitted for the ACT, the highest Matthews Correlation Coefficient (MCC) score measured was 0.55 at an accuracy of 89% and the best AUC iP/R was 68%. Most ACT teams explored machine learning methods, some of them also used lexical resources like MeSH terms, PSI-MI concepts or particular lists of verbs and nouns, some integrated NER approaches. For the IMT, a total of 42 runs were evaluated by comparing systems against manually generated annotations done by curators from the BioGRID and MINT databases. The highest AUC iP/R achieved by any run was 53%, the best MCC score 0.55. In case of competitive systems with an acceptable recall (above 35%) the macro-averaged precision ranged between 50% and 80%, with a maximum F-Score of 55%.ConclusionsThe results of the ACT task of BioCreative III indicate that classification of large unbalanced article collections reflecting the real class imbalance is still challenging. Nevertheless, text-mining tools that report ranked lists of relevant articles for manual selection can potentially reduce the time needed to identify half of the relevant articles to less than 1/4 of the time when compared to unranked results. Detecting associations between full text articles and interaction detection method PSI-MI terms (IMT) is more difficult than might be anticipated. This is due to the variability of method term mentions, errors resulting from pre-processing of articles provided as PDF files, and the heterogeneity and different granularity of method term concepts encountered in the ontology. However, combining the sophisticated techniques developed by the participants with supporting evidence strings derived from the articles for human interpretation could result in practical modules for biological annotation workflows.

OntoGene in BioCreative II.5

We describe a system for the detection of mentions of protein-protein interactions in the biomedical scientific literature. The original system was developed as a part of the OntoGene project, which focuses on using advanced computational linguistic techniques for text mining applications in the biomedical domain. In this paper, we focus in particular on the participation to the BioCreative II.5 challenge, where the OntoGene system achieved best-ranked results. Additionally, we describe a feature-analysis experiment performed after the challenge, which shows the unexpected result that one single feature alone performs better than the combination of features used in the challenge.

Mining of relations between proteins over biomedical scientific literature using a deep-linguistic approach

OBJECTIVE The amount of new discoveries (as published in the scientific literature) in the biomedical area is growing at an exponential rate. This growth makes it very difficult to filter the most relevant results, and thus the extraction of the core information becomes very expensive. Therefore, there is a growing interest in text processing approaches that can deliver selected information from scientific publications, which can limit the amount of human intervention normally needed to gather those results. MATERIALS AND METHODS This paper presents and evaluates an approach aimed at automating the process of extracting functional relations (e.g. interactions between genes and proteins) from scientific literature in the biomedical domain. The approach, using a novel dependency-based parser, is based on a complete syntactic analysis of the corpus. RESULTS We have implemented a state-of-the-art text mining system for biomedical literature, based on a deep-linguistic, full-parsing approach. The results are validated on two different corpora: the manually annotated genomics information access (GENIA) corpus and the automatically annotated arabidopsis thaliana circadian rhythms (ATCR) corpus. CONCLUSION We show how a deep-linguistic approach (contrary to common belief) can be used in a real world text mining application, offering high-precision relation extraction, while at the same time retaining a sufficient recall.

OntoGene in BioCreative II

Background:Research scientists and companies working in the domains of biomedicine and genomics are increasingly faced with the problem of efficiently locating, within the vast body of published scientific findings, the critical pieces of information that are needed to direct current and future research investment.Results:In this report we describe approaches taken within the scope of the second BioCreative competition in order to solve two aspects of this problem: detection of novel protein interactions reported in scientific articles, and detection of the experimental method that was used to confirm the interaction. Our approach to the former problem is based on a high-recall protein annotation step, followed by two strict disambiguation steps. The remaining proteins are then combined according to a number of lexico-syntactic filters, which deliver high-precision results while maintaining reasonable recall. The detection of the experimental methods is tackled by a pattern matching approach, which has delivered the best results in the official BioCreative evaluation.Conclusion:Although the results of BioCreative clearly show that no tool is sufficiently reliable for fully automated annotations, a few of the proposed approaches (including our own) already perform at a competitive level. This makes them interesting either as standalone tools for preliminary document inspection, or as modules within an environment aimed at supporting the process of curation of biomedical literature.

