1. Introduction

Allium L. is one of the largest and most diversified genera in the petaloid-monocotyledonous plant family Amaryllidaceae, currently comprising ca. 1018 accepted species worldwide (Friesen et al., 2021, 2024; Yusupov et al., 2022; Munavvarov et al., 2023). Current species diversity within the genus Allium is increasing because of new species discoveries (Khassanov et al., 2023; Balos et al., 2023; Eker, 2024). Subgenus Melanocrommyum (Webb & Berth.) Rouy is considered to be the second largest subgenus in the genus (Friesen et al., 2006, 2021; Gurushidze et al., 2008, 2010; Fritsch, 2012; Fritsch, 2016). The subgenus contains more than 180 species and subspecies assigned into 21 currently recognized sections, making it one of the most taxonomically diversified subgenera of Allium (Friesen et al., 2021; Yusupov et al. 2021, 2022; Munavvarov et al., 2022).

The classification of subgenus Melanocrommyum has traditionally presented difficulties because earlier divisions relied almost entirely on morphological characteristics. Webb and Berthelot (1844) were the first to investigate the distinctive features of A llium nigrum s.l. and its close relatives from the Canary Islands and hence defined section Melanocrommyum s.s. Koch also investigated A. caspium s.l. and proposed to place it in a new section Kaloprason C. Koch. However, Regel (1875) did not recognize these sections but placed these species into the section Molium s.l., and this treatment was followed by many botanists for more than a century. By the mid-20th century, it became evident that section Molium consisted of two, quite distantly related subgroups with different basic chromosome numbers (x = 7 and x = 8) and other distinguishing morpho-anatomical characteristics (REF). Wendelbo (1969) raised section Melanocrommyum to the subgeneric rank and presented a detailed organizational framework at the sectional level, shedding more light on the abundance of the Old World species and their wide range of morphological variations. Subsequent contributions from Kamelin (1973), Hanelt, 1992, Khassanov and Fritsch (1994), and Seisums (1994) brought depth to the taxonomic resolution of subgenus Melanocrommyum, despite varied perspectives and acceptance of the sections. The most recent classification based on broad sampling was proposed by Fritsch et al. (2010), and a new section (Tulipifolia) was suggested by Friesen et al. (2021) (see Appendix 1). The taxonomic diversity stretches across its large geographical distribution, which ranges from the Mediterranean region and Near and Middle East, north-western China and Pakistan to southern Siberia, with South-western and Central Asia as the primary center of diversity (Gurushidze et al., 2008; Fritsch, 2012).

Molecular phylogenetics has provided us with completely new insights into the phylogenetic relationships within the genus (Friesen et al., 2006; Li et al., 2010). Previous studies conducted by Gurushidze et al. (2008, 2010) based on the Internal Transcribed Spacer region (ITS) of Nuclear Ribosomal DNA (nrDNA) and a small number of chloroplast regions (trnL-trnF, matK) determined that the subgenus is monophyletic, but some sections were paraphyletic or polyphyletic, indicating incongruence between traditional morphological classifications and molecular genetic data (Gurushidze et al., 2008, 2010). High-throughput sequencing technology has the potential to generate genomic data exhibiting substantial sequence variation among closely related congeneric species (Ruhsam et al., 2015; Huang et al., 2016; Liu et al., 2021). Chloroplast genome (plastome) sequencing is an accessible method for resolving complicated species groups (Huang et al., 2016; Joyce et al., 2023; Larson et al., 2023; Masters et al., 2023). Here, we use this approach to re-examine the species within the subgenus and amend its classification. Dating analyses based on nrDNA data from section Decipientia (Friesen et al., 1997) showed that the subgenus originated in the Late Oligocene. However, this study did not include analysis based on cpDNA data and did not investigate the biogeographic routes of the subgenus during its diversification periods.

The aims of our study are (1) to provide a detailed analysis of phylogenetic relationships using chloroplast genome sequences, (2) to perform analyses on divergence time and biogeography to better understand the origin and dispersal of species represents the historical biogeography of the subgenus Melanocrommyum and (3) to identify the primitive morphological characters within the subgenus. To achieve these goals, we generated and analyzed the plastome sequences of 117 accessions representing 107 taxa across 19 sections of the subgenus. Based on these data, we provide new insights into the molecular phylogeny, evolutionary history and biogeography of Allium subgenus Melanocrommyum.

2. Materials and methods 2.1. Taxon sampling and DNA extraction

A total of 117 accessions sampled in this study represent 107 taxa belonging to 19 currently recognized sections of subgenus Melanocrommyum. Six species were also included in this study as outgroup taxa (115 taxa were newly sequenced). The samples used included both collections from recent field trips and herbarium samples. Fresh leaves were collected in natural habitats. Plant voucher specimens were deposited in AIBU, FRU, IBRC, KUN and TASH (Fig. 1). Herbarium materials were taken from the National Herbarium of Uzbekistan, Institute of Botany of the Uzbekistan Academy of Sciences (TASH); the Kunming Institute of Botany, Chinese Academy of Sciences (KUN); Iranian Biological Resource Center Herbarium (IBRC), Guilan University Herbarium, Iran (GUH), Tehran University Herbarium, Iran (TUH), Gatersleben Herbarium of the Genebank Department of the Leibniz Institute of Plant Genetics and Crop Plant Research, Germany (GAT), the collections of the Institute of Biology and Soil Science, Kyrgyzstan (FRU), Abant Izzet Baysal University Herbarium, Turkey (AIBU), and the herbarium of the Ferdowsi University of Mashhad (FUMH). All voucher specimens and GenBank accession numbers are provided in Table S1. Genomic DNA was extracted from leaf tissues with the CTAB (cetyltrimethylammonium bromide) method for high-throughput sequencing (Doyle 1991).

2.2. Illumina sequencing and chloroplast genome assembly and annotation

Genomic DNA was utilized to create a library of 350-bp sequences using the Genomic DNA Sample Prep Kit (Illumina) following the manufacturer's instructions. DNA was subsequently sequenced with 150 paired-end reads on the Illumina NovaSeq X Plus at Beijing Novogene Bioinformatics Technology Co., Ltd., in Beijing, China. We employed the NGS QC Tool Kit with default settings for raw data processing, followed by assembly of clean reads using GetOrganelle (Jin et al., 2020) with optimized parameters "w 0.5 -o -R 30 -t 8 -k 75, 95, 115, 127" (Patel and Jain, 2012). Geneious v.10.0.2 was then used for gene annotation; Allium karataviense (accession number: ncbi-n:NC_057573) was set as the reference genome. The verifying and checking of start, stop codons and intron/exon boundaries for protein-coding genes was conducted manually (Kearse et al., 2012). All annotated cp genome sequences were deposited in the NCBI database (Table S1). The nrDNA sequencing reads were assembled utilizing GetOrganelle v.1.7.4.1 (Jin et al., 2020). Partial nrDNA sequences consisted of the internal transcribed spacer 1, 5.8S ribosomal RNA gene, and the internal transcribed spacer 2 (Table S1).

