https://orcid.org/0000-0003-3284-5326
Figures
Abstract
A ramp (Allium tricoccum Aiton), or wild leek, is a perennial herbaceous plant native to the forests of eastern North America. Like other members of the Allium genus, ramps produce sulfur-containing compounds that give them culinary and medicinal appeal. Ramps reproduce clonally via bulb division and sexually via seed production, but little is known regarding how much genetic diversity is present in ramps and ramp populations. Furthermore, there is an unresolved question regarding species delineation, with as many as four species suggested. We developed four polymorphic microsatellite markers that we used to measure genetic diversity in ramps and found that ramp populations have low to moderate levels of genetic variation and high differentiation and that individual ramp populations vary in clonality and genetic diversity. Finally, we provide the first preliminary genetic evidence supporting the delineation of the purported second ramp taxon, narrow-leaf ramps (Allium burdickii (Hanes) A.G. Jones).
Citation: Nilson SE, Estep MC, Burkhart EP, Ballard H, Houston E, Sitepu BS, et al. (2025) New microsatellite markers distinguish two species of ramps (Allium tricoccum Aiton Complex, Amaryllidaceae) and show variation in clonality and genetic diversity between species and among populations. PLoS One 20(10): e0332086. https://doi.org/10.1371/journal.pone.0332086
Editor: Mojtaba Kordrostami, Nuclear Science and Technology Research Institute, IRAN, ISLAMIC REPUBLIC OF
Received: August 26, 2024; Accepted: August 25, 2025; Published: October 8, 2025
Copyright: © 2025 Nilson et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the paper and its Supporting Information.
Funding: Research funding was provided by Pennsylvania Department of Conservation and Natural Resources (DCNR) Wild Resource Conservation Program Grant 18585 (to SEN, EPB) and the Nobles Family Charitable Fund (to MCE). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Ramps, also known as wild leeks (Allium tricoccum Aiton), are perennial herbs occurring most commonly on mesic forestlands in the midwestern and eastern regions of North America [1–3]. Ramps belong to the genus Allium L., which includes onion (A. cepa L.), garlic (A. sativum L.), leek (A. ampeloprasum L.), and other economically important plants. Containing 550–800 species [4–5], Allium was traditionally submerged in the broadly circumscribed Liliaceae, sensu lato, but molecular phylogenetic studies and treatments now place it in Alliaceae [6] or Amaryllidaceae [7–8].
Increased public interest in ramp collection (i.e., foraging) from wild populations for personal and commercial consumption, along with an increase in commercial trade, has raised conservation concerns over the impact of wild collection in some parts of the range because ramp populations can be slow to recover following harvest [9–10]. Due to these concerns, ramp collection has been restricted to personal use only in some regions of the United States [e.g., 11] and Canada [e.g., 12]. As ramps have become increasingly commercialized, they are now being promoted as a specialty crop for use within agroforestry systems, especially through the practice known as forest farming [2,13]. This cropping system has potential to be both a conservation mechanism and an economic development opportunity [2,14,15].
Ramps emerge from bulbs in the late winter and early spring, before forest canopy leaf-out, to take advantage of higher insolation in the forest understory [16–17]. During the approximately six-week photosynthetic phase, ramp bulbs swell [17–18] and may also begin clonal division [19]. During the summer months, the scape bears small, white flowers arranged in a narrowly conical to nearly spherical umbel. Ramps can reproduce via both self-compatibility and out-crossing [20–21]. Following pollination, the seeds mature in three-seeded capsules during late summer and early fall before the plant enters winter dormancy. One demographic study of A. tricoccum found that sexual reproduction contributes minimally to population growth and that asexual reproduction, a process which can take five to eight years, is the primary mode of population maintenance [19]. However, significant annual variation in seed production and seedling recruitment was also noted [19].
Despite the growing interest in ramps as a wild food and agroforestry crop [13], information regarding ramp genetic diversity remains limited. Genetic diversity gives populations and species the potential to adapt to and persist in changing conditions, such as those brought on by disturbance or climate change [22–24]. Although counter-intuitive, clonal and partially clonal plant populations do have the potential to maintain moderate to high levels of genetic diversity, depending on the amount of diversity present in the founding population, frequency of seedling establishment, and the degree of gene flow into the population via pollen or seed dispersal [25]. The only study of ramp genetic diversity to date used seven isoenzyme systems to document genetic variation in ramps collected at six sites in Quebec [26]. Only two of the seven isoenzyme markers were polymorphic, which led the researchers to conclude that ramp genetic diversity is very low, and that ramps are likely reproducing primarily via clonal reproduction [26]. Since the original study was performed, ramp harvesting has increased and become commercialized [13], creating a need to develop better genetic tools which can be used to measure genetic variation within and among ramp populations across the range of this species. These data can be used to identify populations of conservation concern, shed insight into the relative contributions of asexual and sexual reproduction in ramps at the population and species levels, and help land managers and forest farmers develop ramp harvesting and stewardship guidelines to support ramp conservation.
The development of tools to assess ramp genetics may also help resolve the question of how many “ramp” taxa exist in North America. There has been debate as to whether there is one polymorphic species (A. tricoccum), one species with two varieties (var. tricoccum and var. burdickii Hanes) or two distinct species (A. tricoccum and A. burdickii (Hanes) A.G. Jones) in eastern North America. The existence of two ramp taxa was first proposed by Hanes and Ownbey [27], when they contrasted the burdickii race against A. tricoccum based on morphological, phenological, and habitat differences in specimens from Michigan and Wisconsin. Hanes [28] later formally described var. burdickii. Walker [29] examined populations in New York state and confirmed observations of two taxa. Jones [30] conducted extensive taxonomic studies of the Allium tricoccum complex and documented many macromorphological trait differences and mostly non-overlapping phenologies between the two taxa. Jones concluded that the two taxa represented morphologically, ecologically, and phenologically different species and raised var. burdickii to species rank, as A. burdickii (Hanes) A. G. Jones.
Kauffman [31] and McNeal and Jacobsen [32] retained A. burdickii as a variety of A. tricoccum. Kauffman noted that some specimens appeared to be intermediate or expressed confounding variation patterns that did not fit either taxon, while McNeal and Jacobsen stated that A. burdickii expressed overlapping traits and geographic range with A. tricoccum and was doubtfully distinct. Bell [20] conducted herbarium and field studies like those of Jones [30], but they initially separated plants into two groups based solely on presence or absence of red-purple pigmentation on any part of a plant. Bell also made laboratory observations of morphology and conducted crossing experiments among individuals, populations, and taxa to interpret inter-fertility on living plants from populations in eastern West Virginia, Pennsylvania, and New York. They reported significant overlap of morphological traits and geographic distributions, as well as complete interfertility, between the two taxa, and concluded that only one polymorphic species, A. tricoccum, should be recognized.
