WO2010146357A1 - Oil palm and processes for producing it - Google Patents

Abstract

The preferred form of oil palm for commercial production is termed Tenera. Tenera is heterozygous for a gene at a locus K that controls endocarp thickness: it produces fruit with intermediate endocarp thickness between the two homozygous forms of oil palm, Dura and Pisifera, giving easier processing and a higher yield of oil. The invention provides a method of obtaining a population of Tenera offspring from crossing two Tenera parents, which comprises testing immature progeny from the cross with distinct right and left differentiating molecular markers, and selecting for further cultivation those individuals whose marker inheritance patterns indicate they are likely to be Tenera.

Description

Oil palm and processes for producing it

This invention relates to oil palm, and to processes for producing it. More particularly, it relates to processes for obtaining populations of oil palm consisting mostly or entirely of a desired genetic constitution, for example of Tenera palms.

Background

Oil palm is a tropical crop of major economic importance. It is diploid, with two copies of each chromosome. Thickness of the lignified endocarp of oil palm fruits is a property that has great commercial importance: it should be neither too thin nor too thick. The property is mainly determined by a single gene at a locus we will term K. This gene has two forms, referred to as Sh and sh. The original cultivated oil palms all possessed thick endocarps and are known as 'Dura' types. They are homozygous for the Sh gene, with genotype Sh/Sh. The original cultivated Dura palm varieties exhibited a low oil extraction ratio (OER) in the range of 12-16%. However, a mutant allele of Sh was later discovered, termed sh. This yielded virtually no lignified endocarp in its homozygous form (sh/sh, giving rise to a type of palm known as Pisifera) but in the heterozygous state (Sh/sh) gave an intermediate shell thickness that allowed higher oil extraction and more than doubling of harvestable yields. Such palms are termed 'Tenera¹. Modern commercial oil palm varieties are all of the heterozygous 'Tenera' type. Tenera palms can be produced reliably from crosses between Dura (Sh/Sh) and Pisifera palms (sh/sh), although it is difficult to predict the agronomic performance of the Tenera offspring from that of the parental clones, because of the over-riding importance of shell thickness. The favoured approach is therefore to cross Tenera parents, but this means that only half of the progeny will possess the desired genotype Sh/sh (the others being either Dura, Sh/Sh or Pisifera, sh/sh). There is also the additional problem that the shell thickness phenotype can only be scored after the palms have completed the juvenile phase of development and started to flower and set fruit. This generally requires 2-4 years. The provision of a reliable molecular marker system to predict the genotype at locus K is therefore highly desirable.

Two main oil palm plant species are grown commercially: Elaeis oleifera Kunth and Elaeis guineensis Jacq. The historic origin of the oil palm (Elaeis guineensis) is understood to be West Africa, where it has been cultivated for many years: the species was introduced from West Africa to the Pacific region in the first half of the last century, since when it has been widely cultivated throughout that and other tropical regions. Two countries in which it is currently widely grown are Malaysia and Indonesia.

There are several challenges that impede the production of a reliable molecular marker system for predicting genotype at locus K. First, whilst it is now well- established that shell thickness is under the major influence of a single gene, there is considerable variation in the phenotypes of the three genotype categories (Dura, Tenera and Pisifera) when viewed across a wide range of germplasm (e.g. see Figures 1 and 2) . This is presumed to be due to the action of modifier genes and can complicate accurate distinction between the three types. A second problem is the tendency of the Pisifera types not to flower, or to flower late or intermittently. However, the absence of flowering is not a reliable indication of a Pisifera genotype: some Tenera and Dura types can exhibit a prolonged juvenile phase (and so flower late) or may show only intermittent flowering. Third, when genetic recombination occurs between any potential markers and the locus K primarily responsible for shell thickness, the association between shell thickness phenotype and marker allele is lost. Recombination occurring in the ancestry of the parent of a particular palm could mean that the marker alleles are in reverse phase to the usual condition. Fourth, use of anonymous multi-locus marker systems such as AFLP (Amplified Fragment Length Polymorphism) to generate markers for the locus K seriously complicates application of a particular marker within the profiles of other progenies and usually necessitates conversion to create a Sequence Tagged Site (STS) marker. Fifth, STS markers that are polymorphic in a segregating progeny and so can be mapped as linked to one of the alleles at locus K may nevertheless lack polymorphisms between parents of other progenies and so have no value for Marker-Assisted Selection. Sixth, there can be considerable variability in endocarp thickness between fruits on a single palm and this can change with maturity. Moreover, the first fruits of a palm may be predominantly sterile or difficult to characterise. This property can compromise the accuracy of any mapping study aiming to identify markers tightly linked to the locus K. Thus, there is a significant chance that the presumed location of K may vary considerably. The scoring of the population is likely to change with age (as late flowering palms produce fruits that can be more accurately phenotyped).

When the above factors are considered collectively, there appears little prospect of developing a robust marker system that reliably indicates genotype at locus K across a wide range of progenies without identifying the locus itself or a marker so close to it as to render recombination extremely unlikely. Whilst the locus K is included in the only published map of oil palm (Billotte et al., 2005), the closest flanking marker is an AFLP marker positioned some 4.7 centimorgans (cM) distant from the presumed position of K, with the nearest Sequence-Tagged Site (STS) marker around 23 cM from the inferred position of the gene. Quite apart from the genotyping errors expected from recombination when using either of these markers to predict shell thickness (around 5% for the AFLP marker and 23% for the microsatellite marker), there is the additional problem that phase linkage is very likely to be broken as a consequence of recombination during breeding. In this way, if presence of the AFLP band or a particular microsatellite allele is associated with the dominant Sh allele in the mapping population, recombination at some stage during the breeding of the crop might well mean that the same markers are associated with the recessive sh allele in other populations. There can also be considerable effort required to convert an AFLP marker into a reliable STS marker. Finally, the polymorphism exploited in establishing the linkage of either marker to locus K may well be absent between other breeding parents such that segregation cannot be traced in their progenies. Thus, whilst Billotte et al. (2005) claim that the use of the two AFLP markers in combination would enable "prediction of the variety type directly from the nursery type, with around 99.5% (1 -[0.047 x 0.1 16]) reliability", the reality is very different, with both markers needing to be converted into STS markers before general use and the likelihood of ancestral recombination being extremely high. It follows that there are real problems impeding progress towards the development of a robust and widely applicable marker system to predict the genotype at the locus K.

