Diversity of Sweet Basil Accessions from Croatian National Plant Gene Bank Based on Amplified Fragment Length Polymorphism Markers

Abstract

This study investigates genetic diversity among five morphotypes and five chemotypes of Ocimum basilicum L. (sweet basil) using amplified fragment length polymorphism (AFLP) markers. Conducted on 80 basil accessions from the Collection of Medicinal and Aromatic Plants of the National Plant Gene Bank of the Republic of Croatia, this research aims to enhance the conservation and utilization of sweet basil’s genetic resources. AFLP analysis using extracted genomic DNA revealed high levels of polymorphism, particularly within the True basil morphotype, which displayed 95.6% polymorphic markers. The results showed genetic differentiation between the morphotypes, particularly between the ’green’ and ’purple’ groups, and within certain chemotypes, such as the High-linalool chemotype (Chemotype A). Principal coordinate analysis (PCoA) and Bayesian clustering further highlighted the genetic structures, with some admixture observed, particularly in the Purple basil B morphotype. Analysis of molecular variance (AMOVA) indicated that most of the genetic diversity was between accessions, emphasizing the value of individual variability. These findings underscore the genetic potential within sweet basil accessions, offering valuable insights for future breeding programs aimed at selecting basil cultivars with tailored biochemical and morphological traits suited for pharmaceutical, culinary, and ornamental applications. The study provides an important basis for the conservation and improvement of basil genetic resources.

1. Introduction

The use of medicinal plants ranges from ancient traditions to modern medicine and has a significant impact on human culture and health [1,2]. Ensuring the safety, quality, and efficacy of medicinal plants and herbal medicines is a multifaceted challenge that requires the coordinated efforts of regulators, researchers, industry stakeholders, and consumers [2]. In this context, the conservation and utilization of plant genetic resources through gene banks is crucial to ensure sustainable food security and biodiversity conservation [3,4]. As the world’s population grows, the demand for genetic resources, including landraces, wild and related species, genetic material, advanced breeding material, and modern cultivars will increase, as these resources are essential for improving resilience and profitability in agriculture [4]. To safeguard plant genetic diversity, it is essential to implement effective conservation strategies, including in situ and ex situ methods, alongside cultivation practices and resource management techniques. These efforts help to improve agricultural practices and ensure the sustainable utilization of medicinal plant resources [3,5]. The National Program for Conservation and Sustainable Use of Plant Genetic Resources for Food and Agriculture in the Republic of Croatia supports national development goals, food security, sustainable agriculture, and biodiversity preservation through the conservation and utilization of plant genetic resources [6]. Since 1991, the Croatian Bank for Plant Genes (CBPG) has played a key role in preserving and promoting the sustainable use of these genetic resources and organizes various activities aimed at achieving these objectives [7,8]. Additionally, the Basic Collection of Medicinal and Aromatic Plants (MAP), established in 1998 at the University of Zagreb Faculty of Agriculture (Department of Seed Science and Technology, now the Department of Plant Biodiversity), systematically collects, characterizes, conserves, evaluates, documents, and regenerates MAP plant genetic resources. These resources are integrated into agricultural production and breeding programs, with support from the Working Group on Medicinal and Aromatic Plants, which has been active since 2006 [9,10].

The effective documentation of accessions is essential for the increased use of plant genetic resources in breeding programs, as well-structured information on genetic, phenotypic and passport data enables informed decisions on the breeding and research potential of a specimen [11]. Gene banks face the dual challenges of conserving accessions and managing extensive data that should be accessible through public platforms so that researchers can easily explore genetic diversity [11,12]. As highlighted in the FAO Global Plan of Action, improving links between conservation efforts and user communities is key to maximizing the value of these resources for breeders and researchers worldwide [13]. Recognizing this, the Working Group on Medicinal and Aromatic Plants is dedicated to gathering comprehensive information on their specimens to promote their utility in future breeding efforts.

