BIOSMILE: A semantic role labeling system for biomedical verbs using a maximum-entropy model with automatically generated template features

Background  

Bioinformatics tools for automatic processing of biomedical literature are invaluable for both the design and interpretation of large-scale experiments. Many information extraction (IE) systems that incorporate natural language processing (NLP) techniques have thus been developed for use in the biomedical field. A key IE task in this field is the extraction of biomedical relations, such as protein-protein and gene-disease interactions. However, most biomedical relation extraction systems usually ignore adverbial and prepositional phrases and words identifying location, manner, timing, and condition, which are essential for describing biomedical relations. Semantic role labeling (SRL) is a natural language processing technique that identifies the semantic roles of these words or phrases in sentences and expresses them as predicate-argument structures. We construct a biomedical SRL system called BIOSMILE that uses a maximum entropy (ME) machine-learning model to extract biomedical relations. BIOSMILE is trained on BioProp, our semi-automatic, annotated biomedical proposition bank. Currently, we are focusing on 30 biomedical verbs that are frequently used or considered important for describing molecular events.     

Semantic role labeling for protein transport predicates

Background  

Automatic semantic role labeling (SRL) is a natural language processing (NLP) technique that maps sentences to semantic representations. This technique has been widely studied in the recent years, but mostly with data in newswire domains. Here, we report on a SRL model for identifying the semantic roles of biomedical predicates describing protein transport in GeneRIFs – manually curated sentences focusing on gene functions. To avoid the computational cost of syntactic parsing, and because the boundaries of our protein transport roles often did not match up with syntactic phrase boundaries, we approached this problem with a word-chunking paradigm and trained support vector machine classifiers to classify words as being at the beginning, inside or outside of a protein transport role.     

Enhancing biomedical text summarization using semantic relation extraction

Automatic text summarization for a biomedical concept can help researchers to get the key points of a certain topic from large amount of biomedical literature efficiently. In this paper, we present a method for generating text summary for a given biomedical concept, e.g., H1N1 disease, from multiple documents based on semantic relation extraction. Our approach includes three stages: 1) We extract semantic relations in each sentence using the semantic knowledge representation tool SemRep. 2) We develop a relation-level retrieval method to select the relations most relevant to each query concept and visualize them in a graphic representation. 3) For relations in the relevant set, we extract informative sentences that can interpret them from the document collection to generate text summary using an information retrieval based method. Our major focus in this work is to investigate the contribution of semantic relation extraction to the task of biomedical text summarization. The experimental results on summarization for a set of diseases show that the introduction of semantic knowledge improves the performance and our results are better than the MEAD system, a well-known tool for text summarization.     

A Comprehensive Benchmark of Kernel Methods to Extract Protein–Protein Interactions from Literature

The most important way of conveying new findings in biomedical research is scientific publication. Extraction of protein–protein interactions (PPIs) reported in scientific publications is one of the core topics of text mining in the life sciences. Recently, a new class of such methods has been proposed - convolution kernels that identify PPIs using deep parses of sentences. However, comparing published results of different PPI extraction methods is impossible due to the use of different evaluation corpora, different evaluation metrics, different tuning procedures, etc. In this paper, we study whether the reported performance metrics are robust across different corpora and learning settings and whether the use of deep parsing actually leads to an increase in extraction quality. Our ultimate goal is to identify the one method that performs best in real-life scenarios, where information extraction is performed on unseen text and not on specifically prepared evaluation data. We performed a comprehensive benchmarking of nine different methods for PPI extraction that use convolution kernels on rich linguistic information. Methods were evaluated on five different public corpora using cross-validation, cross-learning, and cross-corpus evaluation. Our study confirms that kernels using dependency trees generally outperform kernels based on syntax trees. However, our study also shows that only the best kernel methods can compete with a simple rule-based approach when the evaluation prevents information leakage between training and test corpora. Our results further reveal that the F-score of many approaches drops significantly if no corpus-specific parameter optimization is applied and that methods reaching a good AUC score often perform much worse in terms of F-score. We conclude that for most kernels no sensible estimation of PPI extraction performance on new text is possible, given the current heterogeneity in evaluation data. Nevertheless, our study shows that three kernels are clearly superior to the other methods.     