2.3. Phylogenetic analyses

Phylogenetic relationships of subgenus Melanocrommyum were inferred from complete chloroplast genome and nrDNA sequence data. Phylogenetic analyses were conducted using Maximum Parsimony (MP), Bayesian inference (BI), and maximum likelihood (ML) methods. The plastome dataset for phylogenetic analysis included newly sequenced cp genomes of 115 taxa and 2 ingroup plus 4 outgroup species (A llium macranthum, A. monanthum, A. paradoxum, A. ursinum) from the National Center for Biotechnology Information (NCBI) (Table S1). For some species, our attempts to assemble nrDNA were unsuccessful; in these cases, we downloaded relevant data from NCBI (Table S2). We used MAFFT to align the sequences (Katoh and Standley, 2013). Poorly aligned regions were excluded from the total alignment using trimAI v.1.2 (Capella-Gutiérrez et al., 2009) with the following parameters: -gt: 0.9; -cons: 0.7; -seqoverlap: 0.6; -resoverlap: 0.8. The plastome phylogenetic tree was constructed based on complete chloroplast genome sequences. PAUP* v.4.10 (Swofford, 2002) was utilized to conduct Maximum Parsimony analysis (MP). All characters were given the same weight, missing gaps were ignored, and character states were considered without any specific order. A heuristic search with TBR branch swapping and the Multrees option was conducted, along with random stepwise addition using 1000 replications. Analyses all utilized the best-fitting models of nucleotide substitutions chosen in jModelTest v.2.1.4 (Darriba et al., 2012) based on the Akaike information criterion (AIC). RAxML v.8.0 (Stamatakis, 2014) was utilized to perform Maximum Likelihood (ML) analyses with the optimal GTR + G model and 1000 bootstrap replicates. Bayesian inference (BI) analyses involved running MrBayes v.3.2 (Ronquist et al., 2012) for 5, 000, 000 generations using ten incrementally heated chains and selecting random trees every 1000 generations. The first 1000 trees were discarded as burn-in. State frequencies were set to fixed, Jukes (Cantor (JC) and among-site rate variation was set to equal. iTOL v.4.2.3 (https://itol.embl.de/) and FigTree v.1.4.2 (Rambaut, 2014) were used to visualize the phylogeny.

2.4. Divergence time estimates

A chronogram of subgenus Melanocrommyum was estimated based on both plastome and ITS sequence data using BEAST 2 v.2.6.6 (Bouckaert et al., 2019). XML command files were generated in BEAUti 2 v.2.6.6 (Bouckaert et al., 2019). The strategies for the analysis were as follows: a relaxed uncorrelated lognormal distribution clock; a Yule process tree prior; and a random starting tree. We carried out 150 million generations of Markov chain Monte Carlo (MCMC) runs, saving samples every 1000 generations. The software TRACER v.1.7.2 (Rambaut et al., 2018) was utilized in assessment of the MCMC samples to check for proper sampling or simulations and whether the effective sample size for all relevant parameters was well above 200. An MCC tree was constructed with TreeAnnotator v.2.6.6 (Bouckaert et al., 2019) that included mean and 95% HPD node ages obtained from a likelihood estimation with 10% of trees burned in and no restriction set for the posterior probability. For the plastome data, as the secondary calibration point, we took the mean crown clade age of the subgenus Microscordum of 15.888 Ma, and standard deviation of 1.0 as reported by Xie et al. (2020). For the ITS data, as the secondary calibration point, we used the mean crown age of the subgenus Microscordum, which is 33.89 Ma (Friesen et al., 2024). The final consensus tree was visualized using FigTree 1.3.1 (Rambaut, 2009).

2.5. Ancestral area and state reconstructions

Reconstruction of ancestral areas and assessment of geographic diversification patterns within Melanocrommyum was conducted using (1) BioGeoBEARS (Matzke, 2014) and (2) The Bayesian Binary MCMC (BBM) method implemented in RASP v.3.2 (Yu et al., 2015) using datasets from the BEAST analysis. We first deleted all outgroup samples from plastid and nuclear datasets BEAST MCMC tree before analysis to represent only the biogeographic range of subgenus Melanocrommyum utilizing the outgroup-removal tool provided by RASP. BBM analysis was computed using 50, 000, 000 generations with 10 chains, samples were saved every 1000 generations. The first 1000 trees were discarded as burn-in. We set state frequencies to fixed, Jukes-Cantor (JC), and among-site rate variation was set to equal. A maximum of three areas occupied per node was set. The best models selected using model comparison of BioGeoBEARS (Matzke, 2014) implemented in RASP were the likelihood version of the BayArea model (BAYAREALIKE) (Landis et al., 2013) and the dispersal-extinction-cladogenesis model (DEC) (Ree and Smith, 2008) (Table S3). A total of six models resulted from the BioGeoBEARS analysis. The biogeographic data for species within subgenus Melanocrommyum was compiled from literature sources, online databases and herbarium specimens (Fritsch and Abbasi, 2013; Fritsch, 2016; Friesen et al., 2021). Six biogeographical areas were chosen based on the geographic range and barriers of subgenus Melanocrommyum: A) Central Asia and Afghanistan; B) Iran; C) Turkey (Asian part); D) Europe; E) Caucasus and adjacent areas; F) Levant and Iraq.

We also traced the evolution of two morphological characters: 1. Leaf type: (A) narrow, (B) moderately broad, (C) broad; 2. Inflorescence shape: (A) fasciculate (to semi-globose), (B) fastigiate to semiglobose, (C) semi-globose to subglobose (D) globose. Fresh materials, herbarium specimens, and literature were all analyzed to obtain the morphological data (Vvedensky, 1935, 1941; Khassanov, 2017; Fritsch and Abbasi, 2013; Fritsch, 2016). The BEAST MCMC tree was used to reconstruct ancestral states of morphological characters implemented in RASP v.4.4 using the ape package (Paradis and Schliep, 2019). Of the three models—equal rates (ER), symmetrical (SYM) and all-rates-different (ARD)—we utilized the ER model because Rasp v.4.4 automatically selected this as the best fitting model for both traits.

3. Results 3.1. Plastome features and phylogenetic analyses

Complete plastomes were recovered for all samples. The plastome size ranged from 149, 653 (A llium kujukense) to 153, 910 (A. cardiostemon) nucleotides and the GC content varied from 36.8 to 37.2% (Table S1). The plastomes within subgenus Melanocrommyum were identical in their structural organization, gene content, and gene arrangement. The gene sets were also identical, each with 4 unique rRNA genes, 33 unique tRNA genes, and 82–85 unique protein-coding genes. Each plastome contained 130 complete coding regions including duplicated genes in the IR and open reading frames (ORFs). However, the infA gene and one copy of the rps19 gene (Ribosomal Protein S19) in the IRb (Inverted Repeat b) region of the plastomes were missing in all samples of Melanocrommyum examined, and also in two closely related outgroup species (A. kujukense and A. oreophilum).

The alignment included a matrix with 148, 452 characters, of which 8610 characters (5.8%) were parsimony informative. Phylogenetic reconstructions of Allium subgenus Melanocrommyum were performed using three chloroplast-based datasets: the complete plastome (Fig. 2), coding sequences (CDS) (Fig. S1), intergenic spacers (IGSs) (Fig. S2) and ITS (Fig. S3).

