Characterization of genome properties in oil palm (Elaeis guineensi s Jacq.) breeding populations for better palm oil yield

Abstract

Oil palm (Elaeis guineensis Jacq.) is the most efficient oil-bearing crop grown in humid tropical parts of the world. A better understanding of the Genomic Selection (GS) results in the populations involved needs a detailed study of their genome properties. This study aimed to characterize the genome properties of two complex oil palm breeding populations, i.e, Deli and La Mé. The present study considered 423 Deli, 140 La Mé, and 380 Deli × La Mé hybrid crosses with a total of 943 genotyped individuals. A total of 7,324 SNPs, including 5, 598 SNPs located on the anchored sequences of the genome, were involved. The LepMAP3 software was used to construct the genetic linkage map. Analyses of linkage disequilibrium (LD) in cM and in Mbp were performed using the PLINK software. Haplotypes sharing and minor allele frequency (MAF) were analyzed. The effective size (Ne) was estimated using the NeEstimator 2.1 software and pairwise of fixation index (Fst) was calculated using the SNPRelate R package. The genetic linkage map constructed included 4, 252 SNPs and spanned 1,778.52 cM, with an average recombination rate of 2.85 cM/Mbp. The LD at r2 = 0.3, considered the minimum to get reliable genomic selection (GS) results, spanned over 1.05 cM/0.22 Mbp in Deli and 0.9 cM/0.21 Mbp in La Mé. The LD decay was faster for Deli than for La Mé. The significant degree of differentiation existing between Deli and La Mé was confirmed by the high Fst value (0.53), the pattern of correlation of SNP heterozygosity and allele frequency among populations, as well as the decrease of persistence of LD and of haplotype sharing among populations with increasing SNP distance. The two populations had low Ne (< 5) although a lower Ne was observed in Deli than in La Mé. In conclusion, the study helped to determine the number of markers to be used for future GS studies in oil palm, showing that 10,000 SNPs would be enough to reach the r2 value of 0.3 in Deli and La Mé. Overall, the results showed strong genetic differentiation between Deli and La Mé, but the level of resemblance between them over short genomic distances likely explained the superiority of GS models ignoring the parental origin of marker alleles over models taking this information into account. Future studies in oil palm should consider population-specific genetic maps, new reference genomes and other breeding populations.