Biodiversity Heritage Library Research Articles

The Pensoft Annotator: A new tool for text annotation with ontology terms

Introduction Digitisation of biodiversity knowledge from collections, scholarly literature and various research documents is an ongoing mission of the Biodiversity Information Standards (TDWG) community. Organisations such as the Biodiversity Heritage Library make historical biodiversity literature openly available and develop tools to allow biodiversity data reuse and interoperability. For instance, Plazi transforms free text into machine-readable formats and extracts collection data and feeds it into the Global Biodiversity Information Facility (GBIF) and other aggregators. All of these digitisation workflows require a lot of effort to develop and implement in practice. In essence, what these digitisation activities entail are the mapping of free text to concepts from recognised vocabularies or ontologies in order to make the content understandable to computers. Aim We aim to address the problem of mapping free text to ontological terms ("strings to things") with our tool for text-to-ontology mapping: the Pensoft Annotator. Methods & Implementation The Annotator is a web application that performs direct text matching to terms from any ontology or vocabulary list given as input to the Annotator. The term 'ontology' is used loosely here and means a collection of terms and their synonyms, where terms are uniquely identified via a Uniform Resource Identifier (URI). The Annotator accepts any of the following ontology formats (e.g. OBO, OWL, RDF/XML, etc.) but does not require the existence of a proper ontology structure (logical statements). We use the ROBOT command line tool to convert any of these formats to JSON. After the upload of a new ontology, the Annotator processes the ontology terms by normalising all exact synonyms and by removing all of the other synonyms (related, narrow and broad synonyms). This is done to limit the number of false positive matches and to preserve the semantic similarity between the matched ontology term and the text. After matching the words in the input text and the ontology term labels, the Pensoft Annotator returns a table of matched ontology terms including the following fields: the identifier of the ontology term, the ontology term label or the label of the synonym, the starting position of the matched term in the text, the term context (words surrounding the matched term in the text), the type of ontology term (class or property), the ontology from which the matched term originates and the number of times a given term is mentioned in the text. The Pensoft Annotator allows simultaneous annotation with multiple ontologies. To better visualise the exact ontology from which a matching term has been found, the terms are highlighted in different colour depending on the ontology. The Pensoft Annotator is also accessible programmatically via an Application Programming Interface (API), documented at https://annotator.pensoft.net/api. Discussion & Use Cases The Pensoft Annotator provides functionalities that will aid the transformation of free text to collections of semantic resources. However, it still requires expert knowledge to use as the ontologies need to be selected carefully. Some false positive matches from the annotation are possible because we do not perform semantic analysis of the texts. False negatives are also possible since there might be different word forms of ontology terms, which are not direct matches to them (e.g. 'wolf' and 'wolves'). For this reason, matched terms can be reviewed and removed from the results within the web interface of the Pensoft Annotator. After removal of terms, they will not be present in the downloaded results. The Pensoft Annotator can be used to annotate biodiversity and taxonomic literature to help with the extraction of biodiversity knowledge (e.g. species habitat preferences, species interaction data, localities, biogeographic data). The existence of some domain and taxon-specific ontologies, such as the Hymenoptera Anatomy Ontology, provides further opportunities for context-specific annotation. Semantic analysis of unstructured texts could be applied in addition to ontology annotation to improve the accuracy of ontology term matching and to filter out mismatched terms. Annotation of structured or semi-structured text (e.g. tables) can be done with better success. A recent example demonstrates the use of the Annotator to extract biotic interactions from tables (Dimitrova et al. 2020). The Annotator could also be used for ontology analysis and comparison. Annotation of text can help to discover gaps in ontologies as well as inaccurate synonyms. For instance, a certain word could be recognised as an ontology term match because it is an exact synonym in the ontology but in reality it might be more accurate to mark it as a related synonym. In addition, annotation with multiple ontologies can help to elucidate links between ontologies.