The ML, MP and BI approaches based on complete chloroplast genome sequences produced highly congruent topologies (Fig. 2). Here, we present the ML phylogenetic tree topology with BI and MP values. Our plastome phylogenetic analyses supported subgenus Melanocrommyum as monophyletic, with 19 sections recognized within the subgenus along with its sister relationship to subgenera Vvedenskya and Porphyroprason. A. kujukense and A. oreophilum, representing subgenera Vvedenskya and Porphyroprason, were the most closely related to Melanocrommyum. Within Melanocrommyum, all phylogenetic reconstructions based on the chloroplast data resolved the same five major lineages (A, B, C, D, E) with maximum support. Most major sections were found to be para- or polyphyletic. Lineage A was resolved as the earliest diverged clade of Melanocrommyum. It contained A. fetisowii (sect. Longibidentata) and A. koksuense (sect. Decipientia). Lineage B contained species of sections Regeloprason, Miniprason and Acmopetala, whose species are distributed in Central Asia (Fig. 2). Lineage C included species of sections Melanocrommyum, Pseudoprason, Acanthoprason and species distributed in western Asia and Europe. Lineage D consisted of species of sections Megaloprason, Asteroprason, Brevicaule, Acanthoprason, Regeloprason, whose species are distributed in Central Asia and Iran. The last and largest lineage, Lineage E, included species of sections Thaumasioprason, Compactoprason, Procerallium, Regeloprason, Acmopetala, Megaloprason, Popovia, Stellata, Aroidea, Verticilata, Kaloprason, whose species are distributed in Iran and Central Asia.

The CDS-based phylogeny (Fig. S1) yielded a largely congruent structure with the plastome tree. All five major lineages are recovered with consistent clustering patterns. However, bootstrap values were slightly lower for some internal nodes, suggesting reduced resolution when only coding regions are analyzed. Despite this, the CDS tree effectively resolved closely related species and supports the major topological splits seen in the plastome dataset. The IGS-based tree (Fig. S2) also recovered the same five major lineages but exhibited variation in internal branch support. Although the backbone of the tree is maintained, several nodes show reduced bootstrap values, especially among early diverging species in lineages A and B. Nonetheless, the IGS dataset provides strong resolution within some terminal clades, such as among species in lineages D and E, suggesting that IGS regions may be more informative for resolving recent divergences or closely related taxa.

Overall, the three datasets yielded congruent major topologies, confirming the robust phylogenetic structure of Melanocrommyum into five well-defined lineages. The complete plastome dataset offered the highest resolution and node support, followed by CDS and IGSs. The ITS-based phylogenetic tree recovered seven distinct clusters (A–G), which largely correspond to the five major plastome lineages but exhibit several topological rearrangements and differences in resolution (Fig. S3).

3.2. Divergence time estimates

The chronogram based on complete chloroplast genome sequences (Fig. S4) revealed that the split of subgenus Melanocrommyum from the closely related subgenera Vvedenskya and Porphyroprason occurred around 7.2 million years ago (Ma) in the Late Miocene (5.2–9.47 Ma; 95% HPD; Fig. S4). However, the nrDNA data suggested that subgenus Melanocrommyum emerged around 24 Ma in the Late Oligocene (18.65–29.43 Ma; Fig. S4). The divergence time estimations based on plastome and nrDNA data also showed that subgenus Melanocrommyum diversified into two lineages around 5.04 Ma (N1; 6.46–3.85 Ma; 95% HPD; Fig. S4) and 18.9 Ma (N2; 14.84–23.3 Ma; 95% HPD), respectively.

One lineage was formed by sections Longibidentata and Decipientia. The crown node ages of those two sections was estimated to be 4.7 Ma based on plastome data (4.61–5.69 Ma; 95% HPD). The latter lineage split into two at about 3.93 Ma (N2; 2.98–5.03 Ma; 95% HPD) emerging as the Central Asian clade of sections Regeloprason, Miniprason, Acmopetala. The split between species distributed in western Asia, Europe and Central Asia was estimated to be 3.52 Ma in the Pliocene (N4; 2.67–4.53 Ma; 95% HPD). The divergence of species found in Turkey, Europe and Iran was estimated to date back to 2.61 Ma in the Pliocene (N5; 1.87–3.43 Ma; 95% HPD). Lineages of species that lived in Central Asia and Iran underwent a bifurcation within the Pliocene epoch (N8; 3.39 Ma; 2.57–4.35 Ma; 95% HPD; Fig. S4 and Table 1).

Based on nrDNA data, sections Longibidentata and Decipientia separated from each other approximately 24 Ma (N1; 18.65–29.43 Ma) and 18.9 Ma (N2; 23.3–14.84 Ma), respectively (Fig. 3 and Table 2). Later the lineage branched into two groups where cluster A included the sections Megaloprason, Asteroprason, Regeloprason and Acantoprason, while the second group included other remaining clusters, at about 12.42 Ma (N4; 15.09–9.95 Ma) (Fig. 3 and Table 2).

3.3. Historical biogeography and trait evolution

The results of ancestral area reconstruction analyses from BioGeoBEARS and the Bayesian Binary MCMC (BBM) method were similar, although there were some differences in the finer details of ancestral area probabilities and event likelihoods (Figs. 4, S5 and S6). However, the results from Bayesian Binary MCMC (BBM) showed higher resolution far more nodes than BioGeoBEARS (Fig. 4). Thus, we focus here on the results from BBM, which provide single-distribution areas for ancestral nodes. The BioGeoBEARS analyses showed BAYAREALIKE and DEC as the best-fit biogeographical models among the six models of BioGeoBEARS for chloroplast and nuclear DNA sequences (Table S3). Therefore, we only present the reconstruction details of BioGeoBEARS under BAYAREALIKE and DEC models (Figs. S5 and S6). The number of nodes in divergence time estimates and ancestral area reconstruction analysis were consistent. A summary of the divergence time estimations ancestral area reconstruction details is shown in Table 1, Table 2. Our results of biogeographic analysis suggest that Central Asia and Afghanistan are probably the origin centers for subgenus Melanocrommyum (area A) (nodes 1, 2, 3, 4 in Fig. 4), and this group subsequently diversified into Iran, Turkey and Europe during the Pliocene and Pleistocene periods (nodes 5, 6, 7 in Fig. 4).

Relatively similar results were shown in terms of the divergence time estimations and ancestral area reconstructions at ten nodes (Table 1). The most probable ancestral area with high relative probabilities (RPs) in BBM and BAYAREALIKE for Node 1 (5.04 Ma) was Central Asia (Area A), as well as Node 2 (Node 1 with 2 dispersal events and Node 2 with 1 dispersal event in BAYAREALIKE model). This pattern was similarly dominated by Central Asia also for Node 3 (1.53 Ma). Iran (Area B) apparently contributed to both Nodes 5 (2.61 Ma), 7 (2.08 Ma) and 9 (1.55 Ma). A more complicated lineage, which includes Central Asia, Iran, and Turkey (Areas B, C, and D), was noticed at Node 6 (1.79 Ma) (with 1 dispersal and 1 vicariance events), whereas Turkey and Europe were discovered in BAYAREALIKE (Areas C and D) for the same node. Node 8 (3.39 Ma) showed some influence from both Central Asia and Iran. The findings thus indicate substantial ancestral contributions, mostly from Central Asia and Iran with a few instances of Turkey plus Europe, implying a complex biogeography across this area (Fig. 4).