Attempto controlled english: a knowledge representation language readable by humans and machines

Attempto Controlled English (ACE) is a knowledge representation language with an English syntax. Thus ACE can be used by anyone, even without being familiar with formal notations. The Attempto Parsing Engine translates ACE texts into discourse representation structures, a variant of first-order logic. Hence, ACE turns out to be a logic language equivalent to full first-order logic. The two views of ACE — natural language and logic language — complement each other, and render ACE both human- and machine-readable. This paper covers both views of ACE. In the first part we present the language ACE in a nutshell, and in the second part we give an overview of the discourse representation structures derived from ACE texts.

Relation mining experiments in the pharmacogenomics domain

The mutual interactions among genes, diseases, and drugs are at the heart of biomedical research, and are especially important for the pharmacological industry. The recent trend towards personalized medicine makes it increasingly relevant to be able to tailor drugs to specific genetic makeups. The pharmacogenetics and pharmacogenomics knowledge base (PharmGKB) aims at capturing relevant information about such interactions from several sources, including curation of the biomedical literature. Advanced text mining tools which can support the process of manual curation are increasingly necessary in order to cope with the deluge of new published results. However, effective evaluation of those tools requires the availability of manually curated data as gold standard. In this paper we discuss how the existing PharmGKB database can be used for such an evaluation task in a way similar to the usage of gold standard data derived from protein-protein interaction databases in one of the recent BioCreative shared tasks. Additionally, we present our own considerations and results on the feasibility and difficulty of such a task.

Detection of interaction articles and experimental methods in biomedical literature

BackgroundThis article describes the approaches taken by the OntoGene group at the University of Zurich in dealing with two tasks of the BioCreative III competition: classification of articles which contain curatable protein-protein interactions (PPI-ACT) and extraction of experimental methods (PPI-IMT).ResultsTwo main achievements are described in this paper: (a) a system for document classification which crucially relies on the results of an advanced pipeline of natural language processing tools; (b) a system which is capable of detecting all experimental methods mentioned in scientific literature, and listing them with a competitive ranking (AUC iP/R > 0.5).ConclusionsThe results of the BioCreative III shared evaluation clearly demonstrate that significant progress has been achieved in the domain of biomedical text mining in the past few years. Our own contribution, together with the results of other participants, provides evidence that natural language processing techniques have become by now an integral part of advanced text mining approaches.

Detecting Protein-Protein Interactions in Biomedical Texts Using a Parser and Linguistic Resources

We describe the task of automatically detecting interactions between proteins in biomedical literature. We use a syntactic parser, a corpus annotated for proteins, and manual decisions as training material. After automatically parsing the GENIA corpus, which is manually annotated for proteins, all syntactic paths between proteins are extracted. These syntactic paths are manually disambiguated between meaningful paths and irrelevant paths. Meaningful paths are paths that express an interaction between the syntactically connected proteins, irrelevant paths are paths that do not convey any interaction. The resource created by these manual decisions is used in two ways. First, words that appear frequently inside a meaningful paths are learnt using simple machine learning. Second, these resources are applied to the task of automatically detecting interactions between proteins in biomedical literature. We use the IntAct corpus as an application corpus. After detecting proteins in the IntAct texts, we automatically parse them and classify the syntactic paths between them using the meaningful paths from the resource created on GENIA and addressing sparse data problems by shortening the paths based on the words frequently appearing inside the meaningful paths, so-called transparent words. We conduct an evaluation showing that we achieve acceptable recall and good precision, and we discuss the importance of transparent words for the task.

Using the OntoGene pipeline for the triage task of BioCreative 2012

In this article, we describe the architecture of the OntoGene Relation mining pipeline and its application in the triage task of BioCreative 2012. The aim of the task is to support the triage of abstracts relevant to the process of curation of the Comparative Toxicogenomics Database. We use a conventional information retrieval system (Lucene) to provide a baseline ranking, which we then combine with information provided by our relation mining system, in order to achieve an optimized ranking. Our approach additionally delivers domain entities mentioned in each input document as well as candidate relationships, both ranked according to a confidence score computed by the system. This information is presented to the user through an advanced interface aimed at supporting the process of interactive curation. Thanks, in particular, to the high-quality entity recognition, the OntoGene system achieved the best overall results in the task.

UZurich in the BioNLP 2009 Shared Task

We describe a biological event detection method implemented for the BioNLP 2009 Shared Task 1. The method relies entirely on the chunk and syntactic dependency relations provided by a general NLP pipeline which was not adapted in any way for the purposes of the shared task. The method maps the syntactic relations to event structures while being guided by the probabilities of the syntactic features of events which were automatically learned from the training data. Our method achieved a recall of 26% and a precision of 44% in the official test run, under “strict equality” of events.