Assessment of genetic diversity and population structure in local alfalfa genotypes using iPBS molecular markers

iPBS retrotransposon markers, in other words, inter-primer binding site markers based on retrotransposon, have been helpful for the determination of genetic diversity in several plants. The study was evaluated using 10 iPBS molecular markers on the level of genetic diversity and genetic structure of alfalfa genotypes. A total of 280 alleles in 50 alfalfa genotypes (48 local genotypes and 2 commercial varieties) were obtained by 267 polymorphic markers with an average of 28 per locus, ranging from 9 to 41 alleles. The rate of polymorphism of the markers ranged from 86.36 to 100% with an average of 93.71%. The average number of polymorphic bands per marker was detected as 26.7. The mean PIC value and Dice's similarity index were calculated as 0.14 and 0.50 respectively. The results of UPGMA analysis, principal coordinate analysis (PCoA), and STRUCTURE demonstrated that the 50 alfalfa genotypes could be classified into 4 subpopulations, namely the Q1, Q2, Q3, and Q4. The Nei’s genetic distances ranged from 0.0121 to 0.0359. iPBS markers and alfalfa genotypes used in this study may be used in studies of alfalfa breeding and germplasm conservation.


Introduction
Alfalfa (Medicago sativa) is an important forage plant widely used in the world since it was cultivated in ancient times (Sakiroglu et al. 2011). Alfalfa has different combinations of genes or alleles due to cross-pollination. Since this contributes to the rich genetic diversity, alfalfa can adapt to various climatic conditions (Quiros and Bauchan 1988). Determination of genetic diversity in alfalfa is a very crucial step for elite alfalfa breeding before hybridization Abstract iPBS retrotransposon markers, in other words, inter-primer binding site markers based on retrotransposon, have been helpful for the determination of genetic diversity in several plants. The study was evaluated using 10 iPBS molecular markers on the level of genetic diversity and genetic structure of alfalfa genotypes. A total of 280 alleles in 50 alfalfa genotypes (48 local genotypes and 2 commercial varieties) were obtained by 267 polymorphic markers with an average of 28 per locus, ranging from 9 and selection. It is also fundamental for developing germplasm conservation strategies (Živković et al. 2012). In particular, it is very important to obtain and identify local genotypes in breeding strategies and germplasm characterization (Keskin et al. 2020). Germplasm sources are precious gene pools that can assess agricultural trait diversity and play significant roles in crop improvement (Yılmaz et al. 2021). There are generally two tools for germplasm characterization. These are morphological and molecular markers. DNA-based molecular markers have advantages over morphological markers such as being unaffected by environmental factors, reliability, and reproducibility (Nadeem 2021). They perform an analysis beyond the limitations of morphological markers by detecting diversity at the DNA level (Ali et al. 2020). In this context, alfalfa breeders commonly use DNA markers thanks to excluding the effects of environmental parameters to determine genetic diversities and similarities within or among populations. Different molecular marker types have been used for studies such as genetic mapping (Yu 2017), gene detection (Guertler et al. 2019), phylogeny/taxonomy (Steele et al. 2010), genetic relationship (Rhouma et al. 2017), genetic improvement (Kumar et al. 2018), molecular assisted selection (Castonguay et al. 2009) of alfalfa. Several molecular marker systems have been applied to evaluate genetic diversity in alfalfa, including amplified fragment polymorphism (AFLP) (Keivani et al. 2010), restriction fragment length polymorphism (RFLP) (Maureira et al. 2004), random amplified DNA (RAPD) (Taški-Ajduković et al. 2014), simple sequence repeats (SSR) (Sakiroglu et al. 2010), inter-simple sequence repeats (ISSR) (Touil et al. 2016), sequence-related amplified polymorphism (SRAP) (Rhouma et al. 2017) and single nucleotide polymorphism (SNP) (Rocher et al. 2015). Recently, different molecular markers based on retrotransposons (RTNs) have been developed and applied in alfalfa such as sequence-specific amplification polymorphism (S-SAP) (Porceddu et al. 2002), interretrotransposon amplified polymorphism (IRAP), and retrotransposon-microsatellite amplified polymorphism (REMAP) (Mandoulakani et al. 2012) and iPBS (Lapina et al. 2012). Inter-primer binding site (iPBS) markers based on retrotransposon have advantages compared with other retrotransposon markers because universal markers depend upon the presence of tRNA as a reverse transcriptase primer binding site, cost-effective and high efficiency (Kalendar et al. 2010;Nadeem et al. 2018). iPBS-retrotransposon markers system has been used in several crops in agriculture science (Andeden et al. 2013;Baloch et al. 2015aBaloch et al. , 2015bYildiz et al. 2015;Shirmohammadli et al. 2018). The present study aimed to evaluate the genetic diversity of 50 alfalfa genotypes using iPBS markers based on retrotransposon and to investigate the genetic differentiation level among alfalfa populations which represented 16 locations in ecological conditions of Igdir.

