Genetic diversity, population structure, and gene flow analysis of lowland bamboo [Oxytenanthera abyssinica (A. Rich.) Munro] in Ethiopia

Abstract Bamboo, a member of subfamily Bambusoideae in the grass family (Poaceae), is one of the most important nontimber forest resources and a potential alternative to wood and wood products. Ethiopian lowland bamboo (Oxytenanthera abyssinica) is an economically and ecologically important species which accounts about 85% of total bamboo coverage in the country. This species is experiencing population decline due to a number of anthropogenic factors. As a foundation step, genetic diversity, population structure, and gene flow analysis of various O. abyssinica populations found in Ethiopia are studied using inter‐simple sequence repeat markers. One hundred and thirty isolates of bamboo belonging to 13 geographically diverse populations were collected for DNA extraction and analysis. Heterozygosity, level of polymorphism, marker efficiency, Nei's gene diversity (H), and Shannon's information index (I) analysis, analysis of molecular variance (AMOVA), analysis for cluster, principal coordinates (PCoA), and admixture analyses were performed based on the markers banding pattern. The results indicated high genetic variation (84.48%) at species level. The H, I, observed and effective number of alleles at the species level were 0.2702, 0.4061, 1.8448, and 1.4744, respectively, suggesting a relatively high level of genetic diversity. However, genetic differentiation at the population level was relatively low. Using grouped populations, AMOVA revealed that most (61.05%) of the diversity was distributed within the populations with F ST = 0.38949, F SC = 0.10486, and F CT = 0.31797. Cluster analysis grouped the populations into markedly distinct clusters, suggesting confined propagation in distinct geographic regions. STRUCTURE analyses showed K = 2 for all populations and K = 11 excluding Gambella population. Using these markers, we found strong evidence that the genetic diversity of the lowland bamboo is associated with distinct geographic regions and that isolates of Gambella Region, with their unique genetic origin, are quite different from other bamboos found in the country.