The present invention seeks to solve the problems outlined above and to provide a tractable means of predicting genotypes at locus K of oil palm (particularly of embryos, seeds and seedlings) within a breeding programme. According to the present invention we provide a method of obtaining populations of oil palm consisting predominantly of Tenera, the method comprising: a) crossing or selling parent Tenera oil palm plants to produce progeny; b) testing immature progeny so obtained with at least two right and left markers flanking locus K (the Sh locus), thereby enabling the determination of inheritance patterns of the markers in tested progeny; c) selecting from the tested progeny individuals whose marker inheritance patterns indicate they are likely to be Tenera; d) growing on such selected individuals. Preferably the resulting population is at least 95% Tenera, particularly about 98% Tenera.

The invention further comprises a method of obtaining populations of oil palm consisting predominantly of Tenera from one or two Tenera parents, the method comprising: a) selfing or crossing the parent or parents to produce progeny; b) testing immature progeny so obtained with a pair of distinct right and left differentiating markers flanking locus K in at least one parent, thereby enabling the determination of inheritance patterns of the markers in tested progeny; c) selecting from the tested progeny individuals whose marker inheritance patterns indicate they are likely to be Tenera; d) growing on such selected individuals.

By a 'differentiating marker' we mean one which is linked with one allele at locus K but not the other, in at least one of the parents. Thus, in the two Tenera genomes shown diagrammatically below, the markers P and Q are both differentiating, but the marker R is not.

Sh P Sh R sh Q sh R A differentiating marker will normally be heterozygous in Tenera offspring, in the absence of recombination.

The marker inheritance pattern of the tested progeny determined in step (c) may be about 50:50 homozygous:heterozygous, when the selected individuals are those that are heterozygous. Preferably the marker inheritance pattern of the tested progeny determined in step (c) is over 90% heterozygous for at least one marker. Preferably the selected individuals are heterozygous for both right and left flanking markers. Preferably the associations between the test markers and the Sh and sh alleles are known for both parents. It is preferred that both markers are within about 13 cM, particularly within about 6 cM of the locus K.

Before the process of our invention can be applied to oil palm germplasm, one needs to find suitable markers linked to the Sh and sh genes. However, this may only need to be done once, so it is not an essential step in the process of our invention.

Conveniently the first stage in the search (Phase 1) is to use a single STS marker (e.g., VSl) that is linked to the locus K (but too distant from it to have practical utility to predict fruit phenotype alone) to accumulate a series of STS markers that flank the locus K. The VSl marker can act as a foundation marker around which others that are also linked to the locus K are assembled in the mapping population. In this way, embryos, seeds and immature plants could be partially characterised prior to shell thickness becoming apparent. The emphasis focuses on saturating a local STS map in the vicinity of locus K prior to the phenotype of all members of the population being available. However, because of the unreliable early phenotypic scoring of this trait, the locus K does not feature centrally in the provision of this local map. Rather, it is built around the STS marker (VSl) which is known to be linked to the locus K. Only when the palms in the population are sufficiently mature is the position of the locus K integrated into the map. In this way, one may address difficulties arising from the late flowering and fruiting of some palms within any mapping population and associated difficulties in reliably scoring palms for Tenera phenotype prior to full maturity. This strategy also provides scope for generating sufficient markers on either side of the desired locus to enable the twinned marker scoring strategy outlined below. The concentration on STS markers allows for ready transferral to other populations and enables linkage phase reversal to be traced in pedigrees and accommodated for (see below).

The second phase (Phase 2) of the process involves the use of flanking markers to infer which members of a progeny can be accurately genotyped as Sh/sh (Tenera types). The third phase is to combine Sh (shell thickness) phenotype data and linkage phase of flanking markers to infer which breeding lines (if any) have undergone recombination between both of the closest markers and locus K, thereby reversing linkage phase in progeny secured from them. This knowledge allows broad application of the STS marker set clustered around the locus K and correct adjustment of the predicted phenotype on the basis of the flanking genotype such that erroneous calls should only arise from (relatively rare) de novo double recombination events

(see below).

The invention will now be described in more detail with reference to the accompanying drawings, in which:

Figure 1 illustrates a scheme for scoring the shell thickness phenotype of oil palm fruits, showing a series of photographs of cross-sections of fruit graded from 0 to 5 in shell thickness;

Figure 2 shows Pisifera fruits containing kernel and embryo (at left), and

Pisifera fruits containing kernel but no embryo (at right);

Figure 3 gives the base sequence of the VSl target locus;

Figure 4 shows plots of normalised fluorescence against temperature in High Resolution Melt (HRM) analysis of amplicons of homozygotes and heterozygotes;

Figure 5 shows plots of the derivatives (dF/dt) of the data shown in Figure 4;

Figure 6 is an HRM normalised graph output for a set of six UTR Eg markers screened against six oil palm genotypes as a mapping pre-screen;

Figure 7 shows a first-order differential plot of the data in Figure 6; Figure 8 is a linkage map of locus K based on 144 individuals of the Bref 0104065 IB mapping population.

Phase 1 : identifying markers linked to VSl and the Sh locus

1.1 Mapping population generation and shell thickness characterisation

A mapping population of 288 individuals was generated from the self-pollination of a known Tenera individual BL10323/104-5. The seedlings of this population, designated Bref 0104065 IB, were allowed to mature in order to determine shell thickness genotype of each individual. This required at least 3 years for each tree to reach flowering and fruiting maturity.

1.1.1 Scoring of shell thickness A minimum of five fruits from each individual in the progeny were evaluated for shell thickness. For this, we generally took 10 ripe fruits (for some individuals we were only able to sample 5-9 fruits) per palm. Fruits were cut laterally into two halves and the thickness of shell and fibre ring scored according to the scale outlined below.

Shell-thickness was scored on a visual scale from 0-5, with 0 being Pisifera type (no shell), 1 being fruit with the thinnest shells and 5, fruits with the thickest shells. The scale is illustrated in Figure 1, which shows a series of fruits, with central kernels (light coloured) surrounded by lignified endocarp (dark) and fleshy mesocarp (intermediate shade). .A score of 3 was allocated to indicate an intermediate condition between Dura and Tenera fruit types. Averaging over the ten fruits, a plant with mean score of less than 1 was deemed Pisifera (provisionally - see 1.1.1.1 below), with 1-3 as Tenera and greater than 3 as Dura.

1.1.1.1 Scoring palms that do not flower or produce rotten fruits

Pisifera palms frequently produce sterile female flower bunches, so-called Pisifera steriles, and some workers have used this as an indication of a Pisifera phenotype.

However, Dura and Tenera types also occasionally produce sterile bunches and so this was considered an unreliable means of phenotype assignation. We therefore deemed a palm as being Pisifera only when a 'Pisifera kernel' or 'Pisifera embryoid' was found in the 'Pisifera inflorescence' after collecting and cutting all fruits from the bunch.

