The research team drew on the Vegetation of Middle Asia Database, extracting 1,525 vegetation plots sampled between 2006 and 2019 in Tajikistan and Kyrgyzstan. Each plot contained a complete list of vascular plant species with estimates of cover, recorded on standardized scales and georeferenced to within about five metres. After excluding plots with more than fifty percent shrub cover and those not typically grazed, the dataset comprised five vegetation types: 527 steppe plots, 489 mire plots, 191 salt marsh plots, 168 tall-forb community plots and 150 pseudosteppe plots. Meadow plots were too few, at just 26 records, and were left out of the analysis.

The scale of the underlying flora is remarkable. Middle Asia hosts roughly 9,300 vascular plant species, and the Mountains of Central Asia hotspot recognized by Conservation International covers most of Tajikistan and Kyrgyzstan. Kyrgyzstan’s flora of around 4,000 species includes about 430 graminoids, of which 120 are endemic or sub-endemic. Tajikistan’s flora of roughly 4,300 species contains 438 graminoids, 70 of them endemics or sub-endemics and 133 nationally endangered. In total, the study recorded 255 graminoid species across the open habitats examined, with steppes the richest at 136 species, or 53.3 percent of the total graminoid flora, followed by pseudosteppes with 94 species.

Relative cover told a subtler story. The highest median relative cover of graminoids appeared in steppes at 47.8 percent, mires at 43.06 percent and pseudosteppes at 39 percent, but the distributions were broad, with no single sharp peak. Grasses of the family Poaceae dominated almost everywhere, mirroring the pattern for all graminoids combined. The striking exception was mires, where sedges took over, with a median relative cover of 31 percent. Rushes contributed least in every vegetation type. Salt marshes and tall-forb communities fell below the 25 percent coverage threshold often used to define grasslands, revealing that not all broadly grass-like Middle Asian habitats are actually graminoid-dominated.

To explain these patterns, the researchers tested four drivers: mean annual temperature, sum of annual precipitation, the Global Aridity Index and livestock density, the last serving as a surrogate for grazing intensity. Temperature and precipitation data came from the CHELSA dataset at 30 arc second resolution, aridity from the Global-PET geospatial dataset, and livestock numbers from the Gridded Livestock of the World database, combining sheep, goats and cattle. The team fitted linear and polynomial models for each vegetation type and variable separately, comparing them by coefficient of determination and root mean-squared error, then applied commonality analysis to split the explained variance into unique and shared contributions of each predictor.

The statistical verdict was clear: curvilinear, polynomial models consistently outperformed simple linear ones, underscoring that graminoid responses to environment are rarely straight-line. Mean annual temperature produced the single best-fitting model of the study for salt marshes, explaining 31.6 percent of variance, with 65.3 percent of the regression effect unique to that variable. Graminoid cover in salt marshes declined as temperature rose, while in pseudosteppes it increased. In steppes the temperature relationship was hump-shaped, peaking between roughly 0 and 10 degrees Celsius, with cover falling off both in the cold conditions above 3,000 metres and in the warmer lowland steppes where dwarf shrubs take hold.

Precipitation, aridity and livestock each left their own fingerprints. Annual precipitation was statistically significant for all five vegetation types, though with modest model fits between 5 and 11 percent of variance, acting negatively in pseudosteppes and tall-forb communities and positively in salt marshes and steppes. Aridity Index models fitted best for pseudosteppes and steppes, and in mires aridity showed the highest unique effect of any variable at 46.2 percent. Livestock density produced its strongest model for pseudosteppes, explaining 23.6 percent of variance, with graminoid cover peaking at around 9,000 animals per square kilometre; steppes showed a similar hump at roughly 4,000 animals, while salt marshes and mires displayed U-shaped patterns with minima near 1,500 and 2,500 animals per square kilometre respectively.

