NetiNeti: discovery of scientific names from text using machine learning methods

OpCitance: Citation contexts identified from the PubMed Central open access articles

OpCitance contains all the sentences from 2 million PubMed Central open-access (PMCOA) articles, with 137 million inline citations annotated (i.e., the "citation contexts"). Parsing out the references and citation contexts from the PMCOA XML files was non-trivial due to the diversity of referencing style. Only 0.5% citation contexts remain unidentified due to technical or human issues, e.g., references unmentioned by the authors in the text or improper XML nesting, which is more common among older articles (pre-2000). PubMed IDs (PMIDs) linked to inline citations in the XML files compared to citations harvested using the NCBI E-Utilities differed for 70.96% of the articles. Using an in-house citation matcher, called Patci, 6.84% of the referenced PMIDs were supplemented and corrected. OpCitance includes fewer total number of articles than the Semantic Scholar Open Research Corpus, but OpCitance has 160 thousand unique articles, a higher inline citation identification rate, and a more accurate reference mapping to PMIDs. We hope that OpCitance will facilitate citation context studies in particular and benefit text-mining research more broadly.

Essential Biodiversity Variables: extracting plant phenological data from specimen labels using machine learning

Essential Biodiversity Variables (EBVs) make it possible to evaluate and monitor the state of biodiversity over time at different spatial scales. Its development is led by the Group on Earth Observations Biodiversity Observation Network (GEO BON) to harmonize, consolidate and standardize biodiversity data from varied biodiversity sources. This document presents a mechanism to obtain baseline data to feed the Species Traits Variable Phenology or other biodiversity indicators by extracting species characters and structure names from morphological descriptions of specimens and classifying such descriptions using machine learning (ML).

A workflow that performs Named Entity Recognition (NER) and Classification of morphological descriptions using ML algorithms was evaluated with excellent results. It was implemented using Python, Pytorch, Scikit-Learn, Pomegranate, Python-crfsuite, and other libraries applied to 106,804 herbarium records from the National Biodiversity Institute of Costa Rica (INBio). The text classification results were almost excellent (F1 score between 96% and 99%) using three traditional ML methods: Multinomial Naive Bayes (NB), Linear Support Vector Classification (SVC), and Logistic Regression (LR). Furthermore, results extracting names of species morphological structures (e.g., leaves, trichomes, flowers, petals, sepals) and character names (e.g., length, width, pigmentation patterns, and smell) using NER algorithms were competitive (F1 score between 95% and 98%) using Hidden Markov Models (HMM), Conditional Random Fields (CRFs), and Bidirectional Long Short Term Memory Networks with CRF (BI-LSTM-CRF).

Multiple annotation for biodiversity: developing an annotation framework among biology, linguistics and text technology

Biodiversity information is contained in countless digitized and unprocessed scholarly texts. Although automated extraction of these data has been gaining momentum for years, there are still innumerable text sources that are poorly accessible and require a more advanced range of methods to extract relevant information. To improve the access to semantic biodiversity information, we have launched the BIOfid project (www.biofid.de) and have developed a portal to access the semantics of German language biodiversity texts, mainly from the 19th and 20th century. However, to make such a portal work, a couple of methods had to be developed or adapted first. In particular, text-technological information extraction methods were needed, which extract the required information from the texts. Such methods draw on machine learning techniques, which in turn are trained by learning data. To this end, among others, we gathered the bio text corpus, which is a cooperatively built resource, developed by biologists, text technologists, and linguists. A special feature of bio is its multiple annotation approach, which takes into account both general and biology-specific classifications, and by this means goes beyond previous, typically taxon- or ontology-driven proper name detection. We describe the design decisions and the genuine Annotation Hub Framework underlying the bio annotations and present agreement results. The tools used to create the annotations are introduced, and the use of the data in the semantic portal is described. Finally, some general lessons, in particular with multiple annotation projects, are drawn.