Ancestral state reconstructions on the chloroplast and nuclear trees yielded the same possible evolutionary sequence for the leaf types of subgenus Melanocrommyum (Figs. S7 and S8). On the plastome tree, the ancestral state was narrow leaves for lineages A, B, and the basal part of lineage, C. Moderately broad leaves subsequently evolved in lineage C and lineage D. Broad leaves evolved in lineage E. The general trend is thus toward increasing leaf width. For the second trait, inflorescence shape, even though the basal species (A. fetisowii, A. koksuense) have semi-globose to subglobose-shaped inflorescences, the ancestral character state for inflorescence shape was fasciculate (to semi-globose) (Fig. S8).

4. Discussion 4.1. Phylogenetic implications based on plastome and ITS sequences

Phylogenetic studies have previously used plastome trnL-trnF genes and internal transcribed spacer regions (ITS) to reconstruct the relationships of subgenus Melanocrommyum (Mes et al., 1999; Gurushidze et al., 2008, 2010; Dubouzet and Shinoda, 1999). These studies confirmed the monophyly of subgenus Melanocrommyum and have also shown that some supposed groupings based on morphological features were polyphyletic or paraphyletic. We also found that the members of large sections of Melanocrommyum were dispersed in different clades, and morphologically unrelated taxa clustered within the same clade (Fig. 2). According to Gurushidze et al. (2008), members of section Regeloprason are in nearly all well-supported clusters along the tree. Here, we confirm that section Regeloprason occurs in all lineages except the western Asia and Europe lineage (lineage C) (Fig. 2). Phylogenetic analysis based on ITS divided Melanocrommyum into a grade clade and core clade (Gurushidze et al., 2008). The latter clade includes seven clusters consisting of 1 or 2 subgroups each. However, our plastome phylogeny had a somewhat different topology, with more phylogenetic resolution and high maximum support values for five lineages (lineages 1–5) (Fig. 2). Gurushidze et al. (2010) also examined 100 species of the subgenus with multiple samples and found 74 chloroplast haplotypes. However, their results suggested that chloroplast haplotype sharing exists among up to 15 species (24 haplotypes) and sometimes a single species has several closely related haplotypes. The tree was constructed using the statistical parsimony network method suggesting 6 haplotype-based lineages, since BI/MP-based trees have a lower resolution and species with multifurcating relationships (Gurushidze et al., 2010).

In this study, we also used more than one sample for some species from different, distant localities (A llium cardiostemon, A. cristophii subsp. cristophii, A. derderianum, A. elburzense, A. isakulii A. karataviense, A. kazerouni, A. macleanii, A. minutiflorum, A. pseudowinklerianum, A. komarowii, A. stipitatum, A. altissimum) (Fig. 2). Lineage A consisted of A. fetisowii (sect. Longibidentata), A. koksuense (sect. Decipientia) and is consistent to the grade clade (ITS) and lineage Ⅰ (trnL-trnF) of Gurushidze et al. (2008, 2010), respectively.

Lineage B included some species of sections Regeloprason, Miniprason, Acmopetala; those species are distributed in Central Asia (Fig. 2). The two samples of A. karatviense plus one accession from NCBI (NC_057573) in our study were all placed in an unusual position on the tree, specifically in the basal part of the Central Asian lineage, lineage B. In contrast, the phylogenetic trees based on ITS and trnL-trnF were part of cluster 5 and lineage Ⅲ, accordingly (Gurushidze et al., 2008, 2010). However, A. karataviense was still in the same lineage as A. sewerzowii, A. severtzovioides, A. costatovaginatum, A. tschimganicum, A. dodecadontum, and A. backhousianum in our study, which is consistent with previous studies (Gurushidze et al., 2008, 2010) (Fig. 2). The two samples of A. pseudowinklerianum (sect. Regeloprason) were placed in lineage B (Fig. 2), which is inconsistent with the ITS-based tree of Gurushidze et al. (2008).

Lineage C included members of sections Melanocrommyum, Pseudoprason, Acanthoprason. These species are supposed to be distributed in western Asia and Europe (Table S1). Most members of these sections were also placed in cluster 1 in the ITS-based tree constructed by Gurushidze et al. (2008). In this lineage, the members of sect. Melanocrommyum were grouped together in one place, except for A. moderense, A. keusgenii, and A. bisotunense, which were placed among the species of sect. Acanthoprason (Fig. 2). This study used two samples of A. cardiostemon, one from Turkey and the other from Iran. However, they occupied different positions on the tree due to molecular differences in their plastomes (a 712 bp long difference and a 0.1% difference in GC content) (Table S1). Most of the species from sections Megaloprason, Asteroprason and Regeloprason were part of cluster 2 in Fritsch et al. (2010) and lineage Ⅴ in Gurushidze et al. (2010) (A. suworowii, A. regelli, A. scotostemon, A. brachyscapum, A. assadi, A. darwasicum, A. hissaricum, A. cristophii, A. ellisii, A. elburzense).

The last large lineage (lineage E) included species from sections Thaumasioprason, Compactoprason, Procerallium, Regeloprason, Acmopetala, Megaloprason, Popovia, Stellata, Aroidea, Verticilata and Kaloprason. These species are also primarily distributed in Iran and Central Asia (Table S1). According to Gurushidze et al. (2008), Fritsch et al. (2010) and Gurushidze et al. (2010) members of these sections belong to clusters 3, 4, 6, 7 and lineage Ⅳ, Ⅵ, respectively (Table 3 and Fig. 2). A detailed comparison of lineages A–E and clusters A–G of the current study to clusters 1–7 based on ITS as well as chloroplast lineages Ⅰ–Ⅵ suggested by previous studies is presented in Table 3.

4.2. Divergence time, biogeographic implications and morphological state evolution

Divergence time estimations of the genus Allium, as a member of Amaryllidaceae, have been subject to various calibration methodologies that utilize both molecular substitution rates and fossil data from other family, as no Allium fossil record exists. Thus, estimated divergence times of the genus Allium have varied, ranging from 12.8 to 64.5 Ma. These discrepancies largely arise from differences in fossil calibration points and the methodologies employed in dating (Table S4). Li et al. (2010) examined the biogeography of the first (EL1) and second (EL2) evolutionary lineages of Allium, including the members of subgenera Melanocrommyum, Vvedenskya and Porphyroprason, based on ITS data utilizing S-DIVA method. Their results suggested that the ancestral distribution area for those subgenera was eastern Asia, since earlier branching subgenera of EL2 (Caloscordum and Anguinum) are mainly distributed in eastern Asia. However, Hauenschild et al. (2017) stated that the ancestors of the second evolutional lineage (Melanocrommyum, Poryphyroprason, and Vvedemskya) probably originated in a region includes Europe and northern Asia around the Miocene and Pliocene periods based on a combined chloroplast gene dataset. According to the divergence time analysis of Allium conducted by Gurushidze (2009) based on ITS sequences, the apparently young crown age of subgenus Melanocrommyum was estimated to be 7–9 Ma (Table S4). However, a recent study of Friesen et al. (2021) demonstrated that subgenus Melanocrommyum has an older crown age of c a. 25 Ma in the late Oligocene. Both of these studies used different methodologies for tracing the date (the former study used secondary calibration from Janssen and Bremer, 2004, the age for Alliaceae s. lat.; the latter used ITS substitution rate, which is 4.13 × 10⁻⁹ substitutions per site per year, Table S4).

