Can DNA Testing Confirm Plant Parentage or Genetic Relatedness? SNP Genotyping vs WGS

DNA data can support plant identity, genetic relatedness and parentage, but these are different conclusions. Choose SNP genotyping or whole-genome sequencing according to the relationship hypothesis, informative loci, candidate relatives and inheritance model; more sequence alone does not make a two-plant comparison definitive.

Key takeaways

• A matching fingerprint, a relatedness estimate and a parentage assignment answer different questions.
• Direct-parentage analysis needs inheritance evidence and consideration of alternative relationships.
• Two samples may support or exclude hypotheses without resolving every candidate or relationship direction.
• Marker informativeness, error and population representation matter more than nominal SNP count alone.
• WGS can broaden variant discovery but cannot replace missing pedigree context or a suitable analytical model.

Figure 1: Different relationship claims require different evidence, even when they begin with the same plant DNA samples.

Name the Relationship Claim

"Are these two plants related?" is a useful starting question, but not yet a complete analysis specification. A breeder may suspect a labeling error, an undocumented cross or shared ancestry. A curator may want to distinguish duplicate accessions from close relatives. Write the intended conclusion before choosing a test: matching profiles, relative genetic sharing, or support for a specified parent–offspring relationship.

This separation prevents an attractive similarity percentage from carrying more meaning than the analysis supports. State which specimens and reference accessions the comparison covers. If the question concerns a named cultivar, the provenance of the authenticated reference is part of the evidence; DNA similarity cannot create that provenance after sampling.

The DNA Fingerprinting Services page is relevant when profile differentiation or identity comparison is the objective. Keep that service context distinct from direct parentage. In Dutch heritage apples, Larsen and colleagues combined genotypic comparisons with collection information and identity review [2], illustrating the value of contextual evidence rather than a stand-alone similarity score.

What Two Samples Can Establish

Two plants are not automatically too few for useful analysis. Their genotypes may reveal substantial mismatch, strong similarity or patterns compatible with a particular close relationship. However, the strength and specificity of the conclusion depend on the species, loci, model and alternatives considered. A result supporting close genetic sharing is not the same as uniquely assigning a named parent.

Imagine an offspring and one suspected parent from an outcrossing diploid crop. Reliable incompatible loci can weigh against direct parentage, while compatible loci support the hypothesis to varying degrees. Yet an unsampled close relative of the suspected parent may also fit. Without the other parent, relevant candidate relatives or broader comparison data, the report may need to retain that alternative rather than claim certainty.

The wording should distinguish three outcomes:

• Excluded under the specified model: reliable evidence conflicts with the tested relationship after relevant error review.
• Compatible or supported: the data fit the relationship, but support is conditional on the model and alternatives.
• Unresolved: data quality, missing candidates or overlapping relationship patterns prevent the requested assignment.

Direction is another separate question. Howard, van de Weg and Luby developed approaches to orient parent–offspring duos using additional inheritance and family information in apples [1]. That is evidence against assuming that ordinary pairwise similarity provides direction. A plant's collection date or current size is also not necessarily its genetic generation, especially for clonally maintained cultivars.

Figure 2: Additional candidates and family information can narrow interpretations that remain ambiguous in a two-sample comparison.

Read Sharing and Inheritance Separately

Genetic sharing can be summarized in several ways. Identity by state describes observed allele sharing without itself establishing the ancestral route by which those alleles were inherited. A kinship or relationship estimate expresses sharing under a particular estimator and scale. A parentage likelihood compares defined hypotheses using genotype probabilities and model assumptions. These outputs should not be renamed interchangeably in a report.

Pairwise similarity and kinship

A large similarity value may reflect shared ancestry, inbreeding, clonality or a marker set with limited variation in the tested population. Relatedness estimates can help rank or characterize comparisons, but their expected values depend on the estimator, population and biological model. Name the method and its reference assumptions before applying a relationship label to any numerical output.

Xu and colleagues used SNP-derived relatedness information alongside recorded pedigrees in peach [5]. The useful lesson is that genotype-based relationships add evidence; the study's specific coefficients are not universal thresholds for a different crop. A PCA cluster or short branch in a diversity dendrogram is likewise not a direct-parentage assignment. The GBS resource for genetic diversity and germplasm characterization addresses that broader population context.

Inheritance compatibility and assignment

Under ordinary diploid inheritance, a parent and offspring should share an allele at each correctly genotyped nuclear locus. Opposite homozygotes conflict with that simple model, but the analysis must consider locus reliability, genotype error and applicable biology. Do not turn one uncertain mismatch into an automatic exclusion. Equally, do not erase inconvenient mismatches simply because the expected pedigree is familiar.