The grazing result carries particular ecological weight for pseudosteppes. These secondary grasslands arose where natural thermophilous open woodlands and scrub, including pistachio groves, were cleared and then continuously grazed, and the new data confirm their anthropogenic origin. Intensive sheep and goat grazing favours grazing-resistant species such as Poa bulbosa, Carex pachystylis and Cynodon dactylon, along with ephemeral grasses like Avena sterilis subsp. ludoviciana, Bromus lanceolatus and Vulpia persica that are dispersed by animals and tolerate drought and heat. Notably, the unique effect of livestock density in pseudosteppes was relatively small at 7.4 percent, but its shared effect with climatic factors, at 57.4 percent, was the highest among all vegetation types, showing that grazing and climate act together in shaping these communities.

By contrast, the natural steppes responded mainly to climate rather than grazing, and heavier grazing there may actually reduce graminoid share in favour of spiny, grazing-tolerant Astragalus species. Mires, fed by seeping groundwater, buffer rainfall variation and so showed weak or no clear climatic signal, although their sedge-dominated vegetation held the highest graminoid share in the driest, high-altitude Pamir settings where long-rhizomatous alpine species prevail. Tall-forb communities, dominated by Apiaceae such as Prangos pabularia and Ferula kuhistanica, resisted explanation altogether; many of their forbs are tasteless or poisonous, so grazers avoid them, suppressing grasses indirectly, and local soil and microclimatic conditions not measured here likely govern their composition.

The study’s broader message is that there is no universal rule governing graminoid abundance in Middle Asian open habitats; each vegetation type follows its own logic, rooted in biogeography, evolutionary history of grazing and local climate. That nuance matters as pressures mount. Ongoing climate change may shift graminoid shares in salt marshes, steppes and pseudosteppes, potentially threatening endemic species, while expanding human and livestock populations drive overgrazing and soil erosion across pastures. Because food and economic security in the region rest almost entirely on livestock, the authors argue that this line of research must continue, extending to alpine meadows and Kobresia mats that remain unanalysed. Understanding where and why grasses and sedges thrive is, in Middle Asia, a question of ecological science and human livelihood alike.

Subject of Research: Distribution and relative cover of graminoids in open vegetation types of Tajikistan and Kyrgyzstan under climatic and grazing gradients

Article Title: Distribution of graminoids in open habitats in Tajikistan and Kyrgyzstan

Article References: Świerszcz, S., Swacha, G., Raduła, M. W., & Nowak, S. (2022). Distribution of graminoids in open habitats in Tajikistan and Kyrgyzstan. Vegetation Classification and Survey, 3, 273-286. https://doi.org/10.3897/vcs.95767

Image Credits: AI Generated

DOI: 10.3897/VCS.95767

Keywords: graminoids, Poaceae, Cyperaceae, steppes, pseudosteppes, salt marshes, mires, grazing, aridity, Middle Asia, Tajikistan, Kyrgyzstan

The genus Imperata itself occupies an interesting position within the grasses. With roughly thirteen species distributed across tropical and subtropical regions, it is best known to agronomists and land managers through Imperata cylindrica, one of the world’s most aggressive weeds, and through I. brasiliensis, which similarly burdens agriculture and livestock production in parts of the Neotropics. The genus departs from the typical architectural plan of the tribe Andropogoneae in a way that has long intrigued agrostologists. Whereas most andropogonoid grasses bear one sessile and one pedicellate spikelet at each node of a rachis that disarticulates at maturity, Imperata produces paired spikelets that are both pedicellate, though with pedicels of unequal length, borne along a tough rachis that persists. Inflorescences are densely clothed in long silky trichomes, giving the flowering spikes their characteristic silvery appearance. These features, combined with a reduction of the androecium to one or two stamens and the absence of lodicules, align the genus with the subtribe Germainiinae, a placement supported by molecular phylogenies that recovered Imperata as sister to the Apocopis–Germainia–Pogonatherum clade. This represents a shift from the older classification, which placed the genus in the broadly circumscribed Saccharinae, and it exemplifies how molecular data have progressively refined the higher-level framework within which anatomical and morphological characters are interpreted.

