Data storage and data re-use in taxonomy—the need for improved storage and accessibility of heterogeneous data

The ability to rapidly generate and share molecular, visual, and acoustic data, and to compare them with existing information, and thereby to detect and name biological entities is fundamentally changing our understanding of evolutionary relationships among organisms and is also impacting taxonomy. Harnessing taxonomic data for rapid, automated species identification by machine learning tools or DNA metabarcoding techniques has great potential but will require their review, accessible storage, comprehensive comparison, and integration with prior knowledge and information. Currently, data production, management, and sharing in taxonomic studies are not keeping pace with these needs. Indeed, a survey of recent taxonomic publications provides evidence that few species descriptions in zoology and botany incorporate DNA sequence data. The use of modern high-throughput (-omics) data is so far the exception in alpha-taxonomy, although they are easily stored in GenBank and similar databases. By contrast, for the more routinely used image data, the problem is that they are rarely made available in openly accessible repositories. Improved sharing and re-using of both types of data requires institutions that maintain long-term data storage and capacity with workable, user-friendly but highly automated pipelines. Top priority should be given to standardization and pipeline development for the easy submission and storage of machine-readable data (e.g., images, audio files, videos, tables of measurements). The taxonomic community in Germany and the German Federation for Biological Data are researching options for a higher level of automation, improved linking among data submission and storage platforms, and for making existing taxonomic information more readily accessible.