The latest taxonomic study of the A. tricoccum complex, by Sitepu [33], examined living plants retrieved from two dozen sites in eight midwestern and Appalachian states. Plants grown in a common garden were observed over two seasons for morphological traits and phenologies following the approach of Jones [30]. Sitepu [33] documented noticeable changes in pigmentation in individuals of A. tricoccum sensu lato over seasons in Athens County, Ohio populations, and confirmed the infrequent presence of entirely unpigmented individuals of A. tricoccum in some populations. Statistical analyses of morphological traits distinguished four taxa: A. tricoccum sensu stricto; A. burdickii sensu stricto from the easternmost Great Plains and Midwest; south green ramps in the Interior Low Plateau of Kentucky and Tennessee and possibly southern Ohio and Indiana; and highland green ramps in the Appalachian Mountains. Phenological studies [33] confirmed that the green ramps taxa initiated and concluded anthesis weeks earlier than A. tricoccum with rare overlap in blooming times. Sitepu [33] concluded that the A. tricoccum complex consisted of four distinct evolutionary species under the Unified Species Concept [34]. Aside from these varied and sometimes conflicting taxonomic studies, field botanists and naturalists have continued to report the occurrence of strictly unpigmented or weakly pigmented plants otherwise identical to typical A. tricoccum, populations that do not match any one taxon fully, and populations that express overlapping blooming times between A. burdickii sensu lato and A. tricoccum.
Recent floristic treatments support A. burdickii sensu lato as a separate species [3,33] (see S1 Fig for key differences). While A. tricoccum is found in northeastern United States (U.S.) and adjacent regions of eastern Canada and in the Great Lakes region, the distribution of A. burdickii in the broad sense has been suggested to occur from Maine to North Dakota and south of New Jersey [1,3,30,35]. However, further south, distribution maps for A. burdickii are not consistent on whether it is present or absent in Virginia and North Carolina [3,35,36], and before this research there were no documented occurrences of A. burdickii in Pennsylvania. Allium tricoccum has two different color extremes and can have either red/purple stems or green/white stems whereas A. burdickii only has green/white stems. The misidentification of A. burdickii as unpigmented (lacking anthocyanin) A. tricoccum and vice versa has likely contributed to unclear distributions of A. burdickii. The sole genetic study on ramps was not able to genetically distinguish A. burdickii from A. tricoccum, however the isoenzyme markers had limited resolution [26]. New genetic methods could allow us to determine if A. burdickii is a unique species. The two additional green/white taxa similar to A. burdickii proposed by Sitepu [33] raise further questions about the diversity of A. burdickii and its separation from A. tricoccum.
Microsatellite markers specific to A. tricoccum have not been developed previously because ramps are not as economically or agriculturally significant as some other Allium species. Microsatellites are short tandem repeats of DNA found throughout prokaryotic and eukaryotic genomes. Microsatellites have a high rate of evolution and are highly diverse in repeat length making them excellent markers for genetic studies. There are twenty-five existing microsatellite markers that have been developed among varieties of A. fistulosum (bunching onion), onion, and garlic that had potential for successful use in A. tricoccum [37–39]. All these markers were screened against a small number of individuals and found to not cross amplify well.
This study sought to 1) identify polymorphic markers that can successfully amplify A. tricoccum DNA, 2) utilize these markers to measure genetic diversity and clonality in ramps and ramp populations; and 3) examine ramp population structure, and determine if there is genetic support for more than a single ramp taxon
Materials and methods
Microsatellite marker development
High quality A. tricoccum DNA was sent to West Virginia University (Morgantown WV, USA) Core Facility for Illumina sequencing. The resulting sequences were trimmed in FastP for quality using default parameters [40]. MSATcommander was utilized to identify microsatellite loci and design primers to amplify these loci [41]. Primers were designed between 19 to 25 base-pairs long, GC content was between 45–55%, and a PIG-tail sequence [42] was added to reverse primers to reduce stutter. One hundred primer pairs were screened against a small number of individuals (7 DNA samples and a negative control) to assess consistent amplification, product size, and to identify polymorphisms. Eighteen primer pairs that showed consistent amplification and expected product size were further assessed by labeling the forward primer with a 5’ fluorochrome (either FAM, VIC, or NED) to facilitate multiplex PCR (S1 Table).
Sampled sites, species identification, and voucher collection
Ramp sites were visited and sampled between 2019–2023 (Fig 1A and S2 Table). Where needed, several visits were made to sites to observe changes in phenology and morphology to make a positive confirmation of which species was being sampled (key differences between the taxa are shown in S1 Fig). In order to capture the breadth of genetic diversity at each site, approximately 20 ramps were randomly sampled per site with a minimum of 20 meters between samples. Typically, this was achieved by walking the perimeter of the site and sampling the edges of each population, and then making a transect through the center. In some cases, ramp populations were scattered across multiple hectares and an effort was then made to sample plants randomly throughout the site. For each sample, a single leaf was harvested and placed in a coin envelope and dried in silica gel. This study did not require IRB approval. The authors were given written and/or verbal consent by private landowners prior to sampling. Permits were not required for sampling on private land. Collection permits (USDA Forest Service 4080; PA DCNR State Forests Permit SFRA-1832, SFRA-1908, SFRA-2009; PA DCNR State Parks Permit 2019–69, 2023–30; PA DCNR Wild Plant Management Permits 17–035, 18–035, 19–035, 20–035, 21–035, 22–035, 23–035, 24–035, 20–761, 19–761, 18–761; and National Parks Service GRSM-2020-SCI-2047) were acquired for public lands prior to sample and voucher collection.
A, Sites were identified as either A. burdickii (orange) or A. tricoccum (blue) at collection. One site appeared to be a mixed collection of A. burdickii and A. tricoccum based on vouchers provided to the researchers and is shown in gray. Collections made at the same location are connected by black lines. B, STRUCTURE results from analysis of clone-corrected data set (n = 276 multilocus genotypes (MLGs) collected at 41 sites) based on four microsatellite markers. Population mean Q scores for K = 2 genetic clusters are denoted in pie graphs. MLGs collected at A. burdickii sites are found in the orange cluster (labeled A. burdickii STRUCTURE cluster), while the second cluster in blue, contained only MLGs collected at A. tricoccum sites (labeled A. tricoccum STRUCTURE cluster). Contains information licensed under the Open Government License – Canada.
While most sites were sampled by the authors, some samples were provided by land managers and botanists who we determined were reliable sources with properly identified populations. If both species were believed to be present at the site, the species were collected separately. The species cooccurred at four localities in Minnesota (MN) (two sites), Indiana (IN), and Pennsylvania (PA). At two of these co-occurring sites the land managers who collected samples on our behalf made multiple collections from different areas of the site. We elected to keep these collections separate in the analysis. When permitted, voucher specimens were collected, pressed, mounted, and deposited at the Carnegie Museum of Natural History (CM), Ohio University (BHO), and Penn State University (PAC) herbaria. Based on voucher specimens, one site ECUN, appeared to be a mixed collection of A. burdickii and A. tricoccum and was not included in A. burdickii vs. A. tricoccum comparisons.