In recent years, the role of molecular techniques in plant breeding improvement and genetic resource conservation programs has become increasingly important [14]. Monitoring and assessing the genetic diversity of medicinal plants is crucial as it provides information on the different germplasm present. To contribute to the future conservation and improvement of medicinal plants, numerous morphological, cytological, biochemical, and molecular markers can be used to assess their diversity [15]. Molecular DNA markers are widely used to answer various questions related to taxonomy, molecular evolution, population genetics, and genetic diversity. They also contribute to studies on gene expression, genetic mapping, and species evolution, providing fast and accurate results [4]. Characterization of plants using DNA markers is an ideal approach for identifying medicinal plant species and distinguishing between populations or varieties of the same species [16]. Amplified fragment length polymorphism (AFLP) markers are highly effective tools for detecting extensive polymorphism, allowing for precise and detailed genetic analyses [14]. The ability to assess genetic diversity in economically significant genotypes is crucial for advancing basil breeding efforts. Understanding the genetic diversity within a species also plays a vital role in its effective conservation, management, and utilization [17]. Numerous studies have utilized molecular markers to successfully evaluate genetic diversity among various basil accessions [17,18,19,20,21,22,23,24,25,26].

The genus Ocimum L. is considered the largest genus in the Lamiaceae family, with over 60 species of aromatic herbs and shrubs, most of which are native to tropical regions [27,28]. Ocimum basilicum L., commonly known as sweet basil, is a prominent species valued for its culinary, medicinal, aromatic, ornamental, and therapeutic properties [23,29]. In traditional medicine, basil is used for its health-promoting properties, including anti-inflammatory, immunomodulatory, anthelmintic, antioxidant, and antimicrobial properties [30,31].

Previous research into the morphological, chemical, and genetic diversity of (O. basilicum L.) has highlighted quantitative group distinctions at various levels. Labra et al. [32] found that the combined analysis of all three levels of diversity yields good results in resolving doubts about taxonomical classification based on nine basil varieties. Bernhardt et al. [23] used RAPD markers alongside morphological and chemical analyses to investigate the variability of basil cultivars. However, the results demonstrated little to no clear correlation between genetic markers and the observed morphological or chemical traits. On the other hand, Moghaddam et al. [17] successfully used AFLP markers to discriminate between five Iranian basil species and in some cases varieties, concluding that O. basilicum has the greatest genetic variation.

The morphological diversity of basil accessions kept in the Collection of Medicinal and Aromatic Plants (University of Zagreb Faculty of Agriculture, Department of Plant Biodiversity) as part of the National Bank of Plant Genes was evaluated using a descriptor list based on the morphological characteristics outlined by the International Union for the Protection of New Varieties of Plants [33]. This analysis grouped the accessions into five distinct morphotypes: True basil, Small-leaf basil, Lettuce-leaf basil, Purple basil A, and Purple basil B [34,35]. Similarly, based on their essential oil composition, the same O. basilicum accessions have been categorized into five chemotypes: High-linalool (Chemotype A), Linalool/trans-α-bergamotene (Chemotype B), Linalool/methyl chavicol (Chemotype C), Linalool/trans-methyl cinnamate (Chemotype D), and High-methyl chavicol (Chemotype E) [34].

It has been more than a decade since the basil collection from the Croatian Bank for Plant Genes was last comprehensively and simultaneously analyzed for morphological, biochemical, and genetic diversity [36]. At that time, the collection was small, with only 27 accessions in the bank’s inventory. Today, the collection comprises 119 accessions with a much larger variety in geographical origin and morphological and biochemical characteristics [10].

We hypothesize that the application of AFLP markers will reveal significant genetic diversity among the sweet basil accessions from the Croatian Bank for Plant Genes. By comparing the genetic data with morphological and biochemical traits, the study aims to gain a comprehensive understanding of the sweet basil’s genetic resources, enhancing their potential for utilization and supporting their conservation efforts.

2. Materials and Methods

2.1. Plant Material and DNA Extraction

This research was carried out on 80 accessions of O. basilicum (

Table 1. List of the Ocimum accession included in this study. Morphotypes listed were determined according to Carović-Stanko et al. [35], and chemotypes were determined according to Varga et al. [34].