In this study, we focused on only the single subgenus, Melanocrommyum, and conducted the divergence time and biogeographic analyses accordingly (Fig. 3, Fig. 4) based on plastome and ITS sequence data. Our findings, which are based broad sampling of species, supports previous findings of younger crown age estimates based on chloroplast data (Gurushidze, 2009; Hauenschild et al., 2017) and older estimates from ITS data (Friesen et al., 2021, 2024). This aligns with the hypothesis of a Central Asian origin (Li et al., 2010), suggesting that the lineage likely originated in the Late Miocene or Oligocene, depending on whether plastome or ITS sequence data are considered (Table S4; Figs. S4 and S5). However, we find the ITS data for dating analysis to be most effective for several reasons. First, previous research conducted on chloroplast markers have revealed considerable variation (Table S4) and discordance with data from nuclear data, therefore making age estimations of these markers less reliable. Second, because they have biparental inheritance and higher mutation rates, ITS data generally exhibit better resolution of species-level phylogenies, which might be more applicable in the assessment of divergence estimations at target group. Therefore, the ITS-based dating approach appears to be more reliable in reflecting evolutionary timelines of this group (Friesen et al., 2024).

Based on our current biogeographic analysis and the available data on paleogeography we suggest the following hypotheses on the biogeographic history of subgenus Melanocrommyum. The rapid exhumation of the Western Tian Shan and Bogda Mountains, which occurred between 20 and 30 Ma, had a strong effect on the geological and climatic landscapes of Central Asia. Hendrix et al. (1994), Dumitru et al. (2001), Wang et al. (2008), and Wang et al. (2018) describe these climatic changes as a phase of active tectonism that caused the subsequent uplift of these mountain ranges, profoundly initiating acidification and dry steppe and semi-desert landscape formations as well as evolution of rich herbaceous xerophytes (Xiang et al., 2017). During the Late Miocene, northern Central Asia experienced a large degree of tectonic activity, particularly the uplift of the Tian Shan and Altai Mountain ranges, interacted with the mid-latitude jet stream. This led to a major reorganization of Central Asia's climate, establishing new seasonal precipitation patterns (Caves et al., 2017). This was also the time (24 or 9.36 Ma) when subgenus Melanocrommyum split from the closely related subgenera Poryphyroprason, and Vvedemskya. The resulting taxa diversified and rapid speciation occurred in Central Asia (Nodes 1–3) (Fig. 4). An early diverging section of subgenus Melanocrommyum was sect. Longibidentata. Two representative members of this section (A. fetisowii and A. chychkanense) are restricted to Central Asia. The dispersion of ancestral Central Asian taxa into the territory of Iran was likely associated with the uplift of the Zagros Mountain range, which started approximately 25 Ma and intensified between 15 and 5 Ma (Mouthereau, 2011).

RASP analysis indicated that ancestral Central Asian taxa dispersed into western Asian territory during the Pliocene or Late Miocene following vicariance events. The effects of this dispersal into western Asia on the diversification and radiation of other xerophytic taxa, such as the genera Haplophyllum, Cousinia and Acanthophyllum, has been documented by Manafzadeh et al. (2014), Djamali et al. (2012), and Mahmoudi-Shamsabad et al. (2021). The dispersion of ancestral taxa into Turkey and parts of Europe was probably associated with global cooling and the Messinian salinity crisis (Herbert et al., 2016; Krijgsman et al., 2024). During the late Miocene, a period of global cooling led large areas of the continents to experience drying and a substantial decrease in ocean temperatures, reaching levels similar to those seen today, between approximately 7 and 5.4 Ma (Herbert et al., 2016). This cooling event likely resulted in a notable decrease in precipitation in the Mediterranean Basin. Furthermore, during the Messinian salinity crisis, when the Mediterranean Sea dried out around 5.97 to 5.33 Ma, arid habitats were created that were suitable for taxa in surrounding areas (Krijgsman et al., 2024). According to the results of our RASP analysis, there was a dispersal event to the territory of Turkey and parts of Europe at Node 6, during the Pleistocene or Pliocene (Table 1, Table 2, Fig. 4). Thus, it is possible that climatic changes played a role in prompting the westward expansion and diversification of the taxa. The radiation and diversification of such xerophytic genera as Helianthemum and Chiliadenus due to global cooling and the Messinian salinity crisis events have been documented by Martín-Hernanz et al. (2021) and Bengtson and Anderberg (2018), respectively. According to Li et al. (2010), after migration from Central Asia to western Asia there were subsequent dispersal events in the reverse direction. Our biogeographic analysis confirms this reverse dispersal, as both phylogenetically basal and advanced groups (sections Regeloprason, Acmopetala) are in Central Asia, suggesting a diverse range of evolutionary stages within the taxa (Fig. 4).

The main centers of species diversity of subgenus Melanocrommyum are considered to be Asia Minor, Southwest Asia and Central Asia (Fritsch et al., 2016). Mapping of the five plastome lineages (A–E) revealed that four of these lineages (B–E) showed some congruence with geographic origins (Fig. 2). Clade B is exclusively confined to Central Asia, which is identified as a place of origin according to our biogeographic analysis (Fig. 2). Plastome lineage B is represented by material from western Asia (Iran, Turkey) and Europe (Bulgaria) (Table S1 and Fig. 2). However, lineages D and E contain samples from both Central Asia and Iran. These findings also confirm re-migration of the subgenus to its historical place of origin (Fig. 4), which is consistent with the findings of Li et al. (2010).

The evolution of key morphological characters in subgenus Melanocrommyum was reconstructed using both nuclear internal transcribed spacer (ITS) (Fig. S7) and plastid genome phylogenies (Fig. S8). These complementary phylogenetic frameworks provide critical insights into the diversification of leaf types and inflorescence structures within the group. Consistent with previous studies highlighting the high morphological plasticity in Allium (Friesen et al., 2006; Gurushidze et al., 2008; Aryakia et al., 2016), our results reveal recurrent patterns of convergence and parallel evolution likely driven by ecological pressures and the independent genetic regulation of traits. Ancestral state reconstruction indicates that narrow leaves (A) and fasciculate to semi-globose inflorescences (A) represent the most likely ancestral conditions for this subgenus (Figs. S7 and S8). Leaf morphology, in particular, is evolutionarily labile, with multiple independent shifts among narrow, moderately broad, and broad leaf forms across the phylogeny. This variation likely reflects a range of ecological adaptations. Whereas broader leaves are commonly associated with shaded or mesic environments in many plant lineages (Sessa and Givnish, 2014), in subgenus Melanocrommyum, they are also widespread among species in arid and montane habitats. In these environments, broad leaves may serve as water storage organs or facilitate rapid photosynthetic activity during brief growing seasons (Gibson, 2012). The evolution of inflorescence shape in subgenus Melanocrommyum reveals clear evidence of convergent patterns, particularly in the repeated emergence of globose forms across phylogenetically distant lineages (Figs. S7 and S8). These compact, spherical inflorescences may confer several adaptive advantages, including enhanced visual signalling to pollinators and improved floral display efficiency (Ohashi and Yahara, 2001; Harder et al., 2004). Their independent origin in multiple clades underscores the evolutionary lability of inflorescence architecture in the subgenus, likely shaped by similar selective pressures related to pollination strategies. This aligns with earlier work highlighting that Allium's floral morphology is highly plastic (Jang et al., 2024).