The Biodiversity Heritage Library Response to the Global COVID-19 Pandemic

The global onset of the COVID-19 pandemic began in January 2020. In February, many academic and research institutions began to move to a telework environment. By March, many borders had been closed and nearly all institutions fully transitioned to telework. With their institutions closed and travel shut down, researchers lacked access to both specimen collections as well as the library and archives material needed to continue productive research. Preliminary research on the impact of COVID-19 on the research enterprise has been published. In the United States, the impact on federally funded research has been documented in Effects of COVID-19 on the Federal Research and Development Enterprise (Morgan and Sargent Jr. 2020). Likewise, What Happens to the Continuity and Future of the Research Enterprise? (Coalition for Networked Information 2020) is an initial appraisal of issues and concerns within the North American academic and research environment. Library and archives staff were limited in their access to collections-based work and needed valuable and impactful telework. The Biodiversity Heritage Library (BHL) was founded as a global, dispersed, and virtual library and thus was well-suited to continue to provide service to teleworking researchers as well as telework tasks for staff at BHL's partner institutions. Examples of the telework projects include enhancing BHL metadata, and uploading BHL illustrations to the BHL Flickr collection. BHL makes critical information available thanks to all of the partners' efforts to digitize for open access. The natural strengths of the BHL virtual organization include social distancing, virtual collaboration and opening up science; the pandemic world is a natural ecosystem for BHL partners. Initial analyses of BHL use have shown a significant increase in year-over-year usage of BHL content. The preliminary statistics have shown an approximate 25% increase in usage of the BHL site. BHL has also collected anecdotal evidence in the form of user comments and direct conversations with research staff at BHL partner institutions. Many national or pan-national organizations and projects were instigated to examine the impact of COVID-19 on library and archives materials. One key project that some BHL partners participated in is Reopening Archives, Libraries, and Museums (REALM) Information Hub: A COVID-19 Research Project. With most BHL partners pausing all digitization around the February-March 2020 time period, outcomes from studies such as REALM will assist partners in understanding when it will be safe for digitization to restart. This presentation will cover how usage of BHL changed from early 2020 to the present, showing trends and significant impact. The pivot of many BHL partner staff to improve BHL content and metadata for their local users as well as the global biodiversity community will be documented. Also covered will be the enhanced impact of digital libraries, such as BHL, in both the short and long terms along with lessons learned for expanded virtual open access for library collections supporting biodiversity and museum research.

Improving Taxonomic Name Finding in the Biodiversity Heritage Library

As the world’s largest open access digital library for biodiversity literature and archives, the Biodiversity Heritage Library (BHL) provides access to over a quarter-million volumes of natural history literature to researchers around the world. One of its services is to index taxonomic names in the collection to allow researchers to locate publications about specific taxa. The Global Names Architecture (GNA) is a system of web services to register, find, index, check and organize biological scientific names. GNA recently developed a new Name Finding algorithm and tool that has been integrated with BHL to improve taxonomic name searches within BHL. In our presentation, we will discuss a brief history of name finding in BHL, development of the Name Finding algorithm, results from implementing the algorithm, and challenges that still await us in the realm of taxonomic name finding in BHL.

Towards a scientific workflow featuring Natural Language Processing for the digitisation of natural history collections

We describe an effective approach to automated text digitisation with respect to natural history specimen labels. These labels contain much useful data about the specimen including its collector, country of origin, and collection date. Our approach to automatically extracting these data takes the form of a pipeline. Recommendations are made for the pipeline's component parts based on some of the state-of-the-art technologies.Optical Character Recognition (OCR) can be used to digitise text on images of specimens. However, recognising text quickly and accurately from these images can be a challenge for OCR. We show that OCR performance can be improved by prior segmentation of specimen images into their component parts. This ensures that only text-bearing labels are submitted for OCR processing as opposed to whole specimen images, which inevitably contain non-textual information that may lead to false positive readings. In our testing Tesseract OCR version 4.0.0 offers promising text recognition accuracy with segmented images.Not all the text on specimen labels is printed. Handwritten text varies much more and does not conform to standard shapes and sizes of individual characters, which poses an additional challenge for OCR. Recently, deep learning has allowed for significant advances in this area. Google's Cloud Vision, which is based on deep learning, is trained on large-scale datasets, and is shown to be quite adept at this task. This may take us some way towards negating the need for humans to routinely transcribe handwritten text.Determining the countries and collectors of specimens has been the goal of previous automated text digitisation research activities. Our approach also focuses on these two pieces of information. An area of Natural Language Processing (NLP) known as Named Entity Recognition (NER) has matured enough to semi-automate this task. Our experiments demonstrated that existing approaches can accurately recognise location and person names within the text extracted from segmented images via Tesseract version 4.0.0. Potentially, NER could be used in conjunction with other online services, such as those of the Biodiversity Heritage Library to map the named entities to entities in the biodiversity literature (https://www.biodiversitylibrary.org/docs/api3.html).We have highlighted the main recommendations for potential pipeline components. The document also provides guidance on selecting appropriate software solutions. These include automatic language identification, terminology extraction, and integrating all pipeline components into a scientific workflow to automate the overall digitisation process.