No.	Accession No. a	Taxon/Cultivar	Country of Origin	Morphotype	Chemotype (A–E) b
S01	MAP02297	‘Albahaca Grande Verde’	Spain	True basil	A
S02	MAP02291	-	Azerbaijan	True basil	A
S03	MAP02293	‘Bavires’	Germany	True basil	A
S04	MAP02328	Genovese Basil—Compatto FT	Canada	True basil	A
S05	MAP01645	‘Envigor’	Canada	True basil	A
S06	MAP02336	Gecofure Basil	Canada	True basil	A
S07	MAP00294	Genovese Basil	Croatia	True basil	A
S08	MAP00558	Genovese Basil	Italy	True basil	A
S09	MAP00331	Genovese Basil	Macedonia	True basil	A
S10	MAP00391	Genovese Basil	Croatia	True basil	A
S11	MAP00298	Genovese Basil	Italy	True basil	A
S12	MAP00005	Genovese Basil	Slovakia	True basil	A
S13	MAP02331	Genovese Basil—‘Superbo’	Canada	True basil	A
S14	MAP00145	Genoveser	Austria	True basil	A
S15	MAP02301	Genoveser grossblättrig	Germany	True basil	A
S16	MAP02303	‘Gigante’	Germany	True basil	A
S17	MAP02282	‘Grand Vert de Genes’	Germany	True basil	A
S18	MAP01642	‘Green Gate’	Canada	True basil	A
S19	MAP00576	Grosses gruenes	Austria	True basil	A
S20	MAP02290	Italian Large-Leafed	Italy	True basil	C
S21	MAP02287	-	Italy	True basil	A
S22	MAP02295	-	Japan	True basil	A
S23	MAP02315	Japan A	Germany	True basil	B
S24	MAP02320	-	Madagascar	True basil	A
S25	MAP02314	Mittelgrossblättriges Grünes	Germany	True basil	A
S26	MAP02337	Genovese Basil—Nufar F1	Canada	True basil	C
S27	MAP02288	Genovese Basil—Nufar F1	Germany	True basil	C
S28	MAP01622	Ohre	Germany	True basil	A
S29	MAP02283	-	Russia	True basil	C
S30	MAP02338	‘Stella’	Canada	True basil	A
S31	MAP00186	Sweet basil	Croatia	True basil	A
S32	MAP00232	Sweet basil	Croatia	True basil	A
S33	MAP00414	Sweet basil	USA	True basil	C
S34	MAP02296	-	Togo	True basil	A
S35	MAP01640	Bush Basil	Canada	Small-leaf basil	B
S36	MAP01632	Comune a Foglia Piccola	Germany	Small-leaf basil	A
S37	MAP00560	‘Fine verde’	Italy	Small-leaf basil	B
S38	MAP02334	Greek Bush Basil	Canada	Small-leaf basil	A
S39	MAP01648	‘Green Globe’	Canada	Small-leaf basil	B
S40	MAP02327	‘Marseilles’	Canada	Small-leaf basil	A
S41	MAP02298	‘Massilia’	Germany	Small-leaf basil	A
S42	MAP02326	‘Medinette’	Canada	Small-leaf basil	A
S43	MAP02300	‘Piccolo’	Italy	Small-leaf basil	A
S44	MAP01641	‘Spicy Globe’	Canada	Small-leaf basil	B
S45	MAP00559	Blistered lettuce-leaf basil	Italy	Lettuce-leaf basil	C
S46	MAP02286	-	Italy	Lettuce-leaf basil	C
S47	MAP02281	Lactucaefolium	Germany	Lettuce-leaf basil	C
S48	MAP01654	var. difforme/Napoletano Basil	Canada	Lettuce-leaf basil	B
S49	MAP01623	var. difforme	Germany	Lettuce-leaf basil	C
S50	MAP00375	var. difforme/‘A foglie di lattuga’	Italy	Lettuce-leaf basil	C
S51	MAP02307	var. difforme	Uzbekistan	Lettuce-leaf basil	C
S52	MAP02321	var. purpurescens	Jemen	Purple basil A	C
S53	MAP02333	‘Magical Michael’	Canada	Purple basil A	A
S54	MAP01658	‘Oriental Breeze’	Canada	Purple basil A	A
S55	MAP02311	-	Romania	Purple basil A	A
S56	MAP02313	var. purpurascens	Uzbekistan	Purple basil A	A
S57	MAP02292	var. purpurascens	China	Purple basil A	C
S58	MAP02302	var. purpurascens	Irak	Purple basil A	D
S59	MAP01629	var. purpurascens/Mexican Basil	Germany	Purple basil A	D
S60	MAP00146	var. purpurascens/no. 3193	Austria	Purple basil A	D
S61	MAP02305	var. purpurascens/Persian Anise-scented Basil	Germany	Purple basil A	D
S62	MAP01644	Anise Basil	Canada	Purple basil B	C
S63	MAP01639	Ararat	Canada	Purple basil B	C
S64	MAP00333	‘Dark Opal’	Russia	Purple basil B	A
S65	MAP00284	‘Dark Opal’	Russia	Purple basil B	A
S66	MAP00004	‘Opal’	Slovakia	Purple basil B	A
S67	MAP00283	Opal-ZS98	Slovakia	Purple basil B	A
S68	MAP02308	‘Licorice’	Germany	Purple basil B	A
S69	MAP01630	‘Metalica’	Germany	Purple basil B	A
S70	MAP01650	‘Osmin’	Canada	Purple basil B	A
S71	MAP02330	‘Purple Delight’	Canada	Purple basil B	A
S72	MAP02306	‘Susambari’	Georgia	Purple basil B	B
S73	MAP02335	‘Sweet Salad’	Canada	Purple basil B	C
S74	MAP01657	Thai Basil ‘Queenette’	Canada	Purple basil B	D
S75	MAP02294	var. purpurascens	Armenia	Purple basil B	C
S76	MAP02289	var. purpurascens	Armenia	Purple basil B	C
S77	MAP02319	var. purpurascens/‘Kardinal’	Germany	Purple basil B	C
S78	MAP02309	var. purpurascens/‘Rothaut’	Germany	Purple basil B	A
S79	MAP02317	var. thyrsiflorum/‘Siam Queen’	Thailand	Purple basil B	E
S80	MAP01643	‘Purple Ruffles’	Canada	Purple basil B	B