| INTRODUC TI ON
Bamboo is a member of the grass family, Poaceae, and constitutes a single subfamily Bambusoideae with 121 genera and 1,662 species (Canavan et al., 2016;Vorontsova, Clark, Dransfield, Govaerts, & Baker, 2016). It is the fastest-growing plant in the world (up to 100 cm per day; Tao, Fu, & Zhou, 2018) and is one of the most important nontimber forest resources and a potential alternative to wood and wood product (Ekhuemelo, Tembe, & Ugwueze, 2018).
Currently, 100 species are commercially cultivated around the world  Bamboo is a multi-purpose plant, with over 10,000 documented uses and applications. It has rapid regeneration capacity and the possibility of annual harvesting within a few years of planting offers significant advantages over the other forest species (Akinlabi, Anane-Fenin, & Akwada, 2017;Diver, 2006). Bamboo has also been proven to address many global challenges and contributes to the following United Nations Sustainable Development Goals: SDG 1 (no poverty), 7 (affordable and clean energy), 11 (sustainable and resilient housing), 12 (efficient use of resources), 13 (address climate change), and 15 (life on land) (Bau & Trinh, 2019;Ekhuemelo et al., 2018;Kaushal et al., 2018;Yuen, Fung, & Ziegler, 2017). Due to its potential for water recharge and mitigation soil erosion, bamboo also provides an opportunity for watershed development and restoration of degraded areas (Kaushal et al., 2019(Kaushal et al., , 2020. Furthermore, bamboo is a fodder for livestock and food for humans contributing to ensuring food security (Andriarimalala, Kpomasse, Salgado, Ralisoa, & Durai, 2019;Choudhury, Sahu, & Sharma, 2012;Halvorson, Cassida, Turner, & Belesky, 2011;Mulatu, Bahiru, Kidane, Getahun, & Belay, 2019;Nongdam & Tikendra, 2014). Bamboo has huge economic potential; the global production and local consumption are worth an estimated 60 billion USD, and the international export of the material is valued at USD 2 billion per annum (International Network for Bamboo & Rattan, 2019).
The major species richness of bamboo is found in Asia-Pacific region, followed by South America. Africa has the fewest number of species is (Bystriakova, Kapos, Lysenko, & Stapleton, 2003), while Europe and Antarctica have no native bamboo species (Zhao et al., 2018). According to the Food and Agriculture Organization's thematic world bamboo resources assessment report and regional remote sensing assessment of bamboo resources in Ethiopia, Kenya, and Uganda, these three countries possess the majority of the bamboo resources in Africa by (Lobovikov et al., 2007;Zhao et al., 2018). Africa is home to 43 species, 40 of which are found mainly in Madagascar, with the remaining three found in mainland Africa (Embaye, 2000). Two indigenous woody bamboo species grow in Ethiopia: the monotypic genus lowland bamboo (Oxytenanthera abyssinica [A. Richard] Munro) and the African Alpine Bamboo, or highland bamboo (Yushania alpina K. Shumann Lin; synonym: Arundinaria alpina, Oldeania alpine K. Schumann). These two species are indigenous to Ethiopia and endemic to mainland Africa (Embaye, 2000;Ensermu, Tamrat, Alemayehu, & Gebremedhin, 2000). Ethiopia contributes more than 1.47 million hectares of bamboo coverage (Zhao et al., 2018), which accounts for about 67% of the total bamboo coverage in the continent and 7% of the global coverage (Embaye, 2000). The lowland bamboo (O. abyssinica) accounts for 85% of the total national coverage, while the highland bamboo (A. alpina) accounts for the remaining 15% (Embaye, 2000;Embaye, Christersson, Ledin, & Weih, 2003;Lobovikov et al., 2007). The lowland bamboo grows in an elevation a range of between 540 to 1,750 m and highland bamboo at a higher elevation above 2,480 m (Zhao et al., 2018).
Assessing the genetic variability of a species within and among different populations is important to devise mechanisms for effective identification, conservation, and multiplication of suitable genetic materials. The genetic diversity of bamboos has not been adequately explored, and relatively limited numbers of molecular finger printing studies have been conducted. The main reason for this exploration of limited genetic diversity is related to the difficulty in assessing the phenotypic variability of clones . Molecular marker techniques such as random amplified polymorphic DNA (RAPD), inter-simple sequence repeats (ISSR), and amplified fragment length polymorphism (AFLP), simple sequence repeat (SSR), expressed sequence tag derived simple sequence repeat (EST-SSR), sequence-related amplified polymorphism (SRAP), restriction fragment length polymorphism (RFLP), and inter-retrotransposon amplified polymorphism (IRAP) have been used for characterization of some bamboo germplasm Ma et al., 2013;Nag et al., 2013;Nilkanta, Amom, Tikendra, Rahaman, & Nongdam, 2017;Tian, Yang, Wong, Liu, & Ruan, 2012;Yang, An, Gu, & Tian, 2012) around the world. ISSR molecular markers are widely used for population genetic analysis of different plants, generating more reliable and reproducible bands than RAPD (Nagaoka & Ogihara, 1997;Zhang & Dai, 2010). Use of ISSR analyses showed K = 2 for all populations and K = 11 excluding Gambella population. Using these markers, we found strong evidence that the genetic diversity of the lowland bamboo is associated with distinct geographic regions and that isolates of Gambella Region, with their unique genetic origin, are quite different from other bamboos found in the country.

K E Y W O R D S
bamboo, genetic differentiation, ISSR primers, Oxytenanthera abyssinica, population structure is also technically simpler, quicker, and more cost-effective (Oumer, Yohannes, Kassahun, Abel, & Endashaw, 2015;Tesfaye, Govers, Bekele, & Borsch, 2014) than RFLP, SSR, and AFLP markers, as no previous sequence information is required to generate DNA amplification products (Mukherjee et al., 2010;Tian et al., 2012). ISSR markers are observed to be highly variable within the species and reveal many more polymorphisms since they use longer primers that allow more stringent annealing temperatures (Hillis, Moritz, & Mable, 1996).
Genetic erosion of bamboo and their wild relatives are accelerating at a higher rate because of human activities such as deforestation, wild fire, overexploitation, and the introduction of exotic species without proper research on the potential impact of genetic pollution and problems associated with the transfer of exotic germplasm (Canavan et al., 2016;Tian et al., 2020;Xu et al., 2020). Starting from 2007, Ethiopia has introduced 23 new bamboo species belonging to seven genera (Mulatu, Alemayehu, & Tadesse, 2016). These species are under multiplication at the Holetta and Gurd-Shola nurseries of the Central Ethiopia Environment and Forest Research Center (CE-EFRC), Addis Ababa. Overall, it is estimated that about 40 bamboo species are introduced to the country and are being multiplied for planting and/or planted in different geographical locations of the country. Such a massive introduction of exotic species will accelerate the genetic erosion of the native bamboo in the country.
About half of the world`s woody bamboo species are vulnerable to extinction as a result of massive forest destruction (Bystriakova, Kapos, & Lysenko, 2004). In addition to the genetic erosion by ex- and is a prime example of human activities contributing to bamboo deforestation in the country. Moreover, lack and/or gap of knowledge on the plant's biology, genetics, and techniques and technologies for value addition constituting a failure to fulfill the plant's economic potential has also led to unsustainable management of the bamboo plant. The lack of research conducted in Ethiopia, especially on the diversity and systematics of O. abyssinica (the species with great ecological and industrial benefit and great coverage in Africa) at DNA level, prompted the commencement of this research. Therefore, in the present study, we used 19 ISSR primers out of 108 initial screenings and 38 final screening of ISSR primers aiming to assess the genetic diversity, population structure, and gene flow analysis of O. abyssinica collected from lowland bamboo growing areas in Ethiopia.