Genomic DNA isolation
Dried leaf tissue (2–3 mg) was weighed and placed into a screw cap microtube with a sterilized steel bead. The tubes were placed into a Bead Bug (45 sec at 400x10) until the tissue was finely ground into a powder. Genomic DNA was isolated from the ground leaf tissue using the Thermo Scientific GeneJET Plant Genomic DNA Purification Mini Kit according to the manufacturer’s directions or using a modified CTAB method [43]. Extracted DNA was quantified with a Nanodrop-1000 Spectrophotometer and agarose gel electrophoresis. High-quality DNA samples were then diluted to 20 ng/ul and were randomly arrayed into 96-well plates.
Amplification of microsatellite loci
Microsatellite loci were amplified using multiplex PCR using five different multiplex mixes containing primers to amplify three to four loci per mix (S1 Table). Each 15 µL reaction contained 7.5 µL of 2X Qiagen Type-it Microsatellite PCR Master Mix, 0.2uM of each primer and 2.5 ng/µL genomic DNA, and dd H2O. PCR was performed in Bio-Rad 96-well plates using MiniAmp™ Plus Thermal Cycler. The PCR cycle used was an initial denaturation of 95°C for 15 min, followed by 94°C denaturation for 30 sec, 58°C annealing for 90 sec, 72°C extension for 90 sec, for 35 cycles and a final extension at 72°C for 10 min, and hold at 4°C. Negative control wells containing no DNA were included in each PCR run. PCR products were diluted 1:25 using dd H2O. 1 µl of the diluted PCR product was added to 10 µL Hi-Di formamide (Applied Biosystems) containing 0.42 µL GeneScan 500 LIZ dye size standard (Applied Biosystems), denatured for 5 min at 37°C, and run on an Applied Biosystems 3730XL DNA Analyzer at the Penn State University Genomics Core Facility (State College, PA USA) to determine microsatellite allele sizes. Chromatographs were viewed and microsatellite allele sizes were scored using the microsatellite plug-in 1.4.7 for Geneious Prime 2023.1.2.
Data analysis
Microsatellite and population genetic diversity statistics were calculated for each locus and site using GenAlEx 6.5 [44] and the poppr package for R [45]. Tests for Hardy-Weinberg equilibrium and the presence of null alleles were performed using poppr. Clones were identified and removed from populations using the clonecorrect command in poppr based on a genetic distance threshold informed by the observed distribution of pairwise genetic distances (S2 Fig), and diversity statistics were calculated for each site; when both species were present at a site, they were analyzed separately. Samples with missing genotype data were removed before clone counting. Spatial autocorrelation analysis was not feasible due to the lack of precise sub-meter spatial data. To assess the populations for linkage disequilibrium (LD), the index of association, rbard, and the p-value for the test of the null hypothesis of no LD between markers was performed in poppr, on both the non-clone-corrected and the clone-corrected data sets. Statistical comparisons between a priori determined A. tricoccum and A. burdickii sites were performed using non-parametric Mann-Whitney U test using GraphPad Prism 9.4.1.
Bayesian clustering analysis using STRUCTURE v.2.3.4 admixture model was used to examine population structure [46]. Two STRUCTURE models were run. The first was run on the entire data set (n = 593), including individuals with incomplete genotype data. The second analysis was performed on clone corrected data and any samples with missing genotype data were removed (n = 276). All STRUCTURE analyses were performed using the admixture model with assumed correlation between alleles. The number of most likelihood populations (K) was set from 1 to 9 with 10 iterations run per K value. 100,000 initial burn-in iterations were followed by 100,000 Markov Chain Monte Carlo (MCMC) replications. STRUCTURE results for multiple iterations of each K were aggregated and visualized using CLUMPAK [47]. The optimal number of clusters was determined using the Evanno Δ K method [48] using CLUMPAK [47]. Genotypes were assigned to the group that had the highest membership coefficient, as long as a coefficient (q) ≥70%, otherwise they were placed in a third group called admixed. Discriminant analysis of principle components (DAPC) was performed in R using clone corrected data on populations with 10 or more individuals with full genotype data. First a Principal Component Analysis was performed in ade4 v1.7-22 [49]. Then the number of clusters were identified and a DAPC was constructed using Adegenet v2.1.10 [50]. To test the robustness of the STRUCTURE and DAPC results, both analyses were repeated on the clone-corrected data using only three of the four loci (AT04 was dropped). To assess the distribution of genetic variation, Analysis of Molecular Variance (AMOVA) was performed using ARLEQUIN version 3.5.2.2 [51] with significance tests based on 1,000 permutations. Variation was partitioned hierarchically among A. burdickii and A. tricoccum groups, among populations within each group, and within populations. AMOVAs were performed on non-clone corrected data which included only individuals with complete genotype data and populations with ten or more individuals. The population ECUN, which appeared to be a mixed collection based on voucher specimens, was not included in the analysis. AMOVA on clone-corrected data could not be performed since some of the populations were entirely clonal and contained a single MLG.
Results
Microsatellite marker diversity
The eighteen primer pairs were initially tested for polymorphism on a subset of 95 individuals from 12 different sites across the sampled range and included both A. tricoccum and A. burdickii samples. While all primer pairs successfully amplified a product, only four loci (22.22%) were polymorphic (S1 Table). These four primer sets were used to genotype a total of 593 ramps collected at 41 sites, and complete genotype data was acquired for 576 individuals. The markers had moderate levels of variability and diversity and high evenness making them suitable for use in this study (Table 1). A total of 26 alleles were amplified and the total number of alleles per locus (Na) ranged from 3–9 alleles (mean = 6.5). Nei’s genetic diversity (He) ranged from 0.46 to 0.79 (mean = 0.64) and evenness (E5) ranged from 0.70 to 0.81 (mean = 0.77, Table 2). Simpson’s index (1949) was also calculated and was identical to He values for all markers. Tests for Hardy-Weinberg Equilibrium (HWE, S1 File) were performed separately for each marker at each site (152 tests total). Non-significant results were observed in 56 tests (36.8%) and significant deviation from HWE was found in 52 tests (34%). The allele was monomorphic in 44 tests (28.9%). Tests for null alleles based on homozygote excess indicated we cannot rule out the possibility of null alleles at each locus (S2 File).