Materials and methods
Plant material 48 alfalfa (M. sativa L.) populations and two commercial alfalfa varieties (Sunter and Kayseri) were used in this study. The populations were collected from 16 locations in the Igdir province of Turkey. The description of the 50 alfalfa genotypes used in the present study was given in Table 1. DNA extraction, PCR amplification, and data analysis The fresh leaves of 50 alfalfa genotypes were collected from each plant for DNA extraction and kept at − 20 °C in zip-locked plastic bags until use. The total DNA was extracted from leaves using the hexadecyltrimethylammonium bromide (CTAB) method (Doyle and Doyle 1990) modified by Gulsen et al. (2005). A total of 10 iPBS primers selected from the literature were carried out for PCR reactions. (Table 2). PCR reactions are comprise of 10 μl volume containing: 4.2 μl dH 2 O, 1 μl primer (10 pmol), 1 μl 10X PCR buffer (750 mM Tris-HCI pH 8.8, 200 mM (NH 4 ) 2 SO 4 , %0.1 (v/v) Tween-20), 1 μl MgCl 2 (25 mM), 0.5 μl dNTP (2 mM), 0.3 μl Taq DNA polymerase (5 U/μl) ve 3 μl DNA (about 20 ng/μl) (Hossein-Pour et al. 2019). The iPBS PCR amplification was carried out in a thermocycler programmed as follows: 1 cycle of 94 °C/30 s pre-denaturation, 42 cycles of amplification (94 °C/25 s pre-denaturation, X °C/ 45 s annealing and 72 °C/1 min extension), and 1 cycles of 72 °C/5 mi sec final extension. PCR products were separated on 3% agarose gel in 1 × TBE buffer at 120 V for 4 h. The fragment patterns were photographed under UV light for further analysis. The several genetic diversity parameters of each iPBS marker were estimated, including the effective alleles number (Ne), Shannon's Information Index (I), fixation index (F ST ), Nei's genetic differentiation index among populations (G st ), gene flow (Nm) and genetic diversity within (H S ) population and overall gene diversity (H T ) using PopGene software 1.32 version. The gene diversity (He) and the polymorphism information content (PIC) were estimated with PowerMarker software ver. 3.25. The F ST (F-statistic) values belonging to iPBS markers were computed by Arlequin software ver. 3.5. The Dice's similarity index were computed and a dendrogram was constructed with UPGMA by NTSYS-pc software version 2.02 (Rohlf 2000). This software was used to calculate correlation (r) between the cophenetic values and the Dice's similarity index, called the Mantel test (Mantel 1967). Firstly, the cophenetic value matrix was calculated from the tree matrix of the iPBS analysis via the COPH module. The goodness fit value (r) between the UPGMA dendrogram and the Dice's similarity index matrix was examined with the MXCOMP module. Moreover, Principal coordinate analysis (PCoA) was performed via Past software to reveal genetic diversity among alfalfa populations. Population structure analysis of alfalfa was conducted with STRU CTU RE v2.3.4 (Pritchard et al. 2000) using iPBS marker data. The best K value (K subpopulations) was determined according to the protocol of Evanno et al. (2005). He and F ST values for populations were obtained from STRU CTU RE software v.2.3.4. In addition, Nm, Shannon's information index, and Ne for each alfalfa population were estimated by PopGene software 1.32 version. This software computed the distances among pairs of subpopulations to investigate the population structure of alfalfa genotypes.