Within this framework, the delimitation of species has remained stubbornly difficult. The principal problem identified by successive treatments is the overlap of morphological characters, particularly between I. brasiliensis and I. tenuis, two species that have traditionally been separated mainly by plant height and inflorescence length. Such continuous, size-based characters are notoriously unreliable in grasses, where environmental conditions, developmental stage, and phenotypic plasticity can inflate or compress measurements across populations. Molecular analyses of Neotropical Imperata have not resolved the problem; instead, they recovered limited resolution among American lineages and placed I. brasiliensis and I. tenuis close together in phylogenetic trees, with neither forming a monophyletic group. When molecular and morphological evidence both point toward a tight evolutionary relationship yet fail to provide clean species boundaries, comparative anatomy and micromorphology become especially valuable, because they can reveal characters that are independent of gross habit and inflorescence architecture and that may reflect stable, genetically anchored differences.

The methodological design of the study reflects a deliberate attempt to capture anatomical variation comprehensively rather than relying on a single tissue or region. For each of the three confirmed species, three herbarium specimens were sampled, and from each specimen one fully developed leaf was sectioned at four positions: the leaf sheath and the basal, median, and apical portions of the blade. Transverse sections at all four positions allowed the researchers to assess tissue composition and structural continuity along the leaf axis, while paradermal sections of the median blade clarified the arrangement of epidermal cell types in surface view. Standardizing descriptions on the median portion of the blade, with deviations from other regions noted when relevant, is a sound practice in grass anatomy, since the blade can vary considerably from base to apex and comparisons among species are most meaningful when made at homologous positions. The laboratory pipeline was equally conventional and rigorous: rehydration in sodium hydroxide followed by prolonged washing, dehydration through an ascending ethanol series, embedding in historesin, sectioning on a rotary microtome at thicknesses of six micrometers for transverse sections and four micrometers for paradermal and culm sections, staining with dilute toluidine blue in acetate buffer, and mounting in synthetic resin for examination under a research-grade light microscope. For micromorphology, leaf samples were rehydrated in a humid chamber for forty-eight hours to restore flexibility, mounted on aluminum stubs, sputter-coated with gold, and examined with a scanning electron microscope. These procedures, drawn from well-established protocols in plant microtechnique, ensure that the resulting descriptions are reproducible and comparable with the substantial body of grass anatomical literature accumulated over more than a century.

That literature provides the interpretive context for the new findings. Comparative leaf anatomy has been a cornerstone of grass systematics since the mid-twentieth century, when standardized descriptive frameworks were developed to make anatomical characters comparable across taxa. Under these frameworks, features such as epidermal cell shape, stomatal complex morphology, trichome types, vascular bundle arrangement, and the structure of bundle sheaths have repeatedly proven their systematic worth. Within Andropogoneae specifically, numerous studies have used these characters to delimit genera and species. For Imperata, however, the anatomical record was sparse and skewed toward a single species. Renvoize’s description of the leaf-blade anatomy of I. cylindrica recognized four orders of vascular bundles, one more than is usually reported for Poaceae, a detail that itself hints at the anatomical distinctiveness of the genus. Subsequent studies of I. cylindrica reinforced the taxonomic value of epidermal and vascular characters. In Brazil, the only prior anatomical work was a description of the rhizome and leaf-blade anatomy of I. brasiliensis published in 1971, leaving the comparative picture across all Brazilian species essentially blank until now.