^a Accession number from The Collection of Medicinal and Aromatic Plants, Zagreb, Croatia available at: cpgrd.hcphs.hr. ^b Chemotypes: (A) Linalool, (B) Linalool/trans-α-bergamotene, (C) Linalool/methyl chavicol, (D) Linalool/trans-methyl cinnamate, and (E) methyl chavicol. Reprinted from Industrial Crops and Products, 109, Filip Varga, Klaudija Carović-Stanko, Mihailo Ristić, Martina Grdiša, Zlatko Liber and Zlatko Šatović, Morphological and biochemical intraspecific characterization of Ocimum basilicum L., 613–614 [34], Copyright (2024), with permission from Elsevier.

) from the Collection of Medicinal and Aromatic Plants of the Department of Plant Biodiversity, University of Zagreb Faculty of Agriculture, Croatia—http://cpgrd.hcphs.hr (accessed on 28 October 2024). Seeds were sown in polystyrene seedling trays in commercial seedling substrate Subs-trate 1 (Klasmann-Deilmann GmbH, Geeste, Germany). Germination and initial plant growth took place in the greenhouse at an average daily temperature of 22.5 °C. Samples of young leaves were collected from well-developed seedlings.

Genomic DNA was extracted from young leaves using a commercial DNA isolation kit (DNeasy Plant Mini Kit, Qiagen GmbH, Hilden, Germany). DNA concentration and quality were assessed using a nanophotometer P330 (Implen^®, Munich, Germany).