Figure 2 shows examples of Pisifera structures with and without embryos.

1.2 VSl marker amplification

The marker VSl derives from genomic DNA sequence of oil palm (Elaeis guineensis). The DNA fragment (Figure 3) contains a sequence polymorphism based on length variation in an internal microsatellite with an AT motif. The sequence is shown in Figure 3 with the internal microsatellite highlighted. 1.2.1 Primer design

Primers were designed using 'Primer 3' (http://frodo.wi.mit.edu/cgi- bin/primer3/primer3_www.cgi). Possible self-complementarity of the primer was avoided using the 'Oligonucleotide Properties Calculator'

(www.basic.nwu.edu/biotools/oligocalc.html^*). The primers developed for VSl in this way are given below: Forward 5 '-3': GAG ATT ACA AAG TCC AAA CC Reverse 5 '-3': TCA AAA TTA AGA AAG TAT GC

1.3 DNA extraction

Around 50 mg leaf tissue was removed from the seedling of each individual of mapping population Bref 0104065 IB. This material was used to extract DNA using the Qiagen 96 DNeasy extraction kit according to the manufacturer's instructions as described below.

1.3.1 Preparation

1. Ensure 100% ethanol has been added to buffers AP3/E and AW

2. Preheat AE and API buffer to 65°C 3. If API buffer has a cloudy appearance, heat to 65°C and shake until the solution becomes clear.

1.3.2. Protocol

1. Add 50 mg plant material into each tube in two collection microtube racks 2. Add one tungsten carbide bead into each microtube

3. Prepare the lysis solution: (400 μl APl+ 1 μl RNAse + 1 μl Reagent DX)/reaction plus 15% of each component

4. Disrupt the sample using a Retsch Mixer Mill 300, 30 Hz for 1.5 min

5. Pulse centrifuge to 3,000 rpm 6. Remove and discard caps, add 130 μl AP2 buffer to each collection microtube

7. Close the microtubes with new caps. Place a clear cover (from step 1) over the 96-well plate. Shake the plate vigorously for 15s. Pulse centrifuge to 3,000 rpm 8. Incubate the racks for 10 min at -20°C

9. Remove and discard the caps. Transfer 400 μl of each supernatant to new plate of collection microtubes. Do not transfer pellet and floating particles. Hold the strips and use the lowest pipette speed. Recover the tungsten beads 10. Add 1.5 x volumes (typically 600 μl) of AP3/E buffer

11. Close the microtubes with new caps and mix vigorously

12. Pulse centrifuge (3,000 rpm) to collect solution

13. Place 96 well plates on top of S-Blocks provided

14. Transfer ImI of sample into each well of the 96 well plate 15. Seal with Airpore Tape sheet and centrifuge for 4 min at 6,000 rpm

16. Add 800 μl of Buffer AW to each sample

17. Centrifuge for 15 min at 6,000 rpm

18. Add 100 μl of buffer AE to each sample and seal with new AirPore sheets

19. Incubate for 1 min at room temperature (15-25°C) 20. Centrifuge for 2 min at 6,000 rpm

21. Store resultant DNA preparations at -20⁰C

1.4 VSl amplification via Polymerase Chain Reaction (PCR)

VSl primers were applied to the DNA isolated from each of the 288 individuals of the Bref 0104065 IB mapping population in order to determine the VSl genotype for each individual.

1.4.1 Reaction mixtures and thermocycling conditions for PCR In all cases, 10 μL of reaction mixture contained the following reagents; 5.0 μL of 2x Biomix (Bioline, 0.5 μL of each primer pair (10 μM), and 1-5 ng of DNA (extracted as above).

The following conditions were used for the Polymerase Chain Reaction (PCR) for VSl: an initial 94°C denaturing step for 5 min followed by 35 cycles of; 94°C for 30 sec, 50⁰C for 30 sec and 72°C for 45 sec, with a final extension step of 72°C for 10 min. 1.4.2 Separation of PCR products by agarose gel electrophoresis

The products of application of VSl primer to each of the 288 individuals of the Bref 0104065 IB mapping population were fractionated by agarose gel electrophoresis in a 3% w/v metaphor TAE gel stained with ethidium bromide and visualised under UV excitation.

1.4.2.1 Reagents

Tris Acetate-EDTA (TAE) buffer: 40 mM Tris-acetate, ImM EDTA, pH 8.3 Bromophenol blue loading buffer. 0.23% (w/v) bromophenol blue, 60 mM EDTA, 40% (w/v) sucrose

Ethidium bromide stain: 1% (w/v) ethidium bromide 100 base pair ladder (Gibco Life Science BRL)

1.4.2.2 Gel preparation 3.0% (w/v) agarose was prepared in TAE buffer heated in a microwave (700 W) for 2 x 1 min at full power to create a clear solution. The gel was cooled to approximately 55°C prior to the addition of ethidium bromide (3.5μl per 100 ml gel). The ends of a suitable gel tray rig were sealed with masking tape and an appropriate number and type of combs placed in position. Combs with 40 x 20 μl wells were most often employed. The gel solution was carefully poured into the prepared tray and allowed to cool for at least 20 min. Combs and tape were then removed and the gel tray submerged into a tank containing 1 x TAE buffer.

1.4.2.3 Sample loading Generally, 5 μl of sample were mixed with 2 μl of bromophenol blue buffer prior to loading. The loading buffer serves two functions: first, it increases the specific gravity of the sample thereby preventing diffusion of DNA from the top of the well into the surrounding buffer, and second, it indicates the progress of products as they migrate through the gel by electrophoresis (the blue dye migrates at approximately the same position as DNA fragments 200 bp in length). To estimate the size of the amplicons, 4 μl of 100 bp Gibco's ladder (Gibco Life Science BRL) were loaded together with the analysed samples.

1.4.2.4 Electrophoresis and visualisation Electrophoresis was performed at 100 Volts in IX TAE buffer for approximately 2.5 h. Following electrophoresis, gels were viewed under Ultra Violet illumination using a UVP Bio-Doc-system. Images of the gels were captured by the UVP Bio-Doc system in jpeg format for scoring.

1.5 Generation of additional markers

STS anchor points were taken from the mapped polymorphic microsatellites described by Billotte et al. (2005). Microsatellite markers used from this set included the microsatellite listed as being closest to the Sh locus (mEgCIR.3275) and all others on linkage group 4 of oil palm (Billotte et al., 2005). In order to develop markers that are closer to Sh and VSl, additional markers were generated in two ways:

1. Microsatellites were identified from EST databases and also by screening a small-insert genomic library for tandem repeat motifs

2. SNP markers were identified by High Resolution Melt (HRM) screening of PCR amplicons generated using locus-specific primers designed from cDNA sequences held on EST databases.