Against this background, the new results are notable both for what they reveal and for what they do not. The most informative characters proved to reside in the leaf blade and, specifically, in its epidermal micromorphology. Imperata brasiliensis is distinguished by dumbbell-shaped guard cells, a feature of particular interest because guard cell shape is a classical character in grass systematics; dumbbell-shaped subsidiary cell complexes and guard cells are associated with certain chloridoid and panicoid lineages, and their presence or absence has been used to diagnose groups at levels well above the species. Finding this character diagnostic within Imperata suggests that stomatal morphology may carry phylogenetic signal relevant to understanding relationships among the American species. Imperata contracta, by contrast, is characterized by a distinctive pattern of long and short epidermal cells, the alternating arrangement of which is another staple of grass epidermal taxonomy. Imperata tenuis stands apart by the absence of cushion-hairs, a trichome type that, when present, appears as small rounded bases bearing short hairs and that has been recorded in other members of the genus. The fact that each of the three species is diagnosed by a different epidermal feature, rather than by a shared gradient of the same feature, strengthens the case that these characters reflect genuine species-level differences rather than intraspecific variation.

Equally informative is the relative uniformity of the leaf sheath and culm anatomy. In many grass groups, culm anatomy contributes useful taxonomic characters, particularly in bamboos and other woody grasses, but in herbaceous andropogonoids the internode tends to be anatomically conservative. The finding that sheath and culm sections were comparatively uniform across the Brazilian species is therefore consistent with broader patterns in the family and reinforces the conclusion that the leaf blade, as the primary photosynthetic organ exposed to the greatest range of selective pressures and displaying the greatest structural diversity, is the most rewarding tissue for taxonomic investigation in this group. This asymmetry between organs is a recurring theme in grass anatomy and helps explain why leaf-blade characters have historically dominated anatomical treatments of the family.

The practical outputs of the study extend beyond the anatomical findings. The authors provide an identification key adapted from earlier treatments and expanded to incorporate the new anatomical and micromorphological characters, updated descriptions, distribution maps generated in a geographic information system from all specimens examined, and illustrations of habit, inflorescences, spikelets, and other diagnostic features prepared by a botanical illustrator. Lectotypes were designated for I. contracta and I. tenuis, an important step in stabilizing the application of these names, since lectotypification fixes the name to a single specimen that conforms to the original protologue and is the best-preserved material from the corresponding type collection. Such typifications matter for a genus in which species boundaries have been contested, because the type specimen is the objective reference point against which all subsequent identifications are measured.

More broadly, the study contributes to a growing appreciation that species delimitation in difficult groups benefits from the accumulation of independent lines of evidence. Molecular phylogenies, gross morphology, anatomy, and micromorphology each capture different aspects of organismal divergence, and when one line of evidence is ambiguous, another may provide the discriminating signal. The case of I. brasiliensis and I. tenuis, in which molecular data place the species close together without resolving them as monophyletic, and in which traditional morphological characters overlap, is precisely the situation in which epidermal micromorphology can be decisive. The dumbbell-shaped guard cells, the distinctive long and short cell pattern, and the presence or absence of cushion-hairs are characters that can be observed with scanning electron microscopy on herbarium material, making them accessible even for rare or poorly collected species where fresh material is unavailable. This accessibility is a practical advantage for tropical floras, where herbarium collections often vastly outnumber opportunities for new fieldwork.

The exclusion of I. minutiflora from Brazil also has biogeographic implications. The species is known from other countries, and comparative specimens from outside Brazil were examined during the study, which confirms that the taxon itself is well characterized and that its absence from Brazil reflects a genuine distributional limit rather than a taxonomic artifact. Understanding where species genuinely occur is fundamental to conservation planning, weed risk assessment, and ecological research, and the correction of erroneous country-level records is a service that careful taxonomic revision provides to all downstream disciplines. Given that Imperata includes some of the world’s most troublesome grassy weeds, accurate species-level taxonomy within the genus has consequences that reach well beyond herbarium walls, informing weed management strategies, biological control programs, and the interpretation of ecological studies that depend on correct species identification.

