Genomic data from two named Camellia sections show no clear genetic divide between them, supporting the authors’ conclusion that the groups should be treated as a cohesive genetic group. The authors also conclude that ovary pubescence, the presence or absence of fine hairs on the ovary, is not sufficient on its own to separate the sections.
The analysis covered 92 Camellia accessions. Of these, 87 belonged to the two focal sections, with 26 assigned to Eriandria and 61 to Theopsis; five accessions from other Camellia sections served as outgroups. The focal material was collected in 11 Chinese provinces between 2017 and 2024.
Young leaves from all 92 accessions were used to build super-GBS sequencing libraries, generating paired-end reads 150 base pairs long. After processing, the researchers retained 35,885 high-quality SNPs, or single-letter DNA differences, on the 15 main chromosomes and excluded unanchored scaffolds.
The genetic picture was mixed, not divided
Principal component analysis, or PCA, compresses genome-wide variation into a small number of axes. Here, the first three axes accounted for 19.4%, 15.2% and 13.8% of the variation. The five outgroups separated from the 87 focal accessions, but Eriandria and Theopsis overlapped rather than forming two distinct clusters.
ADMIXTURE, another way of estimating shared ancestry, gave its lowest cross-validation error when it modeled four ancestral populations. That model still produced no single ancestry component that separated the two sections. The genetic mixture therefore did not resolve into a clean Eriandria group and a clean Theopsis group.
Another measure, FST, summarizes how strongly groups differ genetically. The genome-wide mean was 0.013, while Jost’s D, another differentiation measure, was 0.005. The other reported diversity measures were also close between the sections, pointing to limited genetic differentiation in the sampled material.
The same pattern appeared in a rooted maximum-likelihood tree, a diagram of genetic relationships. The five outgroups sat at the base, while accessions from Eriandria and Theopsis were intermingled across four major clades instead of each section forming its own complete branch. The tree used all 92 accessions and 1,000 ultrafast bootstrap replicates.
One qualification concerns the taxonomic labels. The sample included 36 accessions with uncertain species status. When the researchers repeated the analysis using 51 taxonomically confirmed accessions, the PCA still showed overlap and ADMIXTURE still showed an admixed pattern, although mean FST rose to 0.038 from 0.013 in the full focal dataset. The exact estimate was sensitive to sample composition, but the broad pattern of limited separation remained.
Distance mattered, but only as an association
Geographic distance had a stronger association with genetic distance than elevation did. Among 58 focal accessions with native coordinates, genetic distance correlated positively with geographic distance, with a Mantel correlation of 0.2211 and a reported P value of 0.0010. The association with elevational difference was weaker, with a correlation of 0.056 and P = 0.019. These are correlations from observational data, so they do not establish that geography or elevation caused the genetic patterns.
Hairy ovaries produced leads, not answers
The study then looked for genomic signals associated with ovary pubescence. This comparison involved 28 accessions with pubescent ovaries and 64 with glabrous, or hairless, ovaries. A top 1% windowed FST scan flagged 123 genomic windows and 137 genes at an FST threshold of 0.0668. A second method, XP-CLR, flagged 100 windows and 113 genes; 45 genes appeared in both lists. Because the two scans were not statistically independent, their overlap is descriptive rather than formal cross-validation.
A separate genome-wide association study, or GWAS, tested whether individual SNPs tracked the ovary trait across the 92 accessions. No SNP reached the reported genome-wide significance threshold. Two SNPs were suggestive and corresponded to three candidate genes, but the authors cited limited statistical power at this sample size. The GWAS genomic inflation factor was 0.808.
Combining the three approaches produced 208 unique candidate genes. The researchers say these are exploratory, not validated causes. Average marker spacing was about 84 kilobases, the GWAS found no genome-wide significant signal, and ploidy, the number of chromosome sets, was not independently determined. The authors call for independent population studies and functional assays before the candidates are treated as explanations for ovary pubescence.
A taxonomic case with limits
Taken together, the authors argue that Eriandria and Theopsis should be recognized as a cohesive genetic group, and that ovary pubescence alone is insufficient for drawing a sectional boundary. The conclusion concerns the genomic relationships represented by the accessions analyzed.
The study offers preliminary targets for finding genes associated with the trait, but it does not establish an ecological function for ovary pubescence or show that any candidate gene causes it. The candidate list remains exploratory and requires independent population studies and functional validation.
Paper data and sources
Original title: Genomic evidence supports the merging of Camellia sect. Eriandria and sect. Theopsis: ovary pubescence as a trait of potential ecological relevance
Authors: Lyu M, Chen M, Fang M et al.
Journal/Repository: Frontiers in plant science
Status: Peer-reviewed
First online: 2026-10-09
DOI: Not available
Original paper · Full text