MaSTerClass: a case-based reasoning system for the classification of biomedical terms

MOTIVATION: The sheer volume of textually described biomedical knowledge exerts the need for natural language processing (NLP) applications in order to allow flexible and efficient access to relevant information. Specialized semantic networks (such as biomedical ontologies, terminologies or semantic lexicons) can significantly enhance these applications by supplying the necessary terminological information in a machine-readable form. With the explosive growth of bio-literature, new terms (representing newly identified concepts or variations of the existing terms) may not be explicitly described within the network and hence cannot be fully exploited by NLP applications. Linguistic and statistical clues can be used to extract many new terms from free text. The extracted terms still need to be correctly positioned relative to other terms in the network. Classification as a means of semantic typing represents the first step in updating a semantic network with new terms. RESULTS: The MaSTerClass system implements the case-based reasoning methodology for the classification of biomedical terms.     

MedScan,a natural language processing engine for MEDLINE abstracts

MOTIVATION: The importance of extracting biomedical information from scientific publications is well recognized. A number of information extraction systems for the biomedical domain have been reported, but none of them have become widely used in practical applications. Most proposals to date make rather simplistic assumptions about the syntactic aspect of natural language. There is an urgent need for a system that has broad coverage and performs well in real-text applications. RESULTS: We present a general biomedical domain-oriented NLP engine called MedScan that efficiently processes sentences from MEDLINE abstracts and produces a set of regularized logical structures representing the meaning of each sentence. The engine utilizes a specially developed context-free grammar and lexicon. Preliminary evaluation of the system's performance, accuracy, and coverage exhibited encouraging results. Further approaches for increasing the coverage and reducing parsing ambiguity of the engine, as well as its application for information extraction are discussed.     

Fast and scalable neural embedding models for biomedical sentence classification

Background

Biomedical literature is expanding rapidly, and tools that help locate information of interest are needed. To this end, a multitude of different approaches for classifying sentences in biomedical publications according to their coarse semantic and rhetoric categories (e.g., Background, Methods, Results, Conclusions) have been devised, with recent state-of-the-art results reported for a complex deep learning model. Recent evidence showed that shallow and wide neural models such as fastText can provide results that are competitive or superior to complex deep learning models while requiring drastically lower training times and having better scalability. We analyze the efficacy of the fastText model in the classification of biomedical sentences in the PubMed 200k RCT benchmark, and introduce a simple pre-processing step that enables the application of fastText on sentence sequences. Furthermore, we explore the utility of two unsupervised pre-training approaches in scenarios where labeled training data are limited.

Results

Our fastText-based methodology yields a state-of-the-art F1 score of.917 on the PubMed 200k benchmark when sentence ordering is taken into account, with a training time of only 73 s on standard hardware. Applying fastText on single sentences, without taking sentence ordering into account, yielded an F1 score of.852 (training time 13 s). Unsupervised pre-training of N-gram vectors greatly improved the results for small training set sizes, with an increase of F1 score of.21 to.74 when trained on only 1000 randomly picked sentences without taking sentence ordering into account.

Conclusions

Because of it’s ease of use and performance, fastText should be among the first choices of tools when tackling biomedical text classification problems with large corpora. Unsupervised pre-training of N-gram vectors on domain-specific corpora also makes it possible to apply fastText when labeled training data are limited.

     

An Unsupervised Text Mining Method for Relation Extraction from Biomedical Literature