4.3. Recent diversification based on plastome data, incomplete lineage sorting, hybridization

According to previous studies (Ramdhani et al., 2009, 2011; Kellner et al., 2011; Escobar et al., 2020), there may be several reasons for the non-monophyly of species, including contemporary (recent) diversification (radiation), incomplete lineage sorting, and hybridization. The most influential factor of these might be recent diversification of species, which may also increase the likelihood of incomplete lineage sorting and hybridization. According to Ramdhani et al. (2011), plant genetic markers used for phylogenetic analysis evolve at different rates. But if the rate of sequence divergence has not been in step with morphological evolution, or if the group under study is young, then there can still be instances of non-monophyletic groups, even when various types of markers have been employed. In our study, phylogenetic analysis based on complete chloroplast genome still generated para- or polyphyletic sections of the subgenus (Fig. 2). These findings may reflect a recent diversification process, given that ITS-based dating indicates the majority of these species diverged during the Pliocene-Pleistocene periods (Fig. 3). Furthermore, according to research conducted by Gurushidze et al. (2010) the older (ancestral) haplotypes were more frequent, not only in the basal clade but in all lineages (Ⅰ–Ⅵ), suggesting that the subgenus had indeed undergone recent radiations. This was also confirmed by the ITS-based dating analyses performed by Gurushidze et al. (2010) and Hauenschild et al. (2017) suggesting the crown age of the subgenus as 7–9 Ma and 6.6–10 Ma, respectively. The non-monophyly due to recent radiation (diversification) has also been observed in related taxa, such as the genera Gilliesia and Miersia. Both of these genera, along with the current taxon (subgenus Melanocrommyum), are included in the same subfamily, Allioideae (Escobar et al., 2020). Furthermore, the loss of certain genes in the plastome of Allium, i.e., rps16, infA, rpl22, rps2, and ccsA, has been documented by previous researchers (Omelchenko et al., 2020; Scobeyeva et al., 2021; Munavvarov et al., 2022). Notably, all species of Allium subgenus Melanocrommyum (including the two closely related outgroup species, A. kujukense, A. oreophilum) have lost infA and rps19 gene, and have only one copy in the LSC region, although two copies have been reported in the IRa and IRb regions of all other genomes of Allium (Xie et al., 2020; Scobeyeva et al., 2021; Munavvarov et al., 2022). Previous research on Zingiberaceae has indicated that the loss of one copy of rps19 in Cautleya gracilis was the result of LSC region expansion (Yang et al., 2022). Similarly, the loss of infA and one copy of rps19 in the IRb region are likely derived within Allium subgenus Melanocrommyum, hinting at its relatively recent radiation. However, according to Munavvarov et al. (2022), A. tuberosum A. ramosum, A. ampeloprasum also lost the infA gene. Millen et al. (2001) found that infA was transferred from the chloroplast to the nuclear genome during angiosperm evolution many times, making infA by far the most mobile chloroplast gene known in plants. Furthermore, Yusupov et al. (2022) found the advanced testa characteristics (convex periclinals, large and small verrucae) for the species of the subgenus Melanocrommyum, also implying it had undergone recent diversification.

Numerous studies have demonstrated that incomplete lineage sorting (ILS), which refers to the continuation and preservation of ancestral genetic variation through speciation events, can complicate phylogenetic reconstruction (Gurushidze et al., 2010; Ramdhani et al., 2010, 2011; Escobar et al., 2020). Rapid and recent divergence events decrease the likelihood of lineages being sorted before the completion of cladogenesis (Syring et al., 2007; Escobar et al., 2020). Therefore, incomplete lineage sorting may explain why sections of subgenus Melanocrommyum are para- or polyphyletic. Predicting the behavior of hybrids in phylogenetic reconstruction is highly challenging. Hybrids can lead to a loss of resolution, causing topological changes in regions with weak support in the phylogenetic analyses (Ramdhani et al., 2011; Yao et al., 2015). Thus, hybridization and introgression can lead to non-monophyly at various taxonomic levels (Yao et al., 2015). The capability of the species of Allium subgenus Melanocrommyum to hybridize under artificial conditions has been thoroughly documented by Friis et al. (1997), which listed hybrid cultivars from the combinations of A. macleanii, A. cristophii, A. karataviense, A. stipitatum, A. hollandicum, and A. rosenorum. Although our comparative analysis of plastome lineages (A–E) and ITS clusters (A–G) (clusters 1–7 also in Gurushidze et al., 2008) indicated that species within the clusters are similar, there was a difference in the basality or position of the clusters throughout the phylogenetic trees (Fig. 2 and Table 2). Therefore, it is expected that the para- or polyphyly of sections within subgenus Melanocrommyum is also driven by ancient hybridization. To further clarify the evolutionary relationships within Melanocrommyum, future studies should incorporate high-resolution nuclear genomic data, ideally with multiple individuals sampled per species. Utilizing a large set of single-copy nuclear orthologous genes could provide valuable insights into the intricate evolutionary dynamics of Melanocrommyum. In particular, addressing gene tree discordance, exploring potential reticulate evolution through phylogenetic network approaches, and conducting simulations to assess the influence of incomplete lineage sorting (ILS) would offer crucial strategies for testing the hypotheses of hybridization and ILS that may underlie the phylogenetic patterns observed in this study.

5. Conclusions

The current research aimed to contribute to the knowledge on Allium subgenus Melanocrommyum by constructing its phylogeny, historical biogeography, and morphological character evolution based on plastome and ITS sequences. Plastome analysis identified five well-defined lineages (A–E) with a Central Asian origin dating back to Late Miocene or Oligocene. The geographical events, such as mountain uplift and climate changes, have influenced diversification and dispersal patterns within this subgenus. Our findings confirm previous inconsistencies between morphological and molecular classifications, pointing out recent diversification, incomplete lineage sorting (ILS), and ancient hybridization of the subgenus contributing towards the para- or polyphyly of sections within the subgenus. Furthermore, we found that the ancestral morphology of the subgenus was characterized by narrow-leaved plants with fasciculate to semi-globose-shaped inflorescence. Thus, this comprehensive research provides insight into the biogeographic history of the subgenus Melanocrommyum, suggesting that diversifications within the subgenus probably occurred recently. However, certain gaps still exist between plastome and nuclear data, which might hinder the resolution of phylogenetic relationships. In this regard, further research should be conducted utilizing more detailed nuclear genomic data with multiple samples from each species. This might address these difficulties and enhance the resolution of species-specific relationships within the subgenus.

