Anotación automática GO de productos génicos en SARS-CoV-2
Cargando...
Fecha
Authors
Título de la revista
ISSN de la revista
Título del volumen
Editor
Facultad de Ciencias Agrarias. UNR
Resumen
La anotación de funcionalidades biológicas de productos génicos, RNA y proteí nas, es una tarea crítica en el desarrollo de proyectos de secuenciación genómica. En
el caso de proyectos de genomas virales, estas anotaciones infieren el rol molecular
de estos productos virales de interés durante la infección a sus células diana, indi cando aquellos procesos biológicos en los que están involucrados y constituyen una
herramienta útil para el desarrollo y mejoramiento de tratamientos antivirales. La
velocidad actual a la que se generan nuevas secuencias de RNA y proteínas a partir
de proyectos de secuenciación genómica genera un cuello de botella para los méto dos de anotación tradicionales, basados en estudios experimentales exhaustivos. Este
cuello de botella puede resolverse parcialmente mediante métodos computacionales
de anotación. Es de interés global el estudio de virus y en particular, SARS-CoV-2,
que causa la enfermedad COVID-19 y representa aún una amenaza para la salud
mundial. Los esfuerzos para desarrollar medicamentos y vacunas eficaces frente a
nuevas variantes se ven obstaculizados por el conocimiento limitado de los detalles
moleculares de cómo el SARS-CoV-2 infecta y se propaga. En particular, en es te trabajo se aborda el problema de anotación funcional automática de productos
génicos para SARS-CoV-2 a través de ontologías y aprendizaje computacional. La
ontología funcional de genes utilizada es Gene Ontology (GO) y el método de apren dizaje computacional utilizado se llama Factor Graph GO Annotation (FGGA). Este
método de clasificación jerárquico toma como entrada un conjunto de atributos, ca racterísticas, extraídos desde las secuencias y devuelve un grafo consistente en los
tres subdominios de GO. El proceso de extracción de atributos desde las secuencias
se lo denomina caracterización. En este trabajo, se considera una caracterización
básica que consiste en propiedades fisicoquímicas y una caracterización enriquecida,
desarrollada en este proyecto, que agrega atributos virales. La incorporación de es tos contribuye a mejorar la especificidad de predicción de las funcionalidades GO.
Finalmente, se evalúa el rendimiento de las predicciones GO obtenidas y se compara
los resultados obtenidos sobre 31 productos génicos anotados en forma experimental
en Jungreis et al. (2021). Estos resultados validaron de forma exitosa las anotaciones
existentes curadas manualmente y también generaron nuevas anotaciones in-silico
que fueron avaladas por diversas fuentes bibliográficas disponibles en la actualidad
The annotation of biological functionalities of gene products, RNA and proteins, is a critical task in the development of genomic sequencing projects. In the case of viral genome projects, these annotations infer the molecular role of these viral products of interest during the infection of their target cells, indicating the biological processes in which they are involved and constitute a useful tool for the development and improvement of antivirals treatments. The current rate at which new RNA and protein sequences are generated from genomic sequencing projects creates a bottle neck for traditional annotation methods, based on extensive experimental studies. This bottleneck can be partially resolved by computational annotation methods. The study of viruses is of global interest and, in particular, SARS-CoV-2, which causes the disease COVID-19 and still represents a threat to global health. Efforts to develop effective drugs and vaccines are hampered by limited knowledge of the molecular details of how SARS-CoV-2 infects cells. In particular, this paper ad dresses the problem of automatic functional annotation of gene products for SARS CoV-2 through ontologies and computational learning. The functional gene ontology used is Gene Ontology (GO) and the computational learning method used is called Factor Graph GO Annotation (FGGA). This hierarchical classification method takes as input a set of attributes, characteristics, extracted from the sequences and returns a graph consisting of the three GO subdomains. The process of extracting attributes from the sequences is called characterization. In this work, a basic characterization is considered, which consists of physicochemical properties, and an enriched char acterization, developed in this project, which adds viral attributes. The inclusion of these attributes contributes to improving the specificity of GO functionality pre dictions. Finally, the performance of the obtained GO predictions is evaluated and compared with the results obtained for 31 gene products experimentally annotated in Jungreis et al. (2021). These results successfully validated existing hand-curated annotations and also generated new in-silico annotations that were supported by various currently available literature sources.
The annotation of biological functionalities of gene products, RNA and proteins, is a critical task in the development of genomic sequencing projects. In the case of viral genome projects, these annotations infer the molecular role of these viral products of interest during the infection of their target cells, indicating the biological processes in which they are involved and constitute a useful tool for the development and improvement of antivirals treatments. The current rate at which new RNA and protein sequences are generated from genomic sequencing projects creates a bottle neck for traditional annotation methods, based on extensive experimental studies. This bottleneck can be partially resolved by computational annotation methods. The study of viruses is of global interest and, in particular, SARS-CoV-2, which causes the disease COVID-19 and still represents a threat to global health. Efforts to develop effective drugs and vaccines are hampered by limited knowledge of the molecular details of how SARS-CoV-2 infects cells. In particular, this paper ad dresses the problem of automatic functional annotation of gene products for SARS CoV-2 through ontologies and computational learning. The functional gene ontology used is Gene Ontology (GO) and the computational learning method used is called Factor Graph GO Annotation (FGGA). This hierarchical classification method takes as input a set of attributes, characteristics, extracted from the sequences and returns a graph consisting of the three GO subdomains. The process of extracting attributes from the sequences is called characterization. In this work, a basic characterization is considered, which consists of physicochemical properties, and an enriched char acterization, developed in this project, which adds viral attributes. The inclusion of these attributes contributes to improving the specificity of GO functionality pre dictions. Finally, the performance of the obtained GO predictions is evaluated and compared with the results obtained for 31 gene products experimentally annotated in Jungreis et al. (2021). These results successfully validated existing hand-curated annotations and also generated new in-silico annotations that were supported by various currently available literature sources.
Descripción
Palabras clave
Citación
Aprobación
Revisión
Complementado por
Referenciado por
Licencia Creative Commons
Excepto donde se indique lo contrario, la licencia de este ítem se describe como Attribution-NonCommercial-ShareAlike 4.0 International