Finally, the study demonstrates the continuing value of classical botanical methods in an era dominated by molecular approaches. The combination of field expeditions across Brazilian states, examination of type material and historical collections, light microscopy of sectioned material, and scanning electron microscopy of epidermal surfaces represents a methodological repertoire that has been refined over decades and remains indispensable. The anatomical terminology applied in the study follows the standardized frameworks established for transverse sections and epidermal surface views, ensuring that the descriptions can be compared directly with earlier work on Imperata and with the broader grass anatomical literature. As the flora of Brazil continues to be documented and revised, studies of this kind show how careful morphological and anatomical analysis, anchored to type specimens and supported by modern imaging, can resolve questions that neither molecular data nor field observation alone can settle, and how the resulting characters can be folded back into identification tools that make the taxonomy accessible to the next generation of botanists, ecologists, and land managers.

Subject of Research: Taxonomic reassessment of Brazilian Imperata grasses using leaf anatomy and micromorphology

Article Title: Anatomy and micromorphology of Imperata ( Poaceae , Andropogoneae ) from Brazil and their taxonomic significance

Article References: Pinto, B. L. S., Welker, C. A. D., & Marzinek, J. (2026). Anatomy and micromorphology of Imperata (Poaceae, Andropogoneae) from Brazil and their taxonomic significance. PhytoKeys, 279, 155-178. https://doi.org/10.3897/phytokeys.279.207098

Image Credits: AI Generated

DOI: 10.3897/phytokeys.279.207098

Keywords: Imperata, Poaceae, Andropogoneae, leaf anatomy, micromorphology, taxonomy, Brazil, scanning electron microscopy, species delimitation, grass systematics, Anatomy, taxonomic

The new species belongs to Elymus, the largest genus in the wheat tribe Triticeae, which encompasses more than 150 perennial species distributed across temperate Asia, Europe and the Americas. In most members of this tribe, including typical Elymus, the inflorescence is a simple, unbranched spike bearing one or a few laterally compressed spikelets per node of the central axis. Branching of the spike axis is exceedingly rare. That is precisely what makes E. chaobotryus so conspicuous: its compound spike, 16 to 19 centimetres long, produces 10 to 15 short lateral branches, each 2 to 4 centimetres long, from nearly all nodes of the main rachis. The branches lengthen from apex to base, creating the hierarchical, grape-cluster outline that inspired the species name, a combination of the collector’s given name and the Latinized Greek botrys, meaning a bunch of grapes.

The discovery traces back to 2019, when seeds were first collected in Guashize Township, Tongren City, in southeastern Qinghai Province. Plants raised from those seeds have been grown continuously since 2020 in a forage germplasm nursery near Qinghai Lake, at roughly 3,200 metres elevation. Across multiple years of cultivation, the diagnostic inflorescence architecture remained stable, and in 2025 replicated field plots were established in Haibei Tibetan Autonomous Prefecture to confirm that the grape-like compound spike persists in both nursery-grown and field-grown plants. This multi-year stability was critical evidence, because superficially complex-looking spikes can sometimes result from overlapping structures, spreading awns, flexuous rachises or poor specimen preservation rather than true branching.

To rule out that possibility, the team conducted an extensive herbarium survey, examining 1,321 specimens of Elymus and related taxa from Qinghai, Gansu, Xinjiang, Tibet, Inner Mongolia and Sichuan. This included 660 specimens of E. dahuricus, the species to which the new grass is phylogenetically closest, as well as hundreds of specimens of E. nutans, E. sibiricus and other regionally occurring species. None of the examined specimens possessed the hierarchical compound-spike organisation characteristic of E. chaobotryus. Several showed lax, flexuous or pendulous spikes that looked superficially branched, but in every case the inflorescence retained a single continuous main rachis and lacked lateral spike-bearing branchlets. The apparent complexity was attributable to loose spikelet arrangement, awn spreading, overlapping culms or specimen pressing.

