22.08.2016 Views

Annual Scientific Report 2015

EMBL_EBI_ASR_2015_DigitalEdition

EMBL_EBI_ASR_2015_DigitalEdition

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

Maria-Jesus Martin<br />

Protein Function Development<br />

BSc In Veterinary Medicine, University<br />

Autonoma in Madrid. PhD in Molecular Biology<br />

(Bioinformatics), 2003.<br />

At EMBL-EBI since 1996.<br />

Team Leader since 2009.<br />

millions of proteins in UniProt. Informed by specialist<br />

biocurators, the automated system adds as much useful<br />

information as possible to imported sequences, which<br />

now include domains, signal, transmembrane and<br />

coil regions. We extended UniRule and the Statistical<br />

Automatic Annotation System (SAAS), two systems<br />

for the automatic annotation of large volumes of<br />

uncharacterised proteins. These are now available<br />

through newly implemented interactive web pages,<br />

allowing our users to browse annotation rules. We also<br />

started to work in a service to download and/or use these<br />

rules as a system for genome annotation. We extended<br />

our collaborations with external automatic annotation<br />

communities including the Biofunction Prediction and<br />

Critical Assessment of Function Annotation initiatives,<br />

which will expand our knowledge and use of functional<br />

prediction methods.<br />

In <strong>2015</strong> we further extended the scope of GO<br />

annotation to support annotations to RNA, identified by<br />

RNAcentral identifiers. We made significant changes<br />

to our database and Protein2GO, the web-based GO<br />

curation tool used by UniProt and GO Consortium<br />

curators to contribute annotations to the GOA project,<br />

in order to support a number of changes to annotation<br />

format and rules agreed by the GO Consortium. We<br />

re-engineered our pipeline that verifies the taxonomic<br />

correctness of GO annotations using a much-extended<br />

set of taxonomic constraints that originate from both<br />

GO and other ontologies, principally UBERON.<br />

To make GO protein–protein interaction annotations<br />

available for visualisation in tools such as Cytoscape,<br />

we implemented a PSICQUIC (Proteomics Standards<br />

Initiative Common Query Interface) server, available<br />

through EMBL-EBI’s PSICQUIC portal.<br />

Our team maintains the Enzyme Portal, a resource<br />

that integrates enzyme-related data for all relevant<br />

EMBL-EBI resources and the underlying functional and<br />

genomic data. We re-launched the service, which now<br />

features improved interfaces and functionalities and<br />

provides a one-stop shop for all information available<br />

on enzymes. To further improve the discoverability of<br />

enzyme data, we collaborated with the Web Production<br />

team to refine the enzyme search within the EBI-Search<br />

and EBI Blast sequence search tools.<br />

Future plans<br />

In 2016 we plan to release a protein-sequence feature<br />

viewer that summarises functional sites in the<br />

UniProt web site. We will continue to engage with user<br />

communities working in functional prediction, and<br />

explore methods and data-exchange mechanisms to<br />

improve accuracy and coverage of protein annotations.<br />

We will maintain our focus on usability and engage<br />

with our users to ensure we maintain a global genome/<br />

proteome- and gene-product-centric view of the<br />

sequence space. We aim to expand our collaboration<br />

with the ProteomeXchange resources in the integration<br />

of post-translational modifications in UniProtKB, and in<br />

the provision of experimental, unique peptide mappings<br />

for reference species. We will continue to co-operate<br />

with variation projects such as ExAC to integrate<br />

relevant genome and proteome information.<br />

Restructuring GO electronic annotation pipelines,<br />

principally those based on orthology supplied by<br />

Ensembl, will help us improve the quality of the<br />

projected annotations. We will continue the work<br />

undertaken on behalf of the GO Consortium in <strong>2015</strong><br />

to transition from using UniProt cross-references<br />

rather than MOD-supplied mapping files to map from<br />

“foreign” identifiers to UniProtKB accessions. We also<br />

plan to revise the set of annotation files that we publish<br />

and submit to the GO Consortium. We plan to release<br />

a new QuickGO with re-designed interfaces and new<br />

features to improve the overall user experience. We<br />

will also continue to develop Protein2GO to keep it in<br />

line with changes in annotation strategy agreed by the<br />

GO Consortium, and to introduce additional function<br />

to enhance curators’ workflow. We will make use of<br />

the enhanced set of Web Services provided by the new<br />

QuickGO to provide improved searching capabilities,<br />

and the ability to use any available ECO evidence code.<br />

Following the successful relaunch of the Enzyme Portal<br />

in <strong>2015</strong>, we will expand its functionalities in response to<br />

user needs and create new training activities.<br />

Selected publications<br />

Alpi E, Griss J, et al. (<strong>2015</strong>) Analysis of the tryptic search<br />

space in UniProt databases. Proteomics 15:48-57<br />

Huntley RP, et al. (<strong>2015</strong>) The GOA database: Gene<br />

Ontology annotation updates for <strong>2015</strong>. Nucleic Acids Res.<br />

43:d1057-d1063<br />

Pundir S, Magrane M, Martin MJ, O’Donovan C, UniProt<br />

Consortium (<strong>2015</strong>) Searching and navigating UniProt<br />

databases. Curr. Protoc. Bioinform. 50:1.27.1–1.27.10<br />

UniProt Consortium (<strong>2015</strong>) UniProt: a hub for protein<br />

information. Nucleic Acids Res. 43:d204-d212<br />

<strong>2015</strong> EMBL-EBI <strong>Annual</strong> <strong>Scientific</strong> <strong>Report</strong> 104

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!