Population clonality and diversity
Two hundred and fifteen unique MLGs were detected (S3 Fig), 17% of MLGs were found at multiple sites while 83% were found at a single site (S4 Fig). Population summary statistics were calculated for sites with ten or more individuals with full genotype data (38 sites, n = 562). Clonality varied by site. Two A. burdickii sites (B, CB) and one A. tricoccum site were completely clonal consisting of a single MLG. Expected multi-locus genotypes, (eMLG, expected number of MLGs at the smallest sample size ≥10 based on rarefaction) ranged from 1–10 and clonal richness (R) ranged from 0.06 to 1.00 (Table 2). Compared to A. tricoccum sites, A. burdickii sites had significantly fewer eMLGs (A. tricoccum mean eMLG = 7.16 vs 2.96 for A. burdickii, p = 0.0003) and significantly lower Shannon Index (H, p = 0.0006). However, A. tricoccum sites varied in clonality, with eMLGs ranging from 1–10. In addition to a high frequency of identical genotypes, another expected consequence of clonal reproduction is non-random associations between loci [52–53]. Therefore, tests for LD are commonly employed to determine the extent of clonal reproduction in clonal species [53–54]. We were able to test for LD in populations with three or more MLGs (seven A. burdickii sites, and 21 A. tricoccum sites). The null hypothesis of no LD was rejected for three A. burdickii sites (9755, CC, ECB2) and two A. tricoccum sites (9776, YH) when using the non-clone-corrected data, but the tests were not significant when using the clone-corrected data (S3 Table) which suggests that these populations are reproducing both asexually and sexually.
Simpson’s index (lambda) and Nei’s unbiased gene diversity (He) were calculated for each collection and across all collections for both non-clone-corrected and clone-corrected data (Table 2). Across all sites (both A. tricoccum and A. burdickii collections) ramps had moderate levels of genetic diversity (lambda = 0.982, lambda CC = 0.994, Hexp = 0.645, HexpCC = 0.624). However, there was less diversity at the population level. A. burdickii sites had a significantly lower mean Simpson’s index (lambda = 0.431, lambda CC = 0.568) than A. tricoccum sites (lambda = 0.744, lambda CC = 0.809, p = 0.0007 for non-clone-corrected; p = 0.0005 for clone- corrected). Mean Nei’s genetic diversity was also lower for A. burdickii collections (Hexp = 0.232, HexpCC = 0.311) than for A. tricoccum collections (He = 0.333, HexpCC = 0.367) but the difference was only significant for the non-clone-corrected data (p = 0.02). In addition, A. tricoccum sites showed a greater range in genetic diversity (HexpCC ranged from 0.000 to 0.567) compared to A. burdickii sites (HexpCC ranged from 0.125 to 0.533). Genetic diversity was the lowest at A. tricoccum sites MD, 9759, CT, and H (HexpCC ranged from 0.000 to 0.188) and the highest at sites PU, RPT, SFT, ANF (HexpCC ranged from 0.493 to 0.567). Among A. burdickii sites, the lowest diversity was found at site WP (HexpCC = 0.125) and was the highest at site CC (Hexp = 0.533). The sites that had the most unique MLGs (not found at any other site) were RPT and WE (17 and 13 MLGs) for A. tricoccum and ECB2 and RY (9 and 5 MLGs) for A. burdickii (S4 Fig).
Genetic structure
Bayesian STRUCTURE, DAPC, and AMOVA analyses showed population structure among the ramp sites. STRUCTURE analyses were performed on both the entire dataset and the clone-corrected data set, and in both cases the log probability (ln[Pr(X/K)]) [48], was maximized at ΔK = 2, indicating two genetic groups (Fig 2A). One group, shown in orange, includes all MLGs sampled at A. burdickii sites while the second group, shown in blue, contains only MLGs from A. tricoccum sites, except for one MLG collected at A. burdickii site CC, which could have been a misidentified white color morph of A. tricoccum (Fig 2B). MLGs collected at A. burdickii sites showed little evidence of admixture, but evidence of admixture was present at some A. tricoccum sites, notably sites where A. burdickii co-occurs (9759, 9770, 9776, ECT1, ECT2, EF2) and in southern sites (PU, RPT, SFT, Figs 1B and 2B). Interestingly, some MLGs collected at A. tricoccum sites contained MLGs which clustered with the orange group, with little evidence of admixture. These included sites where A. burdickii is present (9770, 9776, ECT1, ECT2, EFT) and sites where it is not known to be present (TH, TRT, PU, TRH). Site ECUN, which appeared to be a mixed collection of A. tricoccum and A. burdickii based on voucher specimens, contained MLGs clustering in both blue and orange groups and admixed MLGs. While Bayesian STRUCTURE analysis, due to its assumption of Hardy-Weinberg equilibrium, may not be the most appropriate method for examining genetic structure in a partially clonal species [55], it did show that MLGs obtained from A. burdickii populations all clustered into one genetic group (orange) and these genotypes showed little admixture with the second genetic group (blue, Fig 1B and Fig 2B).
A. Delta K values for K = 2-8 determined using the Evanno method [48]; B. CLUMPAK plot of STRUCTURE results for clone-corrected data set (n = 276 MLGs collected at 41 sites) based on four microsatellite markers for K = 2 organized by site. MLGs collected at A. burdickii sites correspond to one cluster (orange) with little admixture with the second genetic cluster (blue) except for one MLG collected at site CC which may have been misidentified at collection. The blue cluster contained only MLGs collected at A. tricoccum sites. Some MLGs collected at A. tricoccum sites were placed in the orange cluster or showed potential admixture with A. burdickii. Admixture was most prevalent at sites where A. burdickii occurred with A. tricoccum (9759, 9770, 9776, ECT2, ECT1, EFT, ECUN) and at southern sites (PU, RPT, SFT, TRH, TH).
Consistent with the STRUCTURE findings, DAPC analysis of the clone-corrected data from populations with n = 10 or more genotyped ramps showed two clusters. After 20 principal components were retained, one cluster contained only MLGs from A. burdickii sites while the second cluster contained all the MLGs from A. tricoccum sites (Fig 3). A. burdickii site CC contained three MLGs, one MLG which was shared by 82% of the CC individuals clustered with A. burdickii, and two MLGs representing 6% and 12% of CC individuals clustered with A. tricoccum. This further supports the STRUCTURE findings that some misidentified A. tricoccum individuals may have been collected at this site. Inconsistent with the STRUCTURE results, DAPC placed two A. burdickii sites (FH and WP) with the A. tricoccum cluster, whilst no A. tricoccum collections were found in the A. burdickii cluster. Site ECUN which appeared to be a mixed collection of A. burdickii and A. tricoccum based on voucher specimens, was placed in the A. tricoccum cluster by DAPC, but contained one MLG that was placed in the A. burdickii cluster. Sensitivity analysis showed the clustering observed in STRUCTURE and DAPC remained consistent even when the AT04 marker was dropped from the analyses (S5 Fig). Consistent with the previous analyses, AMOVA showed that approximately 27.29% of the overall genetic variation is found between A. burdickii and A. tricoccum site groups and high differentiation is present between the two groups (FCT = 0.27293, p < 0.00001, Table 3). Substantial percentages of overall diversity and high differentiation was also partitioned among populations within groups (33.89%, FSC = 0.46607, p < 0.00001) and within populations (38.82%, FST = 0.61180, p < 0.00001).