Beyond the inflorescence, a consistent set of floral and vegetative characters separates the new species from its morphologically similar relatives. Its glumes are lanceolate, 3 to 4 millimetres long, consistently 3-veined, and tipped with very short mucros of less than 0.9 millimetres. The lemmas are comparatively short, 5 to 7 millimetres long and 3-veined, bearing awns of only 3 to 4 millimetres that are densely covered with fine bristles. The leaf blades are conspicuously narrow, just 2 to 5 millimetres wide, with glabrous sheaths and nearly smooth surfaces. The culms are erect, tufted and non-rhizomatous, reaching 80 to 130 centimetres. In E. multiramosus, the other branched-spike species from the plateau, the glumes are distinctly longer at 4 to 7 millimetres with much longer apical awns, the lemmas bear awns of 9 to 12 millimetres, and the leaf blades are broader at 5 to 7 millimetres. E. sibiricus, E. nutans and E. dahuricus var. dahuricus all bear simple, unbranched spikes and differ further in glume and lemma proportions.

Cytogenetic analysis provided an independent line of evidence. Flow cytometry of fresh leaf tissue, using the maize inbred line B73 as an internal standard, estimated a genome size of 11.43 gigabases for E. chaobotryus, closely matching the 11.32 gigabases measured for E. multiramosus and clearly exceeding the 7.10 gigabases of the tetraploid E. sibiricus. Chromosome counts from DAPI-stained root-tip cells confirmed 2n = 42 chromosomes, predominantly metacentric to submetacentric and approximately 5.0 to 8.0 micrometres long. Genomic in situ hybridization, using total genomic DNA of Pseudoroegneria spicata as a probe, identified 14 St-genome chromosomes. Sequential fluorescence in situ hybridization with Oligo-HvCSR and Oligo-pTa535 probes revealed distinct hybridization patterns among the remaining chromosomes, allowing them to be assigned to two groups of 14 each, interpreted as the H and Y genomes. Together, these results support a hexaploid complement of 2n = 6x = 42 and an StHY genomic constitution, the classic allopolyploid signature of Elymus.

Plastome sequencing added the phylogenetic dimension. The complete chloroplast genome of E. chaobotryus is 135,049 base pairs in length with an overall GC content of 38 percent, exhibiting the typical quadripartite architecture of a large single-copy region of 80,661 base pairs, a small single-copy region of 12,772 base pairs, and a pair of inverted repeats of 20,808 base pairs each. A total of 88 genes were annotated. In maximum-likelihood analyses of complete plastome sequences from 56 taxa, conducted in IQ-TREE under the GTR+F+R3 model with 1,000 bootstrap replicates, E. chaobotryus was recovered as sister to E. dahuricus, embedded within the broader Elymus complex. This placement is consistent with the species’ morphological assignment to Elymus and clarifies its maternal evolutionary history.

Particularly noteworthy is what the phylogeny reveals about the branched inflorescences themselves. Despite both possessing strongly compound spikes, E. chaobotryus and E. multiramosus are recovered in different plastome lineages rather than as a single clade. E. multiramosus forms a well-supported sister pair with E. sinosubmuticus in a separate lineage. This pattern indicates that spike branching is not necessarily associated with close maternal relatedness and may have arisen independently in different Elymus lineages. Spike branching therefore remains highly informative for morphological diagnosis but should be interpreted cautiously in phylogenetic inference, a lesson that echoes broader hypotheses in Triticeae suggesting that the unbranched spike may represent a derived condition resulting from reduction of a more highly branched ancestral inflorescence.

The discovery also highlights the plateau’s role as a centre of mountain plant diversity and the ongoing value of integrative taxonomy. The Qinghai-Tibetan Plateau and the adjacent Himalaya-Hengduan Mountains have long been recognized as major centres of diversification for genera such as Gentiana, Pedicularis, Primula and Rhododendron, and recent field work continues to reveal unrecognised St-genome grass taxa there, including E. multiramosus from western Qinghai and Roegneria yenchiana from the Hengduan Mountains. Elymus species are allopolyploids, each possessing one St haplome from Pseudoroegneria combined with additional haplomes from Hordeum, Agropyron, Australopyrum and an enigmatic Y haplome, and this reticulate evolution has long complicated their classification.

