Agritrop
Accueil

How to improve selection decisions in the first replicated yield trial (RYT) of sugarcane selection programs ?

Hoarau Jean-Yves, Barau Laurent, Thong-Chane Audrey, Dumont Thomas. 2018. How to improve selection decisions in the first replicated yield trial (RYT) of sugarcane selection programs ?. In : Book of abstract of the ISSCT Joint 12th Germplasm & Breeding and 9th Molecular Biology Workshops: " Improvement of sugarcane for stress environments". ISSCT, JSSCT. Okinawa : ISSCT, Résumé, 18. ISSCT Germplasm and Breeding and Molecular Biology workshops. 12/9, Okinawa, Japon, 22 Octobre 2018/26 Octobre 2018.

Communication avec actes
[img]
Prévisualisation
Version publiée - Anglais
Utilisation soumise à autorisation de l'auteur ou du Cirad.
B18.pdf

Télécharger (1MB) | Prévisualisation

Résumé : In the context of opportunities of revenue diversification from sugarcane, cane biomass remains the primary criteria of selection considered in variety development programs (VDPs ). Measurements of cane yield (CY) is performed in replicated yield trials (RYTs) which usually start from the middle term of VDPs. Prediction of the genotypic value of candidate varieties (BLUP) for their cane yield is sought after as accurate as possible. In particular, confidence level in selection decision taken in the first R YT stage is crucial to expect for highest genetic gains for CY at the end of selection programs. Before RYTs, most of the initial genotype candidates are discarded in non-replicated stages due to insufficient performance for some traits showing good heritability. However, the first RYT can still contain a relatively large number of candidates, reaching about one to several hundred candidates ( depending on programs). In some fields with a hilly topography a full replicate might involve risks of spatial heterogeneity due to possible differences in soil fertility, depth or humidity. Multidimensional regression spline methods represent a potentially attractive option to correct for potentially complex spatial heterogeneities. Such methods can be implemented in the framework of mixed linear models (REML algorithm). The study aimed to assess the potential of multidimensional regression spline (MRS) methods to improve selection decision in the first RYT stage of eRcal).e program. The MRS methods were applied to four 2 variety trial series of 120 to 138 candidates. These candidates were tested for CY on 15m plots in a first RYT stage in a randomized complete block design (RCBD) in two replicates. In each series, the 30 elite 2 candidates were advanced to the second RYT stage in a RCBD in three replicates on 45m plots. Compared to the conventional RCBD model, MRS methods allowed a reduction of the residual coefficient of variation of CY in the first RYT stage (0.65% to 4.36%), depending on series considered. Correlation between the first and second RYT stages for CY was improved (3 % to 10%) when considering variety BLUPs inferred in the first RYT from MRS data modeling. The set of the highest 30 candidates for CY in the first RYT stage differed from 2 to 10 genotypes when comparing BLUPs inferred· from RCBD and MRS models. These four case studies illustrate opportunities of improved trial precision and selection decision provided by data modelling of CY using MRS approaches.

Mots-clés libres : Replicated yield trials, Spatial heterogeneity, Multidimensional regression spline

Auteurs et affiliations

  • Hoarau Jean-Yves, CIRAD-BIOS-UMR AGAP (REU) ORCID: 0000-0001-9734-4165
  • Barau Laurent, eRcane (REU)
  • Thong-Chane Audrey, eRcane (REU)
  • Dumont Thomas, eRcane (REU)

Source : Cirad-Agritrop (https://agritrop.cirad.fr/590249/)

Voir la notice (accès réservé à la Dist) Voir la notice (accès réservé à la Dist)

[ Page générée et mise en cache le 2019-10-08 ]