MLGs form two clusters corresponding to A. burdickii (left cluster) and A. tricoccum (right cluster). Dots represent MLGs, sites identified as A. burdickii at collection are shown in orange, and sites identified as A. tricoccum at collection are shown in blue, ECUN which appeared to be a mixed collection from voucher specimens, is shown in grey.
Pairwise Gst values were calculated for all sites with ten or more individuals using the non-clone-corrected data (Fig 4). At the four sites where both A. tricoccum and A. burdickii cooccur, the average pairwise Gst between the A. burdickii and A. tricoccum collections was 0.560 and varied from 0.362 (ECB2 vs ECT1 and ECT2) to 0.790 (CB vs CT) indicating significant differentiation (p ≤ 0.001 for all comparisons and significant after Bonferroni correction) at these locations between the A. burdickii and A. tricoccum collections. Mean pairwise Gst for A. burdickii sites was 0.367 versus 0.275 for A. tricoccum sites. Raw genotypic data (individual-by-locus allele scores) are included in the supporting information (S2 File).
Cooler colors indicate lower Gst and less differentiation. Gst values between A. burdickii and A. tricoccum collections at the same sites are highlighted by the black boxes.
Discussion
Ramp genetic diversity and clonality
Despite the growing interest in ramp foraging [13] and the emerging conservation concerns regarding overharvesting of wild populations, little is known about ramp genetic diversity. This information could be valuable to identify germplasm conservation priorities and contribute to future selection and breeding work. To address this need, we developed a set of 18 microsatellite markers and used them to genotype A. burdickii and A. tricoccum accessions collected from midwestern and northeastern sites comprising a significant portion of the ramp’s range in the United States. Only four of the 18 markers were polymorphic (22.2%), which is notably lower than for other Alliums including A. cepa L. (95.2%, [56]), A. fistulosum (66%, [37]), and A. sativum (62.5%, [57]). Due to the limited number of polymorphic markers, we were able to isolate, we recommend that genotyping-by-sequencing methods (e.g., MIG-Seq, [58]) be developed and deployed for future studies of ramp genetic diversity.
We found that A. burdickii populations are more clonal and less diverse than A. tricoccum, but that ramp populations, especially among A. tricoccum, can vary considerably in clonality and diversity. Although these estimates could be biased since only four polymorphic markers were used. It has been established that bulb division is the primary mode of population increase in ramps [19]. One possible explanation of our results is that sexual reproduction and seedling recruitment, even if they contribute negligibly to population increase, may be reduced in A. burdickii populations compared to A. tricoccum. Six out of 13 (46%) A. burdickii sites had fewer than three MLGs while three out of 24 (12.5%) A. tricoccum sites had fewer than three MLGs. Furthermore, the mean eMLG for A. tricoccum (7.16) was more than double that of A. burdickii (2.96). Asexual reproduction is widespread in ramps; samples were collected in a manner to avoid collecting clones, but clones were detected in all A. burdickii populations and the majority of A. tricoccum populations including the most diverse site sampled (PU, Hexp = 0.567). However, only four microsatellite markers were used to discern the genotypes so it is possible that clonality may be overestimated and therefore the differences we observed between A. burdickii and A. tricoccum sites should be confirmed using genotyping-by-sequencing.
Notably, field observations by Jones [30] and field and common garden experiments by Sitepu [33], along with our own observations, agree that A. burdickii produces fewer flowers per umbel than A. tricoccum, reducing the potential for sexual reproduction on a per plant basis for A. burdickii compared to A. tricoccum. A. tricoccum also produces larger seeds than A. burdickii [30,33] which could lead to increased germination and seedling establishment in A. tricoccum versus A. burdickii. Furthermore, A. burdickii leaves emerge later than A. tricoccum and undergo senescence before leaves of A. tricoccum [27,28,30,33], giving A. burdickii a shorter window for photosynthesis, bulb growth, and nutrient storage compared to A. tricoccum. A. burdickii bulbs are smaller than A. tricoccum [30,33] and therefore may have less resources available for flowering and fruiting compared to A. tricoccum. Also, A. burdickii flowers two to three weeks earlier than A. tricoccum and for a shorter period [28,33,59], which may result in less opportunity for sexual reproduction in A. burdickii. Earlier fruiting in A. burdickii could also lead to increased seed predation compared to A. tricoccum, which could further limit seedling recruitment. Habitat differences between A. burdickii and A. tricoccum may also favor increased sexual reproduction in A. tricoccum. Allium tricoccum prefers moister conditions than A. burdickii (32, 34, 60] and soil moisture has been shown to be an important factor in ramp growth and reproduction [61–63]. While mature A. burdickii bulbs are more deeply set and rooted than A. tricoccum, which may allow it to survive in areas of lower soil moisture than A. tricoccum, seed germination and seedling establishment and survival could be lower in these areas than in the moister soils preferred by A. tricoccum. Additional demographic and reproductive studies of both A. burdickii and A. tricoccum are needed to confirm our hypothesis of differential sexual reproduction and seedling recruitment in these species.
We found that ramp populations vary considerably in clonality and genetic diversity. One large A. tricoccum population (MD) was entirely clonal while no clones were detected in three A. tricoccum populations (ANF, RPT, WE), all of which had moderate levels of genetic diversity (Hexp ranged from 0.439–0.560). Our sampling strategy and limited number of markers did not allow us to assess how clonality and diversity is spatially distributed within individual populations but these are important studies to do in the future, especially in relation to ramp density and harvest pressure. Further research is also needed to understand what site factors might be associated with increased diversity and/or less clonality at sites. Abiotic factors such as soil fertility, soil moisture, aspect, and light availability all contribute to photosynthetic capacity, bulb growth, and subsequent nutrient sequestration in the bulb, which can influence flowering, scape survival, and seed production [19,61–63]. Furthermore, significant yearly variation in seed production and seedling recruitment was observed in a dense population of A. tricoccum over a 5-year period at a site in Quebec, including a single mast year when reproductive output was greatly increased after favorable spring conditions [19]. Temporal and spatial differences in soil moisture, plant density, competition, and flower and seed predation could also impact the establishment of new genotypes. Disturbance in space and time can impact the clonal structure of populations by temporarily reducing competition [64–65]. Therefore, low to moderate levels of disturbance, such as careful harvesting and thinning of dense patches could create opportunities for seedling recruitment and increase diversity [64–66]. However more extensive disturbance (e.g., digging large patches, fire, logging, extended drought) could lead to reduced diversity if mortality is high and seedling recruitment is low [64–65]. It is worth noting that the extensive A. tricoccum site MD, which we found to consist of a single MLG has historically been harvested heavily each spring, while the sites with the highest genetic diversity (ANF, SFT, RPT, PU) are in protected areas which either prohibit or limit ramp harvesting.