DNA polymorphism and genetic diversity
A total of 280 alleles were detected by scoring gel images obtained with each of the ten iPBS molecular markers. The number of alleles per locus was estimated between 9 (iPBS-2298) and 41 (iPBS-2074) with an average of 28 alleles per locus. While 267 alleles of these markers were detected as polymorphic, the markers' polymorphism rate ranged from 86.36% for iPBS -2402 to 100% for iPBS-2377 with an average of 93.71% (Fig. 1). The average number of polymorphic bands per marker was determined as 26.7. The polymorphism information index (PIC) value for all the markers was relatively low with an average of 0.14. The highest PIC value was 0.19 for iPBS-2402. Also, iPBS-2377 and iPBS-2087 had the lowest PIC value (0.12). Maximum and minimum Nei's gene diversity (He) of iPBS primers based on retrotransposon were respectively 0.23 (iPBS-2402) and 0.12 (iPBS-2087 andiPBS-2377). Also, the average gene diversity per primer was 0.17. The other parameters such as effective alleles number (Ne), Shannon's information index (I), and F ST were calculated to evaluate the informativeness of each iPBS primer and were given in Table 2. Clustering, principal coordinate, and structure analysis of alfalfa genotypes The Dice's similarity based on iPBS data revealed that 50 alfalfa genotypes used in this study have reasonable genetic distance. The mean Dice's similarity of the genotypes was determined as 0.50. The lowest Dice's similarity coefficient was determined between genotype 11 and genotype 47, with a coefficient of 0.2258. The highest similarity coefficient was determined as 0.8090 between genotypes 19 and 20. According to this information, the most distant genotypes were determined as 11 (Saracli 2) and 47 (Tuzluca 2) genotypes. These genotypes may Fig. 1 DNA profiles of 50 alfalfa genotypes generated with two iPBS markers assist breeders in the crossing programs, but the DNA regions accessed by these markers may not be linked to the main phenotypic traits for alfalfa breeding. From this perspective, iPBS molecular markers can assist breeders in the selection, but they alone are not sufficient. This information should be synthesized with many other molecular and morphological information and breeding programs should be guided with them. The UPGMA dendrogram generated with Dice's similarities showed that 50 alfalfa genotypes were clustered into four main groups (Fig. 2).
When the UPGMA dendrogram is examined, it is seen that genotype 19 and genotype 20 collected from the Tacirli location were clustered closest to each other. Furthermore, the two commercial alfalfa varieties (Sunter and Kayseri genotypes) used in the study were clustered closely in the dendrogram. On the contrary, genotypes clustered farthest from each other in the UPGMA tree with 0.375 genetic similarities were genotype 1 and genotype 24. Considering the overall UPGMA tree, it is seen that the genotypes collected from the same locations are grouped distantly with each other and there is no clustering according to the collected regions. According to the Mantel test, the correlation between the cophenetic values and the Dice's similarities derived from 10 iPBS molecular markers was detected as r = 0.745.
To illustrate genetic diversity, the principal coordinate analysis (PCoA) has also been performed. The two-dimensional graph showed the genetic distribution of alfalfa genotypes (Fig. 3). All alfalfa genotypes were classified similarly to the UPGMA dendrogram. The cumulative sum of the first three pricipal components explains 11.93% of the total variation (Table 3).
The alfalfa data obtained from iPBS markers were used for population analysis with the Bayesian clustering approach by STRU CTU RE software. The result of Structure Harvester showed that maximum delta K was determined at K = 4. According to the results obtained from structure analysis, alfalfa genotypes were divided into 4 subpopulations (Q1, Q2, Q3, and Q4). Individuals with a membership coefficient of 0.8 and more than 0.8 were considered pure; 29 genotypes (95.45%) were assigned as pure (Gurcan et al. 2017). The largest group (Q4) consists of 23 alfalfa genotypes including two commercial varieties.
While five alfalfa genotypes (1, 8, 9, 24, and 37) were assigned to Q2, nine alfalfa genotypes (2, 3, 5, 10, 17, 25, 27, 36, and 39) were grouped into Q3. The remaining twelve alfalfa genotypes (4, 6, 7, 11, 13, 15, 16, 18, 19, 20, 28, and 33) were clustered into Q1, forming the second largest group. Considering 4 subpopulations, it is seen that the membership coefficient of a genotype (genotype 48) was close to each other (Fig. 4). Therefore, it can be said that the alfalfa genotype of 48 has the highest genetic diversity. In many previous studies and reports, model-based STRU CTU RE application has proven or stated to be more informative, robust and reliable (Bouchet et al. 2012;Nowell et al. 2018;Ali et al. 2019;Stift et al. 2019). Therefore, in this study, we care about the reliability of the STRU CTU RE analysis on behalf of the cluster algorithm for comparison.
