The Repository Chemotion: Infrastructure for Sustainable Research in Chemistry**

Abstract

The repository Chemotion provides solutions for current challenges to store research data in a feasible manner. A main advantage of Chemotion is the comprehensive functionality, offering options to collect, prepare, and reuse data with discipline-specific methods and data-processing tools. We describe the development of a repository for chemistry research data (called Chemotion) that provides solutions for current challenges to store research data in a feasible manner, allowing the conservation of domain-specific information in a machine-readable format. A main advantage of the repository Chemotion is the comprehensive functionality, offering options to collect, prepare and reuse data with discipline specific methods and data processing tools. For selected analytical data, automated procedures are implemented to facilitate the curation of the data. Chemotion provides functions for a feasible data publishing process including automated Digital Object Identifier (DOI) generation and workflows for peer reviewing of the submissions, including embargo settings. The described developments were used to establish a research-data infrastructure to build a new community-driven repository as a comprehensive alternative to commercial databases. Main databases and repositories are an important component for scientific work as they allow the access to data obtained by other researchers. Therefore, databases and repositories are information systems that play a pivotal role for sustainable scientific work and they form an important basis for the dissemination and application of our current knowledge.1 An overview of the currently existing databases and repositories covering research data can be obtained by RDA (Research Data Alliance) endorsed services as DataCites re3data2, 3 or FAIRsharing,4 as well as community specific efforts such as the Enabling FAIR Data project of the American Geophysical Union (AGU).5 These services and efforts collect resources and aggregate information from multiple sources, like the repositories themselves, publishers or persistent identifier (PID) systems. While many common chemical databases store aggregated data in form of textual representation, research data repositories are also used to store and share the underlying data in well-structured and machine readable format, often including the raw data from scientific measurements. During the last years, an increasing number of scientists and stakeholders has realized how important research data repositories are in the view of the quality of scientific work, the reproducibility of experiments and transparency of work processes.6, 7 Furthermore, the current progress in data analysis technologies such as machine learning, semantic technologies and big data reveals that the scientific community could benefit tremendously from large data collections which can be used for the development of intelligent information systems. Repositories could collect the data from a vast amount of mostly small size laboratories and groups of scientists that are producing most of the available data, enabling solutions for current problems related to the “long-tail” of science.8, 9 Altogether, research data repositories offer tools and services to address the often mentioned problems of data quality, accessibility and reusability.10, 11 The key role that repositories play for scientific work is also recognized by publishers as many of them recommend to deposit research data in trustworthy infrastructure along with the publication itself.12-16 The preservation and sharing of research data in repositories is not only supported by funding institutions but is also progressively enforced.17-22 The most often recommended general purpose repositories, FigShare,23 Zenodo,24 and Dryad25 achieved high attention during the last years. These discipline agnostic repositories offer quick solutions to data storage and access as the deposition of data can be done with only minimal preparation and costs. In any case, the relevance of these repositories with respect to the availability of machine readable and reusable data for subject specific tool sets and processing is currently limited due to the generic approach. In particular, the need for data being compliant with the FAIR (findable, accessible, interoperable, reusable) data26 principles, reveals that many of the currently available repositories lack necessary options to reuse data in a discipline-specific manner.27 In chemistry, the discipline specific functionality of a repository is especially important, as the required information of a chemical investigation is given in a schematic presentation including structures and reactions. Processing subject specific information to provide tailored services for identification and citation, as well as collection, search and comparison, is crucial to build repositories in chemistry that enable FAIR data, and therefore, a reuse in the scientific community. So far, only a few repositories and databases in chemistry combine open access to research data and the availability of data in a reusable form. All of them are highly specific with a focus on a certain analytical method. As far as we know, the only well-established database in chemistry that has repository functions in terms of the systematic storage of user-provided data files, is the Crystal Structure Database (CSD) of the Cambridge Crystallographic Data Center (CCDC).28, 29 Despite not being Open Access, most of the universities have full access to the CSD due to moderate costs for academic institutions. While most other databases or repositories in chemistry are rarely used by the scientists, the publication of crystal structures via the CSD has been accepted as a standard procedure for scientists and the repository may serve therefore as a best practice model in terms of acceptance by the research community. Only a few other very good yet rarely used examples are available for chemists, including the well-curated databases massbank30 for mass spectrometry data and NMRshiftDB231 for nuclear magnetic resonance shifts. Regarding synthetic