We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
Desempenho dos métodos de estimação genômicos na identificação da resistência do arroz à brusone.
- Authors
Gomes da SILVA NETO, Zeferino; MARTINS FILHO, Sebastião; Souza da SILVEIRA, Lucas; Souza CARNEIRO, Antônio Policarpo; Silva dos SANTOS, Vinicius
- Abstract
In recent years, rice crop losses have increased due to biotic and abiotic stresses, among which brusone, which can result in 100% losses in susceptible rice cultivars. Therefore, it becomes strategic to identify methodologies that select resistant cultivars. In this work, we aimed to use ROC (Receiver operator characteristic) curve analysis and traditional measures to evaluate the performance of genomic estimation models (RR-BLUP, BLASSO and Bayes Cπ) in identifying rice resistance to brusone. The RR-BLUP and Bayes Cπ models were most accurate for the prediction of brusone resistance, while the best runtime was obtained by RR-BLUP. The area under the ROC curve was equivalent to traditional measures to evaluate the accuracy of the models, with the advantage of allowing graphical evaluation. By graphical analysis, BLASSO performed worst at high levels of specificity (>0.75). At lower levels of specificity, the sensitivity of the models was similar. The ROC methodology proved to be a good alternative for the evaluation of genomic prediction models, and can be used for the selection of rice cultivars resistant to brusone.
- Publication
Nativa, 2022, Vol 10, Issue 4, p466
- ISSN
2318-7670
- Publication type
Article
- DOI
10.31413/nativa.v10i4.13448