Ramp taxonomy
Despite long established differences in morphology, habitat, and phenology between A. burdickii and A. tricoccum [27,28,30,33,59,60], there is still debate over whether A. burdickii should be considered a separate species from A. tricoccum [20,26,31,32]. Our study provides the first preliminary genetic support for A. burdickii and A. tricoccum as two distinct species. Both Bayesian STRUCTURE analysis (which assumes HWE) and DAPC (which does not assume HWE) indicate two primary genetically differentiated sets of populations of ramps and that the assignment of MLGs into these genetic populations aligns closely with our species assignments at sample collection, which were based on established morphological and phenological differences between the two species [30,33]. STRUCTURE did place some A. tricoccum collections with the A. burdickii group and suggested potential admixture in populations; however, this was not the case with DAPC analysis where MLGs clearly clustered into two groups. While we are cautious to not overinterpret the STRUCTURE results, given it may not be the most appropriate tool for examining structure in clonal populations [55], it has been proposed based on morphological, habitat, and phenological differences, that additional green ramp taxa may exist [33], some which have more morphological overlap with A. tricoccum and could have been potentially collected in this study. DAPC did place two A. burdickii collections in the A. tricoccum cluster (though STRUCTURE did not). These two populations (WP, FH) are located near each other and were identified as A. burdickii based on morphological criteria by four of the authors. Additional markers or alternative genetic methods may help clarify some of these inconsistencies. High pairwise Gst values between A. burdickii and A. tricoccum collections found at the same sites, and AMOVA showing substantial genetic variation and high differentiation between the taxa further support A. burdickii as a distinct species. It should be noted that sampling of southern ramp populations was limited in this study and therefore may have affected interpretations of geographic genetic structure and admixture. While our conclusions rest on a small number of microsatellite loci and primarily on populations from the northern half of the range of the complex, the signal of genetic differentiation is quite strong. Undoubtedly more genetic markers, denser sampling of geographically separate populations across all taxa, and comprehensive geographic sampling across the ranges of all taxa (especially the southern populations), would provide greater insights into differences in genetic diversity and distribution of genetic variation within and among populations of the taxa. However, it is not likely to reverse the clear pattern of genetic differentiation into two ramps taxa shown here, and sampling in the southeastern and central Appalachian regions may provide further genetic evidence to support additional taxa detected by Sitepu. A more locus-rich genetic approach, such as Mig-SEQ (58), might yield new information regarding the evolutionary status of the intermediates and finer-scale population genetic structure and should also be done to confirm the preliminary genetic results presented here, given the low number of genetic markers used in the study and that the possibility of null alleles could not be ruled out.
Differences in phenology may contribute to reproductive isolation, as A. burdickii flowers with high consistency one or more weeks earlier than A. tricoccum, a trait first observed by Jones [30] and confirmed by Sitepu [33] and by citizen science data collected through iNaturalist [59]. Some overlap in flowering phenology may occur in late June and early July at certain sites [30] but it is unknown whether the species are interfertile. Additionally, while they sometimes co-occur, A. burdickii tends to occur on drier, more upland sites compared with A. tricoccum [30,60].
Guidance for conservation
We recommend that A. burdickii be recognized as a distinct species and should be tracked separately from A. tricoccum by land managers and botanical organizations. Widespread harvesting of A. burdickii has not been reported but the overall low diversity of populations and high clonality suggest populations could benefit from human-assisted seed sowing to promote the establishment of new genotypes. Allium tricoccum populations varied greatly in terms of clonality and genetic diversity, though even the most diverse A. tricoccum populations are only moderately diverse. Therefore, ramp harvesting strategies that minimize the loss of genetic diversity or promote diversity should be encouraged. Such strategies include limiting harvest to bulbs that are undergoing bulb division (e.g., take one, leave one); careful harvesting from dense patches to create patch gaps to promote seedling recruitment (e.g., thinning); transplanting ramps from dense patches to other areas; and collecting and sowing seeds in patch gaps or lower density areas to encourage seedling recruitment. Furthermore, sites we identified that are enriched with unique MLGs might be considered/prioritized for seed banking efforts.
Conclusion
Our data using microsatellite markers, applied primarily to northeastern and midwestern populations, show for the first time that there are two genetic lineages of ramps which correspond to A. burdickii and A. tricoccum. Our data do not support the existence of additional ramp taxa but the limited number of polymorphic markers we found may have underestimated population structure. Reproductive isolation could be maintained in A. burdickii due to differences in phenology, habitat, and reduced sexual reproduction compared to A. tricoccum. We found that ramp populations vary considerably in clonality and genetic diversity, though our estimates may be impacted by the low number of markers used to genotype and that we cannot rule out the possibility of null alleles. Ramp harvesting and stewardship practices that minimize the loss of diversity or increase diversity, such as thinning to promote sexual recruitment and seed sowing, should be promoted. Finally, while microsatellites have been successfully employed in many plant species to measure genetic diversity, due to the low number of polymorphic markers we found, other methods such as genotyping-by-sequencing should be developed and used in future genetic studies of ramps to confirm and expand on the results presented in this manuscript
Supporting information
S1 Fig. Key morphological differences between A. tricoccum and A. burdickii.
Based on Sitepu [37].
https://doi.org/10.1371/journal.pone.0332086.s001
(XLSX)
S2 Fig. Histogram of Pairwise Genetic Distances.
https://doi.org/10.1371/journal.pone.0332086.s002
(DOCX)
S3 Fig. Multi-locus genotype summary by collection (site).
https://doi.org/10.1371/journal.pone.0332086.s003
(PDF)
S4 Fig. Unique multi-locus genotypes summary.
A. Number of unique multi-locus genotypes (MLGs) and B. number of unique MLGs by site for A. tricoccum (B) and A. burdickii (C) collections.
https://doi.org/10.1371/journal.pone.0332086.s004
(TIF)
S5 Fig. Sensitivity/locus drop analysis on clone-corrected MLGs.
https://doi.org/10.1371/journal.pone.0332086.s005
(JPG)
S1 Table. Microsatellite Loci and Primer Information.
https://doi.org/10.1371/journal.pone.0332086.s006
(DOCX)
S2 Table. Collection and Voucher Information.
https://doi.org/10.1371/journal.pone.0332086.s007
(DOCX)
S3 Table. Linkage disequilibrium summary for each collection for both clone-corrected and non-clone-corrected data.
https://doi.org/10.1371/journal.pone.0332086.s008
(XLSX)
S1 File. Hardy-Weinberg Equilibrium and null alleles tests.
https://doi.org/10.1371/journal.pone.0332086.s009
(XLSX)
S2 File. Raw genotypic data and associated metadata.
https://doi.org/10.1371/journal.pone.0332086.s010
(XLSX)
Acknowledgments
We thank the many land managers, landowners, and botanists who helped direct us to sites, allowed access for sampling, or collected samples on our behalf.
References
- 1.