information on reactions and their details for example, the purification as well as information on molecules and their analytical data, a comprehensive repository that supports the deposit and reuse of research data is currently not available. To our knowledge, the ChemSpider Synthetic Pages is so far the only initiative that provides web-tools to deposit information on reactions and processes in chemistry.32 The repository Chemotion was developed to fill the currently existing gap for the preservation and publication of experimental work in synthetic chemistry. The repository supports chemists in their efforts to store data in a reusable manner and provides functions to offer this data in a well-structured and transparent way. The repository serves as infrastructure for scientists to deposit and share data and can be used as a source to obtain chemical data in a machine readable way. The use of the repository is free of charge for data providers and data consumers and the code of the software is available via GitHub33 as Open Source under an AGPL34 license. The repository focusses on synthetic and analytic chemistry including molecules, their properties and spectroscopic data as well as on information related to chemical reactions. With this focus, there are two main instances according to which the data in the repository are structured: samples, which represent unique batches of a molecule, and reactions (Figure 1, I and II). With this, the repository follows general management standards in chemistry where information is kept along with the experimental chemical process. Information and analytical data are assigned to either a reaction or a sample and can be correlated to those instances due to the availability of standardized identifiers such as the reaction InChI (RInChI and RInChIKey). The repository has three types of user roles: (passive) non-registered users, (active) registered users and reviewers. Non-registered users can access the repository and the disclosed data, but they cannot download the given information and datasets. The registered users can download any disclosed information and data files. They have access to certain functions of the repository like the workspace area, offering options for the creation of an own database and including typical, chemistry-specific data management functions. Furthermore, only registered users are allowed to upload or transfer data to the repository. Reviewers have access to the reviewing functions of the repository that allows them to accept and decline submissions, or to ask the authors for a revision. The repository consists of the two main work areas: the workspace area (Figure 1, yellow panel) with functions to prepare, collect and manage data, and the data publication area (Figure 1, orange panel) that includes the reviewing and embargo processes. Both have a listing view, with filtering options, and a detailed view of a selected element. An important feature facilitating the use of the repository is the possibility to transfer data from an Open Source electronic lab notebook (ELN)35, 36 into the workspace area of the repository. The process transfers all required data directly from server to server instead of having the user to manually upload each data sets and enter the metadata to the repository. Both processes, the direct transfer and the data upload are supported by the Chemotion repository (Figure 1, processes A and B). Schematic presentation of the repository architecture, structured according to reactions (I), molecules/ samples (II) and their associated analyses (blue), and the main functionality of the workspace (1) and publication area (2) of the repository (yellow and orange). Options to enter data into the repository, that is, transfer from ELN to repository (A) or upload of data from any data source (B) are marked in green. The workspace area is a private area only accessible to the registered user that allows to upload and manage data. It supports functions to collect, assign and visualize chemical research data (Figure 2, a–c) without the need of additional instruments or software for chemistry specific tasks. Functions such as drawing of structures with a molecule editor or automatic calculations embedded to a reaction table enable an efficient preparation of data for a publication in the repository (Figure 2, d). The discipline specific functions support the preparation of data to conserve the whole set of information for its further reuse or a detailed analysis at a later stage. This is an important advantage of Chemotion compared to generic and multidisciplinary repositories. Without options to store data according to chemistry specific standards and formats, such as the conversion of data to Word documents or PDFs, subject specific information is usually lost and the utility of the remaining data is reduced. The combination of discipline specific functions with diverse management tools (Figure 2, e) in Chemotion allows an efficient organization of textual data and analytical data files according to common documentation and reporting practices in chemistry. The workspace of the repository supports the assignment of metadata to the given datasets by dropdown menus or text fields, depending on the type of information. Users are requested via notification to provide the necessary metadata before submitting to the data publication area. The metadata fields available as dropdown support the vocabularies of the Chemical Reactions Ontology (RXNO)37 and Chemical Methods Ontology (CHMO).38 Beside the preparation of the data for submissions, the workspace area of the repository is also the place where the publication process can be started (Figure 2, f), or ongoing processes, such as the review, can be followed (Figure 2, g and h). Using the workspace is also helpful for registered users who may benefit from functions like the export of data or the visualization of data with a chemistry specific data viewer. In particular, the export options (Figure 2, i) are important to access machine readable data. To the current state, the workspace supports the export of selected information to MS-Excel, and Word document, or SDF,39 but also