Two types of markers
For the study, Dorfling went back to the source, obtaining samples from nucleus plants. These plants represent the original trees from which all plant material of any given cultivar was propagated in SA.
Twenty rootstock cultivars, including M.9 T337, M.7 EMLA, M.793, MM.109, G.757, G.213, G.202, G.222, G.890, G.228 and G.778, as well as two scion cultivars (Granny Smith and Golden Delicious), formed part of the project.
Prins and her team at CenGen developed two fingerprints for each cultivar, one based on a panel of 14 SSR markers and the other based on a panel of 16 SNP markers. Markers are bits of DNA that scientists use to distinguish between individuals.
The researchers were able to reliably identify the different rootstock and scion cultivars using either marker set, at an estimated cost of R50 per sample for SNP testing (provided a large number of samples are submitted), compared with more than R600 for SSR testing (based on 2024 prices).
“SSRs have been the gold standard for genetic fingerprinting, even in human forensics,” says Prins. “They are very reproducible, but quite expensive.”
Tests based on SSR markers are more expensive because they use more reagents, and interpreting the results is knowledge-intensive and time-consuming. The tests based on SNP markers, on the other hand, use smaller reagent volumes and less staff time. However, there’s a place for both technologies.
“SNP testing is designed for high throughput. So, large sample numbers,” explains Prins. “It doesn’t make sense to test small numbers of samples with that system.”
The pitfalls of pooling
The CenGen team also investigated two other cost-saving options. The first was pooling samples from 10 trees, with the idea of identifying and discarding any group of 10 trees that contained one or more incorrectly identified rootstock cultivars.
“Based on what we’ve seen in the lab, I’m against pooling,” says Prins. In her experience, pooled samples are less sensitive, especially when the individuals in the pool are closely related. This could result in some misidentified rootstocks slipping through.
Nonetheless, the project included tests of pooled samples with both SSR and SNP marker sets. The SSR method can be used to test pooled samples, provided the full panel of 14 markers is employed. However, as a cost-saving measure, it’s suboptimal compared to applying the SNP method to individual trees.
When using the SNP method, it wasn’t possible to reliably detect a mixture of rootstocks, especially if only one misidentified rootstock was present in the pool. The second cost-saving option the team evaluated was alternative sample preparation methods. Before the trees can be fingerprinted, their DNA must be extracted. This process can be expensive and time-consuming, but the cheaper methods tend to yield lower-quality DNA.
Prins and her team assessed two DNA extraction methods, CTAB and SDS. The less expensive SDS method proved adequate, so it can help reduce costs for large sample numbers. However, DNA extracted with the SDS method doesn’t store well. If long-term DNA preservation is required, it’s best to use the CTAB method.
Testing in practice
One significant benefit of the project is the creation of a fingerprint database, which allows all samples to be tested against verified material.
“With the previous service, you had to provide a control leaf,” says Dorfling. “So, if you suspected that your rootstock was the wrong one, you had to identify which rootstocks it might be and then go source control material.”
Growers might still be wondering where they’re going to get leaves of their rootstocks, unless they are lucky enough to find suckers. One potential solution is to sacrifice a tree or two by cutting them below the graft union and sampling the resulting regrowth.
If all else fails, testing bark samples is possible, but sufficiently challenging that it doesn’t lend itself to large sample numbers.
Thanks to this Hortgro-funded collaboration between Provar and CenGen, growers and researchers now have two options for confirming the identity of apple rootstocks. If they want to test fewer than 22 trees, the SSR marker set remains their best option. If they want to test more than this, they have the high-throughput SNP option.
“The industry now has access to both marker databases. So, we’ve created technology that’s suited to different needs,” says Prins. “I think that was the value of this project.”
BOX
Making sense of markers The genome of apple trees consist of ±740 million base pairs divided into 17 chromosomes. Think of this as similar to 740 million letters divided into 17 books. The sequence of the letters is mostly the same for all apple trees, but there are small differences, called polymorphisms, between cultivars.
In any genome, some stretches of DNA code for proteins, and some don’t seem to do anything. Changes in coding DNA are frequently lethal, whereas mutations in the non-coding DNA often have no effect. Therefore, the non-coding DNA is generally a good place to find variability.
Much of the non-coding DNA consists of sequences of two to seven base pairs that are repeated about five to 50 times (simple sequence repeats or SSRs). The number of repeats varies between individuals. So, an apple tree may have a sequence consisting of the bases ACA repeated up to 50 times. But in one cultivar, the sequence is always repeated 13 times, and in another cultivar, it’s always repeated 27 times.
The DNA of cultivars can also differ by a single base at a given position (single-nucleotide polymorphisms, or SNPs). This is a little like English versus American spelling: think of the British “minimise” versus the American “minimize”.
Cultivars inherit SSRs and SNPs from their parents. Because some cultivars are the parents of many other cultivars, certain polymorphisms occur in several cultivars. This is why genetic fingerprinting relies on panels of several SSRs or SNPs.
Besides helping identify cultivars, some genetic markers are associated with specific genes or gene variants. Breeders can use these markers to screen parents or offspring for the presence of desirable or undesirable gene variants.
Marker-assisted selection has been game-changing for many plant breeders. In the past, they had to base their selections on observable characteristics, which are not always visible, especially if a plant has only one copy of a gene. Gene expression can also take many years, for example, when a breeder must wait for a tree grown from seed to bear fruit. After recently completing a Hortgro-funded fingerprinting project for apple rootstocks, Provar and CenGen are now collaborating on two new Hortgro-funded projects that also exploit DNA markers. Read more about those on page 63.