The genetic diversity of alfalfa populations (Q1, Q2, Q3, and Q4) was evaluated with parameters such as He, F ST , Ne, I, and Nm (Table 4). Q2 population had furthermore value in terms of He, Ne, Nm ve I, except for F ST . The Nei's gene diversity index (He) in four alfalfa subpopulations varied between 0.1415 and 0.2004, with an average of 0.226 within the population. By calculating the correlation of alleles, the population difference (F ST : Fixation index) was measured and the average F ST values of each subgroup were determined as 0.27 for Q1, 0.18 for Q2, 0.26 for Q3, and 0.30 for Q4. The Q4 had the highest F ST value, meaning that subpopulation Q4 had the most genetically differentiated population of four populations. Q3 subpopulation represented with nine alfalfa genotypes had the lowest Shannon's information index when compared to other populations. The Nm value is a measure of the degree of diversity within the population. The highest Nm value was observed in the Q2 population, while the lowest Nm value was observed in the Q4 population. In addition, the average Nm value for the populations was calculated as 3.105. The Nei's unbiased genetic distances among pairs of sub-populations was measured by STRU CTU RE Software to further examine the alfalfa population ( Table 5). The distances among pairs of sub-populations changed between 0.0181 (Q3-Q4) and 0.359 (Q1-Q2).
Genetic changes in genomes can cause some changes in morphological and physiological characteristics or chemical composition (Jing-Yuan et al. 2018). Thus, the analysis of genetic diversity within Fig. 2 UPGMA dendrogram formed as a result of structure analysis of 50 alfalfa genotypes using Dice's similarity indexes ◂ plant species is crucial for plant breeding and conservation programs. The genetic variation within or among species varies over time. In addition to longtime selection and worldwide cultivation in plants, global warm and climatic changes may lead to variations in the pattern of genetic diversity in plants. The present study proposes that local 48 alfalfa genotypes collected from Igdir province reveal genetic diversity. The iPBS markers have been formerly practically Fig. 3 Principal coordinate analysis among 50 alfalfa genotypes analyzed 10 iPBS markers (+ represents the genotypes belonging to Q1, Δ to Q2, O to Q3, and □ to Q4)  (Shirmohammadli et al. 2018) and alfalfa (Lapina et al. 2012) to evaluate genetic diversity and population structure. To our knowledge, this study is among the first studies in which molecular characterization was performed in alfalfa genotypes using iPBS markers. In our study, the banding pattern obtained using 10 iPBS markers based on retrotransposon of local alfalfa genotypes exhibited a total of 267 polymorphic bands from 280 iPBS bands. The average number of bands per iPBS primers (26.7) was higher than reported by Baloch et al. (2015a), Aydin and Baloch (2019), Andeden et al. (2013), Baloch et al. (2015b), and Belttar et al. (2017). To interpret the genetic variation among alfalfa genotypes, the PIC value of iPBS markers was computed. The average PIC value in this study (0.14) is lower than those reported in previous studies (Baloch et al. 2015a;Gedik et al. 2017;Bulunuz Palaz et al. 2022;Andeden et al. 2013). Mandoulakani et al. (2015) reported that the average PIC value in 80 genotypes belonging to 8 different alfalfa populations using 10 IRAP and 14 REMAP retrotransposon markers were respectively 0.14 and 0.12, which is an approximate value of 0.14 obtained in our study. The PIC value is used to interpret the efficiency of polymorphic loci in explaining genetic variation among genotypes. In the current study, the average PIC value was determined as 0.14, varying between 0.19 and 0.12. These PIC values were lower than reported in previous alfalfa studies using other markers and characterization studies in different plant species using iPBS markers. For example, Rhouma et al. (2017) reported that the average PIC value was 0.901 in 110 alfalfa individuals using ten SRAP primer combinations to detect genetic diversity. Andeden et al. (2013) used 10 iPBS and 10 ISSR markers to demonstrate genetic diversity among chickpea genotypes and found the average PIC value of iPBS primers to be 0.91. The average genetic diversity (he) (0.17), the effective number of alleles (0.26), and Shannon's information index (0.30) for iPBS markers were lower than reported by Karik et al. (2019) and Yildiz et al. (2015), but higher than Borna et al. (2017).