2.2. AFLP Analysis

Amplified fragment length polymorphism (AFLP) is a robust DNA fingerprinting method that integrates the precision of restriction fragment length polymorphism (RFLP) with the versatility of PCR-based techniques by attaching primer-recognition sequences (adaptors) to restricted DNA. In the initial step, restriction fragments are produced using two restriction enzymes, EcoRI and MseI. Double-stranded adaptors are then attached to the ends of these fragments, creating template DNA for two rounds of polymerase chain reactions (pre-selective and selective PCR). The selective PCR step employs fluorescent dye-labeling technology to enable the detection of DNA fragments via capillary electrophoresis using ABI PRISM™ sequencers (Applied Biosystems, Foster City, CA, USA). The AFLP procedure was performed according to the AFLP^® Plant Mapping protocol [37] with some modifications, as described by Carović-Stanko et al. [38].

• For the selection of PCR primer combinations, we followed the recommendations of the manufacturer of the fluorescently labeled AFLP primers (Applied Biosystems, Foster City, CA, USA). Of the 128 possible combinations suggested by the manufacturer, we selected 16 combinations that were complementary to the DNA adapters and PCR primers used in the previous AFLP steps of DNA ligation and pre-selective PCR amplification.
• We generated dozens of DNA fragments in the range of 50–500 base pairs and peak heights above the absolute value of 50 in the computer program GeneMapper (Applied Biosystems, Foster City, CA, USA).
• We had an estimated error rate per primer combination of less than 5%.
• We produced polymorphic fragments that discriminate between different individuals, taxa, or species in this study.

In total, sixteen primer combinations were selected for the final selective amplifications: (1) NED-EcoRI-AAC + MseI-CAA, (2) NED-EcoRI-AC + MseI-CCG, (3) FAM-EcoRI-ACA + MseI-CAC, (4) FAM-EcoRI-ACC + MseI-CTG, (5) PET-EcoRI-ACC + MseI-CCC, (6) PET-EcoRI-ACC + MseI-CCT, (7) PET-EcoRI-ACC + MseI-CGA, (8) PET-EcoRI-ACC + MseI-CGG, (9) VIC-EcoRI-ACG + MseI-CCC, (10) VIC-EcoRI-ACG + MseI-CCT, (11) VIC-EcoRI-ACG + MseI-CGA, (12) VIC-EcoRI-ACG + MseI-CGG, (13) FAM-EcoRI-ACT + MseI-CAA, (14) NED-EcoRI-AGA + MseI-CAC, (15) NED-EcoRI-AGC + MseI-CTG, (16) FAM-EcoRI-AT + MseI-CCG. The fluorescently labeled products from the selective amplification process were analyzed by Macrogen Europe BV (Amsterdam, The Netherlands) using an ABI PRISM 3730XL automated sequencer (Applied Biosystems, Foster City, CA, USA). AFLP fragments were analyzed as peak size and height value tables using GeneMapper 4.0 (Applied Biosystems, Foster City, CA, USA). The amplified fragments were scored based on the presence (1) or absence (0) of homologous bands. Marker selection for the final statistical analysis, as well as error rate determination, was performed using the scanAFLP ver. 1.2 program [39].

2.3. Data Analysis

Molecular diversity within basil morphotypes was assessed by calculating the percentage of polymorphic markers (P), Shannon’s information index [40], and the number of private markers (N_pr). The estimates of Shannon’s information index (Sh) among morphotypes were compared using repeated measures analysis of variance carried out using PROC GLM in SAS v. 9.4 [41]. To assess the prevalence of rare markers within each morphotype and chemotype, the rarity index (RI) [42] was calculated, which is expected to be higher in morphotypes or chemotypes where rare markers are more commonly observed among accessions. Pairwise distances were calculated using Dice’s coefficient [43]. Since chemotype E comprised only one accession, the aforementioned indices were not calculated for said chemotype. The average Dice’s distance among accessions was also computed within each morphotype and chemotype. Principal coordinate analysis (PCoA), based on the Dice distance matrix, was conducted using PAST version 2.01 [44].

To partition the total AFLP diversity, the analysis of molecular variance (AMOVA) [45] was applied, examining (a) variation between morphotype groups (’green’ vs. ’purple’), among morphotypes within these groups, and within individual morphotypes, as well as (b) variation among and within chemotypes. Statistical significance of the variance components was assessed using non-parametric randomization tests with 10,000 permutations, conducted in Arlequin ver. 3.5 [46]. Pairwise comparisons between morphotypes were analyzed using AMOVA, producing ϕ_ST values, which represent the proportion of total variance attributed to differences between two morphotypes and can be interpreted as the average genetic distance between them [47].

