Читать книгу Biotechnology of Fruit and Nut Crops - Группа авторов - Страница 142

The genus Elaeis comprises two species E. guineensis, the commercial African oil palm, and E. oleifera, which is used in oil palm genetic breeding. The publication of both the African oil palm genome assembly and the first draft sequence of its Latin American relative (Singh et al., 2013b) now enables studies on the composition, structure and evolution of palm genomes through the annotation of their repeated sequences.

Like all genome reference sequences, the oil palm genome remains incomplete, and resequencing and annotations are still a work in progress which is continuously enriched and complemented by the community of researchers. Sequencing technologies are moving quickly and sequencing is becoming cheaper, allowing the pyramiding of various different technologies which are now able to analyse longer strands of DNAs, e.g. single-molecule real-time (SMRT) sequencing or synthetic long-read technologies, with increasing precision rates. More than 15 years after its publication, the human genome is still complemented and enriched using new approaches like nanopore sequencing (Jain et al., 2018).

Next-generation sequencing enables the study of the detailed pattern of DNA methylation by sequencing bisulfite-treated DNA (Jaligot et al., 2014; Ong-Abdullah et al., 2015) or the effect of a treatment on histone modifications by comparing the DNA sequences associated with immunoprecipitated histones. DNA methylation can also be detected by using SMRT sequencing (Smulders and De Klerk, 2011). Such methods are able to produce detailed overviews of epigenetic modifications among tissues and between developmental stages, a single genome being able to generate a huge diversity of epigenomes, depending on the tissue and its environment (Gallusci et al., 2017). Such epigenomics approaches pave the way for the detection of epigenetically controlled genes for major functions in plant growth and development (Seymour and Becker, 2017). For the oil palm, the list of applications includes major breeding objectives, e.g. responses to biotic and abiotic stress (including drought or application of fertilizers), determinism of sex ratio, effects of competition for light and nutriments, etc.

Beulé et al. (2015) provided evidence of a congruence in the transpositional history of long terminal repeat (LTR) retrotransposons between E. oleifera and E. guineensis, especially for the Sto-4 family. Also, 583 full-length LTR-retrotransposons were identified in the E. guineensis genome. Results show that these elements are most likely no longer mobile and that no recent insertion event has occurred. Moreover, the analysis of chromosomal distribution suggests a preferential insertion of Copia elements in gene-rich regions, whereas Gypsy elements appear to be evenly distributed throughout the genome (Beulé et al., 2015).

In order to better understand the molecular basis of oil palm chloroplasts, Uthaipaisanwong et al. (2012) characterized the complete chloroplast (cp) genome sequence obtained from 454 pyrosequencing. The oil palm cp genome is 156,973 bp in length consisting of a large single-copy region of 85,192 bp flanked on each side by inverted repeats of 27,071 bp with a small single-copy region of 17,639 bp joining the repeats. The genome contains 112 unique genes: 79 protein-coding genes, 4 ribosomal RNA genes and 29 tRNA genes. By aligning the cp genome sequence with oil palm cDNA sequences, the authors observed 18 non-silent and 10 silent RNA editing events among 19 cp protein-coding genes. Creation of an initiation codon by RNA editing in rpl2 has been reported in several monocots and was also found in the oil palm cp genome. Fifty common chloroplast protein-coding genes from 33 plant taxa were used to construct ML and MP phylogenetic trees.

Wong and Bernardo (2008) compared by simulation the response to phenotypic selection, marker-assisted recurrent selection (MARS), and genome-wide selection with small population sizes in oil palm. They assessed the efficiency of each method in terms of years and cost per unit gain. The authors concluded that for a realistic yet relatively small population size of N = 50 in oil palm, genome-wide selection is superior to MARS and phenotypic selection in terms of gain per unit cost and time.

Genomic selection (GS) is an approach based on MAS which uses markers distributed across the entire genome. Unlike conventional MAS strategies, GS does not require any association between molecular marker and trait generated from linkage mapping and GWAS (Teh et al., 2016). Genomic selection is aimed at accelerating genetic gain in breeding programmes, and such an approach is particularly useful for perennial crops such as the oil palm, which have breeding cycles that are long and tedious, and require large areas for genetic trials.

Kwong et al. (2017) assessed the various methods and marker systems available for genomic selection of the oil palm. This novel approach is clearly not fully developed, and the optimal method for GS is still under debate. The authors evaluated the effect of different marker systems and modelling methods for implementing genomic selection in an introgressed dura family derived from a Deli dura × Nigerian dura (Deli × Nigerian) with 112 individuals. This family is an important breeding source for developing new mother palms for superior oil yield and bunch characters. The traits of interest were fruit-to-bunch (F/B), shell-to-fruit (S/F), kernel-to-fruit (K/F), mesocarp-to-fruit (M/F), oil per palm (O/P) and oil-to-dry mesocarp (O/DM). The marker systems evaluated were SSRs and SNPs. RR-BLUP, Bayesian A, B, Cπ, LASSO, Ridge Regression and two machine learning methods (SVM and Random Forest) were used to evaluate genomic selection accuracy of the traits. The authors concluded that due to high genomic resolution, the use of whole-genome SNPs improved the efficiency of genomic selection dramatically for oil palm and is recommended for dura breeding programmes. Machine learning was found to slightly outperform other methods although it required parameter optimization for the implementation of genomic selection.

Cros et al. (2017) showed that preselection for yield components using genotyping-by-sequencing (GBS) is the first possible application of genomic sequencing in oil palm, enabling an increase of selection intensity, thus improving the performance of commercial hybrids. With genomic selection it is now possible to overcome the constraint of the limited number of individuals evaluated through traditional progeny tests. Indeed, Cros et al. (2017) estimated the accuracy of genomic selection prediction of seven oil yield components using A × B hybrid progeny tests with almost 500 crosses for training and 200 crosses for independent validation. GBS yielded +5000 SNPs on the parents of the crosses. The genomic best linear unbiased prediction method gave genomic predictions using the SNPs of the training and validation sets and the phenotypes of the training crosses. The practical impact of this study was illustrated by quantifying the additional bunch production of the crosses selected in the validation experiment if genomic preselection had been applied in the parental populations before progeny tests.