the of collections as data a and the assigned experimental data files. specific as well as management functions of the workspace in the repository are in of the repository and of the main functions to manage and prepare data for The repository Chemotion the to transfer data from an Open Source ELN to the repository workspace with only a few This was from the by other to collect research data. The of a repository or database and its acceptance in the community on the of use upload and curation of the data to be Therefore, the use of a repository be as automated and user as for the data the for repository with the and of the data. The with the for and reuse those for example, metadata or the data to support use such as or The most feasible to provide sustainable data is to use of data and their transfer to a repository. The data transfer from an ELN to a repository the and transfer of data from a work of scientists and a key to scientists to deposit their data in The data transfer is currently realized for the Open Source Chemotion ELN that serves as an for other in Information for a of the application and methods to be The data transfer is realized via a work area the ELN allowing the user to and transfer the data. the repository and to the ELN the data can be with two additional to and the The transfer includes a and associated data files methods The scientists who to their data can the publication process via data to the repository or the transfer of data from an ELN In data can be and in the repository workspace the is marked as the workspace (yellow datasets can be selected to them along with a reaction or a sample in including transfer of data from the with and without and the reviewing process including by the repository or by additional reviewers. The the publication in the repository are assigned to three main the preparation of the data, of the repository and of and to workspace area publication area The is by to be and by the data The publication is of the process the workspace of the repository (Figure All data can be either with or without an embargo and are to the reviewing that on and of the The can in a to the and the data, the can either accept or the to the for For accepted submissions, the repository a Digital Object Identifier (DOI) so the can be and user is For submissions, the of the is on the to the embargo and the data. While the data are under additional with can be allowing the of the data the publication process is on (Figure The supports access to the data a publication is to a peer where the peer be to also the data to the In to the of the data via the Chemotion publication the are The molecule structures are for their in the database to information on the of the and molecules are registered with their Chemotion in as The of the data the of the data from the private workspace to the publication area of the repository. This transfer is by a publishing that is available molecule or reaction process (Figure For each of data, the user is requested to and fill necessary information such as the data files (Figure and in allowing the of the data. The publication additional information on the of the given (Figure and supports to the by the the process as well as datasets and analyses (Figure The repository Chemotion supports the generation of as a standard persistent which is by The of the is by of the repository including of the molecules, reactions and their and can be to the This the user to the of a before is registered with during the of the data. This allows the authors to use the and a in the Information of their text before the data are to the a to prepare The functions of the repository are only accessible to and Reviewers access the reviewing area a that the current and reviewing processes (Figure have access to the functions via the The can on each information that was by the allowing a detailed on the data and their As as the is the a notification the which may a to the data to be The user has also the to to the and The procedure of and can be as often as necessary and but may also a the data or the manner of not the quality In the the peer reviewing process be supported by automated tools to data to a peer The repository is an basis to such automated curation and reviewing tools as the information on chemical structures that is and can be compared to experimental or the manually given a is by an that supports the quick of the repository the given of and via the with the of and and the data given are or This how tools may the quality and facilitate the curation of data or support the peer process. The publication area of the repository is accessible without any or the need to therefore the data are directly to the its to the of the data are assigned either to a reaction or a molecule also in the publication area. like the of molecules to their as or of a reaction are in all processes of the repository. reactions the and direct to the assigned molecules and their data. an for a of a molecule and the data that are assigned to this An for the of a reaction is given in the Information (Figure The publication of reactions information in terms of the general metadata but in the reaction specific that consists of the of chemical processes. on the given that is, the in the (Figure the repository provides additional and data to the machine and of the data. For molecules, the publication metadata the molecule with the most important molecule identifiers the or of the molecule and its mass (Figure In of a the information by automated processes of the repository consists of the standardized reaction identifiers and (Figure the molecule is in which is used as the database for described molecules, a the direct to the in molecules are by the information that the molecule was not at the of the to the repository (Figure is given with metadata on the and their as well as the that was selected during the (Figure and The of the also this metadata of the repository available to systems. The including a for for molecule or is given along with a unique the repository which is registered for all molecules, reactions and their data (Figure g and of a chemical for which the process is described