The presence of distinct genetic clusters among 80 basil accessions has been evaluated using a Bayesian approach [48], as implemented in BAPS 6.0 [49]. For each accession, the proportion of its genome attributed to different clusters was estimated, with the number of clusters (K) set to a maximum of 10 and each analysis replicated 20 times. The results from the mixture analysis were then used for population admixture analysis [50], with the default settings in order to detect admixture between clusters. Accessions with over 75% of their genome assigned to a single cluster were classified as belonging to that cluster, while those with less than 75% membership in any cluster were considered to be of ‘mixed origin’ [51]. To assess the admixture level of each morphotype, Shannon’s diversity index was calculated based on the cluster membership proportions inferred by BAPS. For an accession with its genome entirely assigned to one cluster, the Shannon index would be zero, while the index reaches its maximum when the genome is equally distributed across all clusters (indicating the highest level of admixture). The total diversity (Sh_Total) of a morphotype was derived from the average membership proportions, and this was compared with the mean within-cultivar diversity (Sh_Mean), calculated by averaging the Shannon’s diversity indices of the accessions within a morphotype. This allowed the decomposition of the admixture level of a morphotype into within-accession (Sh_Mean/Sh_Total) and among-accession [(Sh_Total − Sh_Mean)/Sh_Total] components.

3. Results and Discussion

3.1. Molecular Diversity of Basil Morphotypes and Chemotypes

A total of 16 AFLP primer combinations yielded a total of 2114 polymorphic markers. The number of polymorphic markers per primer pair combination generated ranged from 84 (FAM-EcoRI-ACT + MseI-CAA) to 177 (NED-EcoRI-AGA + MseI-CAC), with an average of 132.13. This level of polymorphism is much higher than in other studies on basil that have also demonstrated significant genetic diversity using AFLP markers [17,32].

The percentage of polymorphic markers (P) varied within morphotypes (

Table 2. Molecular diversity of basil morphotypes and chemotypes based on 80 accessions.

Morphotype						Chemotype
	True basil	Small-leaf basil	Lettuce-leaf basil	Purple basil A	Purple basil B	A	B	C	D
n	34	10	7	10	19	47	8	19	5
P	95.6%	76.96%	72.52%	80.84%	89.4%	99.432%	77.862%	94.986%	61.258%
Sh	0.618	0.576	0.583	0.616	0.631	0.668	0.592	0.676	0.515
Npr	2	0	0	0	4	18	0	0	0
RI	0.942	0.98	1.128	1.034	1.049	0.976	1.018	1.025	1.1
Davg	0.365	0.387	0.391	0.402	0.377	0.391	0.401	0.412	0.376

n—number of accessions; P—percentage of variable markers; Sh—Shannon’s information index; N_pr—number of private alleles; RI—rarity index; D_avg—average Dice’s distance among accessions.

) from 72.52% (Lettuce-leaf basil) to 95.6% (True basil). The high P in True basil may reflect a broad genetic base within this morphotype, possibly linked to its widespread cultivation and use across different regions. In comparison, the relatively lower P in Lettuce-leaf basil might be indicative of a narrower genetic base, potentially due to selection for specific traits in a breeding program [52]. Genetic diversity measured by Shannon’s information index (Sh) was highest in the Purple basil B morphotype (0.631), which was also characterized by the highest number of private markers (N_pr = 4). The high Sh in Purple basil B is consistent with its status as a morphotype with substantial genetic variation, possibly due to its diverse chemical and morphological traits [53]. The identification of private markers in Purple basil B suggests that it may harbor unique alleles, further supporting its genetic distinctiveness. The rarity index (RI) varied from 0.942 (True basil) to 1.128 (Lettuce-leaf basil). The average Dice’s distances ranged from 0.365 (True basil) to 0.402 (Purple basil A). This suggests that Purple basil A may be more genetically distinct compared to other morphotypes, which could be a result of selection for specific traits such as pigmentation or flavor profiles.