Biodiversity at the global scale: the synthesis continues

Traditionally, the generation and use of biodiversity data and their associated specimen objects have been primarily the purview of individuals and small research groups. While deposition of data and specimens in herbaria and other repositories has long been the norm, throughout most of their history, these resources have been accessible only to a small community of specialists. Through recent concerted efforts, primarily at the level of national and international governmental agencies over the last two decades, the pace of biodiversity data accumulation has accelerated, and a wider array of biodiversity scientists has gained access to this massive accumulation of resources, applying them to an ever-widening compass of research pursuits. We review how these new resources and increasing access to them are affecting the landscape of biodiversity research in plants today, focusing on new applications across evolution, ecology, and other fields that have been enabled specifically by the availability of these data and the global scope that was previously beyond the reach of individual investigators. We give an overview of recent advances organized along three lines: broad-scale analyses of distributional data and spatial information, phylogenetic research circumscribing large clades with comprehensive taxon sampling, and data sets derived from improved accessibility of biodiversity literature. We also review synergies between large data resources and more traditional data collection paradigms, describe shortfalls and how to overcome them, and reflect on the future of plant biodiversity analyses in light of increasing linkages between data types and scientists in our field.

Recognition of Latin scientific names using artificial neural networks

PREMISE

The automated recognition of Latin scientific names within vernacular text has many applications, including text mining, search indexing, and automated specimen-label processing. Most published solutions are computationally inefficient, incapable of running within a web browser, and focus on texts in English, thus omitting a substantial portion of biodiversity literature.

METHODS AND RESULTS

An open-source browser-executable solution, Quaesitor, is presented here. It uses pattern matching (regular expressions) in combination with an ensembled classifier composed of an inclusion dictionary search (Bloom filter), a trio of complementary neural networks that differ in their approach to encoding text, and word length to automatically identify Latin scientific names in the 16 most common languages for biodiversity articles.

CONCLUSIONS

In combination, the classifiers can recognize Latin scientific names in isolation or embedded within the languages used for >96% of biodiversity literature titles. For three different data sets, they resulted in a 0.80-0.97 recall and a 0.69-0.84 precision at a rate of 8.6 ms/word.

COPIOUS: A gold standard corpus of named entities towards extracting species occurrence from biodiversity literature

Background Species occurrence records are very important in the biodiversity domain. While several available corpora contain only annotations of species names or habitats and geographical locations, there is no consolidated corpus that covers all types of entities necessary for extracting species occurrence from biodiversity literature. In order to alleviate this issue, we have constructed the COPIOUS corpus-a gold standard corpus that covers a wide range of biodiversity entities. Results Two annotators manually annotated the corpus with five categories of entities, i.e. taxon names, geographical locations, habitats, temporal expressions and person names. The overall inter-annotator agreement on 200 doubly-annotated documents is approximately 81.86% F-score. Amongst the five categories, the agreement on habitat entities was the lowest, indicating that this type of entity is complex. The COPIOUS corpus consists of 668 documents downloaded from the Biodiversity Heritage Library with over 26K sentences and more than 28K entities. Named entity recognisers trained on the corpus could achieve an F-score of 74.58%. Moreover, in recognising taxon names, our model performed better than two available tools in the biodiversity domain, namely the SPECIES tagger and the Global Name Recognition and Discovery. More than 1,600 binary relations of Taxon-Habitat, Taxon-Person, Taxon-Geographical locations and Taxon-Temporal expressions were identified by applying a pattern-based relation extraction system to the gold standard. Based on the extracted relations, we can produce a knowledge repository of species occurrences. Conclusion The paper describes in detail the construction of a gold standard named entity corpus for the biodiversity domain. An investigation of the performance of named entity recognition (NER) tools trained on the gold standard revealed that the corpus is sufficiently reliable and sizeable for both training and evaluation purposes. The corpus can be further used for relation extraction to locate species occurrences in literature-a useful task for monitoring species distribution and preserving the biodiversity.