Use cases for Taxonomic Name Services

Catalogue of Life Plus (CoL+) was started in 2017 as a collaborative project between the Catalogue of Life (COL), the Global Biodiversity Information Facility (GBIF), Naturalis Biodiversity Center, Barcode of Life Database (BOLD), Biodiversity Heritage Library (BHL), Encyclopedia of Life (EOL), Integrated Taxonomic Information System (ITIS) and representatives from natural history museum collections. The participants jointly recognized a need for the wider biodiversity community to coalesce around common and coordinated approaches for mobilizing and inter-linking biodiversity data and information across a range of scales, themes and types. The Catalogue of Life was identified as being the clear choice to deliver taxonomic services to support these partners. Taxonomic services refer to solutions that address challenges inherent to scientific names: the ubiquitous, and often sole, contextual linkage between a biodiversity data object and the taxon to which it is linked. Without access to expert-mediated taxonomic knowledge, biodiversity data networks risk integrating or delivering biodiversity data that is incomplete or may not make taxonomic sense. As a globally comprehensive and expert-mediated taxonomic resource, the Catalogue of Life is well-positioned to provide the underlying data and services to address these challenges and support a broad community of stakeholders. With financial support from the Netherlands Biodiversity Information Facility, the Netherlands Ministry of Education, Science, and Culture as well as technical support from GBIF, CoL+ represents an effort to stabilize the Catalogue of Life checklist products and refine a roadmap for future services -services that must be informed by the needs of users. The COL+ initiative provides an opportunity to document and understand the requirements of both Col current users and a wider future user base. Key to any expansion is ensuring a clear understanding and the appropriate response to user needs. This includes not only the requirements of "taxonomic data consumer" stakeholders, such as those listed above, but also the needs and motivations of the expert taxonomic community behind the current and future Catalogue of Life. This session will focus on reviewing the user requirements of a diverse community of both taxonomic data producers and consumers and organizing these stories into a set of categories from which we will identify clear and useful tools and services. These will establish the basis for a compelling roadmap for the Catalogue of Life as an enabling foundation to an interlinked and global biodiversity data infrastructure.

BHL and Specimen Collection Data: The needle in the Festuca stack

Data contained in the the Biodiversity Heritage Library (BHL) describes collections held in the world's major museums. Finding those collections data, however, remains a challenge. A literal needle in a Festuca stack as some have noted. BHL is actively engaging in incorporating tools (including Digital Object Identifier's (DOI's)and the recently launched full-text search) to make finding and linking to collection specimen information better. Still, it is not easy to find specific collections information in the non-semantically tagged BHL content. This session will call for ideas on how to locate this content.. BHL is an international consortium, making research literature openly available to the world as part of a global biodiversity community. The BHL was created in 2006 as a direct response to the needs of the taxonomic community for access to early literature. The original BHL organizational model, based on United States and United Kingdom partners, provided a template for what is now over 80 global partners. Through this extensive network of Members, Affiliates, and partners, over 56 million pages of biodiversity literature are available through the BHL portal. BHL changes the lives of researchers and assists the work of collections managers. By enhancing daily research at the Smithsonian and Harvard, BHL provides a global network of researchers with an easy-to-use digital library of content and services.

Taxonomy and Systematics in Biodiversity Informatics: Lessons learned from ichthyology and general perspectives

Although still strongly intertwined, taxonomy and systematics are diverging more and more in their paradigms, methods, and agendas: it is not possible to consider them as synonyms anymore. While taxonomy remains an analytical science based on abductive reasoning (trying to find the historical pathways leading to present species delineations), systematics diverged as an information science since the rise of computers: summarizing, organising, and exposing taxonomical, biological, and ecological data, information, and knowledge in the most efficient ways, with respect to various targeted audiences. One could even consider synonymising biodiversity informatics with systematics instead! Schematically, this led to two different types of information systems: one dedicated to pure taxonomic and nomenclatural data; one oriented to record life-traits. Obviously, the latter must be built along reliable taxonomic backbones, therefore the former should have been built before the latter. It did not happen as exemplified in fishes by the Food and Agriculture Organisation (FAO) Fisheries Global Information System (FIGIS) and related, Catalog of Fishes, and FishBase, and later the databases of the International Union for Conservation of Nature (IUCN) and World Register of Marine Species (WoRMS), or for aggregators with Catalogue of Life and e.g., GBIF, Encyclopedia of Life. This has created some confusion for end users, “which system should I use” being their regular question with the invariable answer, “it depends”. In the absence of a formal preexisting taxonomic information system, each biodiversity information system has developed its own way to manage its taxonomic backbone with more or less impact of taxonomists. For fishes, we are not far from reconciling the various systems, and data to knowledge flows are becoming clearer, but it is not without unnecessary extra work. A real breakthrough is necessary to move from a collaboration stage to a cooperative stage, where systems are interconnected (not necessarily integrated) in such a way that the same taxonomic work is not repeated over and over to synchronise the systems. The difficulty is that taxonomic information systems must be designed for the needs of taxonomists, while their resulting classifications and the way they are exposed must fit the needs of systematics/biodiversity systems purposes, and by extension of the rest of scientific domains and the society in general. Conditions for this breakthrough to happen are discussed. The breakthrough does not reside in only one action but rather is the result of multiple simultaneous advances in the theory of taxonomy, its (mathematical?) formalization and informatics implementation, technology (although progress in that domain may be well in advance over others), data entry, networking, and sociology of science. The "potential taxon" concept (Berendsohn 1995) led to important theoretical progresses but its actual implementation lags behind in many systems, due probably to the huge effort of data entry it requires. Data entry is certainly a part that was neglected at the beginning of biodiversity informatics, because it has to be sustained endlessly, while the development of new systems was seen as more rewarding in time-limited frameworks. This has been corrected at least for occurrences and specimens, with the development of national and international digitization programs. Besides, the development of the Biodiversity Heritage Library and text extraction technologies is quite promising for taxon information. As in all complex situations with many interacting dimensions (e.g., fisheries management in ichthyology), progress must be balanced among all dimensions to have effective results for the overall domain. Among others, issues in sociology of sciences, for instance, must be addressed to make significant progresses. In particular the way data, information and knowledge are published, and jobs delineated and careers evaluated, must still be seriously reviewed in the light of information system development.