The same indices were also calculated with chemotype classification applied to studied accessions (Table 2). P varied from 61.258% (Chemotype D) to 99.432% (Chemotype A). Chemotype A was also the only chemotype with private markers identified (N_pr = 18). Chemotype C had the highest values of both Sh (0.676) and RI (1.025), indicating that it harbors a considerable amount of genetic variation and rare alleles, likely due to the complex interaction between genetic and environmental factors that influence chemical profiles. Average Dice’s distances ranged from 0.376 (Chemotype D) to 0.412 (Chemotype C).

Overall, the lowest Dice’s distance (0.179) was observed between accessions S70 and S71 (both accessions belonging to Purple basil B morphotype and chemotype A). The highest Dice’s distance (0.576) was observed between accessions S48 (Lettuce-leaf basil morphotype and Chemotype B) and S63 (Purple basil B morphotype and Chemotype C). Accessions with the lowest genetic similarity could be used as parents in breeding programs, resulting in offspring with higher genetic diversity [19].

In the PCoA based on Dice’s distances between accessions, the first two principal coordinates accounted for 21.8% of the total variance (Figure 1). The first principal coordinate separated accessions belonging to two groups of morphotypes (’green’ vs. ’purple’). Additionally, within the ’green’ group, the majority of accessions from True basil morphotype grouped on the right side of the plot. However, the PCoA failed to separate the other distinct morphotypes within each group or in general. When focusing on chemotypes, a similar trend was observed. The majority of Chemotype A accessions, also classified as True basil morphotype, were positioned on the right side of the plot, while accessions from other morphotypes that were classified as Chemotype A were dispersed across the plot. Accessions belonging to Chemotypes D and E were positioned on the left side of the plot but interspersed with accessions from other chemotypes; thus, PCoA failed to clearly separate chemotypes based on AFLP data. This could suggest that, while there is some genetic distinction between chemotypes, the genetic differences may not be substantial enough to result in clear separation in a PCoA.

3.2. Genetic Structure of Basil Morphotypes and Chemotypes

Clustering of basil accessions in BAPS revealed that the highest log marginal likelihood (−85,621.25) was at K = 4 (Figure 2). The accessions belonging to ’green’ morphotypes (True basil, Small-leaf, Lettuce-leaf) were assigned to Clusters 1 or 2, while ’purple’ morphotypes (Purple basil A and B) were assigned to Clusters 1 or 3 and 4. Clusters 3 and 4 were dominant exclusively in ’purple’ morphotypes. Additionally, Purple basil B was genetically the most admixed, having accessions belonging to all four genetic clusters (Figure 2). These results suggest that while morphotypes cluster into genetically distinct groups (at least on a morphotype group level), there is a significant admixture, particularly in the Purple basil B morphotype. Genetic clustering of chemotypes revealed that accessions from all chemotypes were assigned to Cluster 1. Cluster 2 was dominant in accessions classified as Chemotypes A, B, and C, while Cluster 3 was dominant in accessions classified as Chemotypes B, C, and D. Cluster 4 was dominant exclusively in accessions from Chemotype A. A high degree of observed variability, particularly in the chemotypes, which cannot be clearly distinguished from each other based on the AFLP data, could be a result of environmental factors. In addition to genetic factors (e.g., mutations), environmental factors such as vegetative and flowering stages can influence the essential oil composition [32,54]. Labra et al. [32] also analyzed nine O. basilicum varieties using AFLP markers and concluded that similarity does not necessarily reflect similarity or difference in yield traits, i.e., cultivars can be genetically very similar but differ in oil composition and morphological characteristics. There is also no apparent connection between genetic divergence and geographical origin of the accessions, which is consistent with previous research [23,52].

The initial analysis of molecular variance (AMOVA) revealed that most of the genetic diversity was attributable to differences among accessions within morphotypes (90.06%), while the remaining variability (9.94%) could be explained by differences between accessions within morphotypes (

Table 3. AMOVA analysis for the partitioning of AFLP diversity of 80 Ocimum accessions: (A) among morphotypes and among accessions within morphotypes; (B) between groups of morphotypes, among morphotypes within groups, and within morphotypes; (C) among chemotypes and among accessions within chemotypes.

