Asparagus is a morphologically and ecologically diverse genus with economic and medicinal importance, yet its chloroplast genome had received limited genus-wide attention compared with recent nuclear genomic efforts. The genus belongs to the family Asparagaceae and comprises well over two hundred species distributed across Europe, Africa, and Asia, ranging from the familiar cultivated vegetable with its tender young shoots to prickly desert shrubs, climbing plants, and ornamentals prized for their delicate feathery foliage. Several species are also staples of traditional medicine, most notably Asparagus racemosus, whose roots are used in Ayurvedic practice, and a number of Chinese species employed in herbal formulations. This combination of economic value, taxonomic difficulty, and ecological breadth makes the genus an attractive target for genomic characterization, and it also makes reliable species identification a genuine practical problem, since many Asparagus species look superficially similar and are easily confused in trade.

To close the genomic gap, the team de novo assembled 35 chloroplast genomes: one from newly generated sequencing data and 34 from publicly available sequencing reads downloaded from the National Center for Biotechnology Information. These were integrated with 40 previously published genomes, bringing the total dataset to 75 species. By combining freshly assembled material with existing resources, the study achieved a sampling breadth that no earlier chloroplast-focused work on the genus had reached. The strategy of mining archived sequencing reads is increasingly common in organelle genomics, because chloroplast genomes are present in many copies per cell and can often be recovered reliably even from nuclear-targeted sequencing projects, allowing researchers to extract organellar value from data originally generated for other purposes.

The most fundamental finding concerns genome structure. Every one of the 75 genomes retained the canonical quadripartite architecture typical of flowering plant chloroplasts, consisting of a pair of inverted repeats separated by a large and a small single-copy region. Genome size varied only narrowly, from 155,654 to 157,211 base pairs, and gene content was essentially invariant, with 131 genes representing 113 unique gene species. The boundaries of the inverted repeats were likewise conserved across the genus. This architecture, in which the inverted repeats act as a stabilizing element by buffering the single-copy regions against rearrangement, is broadly shared across angiosperms, but the degree of uniformity observed here is nonetheless notable given the genus’s diversity in habit, morphology, and ecology, from the succulent-cladode ornamentals to the cultivated vegetable. Chloroplast genomes are well known for their structural conservatism relative to nuclear genomes, and the Asparagus data provide a textbook illustration of that stasis.

While the architecture was static, the genes themselves were not all evolving under the same regime. Using codon-based selection analyses, the researchers identified episodic diversifying selection in four loci: ycf1, ccsA, rps8, and rbcL. The strongest signals came from ycf1 and rbcL, and both were independently corroborated by PAML site-model analysis, a complementary maximum-likelihood approach for detecting codon sites evolving under positive selection. The picture was more nuanced for the other two genes. The signal in ccsA, which is involved in cytochrome c biogenesis of the chloroplast respiratory chain, was restricted to just two codon sites but was recovered by PAML branch-site analysis. The gene-wide signal in rps8, a small ribosomal protein gene, was not recovered at individual sites and was also not supported by the branch-site approach, suggesting it should be interpreted with caution.

The functional implications of these selected genes are intriguing. rbcL encodes the large subunit of RuBisCO, the enzyme responsible for carbon fixation in photosynthesis and often cited as the most abundant protein on Earth, while ycf1 is a large, rapidly evolving gene of unknown but essential function whose product participates in the chloroplast protein translocation machinery. Because RuBisCO sits at the very entry point of carbon into the biosphere, even modest changes in its catalytic properties can matter for photosynthetic efficiency, and adaptive evolution at rbcL has been documented in a range of plant lineages, often in connection with shifts in climate, atmosphere, or photosynthetic strategy. That both rbcL and ycf1 show the clearest evidence of episodic diversifying selection across Asparagus species hints that adaptation, possibly related to photosynthetic performance in different environments, has left measurable footprints in the chloroplast genome even as its overall organization has remained frozen in place. The coexistence of a rigid genome scaffold with a few genes under shifting selective pressure is a recurring theme in organelle evolution, and the Asparagus dataset adds a well-sampled genus-level example.