There is also a poignant ecological twist. Field surveys documented seven local populations totaling approximately 36 individuals, all discovered in the corner of the walled courtyard of a herder’s house surrounded by extensive alpine steppe. The herder family had temporarily left the homestead for seasonal pasture rotation. The enclosed courtyard may have provided a small refuge where the species could persist and reproduce, whereas in the open steppe its populations are likely subject to intense grazing pressure, which may help explain why its unusual branched inflorescences had long escaped notice. With so few known individuals and a distribution restricted to a small area of Tongren City, the species’ long-term persistence remains uncertain, and the researchers suggest that compound-spike architecture should be explicitly scored and incorporated into identification keys for Elymus to ensure that such distinctive plants are not overlooked in future surveys.

Subject of Research: Botany

Article Title: Elymus chaobotryus ( Poaceae ), a new species from the north-eastern Qinghai-Tibetan Plateau, China

Article References: Zhang, Y.-C., Xiao, S., Ma, X.-Y., Chen, J.-G., Qin, Y., Liu, Y., Guo, Z.-Y., Zhou, Y.-Q., Zhao, D.-Z., Shan, Q.-S., Lu, H.-Z., & Wei, X.-X. (2026). Elymus chaobotryus (Poaceae), a new species from the north-eastern Qinghai-Tibetan Plateau, China. PhytoKeys, 279, 215-244. https://doi.org/10.3897/phytokeys.279.181697

Image Credits: AI Generated

DOI: 10.3897/phytokeys.279.181697

Keywords: Elymus, Poaceae, Qinghai-Tibetan Plateau, new species, plant taxonomy, cytogenetics, plastome phylogeny, chaobotryus, species, north-eastern, Qinghai-Tibetan, Plateau

Codons are the three-letter DNA words that specify amino acids during protein synthesis. Because the genetic code is degenerate, most amino acids can be encoded by several synonymous codons, yet organisms rarely use these synonyms equally. This unevenness, known as codon usage bias, is far from trivial. It influences how efficiently genes are translated into proteins, how quickly chloroplasts can respond to environmental stress, and even how successfully a foreign gene can be expressed if it is engineered into a plant. Understanding which codons a species prefers, and why, therefore has both evolutionary and very practical implications.

The eight species examined span the diversity of cultivated and related millets: barnyard millet (Echinochloa esculenta), finger millet (Eleusine coracana), proso millet (Panicum miliaceum), little millet (Panicum sumatrense), kodo millet (Paspalum scrobiculatum), pearl millet (Pennisetum glaucum), foxtail millet (Setaria italica) and sorghum (Sorghum bicolor). All belong to the grass family Poaceae, and all are prized for their nutritional value and resilience to drought, making them increasingly important crops in a warming world. Yet, as the authors note, comprehensive analyses of their chloroplast genomic architecture, and of codon usage in particular, had been lacking.

To fill that gap, the team deployed a battery of standard but powerful bioinformatic measures. The relative synonymous codon usage (RSCU) quantifies how often each codon appears compared with what would be expected if all synonyms were used equally. The effective number of codons (ENC) summarizes how many distinct codons a gene actually draws upon, ranging from 20, meaning extreme bias with only one codon per amino acid, to 61, meaning no bias at all. GC content and GC3, the fraction of guanine and cytosine at the third, wobble position of codons, reveal compositional tendencies. Neutrality plots and parity rule 2 (PR2) analysis then help distinguish between two competing forces: mutational pressure, which pushes base composition in a direction dictated by the chemistry of DNA repair and replication, and natural selection, which favors codons that are translated more efficiently or accurately.

The results were strikingly consistent. Across all eight species, the chloroplast genomes showed a strong preference for codons ending in adenine or thymine. GC3 values hovered around 29 to 30 percent, a clear signature of compositional bias toward AT-richness. Thirty high-frequency codons, defined as those with an RSCU greater than 1, were identified in each genome, and every one of them ended in A or T. Some of these favored codons, such as TGT, CCT and TTA, were shared across all eight species, hinting at a common ancestral or functional logic. But many optimal codons differed from species to species, suggesting that each lineage has fine-tuned its translational machinery in its own way.