| Plant material collection and sampling strategy
Young leaves (5-7 in number) from 130 individuals belonging to 13 populations (each population represented by 10 individual bamboo culms) were collected from the natural lowland bamboo (O. abyssinica) growing areas of Ethiopia. Since 64.07% of the country's lowland bamboo grows in the BGR (Zhao et al., 2018), nine populations were collected from this region and the remaining four populations were collected from three other bamboo growing regions. GPS data and altitudinal information for each population are presented in Table 1. Maps of collection sites are shown in Figure 1. 2.2 | DNA extraction, testing gels, DNA normalization, and primer screening A total of 130 O. abyssinica individuals were used in the study. Five to seven young fresh leaves were preserved in zip-lock plastic bag with the appropriate amount of silica gel. An amount of 100 mg of silica gel dried leaves was crushed by a mixer and miller (Retsch Mixer Mill MM 400) in the presence of three 3.2 mm diameter stainless steel beads in a 2-ml sterile centrifuge tube. Genomic DNA was isolated separately using a modified 2% cetyl trimethyl ammonium bromide (CTAB) DNA isolation method at Plant Molecular Biology Laboratory (PMBL) and Plant Genetics Research Laboratory (PGRL) of Addis Ababa University.
One milliliter preheated 2% CTAB extraction buffer (2% CTAB, 100 mM Tris-Base pH 8.0, 25 mM Na 2 -EDTA, 2 M NaCl, 250 mg/ ml PVP and 2% β-mercaptoethanol) was added to the tube containing crushed leaf powder. The dissolved CTAB mix was incubated for 30 min at 65°C. The tubes were gently inverted every 10 min. Seven hundred microliters of the supernatant (clear liquid only) was transferred to a new sterile centrifuge tube using blue pipette tips, which were cut. Seven hundred microliters of chloroform was added, mixed thoroughly, and centrifuged at 16,000 g for 10 min at 26°C. Six hundred microliters of the supernatant was transferred to the new fresh Eppendorf tube, and 60 µl of 3 M sodium acetate (pH 5.2) was added and thoroughly mixed. Six hundred microliters of ice-cold isopropanol was added and gently mixed by inverting the tubes 3-5 times, and then, the tubes were placed in a refrigerator at −20°C for 2 hr. The mix was centrifuged at 16,000 g at 4°C for 5 min to precipitate the DNA.
The supernatant was discarded, and the DNA was washed using 1 ml of 70% ethanol by dissolving the pallet completely in the wash buffer and centrifuged at 16,000 g for 3 min at 4°C. The wash step was repeated by cold absolute ethanol (1 ml), and the pellet was air-dried.
The pellet was dissolved in 60 µl 0.1X TE (10 mM Tris-HCl pH 8.0 and 1 mM EDTA pH 8.0) buffer containing RNase. The concentration and quality of DNA were checked by test gel electrophoresis in 1% agarose and measured by using NanoDrop (Thermo Scientific NanoDrop 2000 Spectrophotometer). Each sample was measured, and those with high DNA quality were used for PCR amplification after normalization of each sample to a concentration of 100 ng/μl. Gel documentation was taken by Bio-Rad Gel Doc™ EZ System Imager.
Thirty-eight primers (35 designed from the University of British Colombia (UBC) and 3 from previous work) out of 108 ISSR primers (100 from UBC and 8 from the literature) were screened for the initial testing of polymorphism and reproducibility. Nineteen primers were found to be reproducible and polymorphic and used for further ISSR-PCR work for the study. Among these 19 ISSR primers, ten were dinucleotide, two were tri-nucleotide, two were tetra-nucleotide, three were penta-nucleotide, and two were 5′ anchored primers. Primers were further categorized into 3′ anchored, 5′ anchored, and unanchored based on anchorage property (Table 2).