Big Data Knowledge of Major Lineages of Life and Priorities for Genomic Research

Genomic science is revolutionizing and accelerating biodiversity research. For collections-based institutions to continue to lead and support biodiversity research, they must adapt to this new reality. Simultaneously, “big data” is accumulating so rapidly that we have unprecedented capacity to plan strategically to use genomics to advance basic and applied science on multiple fronts. For example, seven “big data” sources have the following numbers of records (2018 data): Global Biodiversity Information Facility (GBIF), ~1B; Biodiversity Heritage Library (BHL), ~3.6M; National Center for Biotechnology Information (NCBI), ~220M; Open Tree of Life (OToL), 1.9M; Barcode of Life Data System (BOLD), ~6.3M; Encyclopedia of Life (EOL), ~99K; Global Genome Biodiversity Network (GGBN), ~2M. Collectively, they offer more than 1.2B records on biodiversity. At the scale of species (~2M described, multiple millions undescribed), these data are still too sparse to permit comprehensive conclusions. At the scale of families (i.e. deeper clades of life), the situation is far more promising: about 9,911 families are known, and relatively few are discovered each year. This suggests that at the family rank (and above), our knowledge of life on Earth is reasonably complete. Approximately 160,000 valid and accepted genera exist, but certainly many new genera await discovery and description. Genomics is the fastest way to group species into more inclusive lineages such as genera and families, and is certainly faster than traditional alpha taxonomy. Synergistically, these “big data” answer four important questions at deeper clade levels: What is it? Where is it? What do we know about it? What do we know about its genome? Approximately 4,500 eukaryotic genomes have been sequenced. The converse of what we know is what we do not know, another meaning of “dark taxa.” We can use the distribution and density of big data at deeper clade levels (families, genera) to quantitatively analyze “dark taxa” and therefore to strategically optimize knowledge and preservation of biodiversity at a global scale. Technicalities of the quantitative prioritization scheme are debatable, but some initial, simple scoring systems can help to prioritize lineages for collection and genetic research so as to most efficiently illuminate regions in the tree of life that that are neither preserved, imaged, geo-located, studied, nor known genomically. This analysis presents criteria and goals for collaborating to build a global genomic collection to maximize efficient acquisition of biodiversity genomic knowledge, and identifies the most valuable and highest priority taxa for genomic research.

When Data Management Meets Project Management

Complex projects that collect, curate and analyse biodiversity data are often presented with the challenge of accommodating diverse data types, various curation and output workflows, and evolving project logistics that require rapid changes in the applications and data structures. At the same time, sustainability concerns and maintenance overheads pose a risk to the long term viability of such projects. We advocate the use of flexible, multiplatform tools that adapt to operational, day-to-day challenges while providing a robust, cost efficient, and maintainable framework that serves the needs data collectors, managers and users. EarthCape is a highly versatile platform for managing biodiversity research and collections data, associated molecular laboratory data (Fig. 1), multimedia, structured ecological surveys and monitoring schemes, and more. The platform includes a fully functional Windows client as well as a web application. The data are stored in the cloud or on-premises and can be accessed by users with various access and editing rights. Ease of customization (making changes to user interface and functionality) is critical for most environments that deal with operational research processes. For active researchers and curators, there is rarely time to wait for a cycle of development that follows a change or feature request. In EarthCape, most of the changes to the default setup can be implemented by the end users with minimum effort and require no programming skills. High flexibility and a range of customisation options is complemented with mapping to Darwin Core standard and integration with GBIF, Geolocate, Genbank, and Biodiversity Heritage Library APIs. The system is currently used daily for rapid data entry, digitization and sample tracking, by such organisations as Imperial College, University of Cambridge, University of Helsinki, University of Oxford. Being an operational data entry and retrieval tool, EarthCape sits at the bottom of Virtual Research Environments ecosystem. It is not a software or platform to build data repositories, but rather a very focused tool falling under "back office" software category. Routine label printing, laboratory notebook maintenance, rapid data entry set up, or any other of relatively loaded user interfaces make use of any industry standard relational database back end. This opens a wide scope for IT designers to implement desired integrations within their institutional infrastructure. APIs and developer access to core EarthCape libraries to build own applications and modules are under development. Basic data visualisation (charts, pivots, dashboards), mapping (full featured desktop GIS module), data outputs (report and label designer) are tailored not only to research analyses, but also for managing logistics and communication when working on (data) papers. The presentation will focus on the software platform featuring most prominent use cases from two areas: ecological research (managing complex network data digitization project) and museum collections management (herbarium and insect collections).