The neutrality plot analysis delivered the study’s central evolutionary verdict. If mutation were the dominant force, GC3 would track the overall GC content of genes closely, and the neutrality plot would show a strong diagonal relationship. Instead, the ENC values and neutrality results indicated that natural selection has played the decisive role in shaping codon usage bias in these chloroplast genomes, outweighing mutational pressure. This makes biological sense for an organelle whose genes encode the photosynthetic machinery on which the entire plant depends. Efficient, accurate translation of photosystem proteins, rubisco subunits and ATP synthase components confers a direct fitness advantage, particularly in the harsh, water-limited environments where millets excel.

The PR2, or parity rule 2, analysis added another layer of nuance. Under purely neutral conditions, adenine should equal thymine and guanine should equal cytosine within a DNA strand. Deviations from this parity reveal directional forces at work. In the millet chloroplast genomes, the PR2 plots showed a preference for thymine over adenine and for guanine over cytosine at the third codon position, a pattern consistent with selection acting on translational efficiency rather than with random drift. Correspondence analysis, a multivariate technique that clusters genes according to their codon usage signatures, revealed that most genes share a similar codon usage profile, but a subset of genes deviated noticeably, possibly reflecting differences in gene function, expression level or adaptation to environmental conditions.

Why does this matter beyond evolutionary theory? Chloroplasts have become a favored target for plant genetic engineering. They are maternally inherited in many crops, which reduces gene flow through pollen, they support very high levels of foreign protein expression, and they offer a natural containment strategy. But chloroplast expression depends on the translational preferences of the plastid itself. Codon optimization, rewriting a transgene to match the host’s preferred codons, has been shown in prior work to dramatically enhance expression levels and yield insights into chloroplast translation. The codon preference data assembled for these eight millets provide exactly the reference table that synthetic biologists would need when designing transgenes for millet improvement, whether the goal is stress tolerance, enhanced photosynthesis or nutritional fortification.

The study also contributes to phylogenetics. Codon usage patterns can carry historical signal, and comparative chloroplast genome analyses across the Poaceae have already been used to resolve relationships among grass lineages. By documenting shared and species-specific optimal codons, the researchers have generated a dataset that can inform both evolutionary reconstructions and the identification of conserved regulatory features in millet chloroplasts. Given that millets are increasingly promoted as climate-resilient staples for future food security, any tool that accelerates their improvement is welcome.

For a crop group often described as orphan grains, overlooked by mainstream breeding programs despite their nutritional virtues, this work represents a meaningful step into the genomic era. It confirms that the chloroplast genomes of millets are not passive repositories of sequence but actively curated libraries, their codon choices sculpted by selection to keep the photosynthetic engine running smoothly under stress. As gene editing and chloroplast transformation mature, the humble codon table may turn out to be one of the most valuable things these ancient grains can teach us.

Subject of Research: Codon usage bias in the chloroplast genomes of eight millet species

Article Title: Comprehensive Analysis of Codon Usage Bias Pattern in Chloroplast Genomes of Eight Millet Species

Article References: Nishanth, J. B., Balaji, R., Govindhan, G., & Somasundaram, S. (2026). Comprehensive Analysis of Codon Usage Bias Pattern in Chloroplast Genomes of Eight Millet Species. Indian Journal of Genetics and Plant Breeding, 86(2), 177-189. https://doi.org/10.1007/s44489-026-00018-z

Image Credits: AI Generated

DOI: 10.1007/s44489-026-00018-z

Keywords: codon usage bias, chloroplast genome, millets, RSCU, effective number of codons, GC3 content, neutrality plot, natural selection, translational selection, codon optimization, Poaceae, plant biotechnology