Kartesz JT. The Biota of North America Program (BONAP); Chaple Hill, NC]. Database [Internet]: Allium tricoccum. 2014. [Cited 27 June 2024. ]. https://bonap.net/MapGallery/County/Allium%20tricoccum.png
- 2. Houston E, Burkhart EP, Stark C, Chen X, Nilson SE. Habitat Associated with Ramps/Wild Leeks (Allium tricoccum Ait.) in Pennsylvania, USA: Guidance for Forest Farming Site Selection. Wild. 2024;1(1):63–81.
- 3.
Weakley AS. Flora of the southeastern United States. Application [Internet]: Allium tricoccum. 2024. [Cited 27 June 2024]. https://fsus.ncbg.unc.edu/main.php?pg=show-taxon.php&family=&plantname=Allium+tricoccum&habitat
- 4. Li Q-Q, Zhou S-D, He X-J, Yu Y, Zhang Y-C, Wei X-Q. Phylogeny and biogeography of Allium (Amaryllidaceae: Allieae) based on nuclear ribosomal internal transcribed spacer and chloroplast rps16 sequences, focusing on the inclusion of species endemic to China. Ann Bot. 2010;106(5):709–33. pmid:20966186
- 5.
McNeal Jr DW, Jacobsen TD. Allium. In: Flora of North America Editorial Committee, editor. Flora of North America. New York: Oxford University Press. 2002. p. 224–76.
- 6. Thorne RF. Classification and geography of the flowering plants. Bot Rev. 1992;58:225–327.
- 7. Chase MW, Reveal JL, Fay MF. A subfamilial classification for the expanded asparagalean families Amaryllidaceae, Asparagaceae and Xanthorrhoeaceae. Botanical Journal of the Linnean Society. 2009;161(2):132–6.
- 8.
Stevens PF. Angiosperm Phylogeny Website Version 14. 2017 [Cited July 2024]. http://www.mobot.org/MOBOT/research/APweb/
- 9. Nantel P, Gagnon D, Nault A. Population viability analysis of American ginseng and wild leek harvested in stochastic environments. Conserv Biol. 1996;10:608–21.
- 10. Rock JH, Beckage B, Gross LJ. Population recovery following differential harvesting of Allium tricoccum Ait. in the southern Appalachians. Biological Conservation. 2004;116(2):227–34.
- 11.
Patterson B. Monongahela National Forest: Here’s how you hunt ramps. West Virginia Public Radio. 2020 [cited 2024 July 25]. https://wvpublic.org/monongahela-national-forest-heres-how-you-hunt-ramps/
- 12.
Canadian Broadcasting Corporation Radio-Canada. Quebec man fined for illegal haul of wild leek bulbs. 2023 [cited 2024 July 25]. https://www.cbc.ca/news/canada/ottawa/wild-garlic-fine-vulnerable-species-montebello-quebec-1.6759198
- 13.
Pugh C. 2022. Ramp/leek “culture” in Northern Appalachia: a study of attitudes, behaviors, and knowledge surrounding a non-timber forest product. M.Sc. Thesis, The Pennsylvania State University. 2022. https://etda.libraries.psu.edu/catalog/24361cvp5259
- 14.
Chamberlain J, Beegle D, Connette K. Forest farming ramps. Agroforestry notes. 2014 [cited 27 June 2024]. https://www.fs.usda.gov/nac/resources/publications/index.php#agroforestry-notes
- 15.
Davis J, Persons WS. Growing and marketing ginseng, goldenseal, and other woodland medicinals. British Columbia, Canada: New Society Publishers. 2014.
- 16. Dion P, Bussières J, Lapointe L. Sustainable leaf harvesting and effects of plant density on wild leek cultivation plots and natural stands in Southern Quebec, Canada. Agrofor Syst. 2015;90:979–95.
- 17. Nault A, Gagnon D. Seasonal Biomass and Nutrient Allocation patterns in Wild Leek (Allium tricoccum Ait.), a Spring Geophyte. Bulletin of the Torrey Botanical Club. 1988;115(1):45.
- 18. Nilson SE, Burkhart EP, Jordan RT, Lambert JD. Ramp (Allium tricoccum Ait.) weight differs across the harvest season: implications for wild plant stewardship and forest farming. Agroforest Syst. 2022;97(1):97–107.
- 19. Nault A, Gagnon D. Ramet demography of Allium tricoccum, a spring ephemeral, perennial forest herb. J Ecol. 1993;81:101–19.
- 20.
Bell RA. A reexamination of species boundaries between Allium burdickii and Allium tricoccum (Liliaceae). M.Sc. Thesis, West Virginia University; 2007. https://researchrepository.wvu.edu/cgi/viewcontent.cgi?article=12145&context=etd
- 21. Nault A, Gagnon D. Some aspects of the pollination ecology of wild leek, Allium tricoccum Ait. Plant Species Biology. 1987;2:127–32.
- 22. Jump AS, Peñuelas J. Running to stand still: adaptation and the response of plants to rapid climate change. Ecol Lett. 2005;8(9):1010–20. pmid:34517682
- 23. Hughes AR, Inouye BD, Johnson MTJ, Underwood N, Vellend M. Ecological consequences of genetic diversity. Ecol Lett. 2008;11(6):609–23. pmid:18400018
- 24. Kramer AT, Havens K. Plant conservation genetics in a changing world. Trends Plant Sci. 2009;14(11):599–607. pmid:19748300
- 25. Widén B, Cronberg N, Widén M. Genotypic diversity, molecular markers and spatial distribution of genets in clonal plants, a literature survey. Folia Geobot Phytotax. 1994;29(2):245–63.
- 26. Vasseur L, Gagnon D, Simon JP. Isoenzymatic variability among populations and varieties of wild leek (Allium tricoccum). Biochem Syst Ecol. 1990;18(5):321–4.
- 27. Hanes C, Ownbey M. Some observations of two ecological races of Allium tricoccum in Kalamazoo County, Michigan. Rhodora. 1946;48:61–3.
- 28. Hanes C. Allium tricoccum Ait. var. burdickii. Rhodora. 1953;55:243–4.
- 29.
Walker LE. A study of the variation in Allium tricoccum Ait. Syracuse University. 1961.
- 30. Jones AG. A study of wild leek, and the recognition of Allium burdickii (Liliaceae). Syst Bot. 1979;4(1):29–43.
- 31. Kauffman G. Allium burdickii in the southern Appalachians: does it exist and how rare is it? Chinquapin. 2001;9(1):4.
- 32.
McNeal DW, Jacobsen TD. Allium. Flora of North America North of Mexico. Oxford University Press. 2002. p. 224–76.
- 33.
Sitepu B. An integrative taxonomic study of ramps (Allium tricoccum Ait.) complex. Ohio University. 2018.
- 34. De Queiroz K. Species concepts and species delimitation. Syst Biol. 2007;56(6):879–86. pmid:18027281
- 35.
NatureServe. NatureServe Network Biodiversity Location Data for A. burdickii narrow leaved leek. Accessed through NatureServe Explorer [web application]. Arlington, VA. 2024 [cited June 2024. ] https://explorer.natureserve.org/Taxon/ELEMENT_GLOBAL.2.155701/Allium_burdickii
- 36.