1.5.1. Generation of microsatellite markers

1.5.1.1 Development of microsatellites from ESTs

Microsatellite regions were identified by an in silico screen for the presence of microsatellite motifs among EST sequences listed on the NCBI (http://www.ncbi.nlm.nih.gov/) and the MPOB PalmGenes (http://palmoilis.mpob.gov.my/palmgenes.html) web sites (accessed January 2005).

EST sequences were screened for repeat motifs comprising five units or more in length (e.g. ATATATATAT) using the Gramene SSR search tool (http://www.gramene.org/db/searches/ssrtool; Temnykh et al., 2001). This screen identified 310 potential sites where there was sufficient flanking sequence on either side of the microsatellite array to design a pair of locus-specific primers. Primers were designed around the repeat array using 'Primer3'_v (http://frodo.wi. mit.edu/cgi- bin/primer3/primer3_www.cgi). The reverse primers were modified with the addition of a fluorescent label (NED, HEX or FAM) to enable capillary detection.

1.5.1.2 Development of microsatellites from an oil palm genomic library We also isolated some microsatellites from a small-insert library of oil palm DNA. A genomic library comprising 18,432 colonies and with an average insert size of 600 bp, was constructed from the parental clone (BLl 0323/104-5) of the mapping population. DNA was first extracted using a CTAB-based method (Doyle & Doyle, 1987). Following digestion with restriction enzymes Sau3A I (Promega) and Hind III (Gibco-BRL), the DNA was fractionated by electrophoresis through a 1% w/v agarose gel and fragments 400 bp - 1 kb in length were gel-extracted and purified using a Nucleospin^® Extract column (Machery-Nagel). Purified DNA fragments were then ligated into the pGEM^®-3Z vector (Promega) cut with Hind III (Gibco-BRL) and BamK I (Promega). The ligated vector was transformed into competent XLl -Blue E. coli cells (Stratagene) and cultured on 22.2 cm x 22.2 cm square agar plates. The Qpix2 Robotic Colony Picker (Genetix) was used to identify and transfer transformed colonies directly into 384-well plates containing 60 μL of L.B./ampicillin broth. PCRs were performed on bulks of the colonies to identify those that contain microsatellites (also known as single sequence repeats, SSRs). The PCR screen employed primers that targeted the appropriate microsatellite (both AC and AG motifs) but also contained a short anchor at the 5' or 3' termini (Charters et al, 1996) in association with the M 13 universal primers designed from the vector sequence. PCRs were conducted in 10 μL mixtures comprising Ix PCR buffer (Bioline - 160 mM (NH₄)₂SO₄, 670 mM Tris-Cl (pH 8.8), 0.1% Tween-20), 0.3 raM MgCl₂, 0.4 mM dNTPs, 0.5 μM of each microsatellite primer (forward or reverse used in combination with Ml 3 forward or reverse), 0.5 μM of the Ml 3 primer (either forward or reverse) and 1 U of Taq DNA polymerase (Bioline). PCRs were performed on a Gene Amp^® PCR System 2700 (Applied Biosystems) thermal cycler using an initial 94°C denaturing step for 1 min followed by 35 cycles of; 94°C for 1 min, 1 min annealing at 55°C and 72°C extension for 1 min, then a final extension step of 72⁰C for 7 min. PCR products were fractionated on 1% w/v agarose and visualised under UV light (as described above). Colonies that produced a 'smear' of DNA fragments between 500 bp and 1.5 kb after PCR amplification were selected for sequencing. These positive bacterial clones were sequenced using a 373A ABI Automated Sequencer (Perkin Elmer) using the ABI Prism™ Big Dye™ Sequencing Ready Reaction Kit (Applied Biosystems) using the M 13 Forward primer for sequencing. Primers were designed flanking the identified microsatellite sequences using 'Primer3' (http://frodo.wi. mit.edu/cgi-bin/primer3/primer3 www.cgi) as previously described with the reverse primers modified with the addition of a fluorescent label (NED, HEX or FAM) to enable capillary detection.

1.5.1.1 Assessment of polymorphism In order to be of value for mapping purposes, markers needed to be polymorphic between individuals of mapping population Bref 0104065 IB. The primer pairs obtained by the various methods above were screened for heterozygosity in the parental genotype BLl 0323/104-5 by PCR and capillary electrophoresis (see below).

1.5.1.1.1 PCR

PCR amplification was conducted in a lOμl reaction volume comprising 5.0μl of 2x Biomix (Bioline), 0.5μl of each primer pair (lOμM), and 1-5 ng of DNA (extracted as above). Samples were then subjected to the following thermocycling conditions: an initial 94°C denaturing step for 5 min followed by 35 cycles of; 94⁰C for 30 sec, 50°C for 30 sec and 72°C for 45 sec, with a final extension step of 72°C for 10 min.

1.5.1.1.2 Capillary electrophoresis

The inclusion of fluorescently labelled reverse primers in the PCR mix generates amplicons with a fluorescent label that are detectable via automated capillary electrophoresis. Samples were submitted to Macrogen Inc. (Seoul, Korea) for GeneScan Analysis. Resultant electropherograms were analysed using GeneMapper software, version 4.0 (Applied Biosystems) and allele peaks scored for product size.

1.5.1.2 Application of polymorphic microsatellites The resultant polymorphic microsatellite markers identified in the above screen were applied to a set of 144 individuals from the mapping population Bref 0104065 IB for linkage map assembly. PCR and capillary electrophoresis conditions were applied as described above. 1.5.2. Generation of SNP-based markers

Primer pairs were designed from unique EST transcripts downloaded from the NCBI (http://www.ncbi.nlm.nih.gov/) and the MPOB PalmGenes web sites using Primer3 (http://frodo.wi.mit.edu/cgi-bin/primer3/primer3 www.cgi). These primers were in turn used to generate PCR amplicons of between 150-350 base pairs in length. The products generated by the parent and five randomly selected individuals from the Bref Ol 04065 IB progeny were then screened for the presence of SNP-based polymorphisms using High Resolution Melt Analysis to assess their potential value for linkage mapping. In addition, primer pairs originally developed for microsatellite analysis that yielded amplicons that were monomorphic for size variation within the mapping population following capillary electrophoresis were also screened for the presence of SNPs using the HRM method described below.