"gnparser": a powerful parser for scientific names based on Parsing Expression Grammar

Background

Scientific names in biology act as universal links. They allow us to cross-reference information about organisms globally. However variations in spelling of scientific names greatly diminish their ability to interconnect data. Such variations may include abbreviations, annotations, misspellings, etc. Authorship is a part of a scientific name and may also differ significantly. To match all possible variations of a name we need to divide them into their elements and classify each element according to its role. We refer to this as ‘parsing’ the name. Parsing categorizes name’s elements into those that are stable and those that are prone to change. Names are matched first by combining them according to their stable elements. Matches are then refined by examining their varying elements. This two stage process dramatically improves the number and quality of matches. It is especially useful for the automatic data exchange within the context of “Big Data” in biology.

Results

We introduce Global Names Parser (gnparser). It is a Java tool written in Scala language (a language for Java Virtual Machine) to parse scientific names. It is based on a Parsing Expression Grammar. The parser can be applied to scientific names of any complexity. It assigns a semantic meaning (such as genus name, species epithet, rank, year of publication, authorship, annotations, etc.) to all elements of a name. It is able to work with nested structures as in the names of hybrids. gnparser performs with ≈99% accuracy and processes 30 million name-strings/hour per CPU thread. The gnparser library is compatible with Scala, Java, R, Jython, and JRuby. The parser can be used as a command line application, as a socket server, a web-app or as a RESTful HTTP-service. It is released under an Open source MIT license.

Conclusions

Global Names Parser (gnparser) is a fast, high precision tool for biodiversity informaticians and biologists working with large numbers of scientific names. It can replace expensive and error-prone manual parsing and standardization of scientific names in many situations, and can quickly enhance the interoperability of distributed biological information.

Electronic supplementary material

The online version of this article (doi:10.1186/s12859-017-1663-3) contains supplementary material, which is available to authorized users.

Constructing a biodiversity terminological inventory

The increasing growth of literature in biodiversity presents challenges to users who need to discover pertinent information in an efficient and timely manner. In response, text mining techniques offer solutions by facilitating the automated discovery of knowledge from large textual data. An important step in text mining is the recognition of concepts via their linguistic realisation, i.e., terms. However, a given concept may be referred to in text using various synonyms or term variants, making search systems likely to overlook documents mentioning less known variants, which are albeit relevant to a query term. Domain-specific terminological resources, which include term variants, synonyms and related terms, are thus important in supporting semantic search over large textual archives. This article describes the use of text mining methods for the automatic construction of a large-scale biodiversity term inventory. The inventory consists of names of species, amongst which naming variations are prevalent. We apply a number of distributional semantic techniques on all of the titles in the Biodiversity Heritage Library, to compute semantic similarity between species names and support the automated construction of the resource. With the construction of our biodiversity term inventory, we demonstrate that distributional semantic models are able to identify semantically similar names that are not yet recorded in existing taxonomies. Such methods can thus be used to update existing taxonomies semi-automatically by deriving semantically related taxonomic names from a text corpus and allowing expert curators to validate them. We also evaluate our inventory as a means to improve search by facilitating automatic query expansion. Specifically, we developed a visual search interface that suggests semantically related species names, which are available in our inventory but not always in other repositories, to incorporate into the search query. An assessment of the interface by domain experts reveals that our query expansion based on related names is useful for increasing the number of relevant documents retrieved. Its exploitation can benefit both users and developers of search engines and text mining applications.

Digitising legacy zoological taxonomic literature: Processes, products and using the output

By digitising legacy taxonomic literature using XML mark-up the contents become accessible to other taxonomic and nomenclatural information systems. Appropriate schemas need to be interoperable with other sectorial schemas, atomise to appropriate content elements and carry appropriate metadata to, for example, enable algorithmic assessment of availability of a name under the Code. Legacy (and new) literature delivered in this fashion will become part of a global taxonomic resource from which users can extract tailored content to meet their particular needs, be they nomenclatural, taxonomic, faunistic or other.