United States Department of Agriculture (USDA) Plants Database. Narrowleaf wild leek; Natural Resource Conservation Service [web application]. 2024 [cited June 2024]. https://plants.usda.gov/home/plantProfile?symbol=ALBU2
- 37. Song YS, Suwabe K, Wako T, Ohara T, Nunome T, Kojima A. Development of microsatellite markers in bunching onion (Allium fistulosum L.). Breed Sci. 2004;54:361–5.
- 38. Khar A, Lawande KE, Negi KS. Microsatellite marker based analysis of genetic diversity in short day tropical Indian onion and cross amplification in related Allium spp. Genet Resour Crop Evol. 2010;58(5):741–52.
- 39. Barboza K, Beretta V, Kozub PC, Salinas C, Morgenfeld MM, Galmarini CR, et al. Microsatellite analysis and marker development in garlic: distribution in EST sequence, genetic diversity analysis, and marker transferability across Alliaceae. Mol Genet Genomics. 2018;293(5):1091–106. pmid:29705936
- 40. Chen S, Zhou Y, Chen Y, Gu J. fastp: an ultra-fast all-in-one FASTQ preprocessor. Bioinformatics. 2018;34(17):i884–90. pmid:30423086
- 41. Faircloth BC. msatcommander: detection of microsatellite repeat arrays and automated, locus-specific primer design. Mol Ecol Resour. 2008;8(1):92–4. pmid:21585724
- 42. Brownstein MJ, Carpten JD, Smith JR. Modulation of non-templated nucleotide addition by Taq DNA polymerase: primer modifications that facilitate genotyping. Biotechniques. 1996;20(6):1004–6, 1008–10. pmid:8780871
- 43. Doyle JJ, Doyle JL. A rapid DNA isolation procedure for small quantities of fresh leaf tissue. Phytochem Bull. 1987;19:11–5.
- 44. Peakall R, Smouse PE. GenAlEx 6.5: genetic analysis in Excel. Population genetic software for teaching and research--an update. Bioinformatics. 2012;28(19):2537–9. pmid:22820204
- 45. Kamvar ZN, Tabima JF, Grünwald NJ. Poppr: an R package for genetic analysis of populations with clonal, partially clonal, and/or sexual reproduction. PeerJ. 2014;2:e281. pmid:24688859
- 46. Pritchard JK, Stephens M, Donnelly P. Inference of population structure using multilocus genotype data. Genetics. 2000;155(2):945–59. pmid:10835412
- 47. Kopelman NM, Mayzel J, Jakobsson M, Rosenberg NA, Mayrose I. Clumpak: a program for identifying clustering modes and packaging population structure inferences across K. Mol Ecol Resour. 2015;15(5):1179–91. pmid:25684545
- 48. Evanno G, Regnaut S, Goudet J. Detecting the number of clusters of individuals using the software STRUCTURE: a simulation study. Mol Ecol. 2005;14(8):2611–20. pmid:15969739
- 49. Dray S, Dufour A-B. Theade4Package: Implementing the Duality Diagram for Ecologists. J Stat Soft. 2007;22(4).
- 50. Jombart T. adegenet: A R package for the multivariate analysis of genetic markers. Bioinformatics. 2008;24(11):1403–5. pmid:18397895
- 51. Excoffier L, Lischer HEL. Arlequin suite ver 3.5: a new series of programs to perform population genetics analyses under Linux and Windows. Mol Ecol Resour. 2010;10(3):564–7. pmid:21565059
- 52. Tibayrenc M, Kjellberg F, Arnaud J, Oury B, Brenière SF, Dardé ML, et al. Are eukaryotic microorganisms clonal or sexual? A population genetics vantage. Proc Natl Acad Sci U S A. 1991;88(12):5129–33. pmid:1675793
- 53. de MeeÛs T, Balloux F. Clonal reproduction and linkage disequilibrium in diploids: a simulation study. Infect Genet Evol. 2004;4(4):345–51. pmid:15374532
- 54. Halkett F, Simon J-C, Balloux F. Tackling the population genetics of clonal and partially clonal organisms. Trends Ecol Evol. 2005;20(4):194–201. pmid:16701368
- 55. Grünwald NJ, Everhart SE, Knaus BJ, Kamvar ZN. Best Practices for Population Genetic Analyses. Phytopathology. 2017;107(9):1000–10. pmid:28513284
- 56. Rivera A, Mallor C, Garcés-Claver A, García-Ulloa A, Pomar F, Silvar C. Assessing the genetic diversity in onion (Allium cepa L.) landraces from northwest Spain and comparison with the European variability. New Zealand Journal of Crop and Horticultural Science. 2016;44(2):103–20.
- 57. Cunha CP, Hoogerheide ESS, Zucchi MI, Monteiro M, Pinheiro JB. New microsatellite markers for garlic, Allium sativum (Alliaceae). Am J Bot. 2012;99(1):e17-9. pmid:22203654
- 58. Suyama Y, Matsuki Y. MIG-seq: an effective PCR-based method for genome-wide single-nucleotide polymorphism genotyping using the next-generation sequencing platform. Sci Rep. 2015;5:16963. pmid:26593239
- 59.
iNaturalist community. Observations of Allium tricoccum, Allium burdickii from North America observed on/between June 1, 1976-May 29, 2024; 2024 [cited 29 May 2024]. https://www.inaturalist.org
- 60. Stark CJ, Nilson SE, Burkhart EP. Two Ramp Taxa, Allium tricoccum and A. burdickii, Differ in Abiotic Habitat Characteristics and Floristic Associates in Pennsylvania. Natural Areas Journal. 2023;43(3).
- 61. Vasseur L, Gagnon D. Survival and growth of Allium tricoccum Ait. transplants in different habitats. Biol Conserv. 1993;68:107–41.
- 62. Bernatchez A, Bussières J, Lapointe L. Testing fertilizer, gypsum, planting season and varieties of wild leek (Allium tricoccum) in forest farming system. Agrofor Syst. 2013;87:977–91.
- 63. Dion P-P, Bussières J, Lapointe L. Late canopy closure delays senescence and promotes growth of the spring ephemeral wild leek (Allium tricoccum). Botany. 2017;95(5):457–67.
- 64.
Sebens KP, Thorne BI. Coexistence of clones, clonal diversity, and the effects of disturbance. In: Jackson JBC, Buss LW, Cook RE, editors. Population Biology and Evolution of Clonal Organisms. New Haven, CT: Yale University Press. 1985. p. 357–98.
- 65. Eriksson O. Dynamics of genets in cluonal plants. Trends Ecol Evol. 1993;8(9):313–6. pmid:21236180
- 66. Olejniczak P. Evolutionarily stable allocation to vegetative and sexual reproduction in plants. Oikos. 2001;95(1):156–60.