Mission: Implausible — Revealing rogue marine species in records across biodiversity data platforms

Online biodiversity platforms publish datasets with graphic tools to help with quality control of submitted records, but more could be done to make the data robust for ecological analyses. Attention has focused mostly on automating tools for obvious errors, including misspelled names and synonyms, dates, or coordinates. However, a manual review of species identifications and distributions may uncover improbable records, such as a species reported in an area far from its usual range, or a rare species found in an area that has many more records of a related species. Examples are shown by constructing checklists in the Northwest Atlantic, using information from the World Register of Marine Species (WoRMS, http://www.marinespecies.org) and the Ocean Biogeographic Information System (OBIS, https://obis.org). Reviewing rare species records revealed some misidentifications, but in other instances, the rare species was valid while it was the commonly reported species that needed correction. Confirmations were obtained by comparing records from different regions, but also across platforms, including photos from observers on iNaturalist Canada (https://inaturalist.ca), genetic analyses on Barcode of Life Data systems (BOLD, http://www.boldsystems.org), and literature in the Biodiversity Heritage Library (BHL, https://www.biodiversitylibrary.org). While this exercise succeeded in validating the marine taxa of a region, it is an obvious candidate for automation in three areas: 1) flagging records of improbable taxa in a region, 2) comparing records with different types of information (e.g., specimen photos, genetic groupings, or literature records), and 3) updating users and providers when records get flagged as unusual or are modified. The first approach could be explored using online graphics tools or R software packages (rOpenSci, https://ropensci.org). The second toolset, comparing records across platforms, is partially realized with some linkages already operating between WoRMS, OBIS, BOLD, BHL, iNaturalist, and the Global Biodiversity Information Facility (GBIF, https://www.gbif.org). The third target will be the most difficult to implement, requiring reliable platform cross-linkages and specimen record identifiers to send notifications of changed status of records to both users and the original data source. Ongoing work is discussed on communicating the need to review records across platforms, with the hope that toolsets will be developed to make this task easier.

Furthering Genomic Research Infrastructures: The Global Genome Biodiversity Network

Most successful research programs depend on easily accessible and standardized research infrastructures. Until recently, access to tissue or DNA samples with standardized metadata and of a sufficiently high quality, has been a major bottleneck for genomic research. The Global Geonome Biodiversity Network (GGBN) fills this critical gap by offering standardized, legal access to samples. Presently, GGBN’s core activity is enabling access to searchable DNA and tissue collections across natural history museums and botanic gardens. Activities are gradually being expanded to encompass all kinds of biodiversity biobanks such as culture collections, zoological gardens, aquaria, arboreta, and environmental biobanks. Broadly speaking, these collections all provide long-term storage and standardized public access to samples useful for molecular research. GGBN facilitates sample search and discovery for its distributed member collections through a single entry point. It stores standardized information on mostly geo-referenced, vouchered samples, their physical location, availability, quality, and the necessary legal information on over 50,000 species of Earth’s biodiversity, from unicellular to multicellular organisms. The GGBN Data Portal and the GGBN Data Standard are complementary to existing infrastructures such as the Global Biodiversity Information Facility (GBIF) and International Nucleotide Sequence Database (INSDC). Today, many well-known open-source collection management databases such as Arctos, Specify, and Symbiota, are implementing the GGBN data standard. GGBN continues to increase its collections strategically, based on the needs of the research community, adding over 1.3 million online records in 2018 alone, and today two million sample data are available through GGBN. Together with Consortium of European Taxonomic Facilities (CETAF), Society for the Preservation of Natural History Collections (SPNHC), Biodiversity Information Standards (TDWG), and Synthesis of Systematic Resources (SYNTHESYS+), GGBN provides best practices for biorepositories on meeting the requirements of the Nagoya Protocol on Access and Benefit Sharing (ABS). By collaboration with the Biodiversity Heritage Library (BHL), GGBN is exploring options for tagging publications that reference GGBN collections and associated specimens, made searchable through GGBN’s document library. Through its collaborative efforts, standards, and best practices GGBN aims at facilitating trust and transparency in the use of genetic resources.

Extracting Data from Legacy Taxonomic Literature: Applications for planning field work