Beyond selection, the study mapped the distribution of genetic variation across the chloroplast genome, a question of practical importance for DNA barcoding and species identification. Nucleotide diversity was substantially higher in non-coding regions than in coding regions, consistent with the general pattern in plant chloroplast genomes where intergenic spacers accumulate mutations more freely because they are not constrained by the need to preserve an amino acid sequence. Among coding loci, rps15 and ycf1 stood out as the most variable, while among intergenic spacers the psbC-trnS-UGA, ccsA-ndhD, and rps3-rpl22 regions showed the highest diversity. The authors highlight these six loci as candidate species-identification markers for Asparagus, offering a starting point for researchers seeking to distinguish closely related species, including those traded as medicinal herbs, using short chloroplast DNA segments. Chloroplast barcoding has long been pursued in plants, with mixed success, because many plant groups radiated too recently for standard barcode loci to accumulate diagnostic differences; identifying the most variable regions within a target genus is therefore a standard and sensible strategy for improving discrimination.

Phylogenetic reconstruction from the chloroplast data delivered a largely well-resolved species tree. Maximum-likelihood analysis based on complete chloroplast genomes recovered twelve previously recognized infrageneric clades, with strong support for relationships within each clade and a well-supported backbone. Notably, the tree built from coding sequences alone showed weaker backbone support than the whole-genome tree, an observation that underscores the phylogenetic value of non-coding spacers even though they are individually more variable. Collectively, the spacers contribute a large share of the informative sites in a complete chloroplast genome, and their combined signal can outweigh the noise introduced at any single locus. The recovery of all twelve recognized clades confirms that the chloroplast genome, sampled densely enough, can capture the genus’s internal diversification signal.

The most consequential result, however, is the comparison with nuclear data. The team quantitatively compared their chloroplast tree with a published nuclear species tree for the 59 species shared between the two datasets, using the Robinson–Foulds distance as a measure of topological disagreement. The unnormalized distance was 72, and the normalized Robinson–Foulds value was 0.64, indicating substantial discordance concentrated at the backbone nodes of the tree. In other words, while the chloroplast and nuclear genomes agree on many fine-scale relationships within clades, they tell partly conflicting stories about the deepest splits in the genus’s history.

Cytonuclear discordance of this kind is a well-recognized phenomenon in plant phylogenomics, and the Asparagus result illustrates why it matters. Chloroplast genomes are maternally inherited in most flowering plants and trace a single genealogical lineage, whereas nuclear genomes combine ancestry from both parents and are reshuffled by recombination. Processes such as incomplete lineage sorting, in which ancestral genetic polymorphisms persist through rapid speciation events, hybridization followed by chloroplast capture, or introgression between lineages can each drive the organellar and nuclear trees apart. Rapid ancient radiations, which are precisely the kind of event that tends to generate poorly resolved deep backbones, are especially prone to such discordance because there is little time for lineages to sort cleanly between speciation events. The authors do not adjudicate among these mechanisms in the abstract, but the magnitude of the normalized discordance they report makes clear that a chloroplast-only phylogeny of Asparagus cannot be treated as a complete account of the genus’s evolutionary history.

The study has limitations worth noting. Thirty-four of the 35 newly assembled genomes came from publicly archived reads rather than fresh sampling, so assembly quality depends on the underlying data, and the comparison with the nuclear tree was restricted to the 59 species common to both datasets. The selection signals in ccsA and especially rps8 rest on limited statistical support, with the ccsA signal confined to two codon sites and the rps8 signal not recovered at individual sites or by branch-site models. And while the identification of variable loci as barcoding candidates is a valuable practical output, validating their discriminatory power across the full breadth of the genus, and across populations within species, will require further testing.

Even so, the significance of the work is considerable. By establishing the first genus-wide chloroplast genome resource for Asparagus, the authors have created a reference framework that others can extend as new species are sequenced. The demonstration that structural conservation and gene-specific selection coexist within the same compact genome adds to a growing body of evidence that organelle evolution is not uniformly neutral. And the quantified chloroplast–nuclear discordance provides a concrete empirical baseline for future phylogenomic studies that integrate organellar and nuclear data, an approach the authors explicitly frame as the way forward.

For applied research, the implications reach into conservation genetics and the authentication of medicinal plant materials. Asparagus species are used in traditional medicine systems, and the study was supported by a Chinese government program dedicated to the sustainable use of valuable Chinese medicine resources. Reliable species-level identification is a prerequisite for both conserving wild genetic diversity and ensuring that herbal products contain the correct species, a concern that has grown alongside the expansion of the global herbal trade and the documented problems of adulteration and substitution in commercial plant materials. The variable loci identified here, particularly the high-diversity intergenic spacers, offer practical molecular tools toward those ends, while the genus-wide genome dataset provides the comparative context needed to interpret any single species’s chloroplast sequence in its proper evolutionary setting.