in the repository, is to the data (Figure h). can be assigned to the molecules and reactions by the or by other registered users at any The data assigned to the molecule of analytical information from which are according to the available sample (Figure The for example, from analytical processes, of additional metadata such as the type of the by an embedded (Figure the of each (Figure and the of the data including the types (Figure Data and the analysis of the data embedded the of the datasets. All related data files are accessible via the and can be by registered users (Figure and The data for a in to the data for a molecule, the information of the the reaction table including all and the of the and experimental details on the purification of the reaction that are assigned as of the reaction and their analytical data are also and with the full information (Figure For molecules and metadata to their metadata can be (Figure is to the for an molecule publication in the repository and of the most important information and The repository Chemotion a to and scientific data in the of chemistry with a focus on synthetic chemistry and analytical chemistry. a research data repository for reactions including analytical data was so The are to establish progressively a large database that supports the of chemists in chemical and that the need of machine readable and FAIR data. to the for scientists to provide data have been with Chemotion and its underlying The of a large number of datasets can be in the by the Chemotion repository for the research community at of that the combination of an ELN and a repository can be used to transfer data, in a and manner, from a work to an Open repository. This of an ELN and a repository data transfer is, as far as we know, unique at in the scientific of chemistry. The data is machine readable and functions including the generation of the to databases as well as export and reporting functions allow the comprehensive reuse of the data also for there are with respect to machine and that the repository cannot at the The repository allows the of processes in common and the of all types of data files. For certain types of data, a was that the user to provide a which is requested in the reviewing process. It is recommended to store data in open such as for spectroscopic data in or for in The of these is given to the user as for many other there are currently only a few standards for the of processes and datasets so that the repository can only procedures for the further development of the repository in terms of the supported standards on the chemical community and other to the current are used to establish a best practice procedure for the deposition of data in chemistry and we to with those to for For the reviewing on that not the best practice model depending on the quality and of the data. These along the to best practices and be and to the of and To the acceptance of the repository by data the repository was set with a functionality to serve data providers with functions such as the workspace area, the reviewing and embargo settings. The main repository at the in is open for any of the community as as the or of users the quality and the size of the data files can be This allows the of a reusable database of reactions and data on molecules by the The Open Source code of the repository available at allows scientists to a repository at their as a and to benefit from the developed is not to data to the a procedure the of data and allow to data with the repository in at a later The repository as be with such as a procedure to information on reactions without analytical data, in A process for the of information in combination with a in the repository allow a of the database with additional information on reactions. developments also the of the currently supported metadata information in the metadata the chemistry community specific open access repositories and databases that serve the of scientists with respect to data storage and search for data, their reuse as well as data While the of funding and publishers in terms of data management and the storage and of research data the scientific community problems to these due to a or yet research data We describe the development of a repository for chemistry research data (called Chemotion) that provides solutions for current challenges to store research data in a feasible manner, allowing the conservation of specific information in a machine readable format. A main advantage of the repository Chemotion is the comprehensive functionality, which options to collect, prepare and reuse data discipline specific methods and data processing tools. For selected analytical data, automated procedures are implemented to facilitate the curation of the data. Chemotion provides functions to facilitate the publishing process of data and the of the data. It supports automated Digital Object Identifier (DOI) the of the with and workflows for peer reviewing of the including embargo settings. The described developments were used to establish a research data infrastructure that is at the of including the necessary storage and support to build a new community-driven repository as a comprehensive alternative to commercial databases. The use of the repository is free of charge and the data are and as Open In the software is available at as Open Source and the code can be used by all scientists or to their own repositories with chemistry specific functions that enable sustainable data This project has been by the project and the of and the of the project as Data Center of the It also to the of the project also by the We are very to the of the who to the of the repository. We are very to and for and with respect to the project and to the of the Data for This work was supported by the in and and Data as well as the of We the support by the of by the and the support by the Data at the Center for at of Open access funding and by The authors of As a to our authors and this provides information by the are peer and may be for but are not or support from information be to the The is not for the or functionality of any information by the be to the for the

Topics & Concepts