The wealth of interaction information provided in biomedical articles motivated the implementation of text mining approaches to automatically extract biomedical relations. This paper presents an unsupervised method based on pattern clustering and sentence parsing to deal with biomedical relation extraction. Pattern clustering algorithm is based on Polynomial Kernel method, which identifies interaction words from unlabeled data; these interaction words are then used in relation extraction between entity pairs. Dependency parsing and phrase structure parsing are combined for relation extraction. Based on the semi-supervised KNN algorithm, we extend the proposed unsupervised approach to a semi-supervised approach by combining pattern clustering, dependency parsing and phrase structure parsing rules. We evaluated the approaches on two different tasks: (1) Protein–protein interactions extraction, and (2) Gene–suicide association extraction. The evaluation of task (1) on the benchmark dataset (AImed corpus) showed that our proposed unsupervised approach outperformed three supervised methods. The three supervised methods are rule based, SVM based, and Kernel based separately. The proposed semi-supervised approach is superior to the existing semi-supervised methods. The evaluation on gene–suicide association extraction on a smaller dataset from Genetic Association Database and a larger dataset from publicly available PubMed showed that the proposed unsupervised and semi-supervised methods achieved much higher F-scores than co-occurrence based method.     

BCC-NER: bidirectional,contextual clues named entity tagger for gene/protein mention recognition

Tagging biomedical entities such as gene, protein, cell, and cell-line is the first step and an important pre-requisite in biomedical literature mining. In this paper, we describe our hybrid named entity tagging approach namely BCC-NER (bidirectional, contextual clues named entity tagger for gene/protein mention recognition). BCC-NER is deployed with three modules. The first module is for text processing which includes basic NLP pre-processing, feature extraction, and feature selection. The second module is for training and model building with bidirectional conditional random fields (CRF) to parse the text in both directions (forward and backward) and integrate the backward and forward trained models using margin-infused relaxed algorithm (MIRA). The third and final module is for post-processing to achieve a better performance, which includes surrounding text features, parenthesis mismatching, and two-tier abbreviation algorithm. The evaluation results on BioCreative II GM test corpus of BCC-NER achieve a precision of 89.95, recall of 84.15 and overall F-score of 86.95, which is higher than the other currently available open source taggers.     

The Relation between Thematic Role Computing and Semantic Relatedness Processing during On-Line Sentence Comprehension

Sentence comprehension involves timely computing different types of relations between its verbs and noun arguments, such as morphosyntactic, semantic, and thematic relations. Here, we used EEG technique to investigate the potential differences in thematic role computing and lexical-semantic relatedness processing during on-line sentence comprehension, and the interaction between these two types of processes. Mandarin Chinese sentences were used as materials. The basic structure of those sentences is “Noun+Verb+‘le’+a two-character word”, with the Noun being the initial argument. The verb disambiguates the initial argument as an agent or a patient. Meanwhile, the initial argument and the verb are highly or lowly semantically related. The ERPs at the verbs revealed that: relative to the agent condition, the patient condition evoked a larger N400 only when the argument and verb were lowly semantically related; however, relative to the high-relatedness condition, the low-relatedness condition elicited a larger N400 regardless of the thematic relation; although both thematic role variation and semantic relatedness variation elicited N400 effects, the N400 effect elicited by the former was broadly distributed and reached maximum over the frontal electrodes, and the N400 effect elicited by the latter had a posterior distribution. In addition, the brain oscillations results showed that, although thematic role variation (patient vs. agent) induced power decreases around the beta frequency band (15–30 Hz), semantic relatedness variation (low-relatedness vs. high-relatedness) induced power increases in the theta frequency band (4–7 Hz). These results suggested that, in the sentence context, thematic role computing is modulated by the semantic relatedness between the verb and its argument; semantic relatedness processing, however, is in some degree independent from the thematic relations. Moreover, our results indicated that, during on-line sentence comprehension, thematic role computing and semantic relatedness processing are mediated by distinct neural systems.     

Hash subgraph pairwise kernel for protein-protein interaction extraction

Extracting protein-protein interaction (PPI) from biomedical literature is an important task in biomedical text mining (BioTM). In this paper, we propose a hash subgraph pairwise (HSP) kernel-based approach for this task. The key to the novel kernel is to use the hierarchical hash labels to express the structural information of subgraphs in a linear time. We apply the graph kernel to compute dependency graphs representing the sentence structure for protein-protein interaction extraction task, which can efficiently make use of full graph structural information, and particularly capture the contiguous topological and label information ignored before. We evaluate the proposed approach on five publicly available PPI corpora. The experimental results show that our approach significantly outperforms all-path kernel approach on all five corpora and achieves state-of-the-art performance.     

