Report on the visit by MR Nouy to the breeding unit at FELDA agricultural services Sdn. Bhd. Pusat Perkhidmatan Pertanian Tun Razak from 21 May to 1 June 2007

Nouy Bruno. 2007. Report on the visit by MR Nouy to the breeding unit at FELDA agricultural services Sdn. Bhd. Pusat Perkhidmatan Pertanian Tun Razak from 21 May to 1 June 2007. Montpellier : CIRAD-BIOS, 125 p. N° de rapport : CP_SIC 2110

Résumé : In 2006, CIRAD and FELDA agreed to launch collaboration in the form of appraisal and training visits for Breeding. This visit was the first scheduled under that agreement. It took place from 21 May to 1 June 2007 at the Breeding Unit of FELDA Agricultural Services Sdn. Bhd. Its purpose was to carry out a critical analysis of some trials being monitored by FELDA, and put forward proposals for improving evaluation quality and gene pool management. FELDA manages a very large experimental network which has to deal with 2 major constraints: high soil heterogeneity (undulating land) and high heterogeneity in palm development, which can induce competition effects. Under such conditions, it is complicated choosing an experimental design. For ongoing or completed trials, it would be worth validating or rejecting the results of each trial and, in order to do that, effectively judge the degree of bias (study of competition effects in some trials comparing highly heterogeneous materials; estimation of observation precision). Pending the results of that critical study, general recommendations have been made to limit competition effects (distribution of materials in the different trials according to their characteristics) and to take land heterogeneity into account more (lattice design rather than blocks, a larger number of replications, and smaller unit plots). The observations (yield, bunch quality) carried out in the trials are similar to those carried out by CIRAD in the genetics trials. They are quality observations which could be made more precise if the bunch analysis rate was the same for all the crosses in the same trial. Two major types of analyses are carried out for each trial: a statistical analysis of cross effects, and an analysis of pisifera parent effects. However, the latter type of analysis is only possible if there is no dura effect. Yet those effects are not insubstantial. A more appropriate approach would be: first to check whether there are any significant parent and interaction effects, then calculate the adjusted means. Analyses of the FELDA genetics trials are carried out trial by trial. There is little or no global or synthetic analysis by group of trials or all the trials. Yet it is useful, if not essential, for breeders to have a clear tool for comparing all the crosses and all the parents tested in the different trials. An examination of the trials currently being harvested shows that all the trials can be connected by crosses common to several trials. Likewise, many of the parents are directly or indirectly linked by common partners. Even if those links need to be strengthened, it seems possible to compare all the crosses with the same frame of reference and also to estimate the GCA of the dura and pisifera parents. It has been proposed that these analyses (calculation of the adjusted cross means) and GCA calculations should be the central subject of the training Mrs Junaidah Judin is to receive in France in 2008. To back up the analyses, the mission was used to take stock rapidly of the recombinations being tested in trials. In fact, it seemed important that the breeders (particularly the young FELDA breeders) acquire a clear vision of the way the different origins have been crossed and any changes that need to be made to the mating designs. For instance, in their current form, the recombinations tested make it difficult to effectively explore the available variability: for example, there is a very large number of origins or sub-origins of the dura parents, but their very low representation and the mating designs implemented do not make it possible to assess their mean value and variability. Lastly, the quality of genetic progress transfer was discussed. Much of the genetic progress achieved in the trials is effectively transferred to seed production. That efficiency contributes to the good quality of FELDA seeds. However, given the absence of a tenera parent test, on average, only half the genetic progress o

Auteurs et affiliations

• Nouy Bruno, CIRAD-BIOS-UPR Génétique palmier (FRA)

Source : Cirad - Agritrop (https://agritrop.cirad.fr/541085/)

Métriques

[ Page générée et mise en cache le 2024-02-20 ]