Please use this identifier to cite or link to this item:
Title: Principal Components Analysis and Clustering of EX Situ Oil Palm ( Elaeis guineensis Jacq.) Germplasm
Authors: Li-Hammed Morufat Abimbola
Keywords: principal compenent analysis
cluster analysis
oil palm
Issue Date: May-2014
Publisher: Universiti Sains Islam Malaysia
Abstract: Knowledge of genetic variability and relations among germplasm materials is crucial for selection of promising breeding materials and also to ease their utilization by plant breeders. Hence, this research was undertaken to explore the extent of variability and associations of traits in the MPOB-Nigerian oil palm germplasm using simple statistical tools. Pattern of variability in oder to identify the characters which delineate the germplasm materials was as well studied using principal component analysis and cluster analysis. The results of coefficient of variation indicated that wide variability existed for fresh fruit bunch and its components. Estimates from correlation studies high and significant relations between FFB and majority of the traits except the fatty acid traits. PCA based on all the traits studied showed that first eight PC’s having eigenvalues greater than one accounted for 90.53% of the total variation with FFB, ABW, MNW, MF, KF, SF, ODM, OB, KB, OY, TEP, PCS, RL, LL, LW, LN, LA, LAI, BDM, VDM, TDM, e and f being the most important characters on PC1 contributing to the overall variation. Cluster analysis using Ward’s method and single linkage method grouped all germplasm materials based on all traits into eight groups. Slight variation was found in both methods. However, from Ward’s method Cluster 3 and Cluster 4 contained accessions with high yield traits while Cluster 2 contains accessions with good oil quality traits. Genetic distance was maximum in Cluster 1 and Cluster 2 as well as Cluster 1 and Cluster 8. Thus, selection for charaters with high wide variations could be used for oil palm improvement while hybridization between accessions of different clusters with high cluster means for desired traits and maximum distance could be achieved.
Appears in Collections:Master

Files in This Item:
File Description SizeFormat 
3120239 declaration.pdf1.2 MBAdobe PDFView/Open
3120239 introduction.pdf401.88 kBAdobe PDFView/Open
3120239 chapter 1.pdf175.9 kBAdobe PDFView/Open
3120239 chapter 2.pdf258.4 kBAdobe PDFView/Open
3120239 chapter 3.pdf907.66 kBAdobe PDFView/Open
3120239 chapter 4.pdf809.89 kBAdobe PDFView/Open
3120239 chapter 5.pdf870.93 kBAdobe PDFView/Open
3120239 chapter 6.pdf88.15 kBAdobe PDFView/Open
3120239 reference.pdf365.98 kBAdobe PDFView/Open

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.