To date, most digitisation of taxonomic literature has led to a more or less simple digital copy of a paper original – the output of the many efforts has effectively been an electronic copy of a traditional library. While this has increased accessibility of publications through internet access, the means by which many scientific papers are indexed and located is much the same as with traditional libraries. OCR and born-digital papers allow use of web search engines to locate instances of taxon names and other terms, but OCR efficiency in recognising taxonomic names is still relatively poor, people’s ability to use search engines effectively is mixed, and many papers cannot be searched directly. Instead of building digital analogues of traditional publications, we should consider what properties we require of future taxonomic information access. Ideally the content of each new digital publication should be accessible in the context of all previous published data, and the user able to retrieve nomenclatural, taxonomic and other data / information in the form required without having to scan all of the original papers and extract target content manually. This opens the door to dynamic linking of new content with extant systems: automatic population and updating of taxonomic catalogues, ZooBank and faunal lists, all descriptions of a taxon and its children instantly accessible with a single search, comparison of classifications used in different publications, and so on. A means to do this is through marking up content into XML, and the more atomised the mark-up the greater the possibilities for data retrieval and integration. Mark-up requires XML that accommodates the required content elements and is interoperable with other XML schemas, and there are now several written to do this, particularly TaxPub, taxonX and taXMLit, the last of these being the most atomised. We now need to automate this process as far as possible. Manual and automatic data and information retrieval is demonstrated by projects such as INOTAXA and Plazi. As we move to creating and using taxonomic products through the power of the internet, we need to ensure the output, while satisfying in its production the requirements of the Code, is fit for purpose in the future.

Entity linking for biomedical literature

BACKGROUND

The Entity Linking (EL) task links entity mentions from an unstructured document to entities in a knowledge base. Although this problem is well-studied in news and social media, this problem has not received much attention in the life science domain. One outcome of tackling the EL problem in the life sciences domain is to enable scientists to build computational models of biological processes with more efficiency. However, simply applying a news-trained entity linker produces inadequate results.

METHODS

Since existing supervised approaches require a large amount of manually-labeled training data, which is currently unavailable for the life science domain, we propose a novel unsupervised collective inference approach to link entities from unstructured full texts of biomedical literature to 300 ontologies. The approach leverages the rich semantic information and structures in ontologies for similarity computation and entity ranking.

RESULTS

Without using any manual annotation, our approach significantly outperforms state-of-the-art supervised EL method (9% absolute gain in linking accuracy). Furthermore, the state-of-the-art supervised EL method requires 15,000 manually annotated entity mentions for training. These promising results establish a benchmark for the EL task in the life science domain. We also provide in depth analysis and discussion on both challenges and opportunities on automatic knowledge enrichment for scientific literature.

CONCLUSIONS

In this paper, we propose a novel unsupervised collective inference approach to address the EL problem in a new domain. We show that our unsupervised approach is able to outperform a current state-of-the-art supervised approach that has been trained with a large amount of manually labeled data. Life science presents an underrepresented domain for applying EL techniques. By providing a small benchmark data set and identifying opportunities, we hope to stimulate discussions across natural language processing and bioinformatics and motivate others to develop techniques for this largely untapped domain.

Finding needles in haystacks: linking scientific names, reference specimens and molecular data for Fungi

DNA phylogenetic comparisons have shown that morphology-based species recognition often underestimates fungal diversity. Therefore, the need for accurate DNA sequence data, tied to both correct taxonomic names and clearly annotated specimen data, has never been greater. Furthermore, the growing number of molecular ecology and microbiome projects using high-throughput sequencing require fast and effective methods for en masse species assignments. In this article, we focus on selecting and re-annotating a set of marker reference sequences that represent each currently accepted order of Fungi. The particular focus is on sequences from the internal transcribed spacer region in the nuclear ribosomal cistron, derived from type specimens and/or ex-type cultures. Re-annotated and verified sequences were deposited in a curated public database at the National Center for Biotechnology Information (NCBI), namely the RefSeq Targeted Loci (RTL) database, and will be visible during routine sequence similarity searches with NR_prefixed accession numbers. A set of standards and protocols is proposed to improve the data quality of new sequences, and we suggest how type and other reference sequences can be used to improve identification of Fungi.

