A plant genotyped as AaBb could carry A and B together on one chromosome, or it could carry A and b together. Linkage phase tells you which arrangement you have. That matters when A is the marker you test and B is the allele you want to follow. In one family, A can usually point you toward B. In another, it can usually point you away from it.
Read the two chromosome copies separately
Write the first arrangement as AB / ab. A and B are in coupling. In Ab / aB, they’re in repulsion. The genotype at each locus is the same in both plants. The slash is what adds information. It separates the two copies and tells you which alleles share a chromosome. A haplotype describes that combination along one copy, which is why a list of separate genotype calls may leave the question unanswered.
Imagine crossing each plant to an ab / ab tester. If A and B are close and recombination between them is uncommon, the largest offspring classes generally reflect the combinations already present in the heterozygous parent, assuming similar survival and reliable genotyping. For AB / ab, those classes carry AB and ab from that parent. For Ab / aB, they carry Ab and aB. The tester supplies ab in both crosses, so its contribution is easy to account for.
Check the marker before you use it to discard plants
Start with the actual parents. If their genotypes establish which chromosome each allele came from, record that phase alongside the cross. If you have to infer phase from offspring, use an informative family and enough observations to see whether the expected classes are clearly separated. A small sample, missing genotypes or poor survival in one class can make a phase call less secure. If all four classes turn up at similar frequencies, the counts alone give you little basis for calling phase. Check parental genotypes or use directly phased data before relying on the marker. Where a study reports a phased panel, check whether it covers your parents and interval.
Even a correctly phased marker can be separated from the target by genetic recombination. If the decision is costly, check the trait directly when you can, or look for informative markers on both sides of the region. The useful record is specific enough to hand to someone else: which parent, which marker allele, which target allele and what evidence put the two on the same chromosome.