1.5.2.1 High Resolution Melt analysis High-resolution melt (HRM) analysis is a rapid technique that allows the detection of single nucleotide polymorphisms (SNPs) without the need to sequence individual genotypes or to size- fractionate PCR products (Germer et al., 1999). The technique involves precise monitoring of fluorescence changes caused by the release of an intercalating dye from double-stranded DNA (for example, PCR products) as they dissociate with increasing temperature. Subtle sequence variations of a common template (that may differ by as little as one base in 500 bases) can be distinguished by melting temperature (T_m) shifts. The power of DNA melt analysis lies in the ability to sensitively detect these T_m shifts whilst maintaining a stable, accurate temperature. Recent developments in fluorescence-detection instruments and the use of fully saturating intercalating dyes have made HRM analysis highly sensitive. Now fluorescent melt data can be acquired at a level of 0.1 °C/sec allowing for rapid, high- resolution analysis of complete fragments within minutes (Wittwer et al., 2003; Krypuy et al., 2006). Differences in T_n, can produce subtle changes in melt curves. Heteroduplex associations between imperfectly complementary strands mean that melt curve divergences are typically greater between the heterozygous genotypes and the homozygous genotypes than between the two homozygous genotypes (e.g. Croxford et α/., 2008). 1.5.2.2 Assessment of polymorphism

PCR amplification and subsequent HRM analyses were performed on a Corbett Life Sciences Rotorgene 6000. The intercalating fluorescent DNA dyes SYTO 09 (Invitrogen) or EvaGreen (Quantace) were used as the substrate to monitor both the accumulation of product during PCR and the subsequent product melt on the RotorGene 6000 (software Version 1.7, Corbett Life Science). These dyes are thought to saturate all (or nearly all) available sites within double stranded DNA (Monis et ai, 2005) and provide an accurate assessment of DNA melt status.

PCRs were performed in 10 μl volumes containing: 1.0 ng template DNA, 2x Sensimix (Quantace), 0.5 μM of each primer and 35 μM of SYTO 9 or EvaGreen. A rapid-cycle PCR protocol was conducted in a 72-well carousel using an initial denaturing step of 95 °C for 10 min followed by 35 cycles of: 95°C for 5 sec, 53-58°C for 10 sec and 72°C for 15 sec, then a final extension step of 72°C for 2 min.

High Resolution Melt analysis was conducted by ramping from 70°C to 90⁰C with 0.1 °C incline per 2 s acquisition step and data were attained from the Cycling A- Green channel. Polymorphic markers were identified when profiles yielded detectable differences in T_m between the two homozygotes.

1.5.2.3 Application of polymorphic SNPs

Markers yielding polymorphic profiles (see Figures 4 and 5) were carried forward for linkage analysis by application to a set of 144 individuals of the Bref 0104065 IB mapping population using the protocols described above.

1.5.2.4 Identification of SNP markers potentially linked to VSl

For all runs performed on the Rotorgene 6000, results for the VSl genotype were colour-coded such that markers potentially linked to the locus K could be readily identified. From this, we uncovered 6 SNP markers putatively linked to VSl and locus K (see Figures 4-7). Linkage was confirmed by linkage mapping (see below). 1.6 Linkage map construction

A linkage map was assembled using the polymorphic Sequence Tagged Site (STS) markers identified above which were applied to 144 individuals of the Bref 0104065 IB progeny. The genotypes of these individuals were determined for presence or absence of 250 polymorphic markers using the methods described above (microsatellites by fragment analysis and SNPs by HRM analysis). The resultant data matrix was then assembled and chi-square tests for segregation distortion and linkage mapping calculations were performed using JoinMap^®4 software (van Ooijen & Voorrips, 2001) using the parameters for the outcrossing population - CP. Chi-square analyses were performed at threshold values ofp = 0.05 and skewed markers less than this constraint were omitted from further analysis. The Kosambi mapping function was used to convert recombination frequencies into map distances (Kosambi, 1944). Linkage groups were constructed and marker order determined using a minimum LOD score threshold of 4.0, a recombination fraction threshold of 0.375 and a ripple value of 1.0.

1.7 Results

1.7.1 Linkage between Sh and VSl

In total, 127 of the 288 individuals of the Bref 0104065 IB mapping population flowered, bore fruit and were scored for shell thickness phenotype. The trees that could be phenotyped for shell thickness comprised only Dura (45 individuals) and Tenera (76 individuals) types due to the delayed flowering associated with Pisifera palms.

The entire population was also screened for its VSl genotype and data have been obtained for 275 of the 288 individuals. Of these 275, 33 individuals generated a single amplicon of 170 bp in length containing an ATxIO repeat array; 15 generated a single 178 bp amplicon containing an ATx 14 repeat; and 73 individuals were heterozygous containing both.

Data for both the locus K and VSl were generated for a subset of 121 individuals. If VSl was unlinked to the locus K controlling shell thickness, there would be an expected 1 :2:1 segregation amongst the subset of the progeny that had been assessed for shell thickness. If the two loci are linked, however, the VSl genotypes should be similarly distorted from this expected ratio. Independence of VSl and locus K was therefore tested using a chi-squared test. The result showed a highly significant rejection of the null hypothesis that VSl and Sh are unlinked (^² = 41.41, /XO.0001). VSl can therefore be used to predict Sh and as an anchor point for additional marker development. The results confirm that the heterozygous condition at VSl is associated with the Tenera phenotype for shell thickness (Sh/sh), the Dura phenotype with homozygosity for the short 170 bp allele (Sh) and Pisifera with homozygosity for the longer 178 bp allele (sh).

1.7.2 Generation of microsatellite markers

1.7.2.1. Screen of mapped microsatellite markers

In all, 256 of the 390 microsatellite markers described by Billotte et al. (2005) were screened for heterozygosity in the parental clone BLl 0323/104-5. Only 47 of these yielded two clear allele peaks in the parent and also segregated amongst the mapping progeny. These markers were accordingly deemed to be heterozygous in the parental clone. Genotype data for these markers was generated from 144 of the set of individuals from the Bref 0104065 IB mapping population and used to assemble a linkage map.

1.7.2.2. In silico screen of EST sequences

The in silico screen of >9,500 EST oil palm sequences listed publicly on the NCBI and MPOB web pages uncovered 310 loci with di-nucleotide microsatellites exceeding 5 repeats in length. Of these, 278 generated single amplicons when subjected to PCR and 31 proved to be heterozygous in the mapping population parent (BLl 0323/104-5). Genotype data for these markers were generated and scored from the set of 144 individuals from the Bref 0104065 IB mapping population and passed forward for subsequent use in linkage mapping (see below).

1.7.2.3. Screen of small-insert genomic library

The initial PCR-based screen of the oil palm genomic library revealed 78 candidate clones in which the Inter Simple Sequence Repeat (ISSR) + Ml 3 (plasmid) primer combination generated PCR products. These products were sequenced using the Ml 3 forward primer and 53 were found to contain microsatellites and adequate flanking sequence to allow primer design. When these primers were applied to the parent clone (BLl 0323/104-5, see Section 1.1 above), 18 were found to be heterozygous and so passed forward for mapping (see below).