| ISSR-PCR amplification and gel electrophoresis
Each DNA amplification reaction was performed in a final volume of 10 µl containing 5.5 µl of 2x Taq plus Master Mix (containing Taq DNA polymerase, dNTPs, MgCl 2 , PCR buffer, PCR reaction enhancer, stabilizer, and a blue tracer dye), 3.5 µl of ddH 2 O, 0.5 µl of ISSR primer (0.2 pmol/µl), and 0.5 µl of normalized genomic DNA (100 ng). The blue dye and a stabilizer of 2x Taq plus Master Mix helped to directly load the final products onto a gel for analysis. The thermal profile included pre-PCR denaturation at 94°C for 4 min followed by 35 cycles of denaturing at 94°C for 30 s, annealing for 30 s at the optimized temperature (details are given for each respective primer given in Table 2), extension at 72°C for 1 min, and a final extension at 72°C for 10 min. The PCR products were stored at 4°C until loading on agarose gel electrophoresis. Five microliters of the ISSR-amplification product of each sample was loaded on 1.67% agarose gel and in 0.5x TBE buffer at a constant voltage of 100 V for 1:30 -2 hr. The agarose gel was stained with 3.0 µl ethidium bromide after dissolving the agarose powder in 0.5% TBE buffer. The ISSR bands were visualized and photographed under Bio-Rad Gel Doc™ EZ System Imager that was connected to a PC with Image Lab software and stored for later data scoring. To estimate the molecular sizes of the resolved fragments, a 100 bp DNA marker was used.

| Scoring of bands
ISSR bands were scored manually. According to the weight of the DNA ladder (100 bp), the same weight bands were marked as a line.
The bands that were clearly visible and repeatable on the electrophoresis map were marked as "1," the absence of a band at the same site was marked as "0" and ambiguous bands were considered as a missing data and marked as "?". Intensity variations among fragments having approximately the same molecular size were not considered although in some cases intensity differences of the bands were observed. A binary data matrix was compiled with individuals in the column and the ISSR markers in the row for each primer set and vice-versa according to the requirements of the software. Each amplified fragment was named by the code of the primers across the row and/or column followed by the Arabic numbers starting from the fragment having high molecular weight to the fragments with low-molecular weight. Both the total number of bands amplified by each primer and the number of polymorphic bands were calculated.

| Band pattern frequency, markers efficiency, and gene diversity analysis
On the basis of the recorded band profiles, different software was employed for data analysis. POPGENE version1.32 (Yeh & Boyle, 1999) and GenAlEx (genetic analysis in excel) version 6.5 (Peakall & Smouse, 2012) were used to calculate genetic diversity for each TA B L E 1 Details of codes, geographic locations and GPS coordinates of bamboo collections population as the number of polymorphic loci, percent polymorphism, gene diversity, and Shannon diversity index. GenAlEx6.5 was also used to calculate band patterns on its frequency and polymorphism. The polymorphism information content (PIC), marker index (MI), expected heterozygosity (H), and discriminating power (D) were inferred via iMEC (https://irsco pe.shiny apps.io/iMEC/) online.

| Proportion of genetic variability within and among populations
Analysis of molecular variance (AMOVA) was used to calculate the F-statistics that were used to estimate the proportion of genetic variability found among populations (F ST ), among populations within groups (F SC ), and among groups (F CT ) using Areliquin version 3.01 (Excoffier & Lischer, 2006) and GenAlEx6.503 (Peakall & Smouse, 2012 TA B L E 2 Details of ISSR primers used for diversity analysis in this study

| Clustering and admixture analysis
NTSYS-pc version 2.02 (Rohlf, 2000) was used to generate the unweighted pair group method with arithmetic mean (UPGMA) phenogram, allowing for a sequential, agglomerative, hierarchical, and nested (SAHN) cluster analysis using the similarity matrix and compare the individual genotypes. The neighbor-joining (NJ) method (Saitou & Nei, 1987;Studier & Keppler, 1988) was used to compare individual genotypes and evaluate patterns of genotype clustering using Free Tree 0.9.1.50 Software (Pavlicek, Hrda, & Flegr, 1999) and TreeView (Page, 1996). Patterns of genetic variation among individual samples were also further examined in three dimensions with the help of principal coordinate analysis (PCoA) on the basis of Jaccard's coefficients of similarities, which was calculated using PAST software version1.18 (Hammer, Harper, & Ryan, 2001