Using contextual and lexical features to restructure and validate the classification of biomedical concepts

Background  

Biomedical ontologies are critical for integration of data from diverse sources and for use by knowledge-based biomedical applications, especially natural language processing as well as associated mining and reasoning systems. The effectiveness of these systems is heavily dependent on the quality of the ontological terms and their classifications. To assist in developing and maintaining the ontologies objectively, we propose automatic approaches to classify and/or validate their semantic categories. In previous work, we developed an approach using contextual syntactic features obtained from a large domain corpus to reclassify and validate concepts of the Unified Medical Language System (UMLS), a comprehensive resource of biomedical terminology. In this paper, we introduce another classification approach based on words of the concept strings and compare it to the contextual syntactic approach.     

Sentence alignment using feed forward neural network

Parallel corpora have become an essential resource for work in multi lingual natural language processing. However, sentence aligned parallel corpora are more efficient than non-aligned parallel corpora for cross language information retrieval and machine translation applications. In this paper, we present a new approach to align sentences in bilingual parallel corpora based on feed forward neural network classifier. A feature parameter vector is extracted from the text pair under consideration. This vector contains text features such as length, punctuate score, and cognate score values. A set of manually prepared training data has been assigned to train the feed forward neural network. Another set of data was used for testing. Using this new approach, we could achieve an error reduction of 60% over length based approach when applied on English-Arabic parallel documents. Moreover this new approach is valid for any language pair and it is quite flexible approach since the feature parameter vector may contain more/less or different features than that we used in our system such as lexical match feature.     

The textual characteristics of traditional and Open Access scientific journals are similar

Background  

Recent years have seen an increased amount of natural language processing (NLP) work on full text biomedical journal publications. Much of this work is done with Open Access journal articles. Such work assumes that Open Access articles are representative of biomedical publications in general and that methods developed for analysis of Open Access full text publications will generalize to the biomedical literature as a whole. If this assumption is wrong, the cost to the community will be large, including not just wasted resources, but also flawed science. This paper examines that assumption.     

Principal Semantic Components of Language and the Measurement of Meaning

Metric systems for semantics, or semantic cognitive maps, are allocations of words or other representations in a metric space based on their meaning. Existing methods for semantic mapping, such as Latent Semantic Analysis and Latent Dirichlet Allocation, are based on paradigms involving dissimilarity metrics. They typically do not take into account relations of antonymy and yield a large number of domain-specific semantic dimensions. Here, using a novel self-organization approach, we construct a low-dimensional, context-independent semantic map of natural language that represents simultaneously synonymy and antonymy. Emergent semantics of the map principal components are clearly identifiable: the first three correspond to the meanings of “good/bad” (valence), “calm/excited” (arousal), and “open/closed” (freedom), respectively. The semantic map is sufficiently robust to allow the automated extraction of synonyms and antonyms not originally in the dictionaries used to construct the map and to predict connotation from their coordinates. The map geometric characteristics include a limited number (∼4) of statistically significant dimensions, a bimodal distribution of the first component, increasing kurtosis of subsequent (unimodal) components, and a U-shaped maximum-spread planar projection. Both the semantic content and the main geometric features of the map are consistent between dictionaries (Microsoft Word and Princeton''s WordNet), among Western languages (English, French, German, and Spanish), and with previously established psychometric measures. By defining the semantics of its dimensions, the constructed map provides a foundational metric system for the quantitative analysis of word meaning. Language can be viewed as a cumulative product of human experiences. Therefore, the extracted principal semantic dimensions may be useful to characterize the general semantic dimensions of the content of mental states. This is a fundamental step toward a universal metric system for semantics of human experiences, which is necessary for developing a rigorous science of the mind.     