1.7.3 High Resolution Melt (HRM) screens for SNP-based markers

In total, amplicons from 1,243 loci were screened for HRM melt profile polymorphisms. From this we found 221 polymorphic profiles that were used for mapping purposes. Figures 6 and 7 are respectively the HRJVI normalised graphs and first order differential outputs for a set of six UTR Eg markers screened against six oil palm genotypes as a mapping pre-screen. The HRM assay was set up with individual traces coded for VSl as: green (solid line) = heterozygote (predicted Tenera), red (dotted line) = homozygote (predicted Pisifera), blue (dashed line) = homozygote (predicted Dura). The co-segregation of melt profiles with VSl colour codes indicates a potential link between the two markers. In the normalised melt curve (Figure 6) the decrease in fluorescence is associated with the release of dye as complementary DNA strands separate with gradual temperature increases. Note that the melt curves separate into three distinct types associated with the three genotypes. First order differential curves of the normalised melt are shown in Figure 7.

The nature of the polymorphism of some of these markers was such that allele types could be scored directly (see Figures 4-7). However, some markers required an additional heteroduplex step in order to separate homozygotes aa and bb. After the initial PCR and melt step, the samples are spiked with the product of one known homozygote and re-melted. The addition of homozygote aa sequence to samples that have the homozygote bb sequence results in an artificially derived heterozygous product (ab) that will then melt in the same manner as the heterozygotes identified in the initial run. The addition of homozygote aa sequence to samples that are also homozygote aa will result in a sample that remains homozygous.

1.7.4 Local mapping around Sh

Of the 319 polymorphic potential STS markers described above, 267 segregated according to Mendelian expectations for an F₂ cross-pollinated population (1 :2:1) at the stringent chi-square threshold of P = 0.05. Of these, 186 formed clear segregation associations with each other dividing into the expected 16 linkage groups at a Limit of Detection (LoD) threshold value of 4.0. This represents the haploid base chromosome number of the species {Elaeis guineensis, x = 16). The linkage group containing the locus K (Linkage Group 4) consisted of 13 STS markers with the closest marker to locus K (UTREg631b) being 2.5 cM away. This linkage group is shown in Figure 8, which is a local linkage map of the locus K based on 144 individuals of the Bref 0104065 IB mapping population. It shows the position of STS markers around the shell thickness locus K (marked WA' on the map): markers used to screen for locus K are indicated by the right brace: } . In all, seven STS markers were sited within 13 cM of the inferred map position of locus K (Figure 8). These markers were carried forward to develop an automated system for screening for Tenera plants (Phase 2).

Phase 2. Use of flanking markers to assign phenotype

A key limitation of using a single pair of flanking markers to assign phenotype is that there is a high probability that many of the parental combinations of interest will lack polymorphisms for at least one of the two marker loci. It was for this reason that a set of STS flanking markers was developed on either side of locus K. Application of this set of flanking markers across 506 parent combinations revealed that there were polymorphic markers on both sides of locus K for 199 (39%) of the progenies tested. All but 6 of the remaining progenies had at least one polymorphic marker on one side. Thus, the use of SNP markers flanking locus K on one side only would provide insufficient coverage across a commercial breeding programme.

The commercial impetus for identifying a marker system that predicts Tenera phenotypes stems largely from the predominant practice of crossing two Tenera parents to generate elite Tenera offspring. The need is to diagnose the desirable immature Tenera from the undesirable Dura and Pisifera offspring at the nursery stage, or better yet in the seed before planting. For this, if we demonstrate that the Tenera parent(s) are heterozygous for two appropriate flanking markers, offspring lacking a single recombination between the flanking markers will also be heterozygous for both markers. We therefore may use the seven markers flanking either side of locus K across a wide range of parent combinations in order to distinguish which genotypes are heterozygous (associated with Tenera phenotype) for markers on either side of locus K using High Resolution Melt analysis.

The following Examples illustrate the invention. All oil palm germplasm used was fromplantation stock grown at Sumatra Bioscience in Indonesia.

EXAMPLE 1

As an example, we crossed two parents (BLl 0660/44-04 x BLl 0660/44-04) that were heterozygous for markers sited either side of locus K, viz: markers EgSNP323 and UTREg621 (shown in Figure 8), to give a small progeny of 20 plants (Bref 06039096B). The progeny were tested for heterozygosity at the marker sites, using HRM analysis. Results are shown in Table 1 below. As expected, half (10) of the progeny were heterozygous at both marker sites, that is on both sides of locus K, indicating the absence of a single recombination event in these individuals (Table 1). Two single recombinants were also noted for marker EgSNP323 and one for UTREg621. The remaining plants were deemed to be either Dura or Pisifera types. The plants exhibiting heterozygosity for both flanking markers (and so deemed to be of the desired Tenera type) were retained and grown on, thereby obtaining a population that could be verified at maturity as exclusively Tenera. All other plants were discarded as either Dura, Pisifera or recombinants. Here, the expected error in Sh genotype binning attributable to double recombination was calculated as 0.27% (0.025 x 0.107). The error rate for other progenies depends on the markers used and may for example be as high as 1.27% (0.099 x 0.128) with markers Eg448 and VSl.

Table 1 Segregation of genotype of 20 progeny plants (Bref 06039096B) derived from a cross between two parents (BLl 0660/44-04 x BLl 0660/44-04) for markers EgSNP323 and UTREg631 as revealed by High Resolution Melt analysis.

Note that plant numbers 02, 03, 06-09, 11, 13, 18 and 20 are heterozygous for both flanking markers and so deemed as Tenera type (T in Genotype column). Plants 04, 12 and 19 are single recombinants (R in Genotype column). Remaining plants are either Dura or Pisifera. Only Tenera plants were retained.

EXAMPLE 2

Practical efficacy of using flanking markers to predict shell thickness phenotype was demonstrated using progenies that had been phenotypically assessed for shell thickness. Offspring of the Bref 0104065 IB population (a selfed progeny of BLl 0323/104-05) were used for this purpose.

In this progeny there was a high level of concordance between the phenotypic scores for fruit shell thickness and the genotype assignments from the flanking markers EgSNP323 and UTREg631, with all but 4 of the 62 progeny indicating identical genotype for locus K. Results are shown in Table 2.