Traditional taxonomic publications have served as a biological data repository accumulating vast amounts of data on species diversity, geographical and temporal distributions, ecological interactions, taxonomic relations, among many other types of information. However, the fragmented nature of taxonomic literature has made this data difficult to access and use to its full potential. Current anthropogenic impact on biodiversity demands faster knowledge generation, but also making better use of what we already have. This could help us make better-informed decisions about conservation and resources management. In past years, several efforts have been made to make taxonomic literature more mobilized and accessible. These include online publications, open access journals, the digitization of old paper literature and improved availability through online specialized repositories such as the Biodiversity Heritage Library (BHL) and the World Spider Catalog (WSC), among others. Although easy to share, PDF publications still have most of their biodiversity data embedded in strings of text making them less dynamic and more difficult or impossible to read and analyze without a human interpreter. Recently developed tools as GoldenGATE-Imagine (GGI) allow transforming PDFs in XML files that extract and categorize taxonomically relevant data. These data can then be aggregated in databases such as Plazi TreatmentBank, where it can be re-explored, queried and analyzed. Here we combined several of these cybertaxonomic tools to test the data extraction process for one potential application: the design and planning of an expedition to collect fresh material in the field. We targeted the ground spider Teutamus politus and other related species from the Teutamus group (TG) (Araneae; Liocranidae). These spiders are known from South East Asia and have been cataloged in the family Liocranidae; however, their relations, biology and evolution are still poorly understood. We marked-up 56 publications that contained taxonomic treatments with specimen records for the Liocranidae. Of these publications, 20 contained information on members of the TG. Geographical distributions and occurrences of 90 TG species were analyzed based on 1,309 specimen records. These data were used to design our field collection in a way that allowed us to optimize the collection of adult specimens of our target taxa. The TG genera were most common in Indonesia, Thailand and Malaysia. From these, Thailand was the second richest but had the most records of T. politus. Seasonal distribution of TG specimens in Thailand suggested June and July as the best time for collecting adults. Based on these analyses, we decided to sample from mid-July to mid-August 2018 in the three Thai provinces that combined most records of TG species and T. politus. Relying on the results of our literature analyses and using standard collection methods for ground spiders, we captured at least one specimen of every TG genus reported for Thailand. Our one-month expedition captured 231 TG spiders; from these, T. politus was the most abundant species with 188 specimens (95 adults). By comparison, a total of 196 specimens of the TG and 66 of T. politus had been reported for the same provinces in the last 40 years. Our sampling greatly increased the number of available specimens, especially for the genera Teutamus and Oedignatha. Also, we extended the known distribution of Oedignatha and Sesieutes within Thailand. These results illustrate the relevance of making biodiversity data contained within taxonomic treatments accessible and reusable. It also exemplifies one potential use of taxonomic legacy data: to more efficiently use existing biodiversity data to fill knowledge gaps. A similar approach can be used to study neglected or interesting taxa and geographic areas, generating a better biodiversity documentation that could aid in decision making, management and conservation.

It’s Not Always FAIR: Choosing the Best Platform for Your Biodiversity Heritage Literature

The Biodiversity Heritage Library (BHL) is a global consortium of over 460 contributors who have specifically chosen the BHL as the online platform to host their digitised biodiversity heritage literature. In 2018, staff from BHL Australia (a project funded by the Atlas of Living Australia and hosted by Museums Victoria) travelled across the Tasman Sea to work with the Auckland War Memorial Museum (Tāmaki Paenga Hira) to start a new branch of the BHL – BHL New Zealand. The Auckland War Memorial Museum has a novel approach when it comes to its own online content. They believe that the best way to amplify their collections is not via their own website, but via external collaborations. Acknowledging that their own website cannot best serve the many different types of content they curate and create, they seek the external platforms that can – platforms that have the specific functionality required to make each type of content findable, accessible, interoperable and reusable (FAIR), and that will reach and impact the greatest audience. When the Auckland Museum was seeking the best platform for their biodiversity heritage literature, they were attracted by the BHL’s functionality (such as Full Text Search, Optical Character Recognition and Global Names Recognition and Discovery); the extended discoverability and interoperability enabled via the BHL’s partnerships (such as with the Encyclopedia of Life, the Global Biodiversity Information Facility, and the Consortium of European Taxonomic Facilities); and BHL’s massive audience. However, the BHL is not the only platform hosting biodiversity heritage literature. Commercial publishers are rapidly uploading the back issues of the journals they host. There are now thousands of out-of-copyright journal articles on websites such as Wiley Online Library and Taylor & Francis Online. Historic literature is not the primary focus for these publishers, but the functionality they provide for journal articles has brought the historic publications they host into the modern linked network of scholarly research. Every article on their websites, whether it was published 200 years ago or today, receives a Digital Object Identifier (a DOI). A DOI is a unique permanent identifier assigned to a piece of online content (usually) at the time of its publication – a persistent link that should be included in every citation of that piece of content. In an ideal world, the DOI system would enable readers to link from article to article in a never-ending trail of knowledge. However, a significant proportion of our historic literature still sits outside this linked network, representing dead ends (or at least annoying setbacks) in the quest for information. The other invaluable benefit of DOIs is that they provide citation metrics, allowing authors, publishers and institutions to track the use and impact of their publications. The historic literature is the foundation upon which our understanding of biodiversity is based. It is critical that this foundation is easily discoverable, openly accessible, persistently linkable and unambiguously trackable. The BHL – the world’s largest online library of biodiversity heritage literature – has an enormous role to play in ensuring that this literature is brought into the DOI system, and that publishers and researchers continue to have the option of an online platform that not only provides the best functionality for biodiversity heritage literature, but also open access to it.