BioIE: extracting informative sentences from the biomedical literature

SUMMARY: BioIE is a rule-based system that extracts informative sentences relating to protein families, their structures, functions and diseases from the biomedical literaturE. Based on manual definition of templates and rules, it aims at precise sentence extraction rather than wide recall. After uploading source text or retrieving abstracts from MEDLINE, users can extract sentences based on predefined or user-defined template categories. BioIE also provides a brief insight into the syntactic and semantic context of the source-text by looking at word, N-gram and MeSH-term distributions. Important Applications of BioIE are in, for example, annotation of microarray data and of protein databases. AVAILABILITY: http://umber.sbs.man.ac.uk/dbbrowser/bioie/     

Negation’s Not Solved: Generalizability Versus Optimizability in Clinical Natural Language Processing

A review of published work in clinical natural language processing (NLP) may suggest that the negation detection task has been “solved.” This work proposes that an optimizable solution does not equal a generalizable solution. We introduce a new machine learning-based Polarity Module for detecting negation in clinical text, and extensively compare its performance across domains. Using four manually annotated corpora of clinical text, we show that negation detection performance suffers when there is no in-domain development (for manual methods) or training data (for machine learning-based methods). Various factors (e.g., annotation guidelines, named entity characteristics, the amount of data, and lexical and syntactic context) play a role in making generalizability difficult, but none completely explains the phenomenon. Furthermore, generalizability remains challenging because it is unclear whether to use a single source for accurate data, combine all sources into a single model, or apply domain adaptation methods. The most reliable means to improve negation detection is to manually annotate in-domain training data (or, perhaps, manually modify rules); this is a strategy for optimizing performance, rather than generalizing it. These results suggest a direction for future work in domain-adaptive and task-adaptive methods for clinical NLP.     

Bio-SCoRes: A Smorgasbord Architecture for Coreference Resolution in Biomedical Text

Coreference resolution is one of the fundamental and challenging tasks in natural language processing. Resolving coreference successfully can have a significant positive effect on downstream natural language processing tasks, such as information extraction and question answering. The importance of coreference resolution for biomedical text analysis applications has increasingly been acknowledged. One of the difficulties in coreference resolution stems from the fact that distinct types of coreference (e.g., anaphora, appositive) are expressed with a variety of lexical and syntactic means (e.g., personal pronouns, definite noun phrases), and that resolution of each combination often requires a different approach. In the biomedical domain, it is common for coreference annotation and resolution efforts to focus on specific subcategories of coreference deemed important for the downstream task. In the current work, we aim to address some of these concerns regarding coreference resolution in biomedical text. We propose a general, modular framework underpinned by a smorgasbord architecture (Bio-SCoRes), which incorporates a variety of coreference types, their mentions and allows fine-grained specification of resolution strategies to resolve coreference of distinct coreference type-mention pairs. For development and evaluation, we used a corpus of structured drug labels annotated with fine-grained coreference information. In addition, we evaluated our approach on two other corpora (i2b2/VA discharge summaries and protein coreference dataset) to investigate its generality and ease of adaptation to other biomedical text types. Our results demonstrate the usefulness of our novel smorgasbord architecture. The specific pipelines based on the architecture perform successfully in linking coreferential mention pairs, while we find that recognition of full mention clusters is more challenging. The corpus of structured drug labels (SPL) as well as the components of Bio-SCoRes and some of the pipelines based on it are publicly available at https://github.com/kilicogluh/Bio-SCoRes. We believe that Bio-SCoRes can serve as a strong and extensible baseline system for coreference resolution of biomedical text.     

RelEx--relation extraction using dependency parse trees

MOTIVATION: The discovery of regulatory pathways, signal cascades, metabolic processes or disease models requires knowledge on individual relations like e.g. physical or regulatory interactions between genes and proteins. Most interactions mentioned in the free text of biomedical publications are not yet contained in structured databases. RESULTS: We developed RelEx, an approach for relation extraction from free text. It is based on natural language preprocessing producing dependency parse trees and applying a small number of simple rules to these trees. We applied RelEx on a comprehensive set of one million MEDLINE abstracts dealing with gene and protein relations and extracted approximately 150,000 relations with an estimated performance of both 80% precision and 80% recall. AVAILABILITY: The used natural language preprocessing tools are free for use for academic research. Test sets and relation term lists are available from our website (http://www.bio.ifi.lmu.de/publications/RelEx/).