Table 2

Sh genotype determined by phenotype of the fruit (phenotype column) and by genotype of markers EgSNP323 and UTREg631 (Genotype heteroduplex column)

K) Ul

K)

K)

Legend: D = Dura; P = Pisifera; T = Tenera; PS = Pisifera with only unfertile flowers; PP - Pisifera with only paithenocarpic fruits; PL - Pisifera with fruit with an empty embryo sac (a hole where the kernel should have been); PK = Pisifera with fruit with a kernel but no embryo; PE = Pisifera with fruit with a kernel and with an embryo.

Phase 3. Use of shell thickness phenotype and linkage phase of flanking markers to infer recombination in historical breeding lines

The outstanding major issue to be addressed relates to the possibility of historical recombination in previous generations reversing the linkage phase of markers on one or both sides of the locus K in some parents. To illustrate:

Imagine two flanking markers A and B in positive linkage phase with Sh, so a parent Tenera palm has the following genotype:

A Sh B a sh b

A recombination (between A and sh in this instance to change a-sh-b to A-sh-b) in the ancestry of another parent palm may give a second Tenera with genotype:

A sh b a Sh B

So both palms appear heterozygous at A and B and are Teneras

Allowing no recombination between A and B in the cross between these two Tenera palms (as will happen for most of the progeny), the offspring will be:

A Sh B (Tenera) sh

A Sh B (Dura) a Sh B

sh b ( Pi sifera) sh

a_ sh _b (Tenera) a Sh B

Here, the offspring are always homozygous for one flanking marker but heterozygous for the other. This property allows us to identify these situations simply from segregation patterns although it also means that we do not know which marker is predicting phenotypic shell thickness (in this case it is B).

EXAMPLE 3

Progeny 06039054B (created from a cross between BLOl 34/89-20 and BLOl 34/89- 20) illustrate the point, with only one of twenty progeny being heterozygous at both 1 loci UTREg631b and Eg448, the remainder being homozygous at one or the other (Table 3).

Table 3

Distribution of heterozygous genotypes in progeny 06039054B for loci UTREg631b and Eg448 that flank the Sh locus

This problem is easily identified in those cases where ancestral recombination means that only one of the homologous chromosomes is affected. In such cases the Tenera parent itself will be homozygous on one side of locus K, thereby revealing the ancestral recombination event leading to phase change. To illustrate, consider the situation where a Tenera carrying an ancestral recombinant on one side of locus K (here a-sh-b recombined on one chromosome to form A-sh-b) is crossed with another Tenera carrying an ancestral recombination event on the other side of the locus (here, A-Sh-B recombined on one chromosome to form A-Sh-b). In this case then, the two Tenera parents being crossed are:

A Sh B

A sh b

X A Sh b a sh b

In this instance, both parents are easily identified as phase change recombinants by the homozygous status of one of the two flanking markers. Thus, it is only where both homologues are recombinants that problems occur because Teneras appear also to be heterozygous in flanking loci (as they do in de novo recombination)

The problem is most extreme when two Tenera parents have historical recombination events on both chromosomes and both sides of Sh:

A sh b (Tenera) a Sh B

and

A Sh b (Tenera) a sh B

Here, the offspring are: _sh_ _b Sh b

A_ _sh_ b a sh B

Sh B

A Sh b

a Sh B a sh B

Now, all predictive power is lost. The offspring that are heterozygotes on both flanks are Dura or Pisifera types whereas the true Teneras are homozygous for both flanking markers.

Both situations can be resolved by reference to the segregation of offspring arising from crosses where the linkage phase of one of the parents is known through phenotype characterisation. Linkage phase can also be detected by direct genotyping in haploids and pollen as described in Phase 4.

Single recombinants can be identified by reference (as above) to crosses with a parent known to be in linkage phase by phenotype characterisation. These single recombinants will be one of two types (recombination A-Sh or Sh-B). Any cross between these parental types and a characterised single recombinant type will either be a like-with-like single recombinant cross or a complementary recombinant cross.

These segregate in very different manners (see below) and so can be used to infer the site of recombination in the unknown single recombinant.

First, where it is a like-for-like cross: A_ _sh_ Ja a Sh B

X

A_ _sh_ _b a Sh B

A_ b A sh b

sh b a Sh B

Sh B

A sh b

a_ Sh B a Sh B

Here, half the offspring are homozygous at both ends and half are heterozygous at both ends.

Second, a complementary recombinant cross:

A sh b a Sh B

x

A sh B a Sh b

A_ _sh_ _b A sh B

A_ _sh_ _b a Sh b

Sh B A sh B

a Sh B a Sh b

Here, the heterozygosity of the flanking markers are in repulsion so that when one flanking marker is heterozygous the other is always homozygous.

In the situation where a characterised single recombinant parent is crossed with an uncharacterised double recombinant, we expect the following:

A sh b a Sh B

X

a Sh b

A sh B

A sh b a Sh b

A sh b A sh B a_ Sh B a Sh b

Sh B

A sh B

Thus, in the absence of recombination, here again the heterozygosity of flanking markers are in repulsion and offspring are always heterozygous for one flanking marker but homozygous for the other.

Therefore, by reference to the genotype of the Tenera parents, and also by the segregation behaviour of flanking markers in historical crosses involving a small number of linkage-characterised genotypes, it is possible to reconstruct linkage phase of all plants in a well-characterised breeding scheme with a complete and accurate pedigree history. Where data are incomplete or absent, direct means are required to establish linkage phase, as discussed below.

Phase 4: direct methods for determining linkage phase

Linkage phase is traditionally determined by analysing segregation ratios in progeny and comparing these with parental types. This however is time-consuming as it requires the development of large progeny populations. Two novel methods are as follows:

4.1 Linkage phase determination through haploid plant analysis Haploids plants in oil palm have recently been developed: see PCT WO2008/114145. The method there described may be used to generate a population of several hundred haploids, which are then available for linkage phase determination. Haploids carry only one copy of a chromosome as opposed to normal diploids that carry two copies of each chromosome (a pair of homologous chromosomes). An advantage haploids have in carrying just one copy of a chromosome pair is that linkage phase is more easily determined: there are no confounding effects of alternative alleles carried on the second chromosome, as is the case in normal diploid plants.

4.1.1 Linkage phase determination in Dura oil palm haploids Haploid plants are selected from crosses between Dura (Sh/Sh) x Pisifera (sh/sh) crosses in which alleles for flanking markers (as well as the Sh gene) itself are homozygous. For example:

Dura Pisifera

A Sh B X a sh b A Sh B a sh b

In such crosses the scoring of marker genotypes is unequivocal. Also, haploids obtained from these crosses are predicted to be all Dura types and will therefore be composed entirely of AB genotypes (Pisifera ab types are expected to be absent. Once maternal inheritance is confirmed for a sample of 10 - 20 haploids from unequivocal scorings we can be confident in correctly assigning all flanking marker haplotypes from Dura x Pisifera crosses. Thus a complete list of Dura haplotypes with respect to Sh flanking markers can be determined. Provided the Duras used to produce haploids are the same or related to those used in breeding the haplotype list produced will cover suitable elite germplasm.