Acknowledgements

We thank Dr. David E. Boufford from the Harvard University Herbaria (U.S.A.) and Dr. Richard Ree from the University of Chicago for editing the English. We also thank Dr. Dörte Harpke, Dr. Frank R. Blattner, Dr. Nikolai Friesen, Dr. Bektemir Osmonali, Dr. Mehmet M. Balos for their help providing leaf material of several taxa. This study was supported by grants from the Second Tibetan Plateau Scientific Expedition and Research (STEP) program (2019QZKK0502), the state research project 'Taxonomic Revision of Polymorphic Plant Families of the Flora of Uzbekistan' (FZ-20200929321) and the State Programs for 2021–2025 'Grid mapping of the flora of Uzbekistan' and the 'Digital Nature. Development of a digital platform for the flora of Central Uzbekistan', implemented by the Institute of Botany of the Academy of Sciences of the Republic of Uzbekistan for the period 2025–2029; National Natural Science Foundation of China (32322006), the Key Projects of the Joint Fund of the National Natural Science Foundation of China (U23A20149), the R & D Program of Yunnan Province (202103AF140005). This publication has also been produced within the framework of the Grant No. PRIM 01-73 "The modernization of the Institute of Botany of the Academy of Sciences of the Republic of Uzbekistan", funded under the MUNIS Project, supported by the World Bank and the Government of the Republic of Uzbekistan. The statements do not necessarily reflect the official position of the World Bank and the Government of the Republic of Uzbekistan.

CRediT authorship contribution statement

Ibrokhimjon Ergashov: Writing – original draft, Visualization, Investigation, Formal analysis. Ziyoviddin Yusupov: Writing – review & editing, Resources, Methodology. Alireza Dolatyari: Writing – review & editing, Investigation. Mina Khorasani: Writing – review & editing, Resources. Ismail Eker: Writing – review & editing, Resources. Nazgul Turdumatova: Writing – review & editing, Resources, Investigation. Georgy Lazkov: Writing – review & editing, Resources. Farruhbek Rasulov: Methodology, Writing – review & editing. Hang Sun: Supervision, Conceptualization. Tao Deng: Supervision, Conceptualization. Komiljon Tojibaev: Supervision, Conceptualization.

Declaration of competing interest

The authors declare there is no conflict of interest regarding this manuscript. All the authors agreed to submit this manuscript.

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi.org/10.1016/j.pld.2025.04.009.

Appendix 1. The conspectus of Allium subgenus Melanocrommyum (Webb & Berthel.) Rouy