Introduction to the Symposium: Improving access to hidden scientific data in the Biodiversity Heritage Library

This will be a short introduction to the symposium: Improving access to hidden scientific data in the Biodiversity Heritage Library. The symposium will present examples of how the Biodiversity Heritage Library (BHL) collaborates across the international consortium and with community partners around the world to help enhance access to the biodiversity literature. Literature repositories, particularly the BHL collections, have been recognized as critical to the global scientific community. A diverse global user community propels BHL and BHL users to develop access tools beyond the standard “title, author, subject” search. BHL utilizes the Global Names Recognition and Discovery (GNRD) service to identify taxonomic names within text rendered by Optical Character Recognition (OCR) software, enabling scientific name searches and creation of species-specific bibliographies, critical to systematics research. In this symposium, we will hear from international partners and creative users making data from the BHL globally accessible for the kinds of larger-scale analysis enabled by BHL’s full-text search capabilities and Application Program Interface (API) protocols. In addition to taxonomic name services already incorporated in BHL, the consortium has also begun exploring georeferencing strategies for better searching and potential connections with key biodiversity resources such as the Global Biodiversity Information Facility (GBIF). With many different institutions around the world participating, the ability to work together virtually is critical for a seamless end product that meets the demands of the international community as well as the needs of local institutions.

Text-mining BHL: towards new interfaces to the biodiversity literature

The taxonomic literature is one of the largest resources of information on biodiversity, both current and in the past. Unlike many scientific disciplines this literature remains perpetually relevant as successive taxonomic work builds upon those earlier foundations. Projects such as the Biodiversity Heritage Library (BHL) have greatly increased access to that literature, as have numerous independent digitisation efforts by museums, herbaria, and publishers. But the focus of this access has been human readers, with limited use of text mining tools, mostly focussed on extracting taxonomic names. This talk explores other kinds of data that can be extracted from text on BHL and elsewhere, focusing on taxonomic names, geographic localities and specimen codes in the context of the BioStor project (https://biostor.org, Page 2011). The problem of finding taxonomic names in text has been well studied (e.g., Akella et al. 2012), and new BHL content is continuously indexed by names. Despite this, there is only weak linkage between taxonomic name databases and BHL. Even projects that create these links (e.g., BioNames, Page 2013) do not enable links in the reverse direction. In other words, a BHL reader is unaware whether the appearance of a name on a page is the first publication of that name, nor are they told of the fate of a name in subsequent research. The absence of these links reduces the value of BHL to working taxonomists. In addition to taxonomic names, a typical taxonomic paper often contains specimen codes. Extracting these from text and linking them to digital representations, such as occurrence records in GBIF, opens up the possibility to provide detailed provenance for occurrence data, as well as citation-based metrics for the utility of natural history collections. Taxonomic papers are also often rich in geographic information. A simple method for extracting locality information from text is to search for latitude and longitude coordinates, and BioStor currently does this. To date some 83,000 individual point localities have been extracted (Fig. 1 ). These are used to provide a simple geographic search interface in BioStor, and are also harvested by JournalMap (Karl et al. 2013). But these localities are not linked to the original location in the source text, nor are they linked to any associated specimens, so they cannot be interpreted as occurrences that could be harvested by GBIF. If the goal is to contribute to GBIF then we need tools that can parse locality information and link that to associated specimens. A general framework for handling data on taxonomic names, specimens, and geographic localities in text is to treat them as annotations (Batista-Navarro et al. 2017). By modelling annotations using the Web Annotation Data Model (https://www.w3.org/TR/annotation-model/ ) we can incorporate these annotations into biodiversity knowledge graphs (Page 2016). We can also combine these annotations with new standards for describing digitised content, such as the International Image Interoperability Framework (IIIF, https://iiif.io). The implications of this approach for developing new interfaces to the biodiversity literature will be discussed.

Finding scientific names in Biodiversity Heritage Library, or how to shrink Big Data