4.1.2 Linkage phase determination in Pisifera oil palm haploids

In addition to Dura haploids, Pisifera haploids have been developed from Tenera x Tenera crosses. As in 4.1.1 maternal inheritance may be demonstrated and following on from this a list of Pisifera haplotypes may be compiled.

4.2 Linkage phase determinatin from pollen

Pollen grains provide another source of haploid material that can be analysed to determine linkage phase. Multiplex DNA analysis from single pollen grains can differentiate between the various linkage phases that cannot be determined in material heterozygous for one or more flanking markers. First, single pollen grains are placed into PCR tubes and PCR amplification is carried out using two sets of primer pairs, one for each flanking marker. Secondly, samples of the multiplex DNA amplification are aliquoted into two or more separate PCR tubes and PCR amplication is carried out, but this time for one specific marker. Thirdly, the results for each single pollen grain are compared. The scheme below provides a worst case scenario in which four Tenera palms have the same genotype (A/a, B/b), but have different linkage phases that cannot be determined only from diploid tissues:

Tenera 1

Genotype of palm Genotypes among pollen A Sh B AB or ab a sh b

Sh Ab or aB sh B

Sh ab or AB sh B

a Sh B aB or Ab A sh b

If no recombination takes place between the flanking markers the two types of pollen produced are expected to segregate 50:50. Tightly linked markers will produced few recombinants and these can be ignored in determining haplotype phase. The haplotype phase of the markers can then be related to the Sh or sh allele present by genotyping the parents of the Teneras. The same process can be applied to Dura and Pisifera palms, but here the situation is made easier by the fact these are homozygous for the shell thickness gene; Duras are homozygous Sh/Sh, Pisiferas are homozygous sh/sh.

Glossary of terms

Gene symbols are in given in italics, phenotypic symbols are in roman script, thus: Sh is the gene symbol for shell thickness, capital first letter indicates thick shell is dominant

Sh/Sh refers to the genotype of Dura, thick shelled sh/sh refers to the genotype of Pisifera, shell-less Sh/sh refers to the genotype of Tenera, thin shelled

Sh/Sh refers to the phenotype of Dura, thick shelled sh/sh refers to the phenotype of Pisifera, shell-less

Sh/sh refers to the phenotype of Tenera, thin shelled

References

Billotte N. et al. (2005) Microsatellite-based high density linkage map in oil palm (Elaeis guineensis Jacq.) Theoretical and Applied Genetics. 110, 754-765.

Charters YM, Robertson A, Wilkinson MJ, Ramsay G (1996) PCR analysis of oilseed rape cultivars {Brassica napus L. ssp. oleifera) using 5'-anchored simple sequence repeat (SSR) primers. Theoretical and Applied Genetics, 92, 442-447.

Croxford AE, Rogers T, Caligari PDS, Wilkinson MJ (2008) High-Resolution melt analysis to identify and map sequence-tagged site anchor points onto linkage maps: a white lupin (Lupinus albus) map as an exemplar. New Phytologist DOI 10.1111/J.1469- 8137.2008.02588.

Doyle JJ, Doyle JL (1987) A rapid DNA isolation procedure for small quantities of fresh leaf tissue. Phytochemical Bulletin, 19, 11-15.

Germer S, Higuchi R (1999) Single-tube genotyping without oligonucleotide probes. Genome Research 9: 72- 78.

Kosambi DD. 1944. The estimation of map distances from recombination values. Annual Eugenics 12: 172-175. Krypuy M, Newnham GM, Thomas DM, Conron M, Dobrovic A (2006) High resolution melting analysis for the rapid and sensitive detection of mutations in clinical samples: KRAS codon 12 and 13 mutations in nonsmall cell lung cancer. BMC Cancer 6: 295

Monis PT, Giglio S, Saint CP (2005) Comparison of SYTO-9 and SYBR Green I for real-time PCR and investigation of the effect of dye concentration on amplification and DNA melting curve analysis. Analytical Biochemistry 340: 24-34.

Temnykh S. DeClerck G. Lukashova A. Lipovich L. Cartinhour S. McCouch S. 2001. Computational and experimental analysis of microsatellites in rice (Oryza sativa L.): frequency, length variation, transposon associations, and genetic marker potential. Genome Research 11: 1441-1452.

Van Ooijen JW, Voorrips RE (2001) JoinMap 3.0, Software for the calculation of genetic linkage maps. Wageningen, The Netherlands: Plant Research International.

Wittwer CT, Reed GH, Gundry CN, Vandersteen JG, Pryor RJ (2003) High-resolution geno typing by amplicon melting analysis using LCGreen. Clinical Chemistry 49: 853-860.

Claims

Claims

1. A method of obtaining populations of oil palm consisting predominantly of Tenera, the method comprising: a) crossing or selfing parent Tenera oil palm plants to produce progeny; b) testing immature progeny so obtained with at least two right and left markers flanking locus K, thereby enabling the determination of inheritance patterns of the markers in tested progeny; c) selecting from the tested progeny individuals whose marker inheritance patterns indicate they are likely to be Tenera; d) growing on such selected individuals.

2. A method as claimed in claim 1 in which the resulting population is at least 98.5% Tenera, preferably 99.5% Tenera.

3. A method as claimed in claim 1 of obtaining populations of oil palm consisting predominantly of Tenera from one or two Tenera parents, the method comprising: a) selfing or crossing the parent or parents to produce progeny; b) testing immature progeny so obtained with a pair of distinct right and left differentiating markers flanking locus K in each parent, thereby enabling the determination of inheritance patterns of the markers in tested progeny; c) selecting from the tested progeny individuals whose marker inheritance patterns indicate they are likely to be Tenera; d) growing on such selected individuals.

4. A method as claimed in claim 3 in which the marker inheritance pattern of the tested progeny determined in step (c) is about 50:50 homozygous :heterozygous, and the selected individuals are those that are heterozygous.

5. A method as claimed in claim 4 in which the marker inheritance pattern of the tested crosses determined in step (c) is over 90% heterozygous for at least one marker.

6. A method as claimed in claim 5 in which the selected individuals are those that are heterozygous for a marker linked to the same allele in both parents.

7. A method as claimed in any of claims 1-6 in which both markers are within 13 centimorgans of locus K.

8. A method as claimed in any of claims 1-7 in which both markers are within 6 centimorgans of locus K.

9. A method as claimed in any of claims 1-8 in which the associations between the markers and the Sh and sh alleles are known for both parents.