| ISSR marker banding patterns
Based on the results from gel pictures taken for each primer, the pattern of DNA amplification obtained was clear and reproducible ( Figure 2) even though scoring of bands and interpretation of certain gel pictures were challenging. The size of the band generated ranged from 100 to 1,700 bp ( Based on the anchorage property of primers, anchored primers produced 221 bands and unanchored primers produced 127 bands. 3′ anchored (are also dinucleotides) produced 185 bands, whereas 5′ anchored primers produced 56 bands (Table 3).

| Analysis of markers efficiency
The highest expected heterozygosity ( and D values were observed in tri-nucleotides and 5′ anchored primers. Again, dinucleotide and anchored primers showed higher D values than other repeat motif types and unanchored primers (Table 4).

| Band pattern and heterozygosity
The highest number of band patterns was observed in the Konta population (227) (Table 5).

| Genetic polymorphism and Shanon's information index
Among the thirteen populations, the Guba population was identi-

| Analysis of molecular variance
Analysis of molecular variance (AMOVA) was carried out in two phases; the first phase focused on the entire population's overall loci by considering them as one geographic region and the second phase separated populations into seven groups based on administrative Zones. The analysis was carried out by computation of the distance between "haplotypes," each individual's data pattern as one "haplotype" and computing variance components for each level (Excoffier & Lischer, 2006).

| Principal coordinate (PCoA) analysis
The data obtained from 19 ISSR primers were used in PCoA analysis using Jaccard's coefficients of similarity for grouping of individuals and clustering of O. abyssinica using three coordinates (Figure 6a

| Admixture analysis
The Bayesian approach-based assignment of the 130 individual clumps to different populations and determination of their population structure (Evanno et al., 2005), using STRUCTURE outputs,

| D ISCUSS I ON
The ultimate goal of this study is to develop genomic tools and re-

| ISSR markers for the genetic polymorphism in O. abyssinica populations
Out of the total 348 scorable bands produced with a total of 19 primers, 294 bands were polymorphic. In terms of the number of polymorphic  to measure the extent to which specific markers contribute to this inference (Amom et al., 2020;Rosenberg, Li, Ward, & Pritchard, 2003).
Several approaches have been developed previously for measuring polymorphism information, but a user-friendly platform to calculate this information is missing or otherwise inaccessible (Nagy et al., 2012).
The iMEC (marker efficiency calculator) created by a group of researchers (Amiryousefi, Hyvönen, & Poczai, 2018) is coded in R and is available as a web application helps to detect markers for lots of genetics researches. For the present study, the highest PIC, MI, H, and D were Ng and Tan (2015), anchored and dinucleotide repeat primers showed more polymorphism and discriminating power than unanchored primers. It is noteworthy that dinucleotide repeats, anchored either at the 3′ or 5′ end, usually revealed high polymorphism and the primers anchored at the 3′ end gave clearer banding patterns compared to those anchored at the 5′ end (Joshi, Gupta, Aggarwal, Ranjekar, & Brar, 2000;Pradeep Reddy, Sarla, & Siddiq, 2002;Tarinejad et al., 2015).

| Genetic differentiation and population structure
The high number of alleles and high polymorphism are very important for the correct estimation of the genetic diversity of germplasm. The degree of polymorphism showed the extent of diversity and effectiveness of the markers (Chesnokov & Artemyeva, 2015), and consequently, polymorphic information is related to expected heterozygosity and is usually determined from allele frequency. In the present study, the largest NPL (164 and 158), the highest PPL       (Miyazaki, Ohnishi, Takafumi, & Hiura, 2009).