Database URL:http://www.ncbi.nlm.nih.gov/bioproject/PRJNA177353

Knowledge extraction and semantic annotation of text from the encyclopedia of life

Numerous digitization and ontological initiatives have focused on translating biological knowledge from narrative text to machine-readable formats. In this paper, we describe two workflows for knowledge extraction and semantic annotation of text data objects featured in an online biodiversity aggregator, the Encyclopedia of Life. One workflow tags text with DBpedia URIs based on keywords. Another workflow finds taxon names in text using GNRD for the purpose of building a species association network. Both workflows work well: the annotation workflow has an F1 Score of 0.941 and the association algorithm has an F1 Score of 0.885. Existing text annotators such as Terminizer and DBpedia Spotlight performed well, but require some optimization to be useful in the ecology and evolution domain. Important future work includes scaling up and improving accuracy through the use of distributional semantics.

BioNames: linking taxonomy, texts, and trees

BioNames is a web database of taxonomic names for animals, linked to the primary literature and, wherever possible, to phylogenetic trees. It aims to provide a taxonomic “dashboard” where at a glance we can see a summary of the taxonomic and phylogenetic information we have for a given taxon and hence provide a quick answer to the basic question “what is this taxon?” BioNames combines classifications from the Global Biodiversity Information Facility (GBIF) and GenBank, images from the Encyclopedia of Life (EOL), animal names from the Index of Organism Names (ION), and bibliographic data from multiple sources including the Biodiversity Heritage Library (BHL) and CrossRef. The user interface includes display of full text articles, interactive timelines of taxonomic publications, and zoomable phylogenies. It is available at http://bionames.org.

Utilizing descriptive statements from the biodiversity heritage library to expand the Hymenoptera Anatomy Ontology

Hymenoptera, the insect order that includes sawflies, bees, wasps, and ants, exhibits an incredible diversity of phenotypes, with over 145,000 species described in a corpus of textual knowledge since Carolus Linnaeus. In the absence of specialized training, often spanning decades, however, these articles can be challenging to decipher. Much of the vocabulary is domain-specific (e.g., Hymenoptera biology), historically without a comprehensive glossary, and contains much homonymous and synonymous terminology. The Hymenoptera Anatomy Ontology was developed to surmount this challenge and to aid future communication related to hymenopteran anatomy, as well as provide support for domain experts so they may actively benefit from the anatomy ontology development. As part of HAO development, an active learning, dictionary-based, natural language recognition tool was implemented to facilitate Hymenoptera anatomy term discovery in literature. We present this tool, referred to as the 'Proofer', as part of an iterative approach to growing phenotype-relevant ontologies, regardless of domain. The process of ontology development results in a critical mass of terms that is applied as a filter to the source collection of articles in order to reveal term occurrence and biases in natural language species descriptions. Our results indicate that taxonomists use domain-specific terminology that follows taxonomic specialization, particularly at superfamily and family level groupings and that the developed Proofer tool is effective for term discovery, facilitating ontology construction.

Applications of natural language processing in biodiversity science

Centuries of biological knowledge are contained in the massive body of scientific literature, written for human-readability but too big for any one person to consume. Large-scale mining of information from the literature is necessary if biology is to transform into a data-driven science. A computer can handle the volume but cannot make sense of the language. This paper reviews and discusses the use of natural language processing (NLP) and machine-learning algorithms to extract information from systematic literature. NLP algorithms have been used for decades, but require special development for application in the biological realm due to the special nature of the language. Many tools exist for biological information extraction (cellular processes, taxonomic names, and morphological characters), but none have been applied life wide and most still require testing and development. Progress has been made in developing algorithms for automated annotation of taxonomic text, identification of taxonomic names in text, and extraction of morphological character information from taxonomic descriptions. This manuscript will briefly discuss the key steps in applying information extraction tools to enhance biodiversity science.