Genetic similarity was computed for the variation estimation of the genotypes among the 50 local alfalfa genotypes and the mean genetic similarity found as 0.50, ranging from 0.22 to 0.80. A wide range of similarity coefficients shows to be highly  polymorphic of iPBS-retrotransposon-based markers at the DNA level. Our results indicated that the range of Dice similarity coefficient was wider than in previous studies carried out using several marker types to reveal genetic diversity in different alfalfa populations (Mandoulakani et al. 2012;Bolourchian et al. 2013;Rashidi et al. 2013). This may be due to the use of different marker systems and different alfalfa genotypes. According to genetic relationship analysis, our results showed that 48 local alfalfa genotypes compared to commercial varieties have moderately genetic variation. Considering genetic similarity coefficients, genotype 11 and genotype 47 which are most genetically distant from each other can be selected as parents in future breeding studies. UPGMA dendrogram was seen not to separate 50 analyzed genotypes into main group local alfalfa varieties. The UPGMA dendrogram constructed using the Dice's similarity showed that 50 alfalfa genotypes were divided into four groups assumed with STRU CTU RE analysis. In addition, the Dice's similarity coefficient and UPGMA dendrogram also revealed that there was a reasonable genetic diversity among alfalfa genotypes collected from Igdir province. According to the Mantel test, the correlation coefficient (r) between the cophenetic values derived from the UPGMA dendrogram and the Dice's similarity matrix was 0.7453. Anbaran et al. (2007) determined the correlation coefficient value as r = 0.88 in their study to determine genetic diversity with SSR markers in 6 alfalfa genotypes cultivated in different geographical regions of Iran. Our Mantel test result revealed a moderate cophenetic correlation (r = 0.74), indicating a reasonable fit between the dendrogram and similarity matrix. To further evaluate relationships among genotypes and population structure, PCoA, UPGMA and STRU CTU RE analysis were performed using data generated from iPBS markers. These two analyses were classified into four groups (Q1, Q2, Q3, and Q4) to alfalfa genotypes consistent with the result of UPGMA analysis except for slight differences. The output of population genetic structure analysis of local alfalfa genotypes is presented in Fig. 4. When based on the ΔK method, the highest peak value was obtained at K = 4 displaying the presence of four subpopulations of alfalfa genotypes. Alfalfa genotypes having a membership coefficient less than 0.79 were regarded as admixture. 21 genotypes in the present study demonstrated an admixture membership pattern. To analyze the genetic variability of the 50 alfalfa genotypes, the He, F ST , Ne, I, and Nm values were calculated by PopGene 1.32 for the alfalfa populations. The average value of He, F ST , Ne, I, and Nm for four populations was respectively 0.1673, 0.25, 1.25, 0.24, and 3.10 (see Table 4). We showed that Nei's genetic identity and distance were different among the assumed four populations ( Table 5). The genetic identity changed between 0.9647 and 0.9820. Also, the genetic distance ranged from 0.0121 to 0.0359, displaying that there was a high degree of distances among the different populations. These values are comparable with results reported by Živković et al. (2012), which evaluated genetic diversity among alfalfa genotypes using RAPD molecular markers. According to Nei's gene diversity and Shannon information index, our results indicated that the genetic differentiation in the assumed four populations was relatively low. Moreover, the F ST value indicating population differentiation measurements changed between 0.18 and 0.30. Q4 showed the highest average value of 0.30, followed by Q1 with 0.27, Q3 with 0.26, and Q2 with 0.18. Taken together with PCoA and population structure analysis, these results supported that Q4 was the most diverse one among the groups of alfalfa genotypes.

Conclusion
As far as we know, the genetic diversity studies in alfalfa by using iPBS molecular markers are limited. We successfully performed molecular characterization of alfalfa genotypes using iPBS markers. The results showed that the alfalfa genotypes create a significant genetic variation. The two of these genotypes can be potential candidates to obtain new varieties thanks to the highest cofficient. Consequently, iPBS molecular markers can be a useful tool in identifying parents with high genetic diversity in alfalfa breeding.