A parentage analysis asks how the observed evidence compares with alternatives, not only whether a pair shares alleles. Depending on the method, allele frequencies, genotyping error and the proportion of possible parents sampled influence the interpretation. Confidence must be defined within that analysis, not borrowed from another crop's validation study or inferred from the number of SNPs printed on an assay label.

Check the interpretation layer explicitly:

• observed allele sharing is not automatically shared inheritance;
• estimator values need their own scale and assumptions;
• parentage support is conditional on tested alternatives.

Count Informative Loci, Not Labels

The question "How many SNPs are enough?" has no universal answer across plants and relationship tasks. A marker can be technically reliable yet contribute little discrimination in a highly uniform breeding group. Thousands of tightly linked loci do not provide thousands of independent observations. A useful design considers variation among relevant candidates, genomic distribution, assay reliability and the specific alternatives that need separating.

Screen the proposed marker set against representative germplasm. Include likely parents and close relatives where available, not only unrelated accessions that are easy to distinguish. Review call rate and replicate agreement by sample and locus, and inspect whether dropout occurs preferentially in one genetic group. Such missingness can weaken both comparison and exclusion evidence.

Marker review should document:

• the loci and genomic coordinates used, including the reference build;
• observed variation in the relevant population or candidate set;
• locus quality, sample missingness and independent replicate evidence;
• linkage treatment and the assumptions behind combined evidence;
• any excluded loci and the reasons for exclusion;
• performance against the closest plausible alternative relationships.

The Custom SNP Microarrays page offers a context for discussing marker-content requirements. It does not establish that any requested plant relationship can be resolved by a standard panel. Agree the species, germplasm representation and interpretation objective before treating a custom genotype resource as a parentage assay.

Choose SNP Genotyping or WGS

SNP genotyping is a broad category: fixed arrays and targeted sequencing can both produce genotypes at selected loci. WGS instead sequences across the genome and can support discovery of additional variants, subject to coverage, reference suitability and calling quality. The decision is not simply "few markers versus all DNA." Both routes ultimately need trustworthy, relevant genotype evidence and an appropriate relationship analysis.

A selected SNP set is sensible when it is informative for the submitted material, reproducible and compatible with relevant references. It may avoid collecting data that do not affect the requested decision. WGS becomes useful when marker discovery, broader variation or later research reuse is also needed. These are scope decisions, not promises of a universally cheaper or more conclusive method.

For Whole Genome Resequencing, discuss the reference assembly, callable regions, genotype confidence and intended analysis. A high nominal SNP count can include unreliable calls or redundant information. In low-depth data, missing calls and imputation require particular caution: a reference-based inference is not automatically suitable for a decisive inheritance-exclusion claim, especially when the samples are poorly represented in that reference.

Figure 3: The relevant comparison is usable genotype evidence, not nominal genome coverage or marker count alone.

The existing LC-WGS, GBS and SNP-array selection resource addresses broader breeding-platform choices. Parentage adds a narrower requirement: the selected evidence must distinguish the hypothesized relationship from plausible alternatives. A platform suitable for exploratory structure analysis is not thereby validated for every direct-parentage conclusion.

Add Candidates and Population Context

Candidate parents and reference populations serve different purposes. Candidate samples allow direct comparisons with plausible contributors to an offspring. Population information can support allele-frequency estimates or place sharing in a broader genetic context. Some estimators use different reference assumptions, so specify which information the planned method actually needs rather than saying every analysis requires the same reference panel.

When possible, include both suspected parents, other plausible parents, known relatives and relevant comparison accessions. Record whether the candidate list is intended to be exhaustive or merely convenient. The strongest-scoring sampled candidate may not be the true parent if the true contributor was never included. Close relatives of a candidate deserve attention because they can be more difficult alternatives than unrelated plants.

For collection-scale work, the Kin Recognition page provides a route for discussing relatedness objectives and marker options. Its current description is SSR-based, so do not assume it promises a standardized SNP parentage pipeline. State the requested genotyping and statistical scope explicitly, including which relationships must remain unresolved if the available material is insufficient.

Keep three inputs distinct in the project brief:

• candidate parents and their close relatives;
• population information used by the selected estimator;
• known-family controls used to assess relationship interpretation.

Account for Plant Reproductive Biology

Ordinary outcrossing diploid expectations cannot simply be copied into every crop. Selfing and inbreeding alter genetic-sharing patterns, while vegetative propagation can preserve highly similar genotypes across different specimen histories. Apomictic reproduction raises another interpretive question. The analysis must distinguish the reproductive model being tested from a convenient pedigree label.