1. Section Longibidentata (R.M. Fritsch) R.M. Fritsch

1.1. Allium chychkanense R.M. Fritsch,

1.2. Allium fetisowii Regel

2. Section Tulipifolia R.M. Fritsch & N. Friesen

2.1. Allium robustum Kar. & Kir.

2.2. Allium tulipifolium Ledeb.

3. Section Decipientia (Omelczuk) R.M. Fritsch

3.1. Allium decipiens Fisch. ex Schult. & Schult.f.

3.2. Allium grande Lipsky

3.3. Allium koksuense R.M. Fritsch, N. Friesen & S.V. Smirn.

3.4. Allium lepsicum R.M. Fritsch, N. Friesen & S.V. Smirn.

3.5. Allium quercetorum (Seregin) Seregin

3.6. Allium subscabrum (Regel) R.M. Fritsch

3.7. Allium viridulum Ledeb.

4. Section Regeloprason Wendelbo

4.1. Allium balkhanicum (R.M. Fritsch & F.O. Khass.) R.M. Fritsch

4.2. Allium cathodicarpum Wendelbo

4.3. Allium chodsha-bakirganicum Gaffarov & Turak.

4.4. Allium cupuliferum Regel

4.5. Allium nuratavicum (R.M. Fritsch & Beshko) Beshko

4.6. Allium darwasicum Regel

4.7. Allium hissaricum Vved.

4.8. Allium iliense Regel

4.9. Allium intradarvazicum R.M. Fritsch

4.10. Allium isakulii R.M. Fritsch & F.O. Khass.

4.11. Allium lipskyanum Vved.

4.12. Allium pseudowinklerianum R.M. Fritsch & F.O. Khass.

4.13. Allium regelii Trautv.

4.14. Allium sochense R.M. Fritsch & U. Turakulov

4.15. Allium subkopetdagense (R.M. Fritsch & F.O. Khass.) R.M. Fritsch

4.16. Allium victoris Vved.

4.17. Allium winklerianum Regel

5. Section Melanocrommyum Webb & Berthel

5.1. Allium aschersonianum Barbey

5.2. Allium asclepiadeum Bornm.

5.3. Allium atropurpureum Waldst.

5.4. Allium bisotunense R.M. Fritsch

5.5. Allium cardiostemon Fisch. & C.A. Mey

5.6. Allium chrysantherum Boiss. & Reut.

5.7. Allium colchicifolium Boiss.

5.8. Allium crameri Aschers. & Boiss.

5.9. Allium cyrilli Ten.

5.10. Allium donmezii Mutlu & Karakuş

5.11. Allium dumetorum Feinbr. & Szelub.

5.12. Allium eginense Freyn

5.13. Allium efeae Özhatay & İ.Genç

5.14. Allium elmaliense Deniz & Sumbul

5.15. Allium fedtschenkoi Nábělek

5.16. Allium haussknechtii Nábělek

5.17. Allium karamanoglui Koyuncu & Kollmann

5.18. Allium keusgenii R.M. Fritsch

5.19. Allium kharputense Freyn & Sint.

5.20. Allium lachnophyllum Paine

5.21. Allium libani Boiss.

5.22. Allium lycaonicum Siehe

5.23. Allium mariae Bordz.

5.24. Allium moderense R.M. Fritsch

5.25. Allium mozaffarianii Maroofi & R.M. Fritsch

5.26. Allium multibulbosum Jacq.

5.27. Allium muratozelii Armagan

5.28. Allium nemrutdaghense Kit Tan & Sorger

5.29. Allium nigrum L.

5.30. Allium noëanum Reut. ex Regel

5.31. Allium olivieri Boiss.

5.32. Allium orientale Boiss.

5.33. Allium rhetoreanum Nábělek

5.34. Allium rothi i Zucc.

5.35. Allium saralicum R.M. Fritsch

5.36. Allium schisticola R.M. Fritsch

5.37. Allium shatakiense Rech.f.

5.38. Allium shirnakiense L.Behçet & Rüstemoğlu

5.39. Allium stenopetalum Boiss. & Kotschy ex Regel

5.40. Allium straussii Bornm.

5.41. Allium sultanae-ismailii Yildirim

5.42. Allium tele-avivense Eig

5.43. Allium tubergenii Freyn

5.44. Allium urmiense Kamelin & Seisums

5.45. Allium vinicolor Wendelbo

5.46. Allium woronowii Miscz. ex Grossh

6. Section Acanthoprason Wendelbo

6.1. Allium akaka subsp. akaka S.G. Gmel. ex Schult. & Schult. f. in Roemer & Schult.

6.2. Allium akaka subsp. bozgushense R.M. Fritsch

6.3. Allium alamutense Razyfard, Zarre & R.M. Fritsch

6.4. Allium alekii R.M. Fritsch & M. Agabab

6.5. Allium austroiranicum R.M. Fritsch

6.6. Allium breviscapum Stapf

6.7. Allium chlorotepalum R.M. Fritsch & M. Jaeger

6.8. Allium derderianum Regel

6.9. Allium egorovae M.V. Agab. & Ogan

6.10. Allium graveolens (R.M. Fritsch) R.M. Fritsch

6.11. Allium haemanthoides Boiss. & Reut. ex Regel

6.12. Allium hamedanense R.M. Fritsch

6.13. Allium iranshahrii R.M. Fritsch

6.14. Allium kurdistanicum R.M. Fritsch & Maroofi

6.15. Allium kuhrangense Akhavan, Saeidi & R.M. Fritsch

6.16. Allium latifolium Jaub. & Spach

6.17. Allium mahneshanense Razyfard, Zarre & R.M. Fritsch

6.18. Allium materculae Bordz.

6.19. Allium minutiflorum Regel

6.20. Allium ramazanicum Parsa

6.21. Allium sabalense R.M. Fritsch

6.22. Allium sahandicum R.M. Fritsch

6.23. Allium shelkovnikovii Grossh.

6.24. Allium subakaka Razyfard & Zarre

6.25. Allium ubipetrense R.M. Fritsch

6.26. Allium vasilevskajae Ogan

6.27. Allium zagricum R.M. Fritsch

7. Section Pseudoprason (Wendelbo) K. Perss. & Wendelbo

7.1. Allium hooshidaryae Mashayekhi, Zarre & R.M. Fritsch

7.2. Allium koelzii (Wendelbo) K. Perss. & Wendelbo

7.3. Allium sanandajense Maroofi & R.M. Fritsch

8. Section Asteroprason R.M. Fritsch

8.1. Allium aladaghense Memariani & Joharchi

8.2. Allium cristophii subsp. cristophii Trautv.

8.3. Allium cristophii subsp. golestanicum R.M. Fritsch

8.4. Allium elburzense Wendelbo

8.5. Allium ellisii Hook. f.

8.6. Allium helicophyllum Vved.

8.7. Allium kuhsorkhense R.M. Fritsch & Joharchi

8.8. Allium monophyllum Vved. ex Czerniak.

8.9. Allium parhamii Memariani

8.10. Allium pseudobodeanum R.M. Fritsch & Matin

9. Section Stellata (F.O. Khass. & R.M. Fritsch) R.M. Fritsch

9.1. Allium taeniopetalum Popov & Vved.

9.2. Allium mogoltavicum Vved.

9.3. Allium turakulovii (R.M. Fritsch & F.O. Khass.) F.O.Khass. & Yusupov

10. Section Megaloprason Wendelbo s.str. str.

10.1. Allium assadii Seisums

10.2. Allium brachyscapum Vved.

10.3. Allium esfahanicum R.M. Fritsch

10.4. Allium fibriferum Wendelbo

10.5. Allium insufficiens Vved.

10.6. Allium kopsedorum R.M. Fritsch

10.7. Allium kwakense (R.M. Fritsch) R.M. Fritsch

10.8. Allium rosenbachianum Regel s.g str.

10.9. Allium sarawschanicum Regel

10.10. Allium schugnanicum Vved.

10.11. Allium scotostemon Wendelbo

10.12. Allium suworowii Regel

11. Section Miniprason R.M. Fritsch

11.1. Allium karataviense subsp. karataviense Regel,

11.2. Allium karataviense subsp. henrikii Rukšāns

12. Section Acmopetala R.M. Fritsch

12.1. Allium aflatunense B. Fedtsch.

12.2. Allium alaicum Vved.

12.3. Allium arkitense R.M. Fritsch

12.4. Allium backhousianum Regel

12.5. Allium bekeczalicum Lazkov

12.6. Allium calocephalum Wendelbo

12.7. Allium costatovaginatum Kamelin & Levichev

12.8. Allium dasyphyllum Vved.

12.9. Allium dodecadontum Vved.

12.10. Allium kurdaicum Bajtenov

12.11. Allium pangasicum Turak

12.12. Allium saposhnikovii Nikitina

12.13. Allium schachimardanicum Vved.

12.14. Allium severtzovioides R.M. Fritsch

12.15. Allium sewerzowii Regel

12.16. Allium tashkenticum F.O. Khass. & R.M. Fritsch

12.17. Allium tokaliense Kamelin & Levichev

12.18. Allium tschimganicum O. Fedtsch.

12.19. Allium vvedenskyanum Pavlov

12.20. Allium zergericum F.O. Khass. & R.M. Fritsch.

13. Section Verticillata Kamelin

13.1. Allium verticillatum Regel

13.2. Allium viridiflorum Pobed.

14. Section Compactoprason R.M. Fritsch

14.1. Allium giganteum Regel

14.2. Allium isfairamicum O. Fedtsch.

14.3. Allium komarowii Lipsky

14.4. Allium macleanii Baker

14.5. Allium majus Vved.

14.6. Allium trautvetterianum Regel.

15. Section Procerallium R.M. Fritsch

15.1. Allium altissimum Regel

15.2. Allium bakhtiaricum Regel

15.3. Allium botschantzevii Kamelin

15.4. Allium hollandicum R.M. Fritsch

15.5. Allium jesdianum subsp. angustitepalum (Wendelbo) F.O. Khass. & R.M. Fritsch

15.6. Allium jesdianum subsp. jesdianum Boiss. & Buhse

15.7. Allium kazerouni Parsa

15.8. Allium orientoiranicum Neshati, Zarre & R.M. Fritsch

15.9. Allium pseudohollandicum R.M. Fritsch

15.10. Allium remediorum (R.M. Fritsch) R.M. Fritsch

15.11. Allium rosenorum R.M. Fritsch

15.12. Allium stipitatum Regel

16. Section Aroidea F.O. Khass. & R.M. Fritsch

16.1. Allium aroides Popov & Vved.

17. Section Acaule R.M. Fritsch

17.1. Allium hexaceras Vved.

18. Section Popovia F.O. Khass. & R.M. Fritsch

18.1. Allium gypsaceum Popov & Vved

19. Section Thaumasioprason Wendelbo

19.1. Allium caroli-henrici Wendelbo

19.2. Allium cuculatum Wendelbo

19.3. Allium khozratense R.M. Fritsch

19.4. Allium mirum Wendelbo

20. Section Kaloprason K. Koch

20.1. Allium alexeianum Regel

20.2. Allium bucharicum Regel

20.3. Allium caspium subsp. baissunense (Lipsky) F.O. Khass. & R.M. Fritsch

20.4. Allium caspium subsp. caspium (Pall.) M. Bieb

20.5. Allium decoratum Turginov & Tojibaev

20.6. Allium hindukuschense Kamelin & Seisums

20.7. Allium nevskianum Vved. ex Wendelbo

20.8. Allium protensum Wendelbo

21. Section Brevicaule R.M. Fritsch

21.1. Allium chitralicum F.T. Wang & Tang

21.2. Allium eugenii Vved.

21.3. Allium sergii Vved.