Congress detail

Authors: Spetale, Flavio; Murillo, Javier; Bulacio, Pilar; Tapia, Elizabeth.

Resumen: BACKGROUND: A large part of the RNA in eukaryotic organisms is non-coding, that is, its final product is not a protein. Unlike proteins, the functional information of lncRNAs (noncoding RNAs with +200 nucleotides) is only weakly reflected at their primary sequence level. The study of lncRNA is relevant because they participate in gene regulation and activation. Gene Ontology (GO)provides a controlled vocabulary of terms for describing gene properties and their relationships in a species-independent manner. GO includes three sub-ontologies, biological process (BP), molecular function (MF), and cellular component (CC). Taking into account that the experimental characterization of lncRNA is hard due to the increasing instability of RNA molecules with their length, the development of in-silico methods is crucial to accelerate their study.RESULTS: In this work, we present a supervised computational method,graph-based, for the automatic GO prediction of lncRNAs. The method, calledFGGA-GO + , aims for consistent lncRNA GO annotations across the three GOsub-ontologies domains. Preliminary results for lncRNAs from the model organism Zebrafish (Danio rerio) on the BP and MF GO domains confirm the feasibility of our proposal. For the sake of simplicity, lncRNAs were characterized by their RFAM domains.CONCLUSIONS: The joint GO BP and MF annotation of lncRNA with FPGA-GO + yields an acceptable 62% precision when considering 300 GO terms.A similar study with proteins involving a PFAM characterization yields 74%precision. The observed precision gap confirms that for the automatic annotation of lncRNA more than primary sequence information must be provided. Taking these baseline studies, a new characterization of lncRNA molecules, focused on their secondary structure and suitable for machine learning purposes,is ongoing. The final goal is to improve the precision of automatic GO annotation of lncRNAs to that of proteins.

Meeting type: Conferencia.

Type of job: Resumen.

Production: GO BP-MF Prediction of lncRNAs on Zebrafish.

Scientific meeting: X International Conference on Bioinformatics 10th Anniversary of SoIBio.

Meeting place: Montevideo.

Organizing Institution: Sociedad Iberoamercana de Bioinformatica & Master in Bioinformatics PEDECIBA-UdelaR Uruguay.

It's published?: Yes

Publication place: Montevideo

Meeting month: 10

Link: here