What's happening at IRRI

Flat View
By Year
Monthly View
By Month
Weekly View
By Week
Daily View
Today
Jump to month
Jump to month
Search
Search

Daily View

Wednesday, July 26, 2017
Wednesday, July 26, 2017
  • 02:00pm - 03:00pm  PB & GB Joint Division Seminar by Eliana Monteverde PB and GB Joint Division Seminar
    (Please click “Add to calendar” to mark it in your calendar. We will not circulate reminders)

    Title: Genome wide association mapping and genomic selection for grain quality traits in
    Uruguayan indica and tropical japonica rice

    Abstract
    In some of the major rice-consuming countries there has been a shift from quantity to quality in terms of consumer demand for rice. In these regions, grain quality must be a major focus for rice improvement, mostly for high quality rice exporting countries, which have to increase their rice production with a concomitant increase in grain quality. The most promising way to accomplish this goal is through the application of genomic tools and statistical approaches that are designed to enhance the rate of genetic gain. This study aimed at identifying genomic regions associated to grain quality traits and exploring opportunities for genomic selection of grain quality and agronomical traits in Uruguayan rice. We used two rice elite breeding populations for both GWAS and GS studies, consisting of 324 indica and 308 tropical japonica lines evaluated for yield after milling, percentage head rice, grain chalkiness, grain yield and plant height. A total of 22 QTL for grain quality traits were detected. Genome annotation was performed identifying five major genes involved in starch synthesis and 17 candidate genes associated either with starch metabolism, cell wall synthesis and grain size that co-located with the GWAS-QTL. We compared the prediction ability of single-environment (SE) and multi-environment (ME) genomic prediction models, where the genetic effects are assessed by the Kronecker product of variance-covariance matrices of genetic correlations between environments and genomic kernels (Genomic Best Linear Unbiased Predictor, GBLUP and the Gaussian Kernel, GK). Percent change in mean prediction accuracy of ME models was higher than for SE models for both kernel methods in both datasets. Results confirmed the superiority of multi-environment over single environment prediction models for all traits when genetic correlations between environments are high.

    Eliana Monterverde :: IRRI Events