| Levels of genetic diversity among and within populations
This may follow a deficit of genetic variation within a population (Wong, 2004).
Our results of AMOVA with respect to grouping and without grouping populations are consistent with several works on bamboo. A study by Li et al. (2020) on IRAP marker-based genetic diversity and population structure of 58 Phyllostachys accessions (Asian bamboo) using 16 primers reported that 75% of variation was within the population and 25% among the populations. The Many factors can determine the genetic structure and differentiation of plant populations, including reproductive biology, natural selection, genetic isolation or genetic drift, geographic distribution range, and gene flow (Hogbin & Peakall, 1999;Loveless & Hamrick, 1984;Schaal, Hayworth, Olsen, Rauscher, & Smith, 1998).
Many ISSR, RAPD, and sequence-tagged microsatellite sites (STMS)based genetic analyses showed that long-lived, out-crossing taxa retained most of their genetic variability within populations (Meena et al., 2019;Nilkanta et al., 2017;Nybom, 2004). The woody bamboos have a long vegetative phase of 20-150 years and typically a species with one of the greater longevities of the grass family (Ma et al., 2013). As one of those critical influences, the out-crossing of a plant species tends to explain 10%-20% of the genetic variation among populations, whereas the selfing of a species leads, on average, to 50% variation between populations (Miyazaki et al., 2009). Oxytenanthera abyssinica can reproduce via seed in the wild, although this phenomenon is rare, and the rate of seed setting is low (Embaye, 2000;Ensermu et al., 2000;Zhao et al., 2018).
Furthermore, studies on floral biology have indicated that O. abyssinica is likely anemophilous and prone to being an out-crosser (Thakur, Barthwal, & Ginwal, 2016), which also was supported by the genetic differentiation (G ST = 0.2442) that was similar to the average of out-crossing species (G ST = 0.22) (Nybom, 2004).

| Clustering and admixture analysis
ISSR markers analysis for the genetic relationship study of five native and economical important bamboos (Amom et al., 2020), for the genetic diversity and structure of Dendrocalamus hamiltonii (Meena et al., 2019) and for the population genetic study of Melocanna baccifera (Roxb.) Kurz (Nilkanta et al., 2017), revealed that most of the populations were clustered in accordance with their geographical distribution and location. In our results, populations were also clustered in accordance with their geographical distribution and location. Most primer combinations in terms of nucleotide repeats and anchorage property showed almost similar clusters the overall assessment. But tri-and penta-nucleotide repeat ISSR primer and anchored primers resulted in the poorest tree topology. The reasons that tri and penta-nucleotides reveal the poorest information may be due to the availability of SSRs in small numbers when compared to other repeats motifs (Zhao et al., 2015).
Admixture results via STRUCTURE and web-based data retrieval from structure harvester and CLUMPAK with and without Gambella samples showed different delta K values (K = 2 and K = 11) with an optimum population structure consisting of two and 11 sub-populations. The delta K = 11 value is in agreement with the result of XLSTAT clustering the Ethiopian lowland bamboo into their geographic location. This implies that samples collected from the Gambella Region are different from others and may indicate a possible existence of additional bamboo species in the country.

| CON CLUS ION
The aim of the present study was to explore the extent of genetic diversity, population structure, and gene flow analysis of O. abyssinica populations and to devise mechanisms for effective identification, conservation and sustainable use of bamboo resources in Ethiopia using ISSR markers. Although a relatively high level of genetic diversity exists at species level, and a relatively low genetic differentiation was observed at the population level, most of the diversity was distributed within the populations, and cluster analysis grouped the populations into sharply distinct clusters, all of which could be attributed to the plant's cross-pollination nature and long-standing presence in the area. Using these ISSR markers, we found strong evidence linking geographic origin with diversity. The Metekel Zone in particular is in need of a conservation strategy, as it was found to have the most diverse population and samples from the Gambella Region were found to be different from those taken from others regions, indicating the availability of additional bamboo species in the country. Though the present investigation yielded some information on the genetic diversity of O. abyssinica populations in Ethiopia using ISSR markers, there is a need for further critical work involving molecular markers giving greater genome coverage to improve our understanding of genetic diversity in bamboo at the species and population level.

ACK N OWLED G M ENTS
The authors would like to thank the regional and district level natural resource experts for their contribution to the collection of samples from naturally grown bamboos. We also would like to thank CE-EFRC for providing us with chemicals and with vehicle for collecting samples across the country under what were often difficult conditions.

CO N FLI C T O F I NTE R E S T
The authors declare no conflict of interest.

O PE N R E S E A RCH BA D G E S
This article has been awarded <Open Data, Open Materials, Preregistered Research Designs> Badges. All materials and data are publicly accessible via the Open Science Framework at https://doi. org/10.5061/dryad.b5mkk wh8z.

DATA AVA I L A B I L I T Y S TAT E M E N T
Oxytenanthera abyssinica (A. Rich.) Munro; lowland bamboo (Poaceae, Bambusinea) in Ethiopia: Genetic diversity, population structure and gene flow analysis submitted with https://doi.org/10.5061/dryad. b5mkk wh8z.