Polyploid material adds allele-dosage and inheritance-model requirements. Howard and colleagues' triploid apple study used ploidy-aware SNP interpretation to investigate pedigrees [4]. It shows why forcing diploid genotype calls onto different ploidies can mislead relationship analysis; it does not provide a rule that all triploids in all species have the same reproductive behavior.

Van Rossum's plant-translocation study distinguishes parentage, sibship and selfing in an ecological setting [3]. It supports treating mating context as part of the question, not as a detail to add after genotyping. Do not present this work as validation of a CD Genomics SNP panel or transfer its population-specific outcomes into a service guarantee.

Protect Identity Before Interpretation

A technically sound relationship estimate can still answer the wrong question if labels, tissues or accession references are unreliable. Use permanent specimen identifiers and record source, collection date, propagation history and any grafted material. Sampling a rootstock instead of a scion changes which genotype is tested. Mixed tissue or pooled individuals also changes the meaning of a supposedly pairwise comparison.

Retain enough material for a repeat where practical and preserve the link between specimen and DNA extraction. Record failed or replaced samples without overwriting their histories. For seed-production questions, the separate maize line-purity and parent-verification resource covers the relevant QC context. A relationship result for two individual plants does not establish purity throughout a seed lot.

Keep provenance corrections separate from analytical conclusions. If an accession label changes during review, retain the original label, the replacement, the evidence and the person authorizing the correction. Otherwise a revised name can make the final pedigree appear to confirm a hypothesis that was never actually tested. The same principle applies when replacement DNA is supplied after a failed extraction.

Before interpreting the genotype comparison, check:

• the sampled plant part and whether grafting or mixed material is relevant;
• consistency of specimen, extract and reference-accession identifiers;
• whether a technical repeat confirms a questionable genotype pattern;
• any changes to the expected pedigree made after results were seen.

Request a Bounded Relationship Report

Agree the report wording before ordering. A useful package separates observed genotypes, technical QC, relationship estimates and the final inference. It should identify the tested model and candidates, show exclusions and unresolved alternatives, and allow another researcher to understand why a relationship label was assigned or withheld. A pedigree graphic alone is insufficient.

Request these components where relevant:

• specimen and DNA manifest with accession provenance;
• locus manifest, reference build and genotype or dosage files;
• sample and locus QC, repeat evidence and exclusion records;
• named sharing estimator and parentage model with parameters;
• candidate comparison and inheritance-mismatch evidence;
• conditional conclusion, alternatives and additional samples that could help.

Figure 4: A defensible relationship report preserves both its supporting evidence and its interpretation limits.

Send CD Genomics the species, known ploidy, suspected relationship, number of specimens, candidate parents or relatives, sample format and available reference data. Describe whether the goal is fingerprint matching, pedigree review or broader relatedness research. Agree feasibility and analytical outputs before production; sequencing capability alone is not a promise that the proposed relationship can be assigned.

Plant Relationship Analysis FAQ

References

1. Howard NP, van de Weg E, Luby JJ. A new method to reconstruct the direction of parent-offspring duo relationships using SNP array data and its demonstration on ancient and modern cultivars in the outcrossing species Malus × domestica. Horticulture Research. 2022;9:uhab069. doi:10.1093/hr/uhab069.
2. Larsen B, van Dooijeweert W, Durel C-E, et al. SNP genotyping Dutch heritage apple cultivars allows for germplasm characterization, curation, and pedigree reconstruction using genotypic data from multiple collection sites across the world. Tree Genetics & Genomes. 2024;20(4):21. doi:10.1007/s11295-024-01655-9.
3. Van Rossum F. Sibship and parentage reconstruction as genetic tool for designing and monitoring plant translocations. Restoration Ecology. 2023;31(3):e13726. doi:10.1111/rec.13726.
4. Howard NP, Micheletti D, Luby JJ, et al. Pedigree reconstruction for triploid apple cultivars using single nucleotide polymorphism array data. Plants, People, Planet. 2023;5(1):98–111. doi:10.1002/ppp3.10313.
5. Xu Y, Yu Y, Qi X, et al. The PeachSNP170K array facilitates insights into a large-scale population relatedness and genetic impacts on citrate content and flowering time. Communications Biology. 2025;8(1):854. doi:10.1038/s42003-025-08144-2.

This article and the described services are intended for agricultural and biological research. They do not provide clinical diagnosis, treatment decisions or individual health assessment.