The European Bioinformatics Institute in 2018: tools, infrastructure and training

Abstract

The European Bioinformatics Institute (https://www.ebi.ac.uk/) archives, curates and analyses life sciences data produced by researchers throughout the world, and makes these data available for re-use globally (https://www.ebi.ac.uk/). Data volumes continue to grow exponentially: total raw storage capacity now exceeds 160 petabytes, and we manage these increasing data flows while maintaining the quality of our services. This year we have improved the efficiency of our computational infrastructure and doubled the bandwidth of our connection to the worldwide web. We report two new data resources, the Single Cell Expression Atlas (https://www.ebi.ac.uk/gxa/sc/), which is a component of the Expression Atlas; and the PDBe-Knowledgebase (https://www.ebi.ac.uk/pdbe/pdbe-kb), which collates functional annotations and predictions for structure data in the Protein Data Bank. Additionally, Europe PMC (http://europepmc.org/) has added preprint abstracts to its search results, supplementing results from peer-reviewed publications. EMBL-EBI maintains over 150 analytical bioinformatics tools that complement our data resources. We make these tools available for users through a web interface as well as programmatically using application programming interfaces, whilst ensuring the latest versions are available for our users. Our training team, with support from all of our staff, continued to provide on-site, off-site and web-based training opportunities for thousands of researchers worldwide this year.

Figure 1.

EMBL-EBI data resources. To the left are deposition resources that store primary data submitted by research scientists as well as ontologies and literature resources that span the entire research effort. To the right of the arrow are added-value knowledgebases ordered according to biological scale, from genes to proteins, structures, chemistry and systems. Updates in this NAR Database issue include ArrayExpress (26), BioSamples (27), ChEMBL (28), Complex Portal (29), ENA (30), Ensembl (9), GenCode (31), GWAS Catalogue (32), HGNC (33), InterPro (8), Open Targets (34), PDXfinder (35), Pfam (7), PRIDE (36), RNACentral (37), SIFTS (6) and UniProt (38). Not all EMBL-EBI resources are shown on the figure. For a complete list see https://www.ebi.ac.uk/services.

Figure 2.

(A) Data accumulation at EMBL-EBI by data type: nucleotide sequences, mass spectroscopy and microarray. (B) Data accumulation by archive: PRoteomics IDEntifications (PRIDE) (36), European Genome-Phenome Archive (EGA) (39), ArrayExpress (AE) (26), European Nucleotide Archive (ENA) (30), Protein Data Bank in Europe (PDBe) (40) and MetaboLights (16). The y-axis for both charts is logarithmic, so most data types are not just growing, but are growing at in increasing rate. In all data resources shown here growth rates are predicted to continue increasing, with notable sustained exponential growth in PRIDE and MetaboLights.

Figure 3.

Installed raw storage at EMBL-EBI. The chart shows total installed data storage at EMBL-EBI, including multiple backups for all data resources as well as unused space to handle submissions in the immediate future. The total volume of a single copy of all data resources is roughly 20–25% of the installed storage capacity. Data points (not shown) are the end of each calendar year, thus the range of the x-axis is 31 December 2012 through 31 December 2018. Data for end of 2018 are projected based on planned procurement. In 2017, we procured a high volume of disk space at good value that increased capacity substantially, requiring relatively less procurement in 2018. This approach allowed us to utilize our infrastructure budget efficiently.