(Prediction of Combining Ability and DxT Progeny Keragaannce of Oil Palm (Elaeis guineensis Jacq.))

Abstract

Best linear unbiased prediction (BLUP) is a method designed to predict genetic values ​​from imbalanced data. BLUP analysis that utilizes kinship information can reduce bias due to data imbalance. This study aimed to estimate genetic parameters and general combining ability of parents through progeny testing and validation of prediction models for untested crosses.  A progeny testing was planted in 2008 in a randomized complete block design with 5 replications. Observations were made on 10 characteristics observed in 2015 - 2017. The results showed that mesocarp/fruit and kernel/fruit characteristics were more controlled by tenera parents. Based on the estimation of the general combining ability of dura parents, BJ5686D and BJ5678D can be used to produce varieties with superior characteristics of productivity and oil yield.  Tenera parent, AP141T can be used to produce superior varieties with good oil-related traits. The BJ5819T and BJ5817T parents can be used to produce varieties with slow height increment and high oil-related traits. Special combining abilities were not significant for all characteristics. The results of the 2-fold cross-validation showed that the accuracy for number of bunch , fresh fruit bunches, CPO production, rachis length, core/fruit, dry oil/mesocarp, and oil/bunches was 0.52 – 0.88.  This information can be utilized in subsequent breeding programs.