The Biodiversity Heritage Library contains 57 million pages of biological information. The majority of this information is a scanned and digitized non-structured text. This "raw" text is hard to access by computers or humans, without the addition of rich metadata. Recent improvements in natural language processing (NLP) and machine learning (ML) promise to facilitate the creation of such metadata. One obvious approach to improve BHL usability is to extract and provide an index of scientific names thereby enabling biologists to find useful information easier and faster. The Global Names Architecture (GNA) detects, verifies, collects, and indexes scientific names from many sources. Six years ago GNA developers created an index of the scientific names in the BHL by parsing every page one by one. This took 45 days to accomplish. Almost immediately BHL users began to find problems in the index and suggest improvements. However, the cost of repeating such a gigantic job was insurmountable and as a result the index remained nearly unchanged for 6 years. Two problems were at the heart of dealing with the “Big Data” of the BHL, the time it took to transfer the raw data prior to processing, and the computational time it took to detect the names themselves.To solve these problems we could either throw more hardware resources into the problem (expensive), or find ways to dramatically improve performance of the tasks (cheaper). We decided to achieve our goal by utilizing hardware more effectively, and by using fast, scalable programming languages. We wrote several Open Source applications in Go and Scala to detect candidate scientific names then verify them as names by comparing them to 27 million scientific name-strings aggregated by GNA. We were able to speed up data mobilization from 24 hours to 11 minutes, and decrease the time for name detection from 35 days to 5 hours. Name-verification time decreased from 10 days to 9 hours. Overall our computing requirements shrank from 4 high-end servers to one modern laptop. As a result we achieved our goal and indexed BHL in only 14 hours and unlocked the reality of iterative improvements to the scientific name index. We also wanted to make it possible to study BHL data in its entirety remotely, in real-time. We created an HTTP2 service that is able to stream gigantic amounts of BHL textual data together with scientific names to a researcher. Sending the text of 50 million pages with an associated 250 million name occurrences takes ~5 hours. For comparison, simply copying BHL text data from Smithsonian Institute to University of Illinois using more traditional methods took us 10 days. What do we hope to achieve with these tools as next steps? To make it possible for everyone to make new discoveries by computing in real-time across the complete BHL text. For example 20% of all names in BHL are abbreviated, and, as a result, very poorly searchable given their existing full-text indexing. We plan to develop algorithms to expand abbreviated genera reliably. Digitized texts contain huge amounts of character recognition mistakes. The tools might help to detect badly digitized pages and mark them for re-digitization. Tools can help to extract scientific names that are identical to "normal" words, such as "Atlanta", or "America", to find common names in texts, and to localize information on locations, adding new search contexts. Finally, we are exploring tools that allow researchers to stream such results back to source thereby growing the “Big Data” and ultimately improving the BHL’s end-user experience.

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.

The Biodiversity Heritage Library: Empowering Discovery through Free Access to Biodiversity Knowledge

The advancement of knowledge about life on the planet—its origins, preservation, and loss of species and environments—is dependent on access and reference to library collections. The Biodiversity Heritage Library (BHL) is a global digital library that serves the biodiversity research community, as well as a widening circle of those interested in learning more about life. Through an international consortium of natural history and botanical libraries and in close collaboration with researchers, bioinformaticians, publishers, and information technology professionals, BHL has democratized access to biodiversity information and revolutionized research worldwide, allowing everyone, everywhere to study and explore life on Earth.

Outreach Strategies to Engage Citizen Scientists: Insights from the Biodiversity Heritage Library

How do you successfully engage volunteers in citizen science projects? In recent years, citizen science has grown considerably in popularity, resulting in rapid increases in the number of citizen science and crowdsourcing projects and providing cost-effective means for scientists to gather more data over broader spatial ranges to tackle research questions in a wide variety of scientific, conservation, and environmental fields Bonney et al. 2016, Aceves-Bueno et al. 2017. While the proliferation of such projects has produced a growing abundance of citizen scientist-generated data and published research informed by citizen science methods Follett and Strezov 2015, this also means that volunteers have a greater number of projects competing for their time. When faced with an increasingly-crowded landscape, how can you generate interest in a citizen science or crowdsourcing project and maintain contributions over the project’s lifetime? The Biodiversity Heritage Library (BHL) supports a variety of citizen science and crowdsourcing projects, from transcribing field notes to tagging scientific illustrations with taxonomic names on Flickr and enhancing data for 19th century periodicals through its Zooniverse-based Science Gossip project. Through a variety of outreach strategies including collaborative social media campaigns, partnerships with citizen science communities, and interactive incentives, BHL has successfully engaged volunteers with diverse projects to enrich the library’s data and increase discoverability of its collections. This presentation will discuss outreach strategies for citizen science projects that BHL has undertaken to further support research initiatives with our content. In addition, the presentation will share lessons-learned and offer suggestions that attendees can apply to their own citizen science engagement efforts.

Improving Search Efficiency in the Biodiversity Heritage Library Corpus

Biodiversity literature and archival collections are not only indispensable in taxonomic research, they provide crucial information for understanding of museums’ natural history collections. Literature and archives document collecting events resulting in specimen collections, contain original descriptions based on those specimens, and provide a wealth of other contextual information for the study of life on earth. The Biodiversity Heritage Library is committed to improving research efficiency by providing open access to a growing body of biodiversity literature and archives. While descriptive metadata is widely available for both specimen collections (i.e., DarwinCore) and literature (i.e., MARCXML), connections between the two collection types cannot generally be found at these descriptive levels thus hindering efficient discovery of relevant materials. The integration of name finding services, powered by Global Names Architecture, provides a significant value-add through page-level access to mentions of a given taxon name. Yet how might one search based on a museum code, a common name, or a place name? This presentation will share how BHL’s top technical priorities for 2018 will help facilitate more efficient searching and discovery of information in the pages of the BHL corpus. Specifically, updates on BHL’s top two priorities – implementation of full text search and incorporation of available crowdsourced transcriptions—will be covered.