Analysis	Source of Variation	df	Variance Components	Percentage of Variation	φ-Statistics	p (φ)
(A)	Among morphotypes	4	34.54	9.94%	0.099	<0.0001
	Within morphotypes	75	312.85	90.06%
(B)	Between groups of morphotypes (‘green’ vs. ‘purple’)	1	19.88	5.61%	0.056	<0.0001
	Among morphotypes within groups	3	21.85	6.16%	0.065	<0.0001
	Within morphotypes	75	312.85	88.23%	0.118	<0.0001
(C)	Among chemotypes	3	14.76	4.30%	0.043	<0.0001
	Within chemotypes	75	328.63	95.70%

). These results are similar to those obtained by Malav et al. [20] when analyzing accessions of Ocimum tenuiflorum L. by using a combination of RAPD and ISSR. AMOVA taking into account morphotype groups revealed that 5.61% of genetic variance could be explained by differences between ’green’ and ’purple’ morphotype groups, 6.16% by differences between morphotypes within the groups, while the majority of genetic variance (88.23%) was attributable to differences among accessions within morphotypes. Also, AMOVA showed that most of the genetic diversity was attributable to differences among accessions within chemotypes (95.70%). The relatively low percentage of variance explained by differences between chemotypes suggests that chemotypic differences may not be strongly associated with overall genetic differences.

The average pairwise φ_ST value between morphotypes was 0.095, ranging from 0.026 (between Purple basil A and B morphotypes) to 0.126 (between Small-leaf and Purple basil B morphotypes). All φ_ST values were significant (p < 0.0001). The average pairwise φ_ST value between chemotypes was 0.052, ranging from 0.029 (between Chemotypes B and C) to 0.089 (between Chemotypes A and C). All φ_ST values were significant (p < 0.0001). These results suggest that there is moderate genetic differentiation between chemotypes, though it is less pronounced than between morphotypes.

Gene banks play a crucial role in conserving plant accessions and making them accessible to plant breeders and researchers [13]. The more detailed the information that is gathered on materials preserved in the base collections is, the more effectively the active collections can be customized to the needs of users [12]. Comprehensive, well-organized information is essential to assess the true value of a resource, especially in terms of its potential for breeding and research. This wealth of information covers various aspects, including data essential for optimal collection management, basic gene bank records (such as passport data), and phenotypic and comprehensive genetic data [11].

The great diversity of basil populations at the morphological, chemical, and genetic levels significantly contributes to the preservation of plant genetic resources, while also providing valuable material for the development of new, improved cultivars. This diversity allows for the preservation of a broader range of genetic traits within the species, which is crucial for resistance to diseases, adaptation to climate changes, and other challenges [55,56]. Additionally, such diversity is essential for the selection and breeding of new cultivars that meet specific needs, such as improved medicinal properties, resistance, or adaptation to certain growing conditions. This comprehensive information will enable breeders to increase the use of plant genetic resources in breeding programs.

4. Conclusions

The results of this study have shown that there is considerable genetic variability among the basil accessions analyzed. At the genetic level, a clear distinction between two groups of morphotypes (‘green’ vs. ‘purple’) was recognized, indicating a certain separation between these groups. However, the remaining accessions show a high degree of variability, especially in the chemotypes, which cannot be clearly distinguished from each other using the AFLP data. As more advanced techniques for genetic analysis than AFLPs are now available (such as a whole range of NGS methods), employing them in future research could potentially yield better results in efforts to resolve complex history of cultivation and breeding of basil. This high variability, especially at the chemotype level, highlights the diverse genetic background within the species. This diversity is a valuable resource for breeding programs as it provides a solid basis for the development of new cultivars that can combine desirable morphological and biochemical traits. These results emphasize the potential for the selection and cultivation of basil cultivars tailored to specific needs of the pharmaceutical, cosmetics, and food industries. Therefore, this genetic diversity represents an essential starting point for future advances in basil breeding and cultivar improvement.