Pre-Breeding of Promising Coffea canephora Genotypes

Abstract

This study evaluated the genetic diversity of 33 Coffea canephora genotypes through morphophysiological and molecular analyses, aiming to identify promising genotypes for pre-breeding purposes in the southern region of Espírito Santo, Brazil. Cutting-propagated seedlings were evaluated 120 days after planting, considering height, stem and crown diameter, number of leaves, fresh and dry shoot and root weight, chlorophyll content, and root characteristics. Molecular analysis was performed on 32 genotypes; one was excluded due to absent DNA, and 18 ISSR markers were used. Morphological data were analyzed by ANOVA, Scott–Knott’s mean test, principal component analysis, and cluster analysis. The results revealed significant diversity among genotypes. The first two principal components explained 75.5% of the total variability. Genotypes 2, 3, 4, 5, 6, 10, 32, and 33 stood out as those that produced the most vigorous seedlings. Molecular analysis also revealed genetic diversity among genotypes, with the formation of 16 groups, while the morphophysiological analysis revealed four groups. The Mantel test demonstrated a small but significant positive difference (r = 0.228; p = 0.018) between the genetic and morphophysiological distances of the genotypes. This diversity indicates that the genotypes evaluated are promising for use in C. canephora breeding programs.

1. Introduction

Coffea canephora, comprising conilon and robusta coffee, either alone or in hybrids, and hereinafter referred to as canephora coffee, plays a crucial role in the Brazilian economy, contributing to agricultural productivity and meeting the growing global demand for the bean [1,2]. The state of Espírito Santo is responsible for more than 60% of Brazilian C. canephora production, with the species being favored due to its resistance to biotic and abiotic factors and productivity potential [3]. Furthermore, the cultivation of conilon coffee supports local economies through job creation, income distribution, and export opportunities [4].

The continuous improvement of coffee bean production and quality is a key objective in modern coffee cultivation. Among the critical factors influencing successful crop establishment is the use of high-quality seedlings [5]. In this context, adequate planning at all stages of cultivation is essential, especially during the stage of crop establishment and formation, as errors during this period can compromise productivity and plant behavior when submitted to stress conditions due to biotic and abiotic factors [6]. Furthermore, planting vigorous seedlings contributes to establishing efficient production, reducing replanting costs and ensuring rapid initial plant growth [7].

The morphological and physiological variables of coffee seedlings are important indicators of quality and vigor [7] and are useful in the early selection of superior materials. These variables can be studied using multivariate analysis techniques, including principal component analysis (PCA) and cluster analysis, which allow the identification of patterns and the distinction between different genotypes [8,9,10,11].

In addition, molecular markers enable the analysis of genetic diversity and the interrelationships between the DNA of genotypes [12]. Among them, ISSR (Inter-Simple Sequence Repeat) markers stand out for being broad-based markers that demonstrate high degree of polymorphism. This technique is fast, reproducible, and does not require prior sequence information, allowing its application across a wide range of plant species [13]. Therefore, ISSR markers have become a valuable tool for molecular analyses and have been widely used in studies of the genetic diversity of C. canephora [14,15,16,17,18].

Genetic diversity is essential for C. canephora breeding programs, as the plant reproduces by allogamy and is self-incompatible. Therefore, genetic variability is essential for successful crop pollination, and breeding strategies will allow for increased resistance to biotic and abiotic stresses, improving coffee yield and quality [5] among numerous other economically important characteristics.

The characterization of genetic resources demonstrates the potential for developing base collections that can facilitate plant breeding [19]. Furthermore, genetic diversity influences the nutritional balance and agronomic characteristics of coffee, impacting bean yield and quality [20]. In general, preserving this diversity is essential for the sustainability and long-term progress of coffee breeding programs, especially in the context of climate change and the increasing challenges faced by agriculture [21].

In this context, pre-breeding is a fundamental step in the breeding of C. canephora. It involves broadening the genetic base, selection and use of genetic materials with desirable traits, thereby enabling the development of clones better adapted to specific regional conditions. This, in turn, supports improvements in crop productivity, quality, and overall sustainability [22]. This step is particularly important for perennial species like coffee, whose long cycle and genetic complexity make the development of improved cultivars a time-consuming and costly process [1].

Despite the growing number of studies on C. canephora diversity, most have focused on commercial cultivars or those from large producing regions [1,23,24], leaving a gap in the genetic characterization of local genotypes in the southern region of Espírito Santo. This study aims to integrate morphophysiological and molecular approaches to assess the diversity of promising regional genotypes at the seedling stage. This pre-breeding approach, focused on non-commercial genotypes of local origin, is crucial because it allows the identification of genetic material with potential local adaptation traits and superior vigor [25]. The results contribute to expanding the genetic base available for pre-breeding programs and support the development of regionally adapted C. canephora clones.

Therefore, this study aimed to evaluate the genetic diversity of 33 C. canephora genotypes based on morphophysiological and molecular variables of seedlings for pre-genetic improvement purposes, aiming at the selection of promising genotypes adapted to the soil and climate conditions of the southern region of Espírito Santo, Brazil.

2. Materials and Methods

2.1. Morphological Analysis of Conilon Coffee Seedlings

2.1.1. Experimental Location

The experiment was conducted in a coffee seedling propagation nursery installed on the property of a partner farmer in the southern region of Espírito Santo, Brazil, in the rural area of the municipality of Castelo (latitude 20°36′13″ S, longitude 41°11′05″ W).

2.1.2. Genetic Material

Thirty-three conilon coffee genotypes (Coffea canephora Pierre Ex. A. Froehner) were used to conduct the experiment. Of these, 31 genotypes (clones) were selected from a partner coffee grower in southern Espírito Santo (genotypes 1 to 31), all of which were promising and unpublished, and 2 commercial genotypes, here considered as controls (genotypes 32 and 33), known as A1 and P2 [26].

2.1.3. Seedling Propagation and Experimental Design

Seedlings used were produced from cuttings of the 33 genotypes, obtained from adult orthotropic branches and removed from the crop when plants presented good phytosanitary and nutritional characteristics. After removal, cuttings were sent to the seedling nursery for standardization [27], with plagiotropic (productive) branches being eliminated just above the insertion, by cutting two-thirds of the leaf blade of the two leaves at each node, discarding the internodes at the ends. Cuttings were individualized, with a bevel cut being made 3 cm below the leaf insertion and a horizontal cut 1 cm above this insertion. They were then planted in conventional coffee substrate established by Ferrão et al. (2012) [28], in polyethylene bags, and arranged in a completely randomized design (CRD) with eight replicates. The seedlings were grown in a seedling nursery with 60% shade and a sprinkler irrigation system.

2.1.4. Morphophysiological Assessments

Seedling assessments were performed 120 days after planting the cuttings, based on the following variables in

Table 1. Morphophysiological variables analyzed in C. canephora seedlings and respective measurement methods.

Variable	Acronym	Measurement
Plant height	Hei	Measurement with graduated ruler
Stem diameter	SD	Measured with a digital caliper in the middle third of the stem
Crown diameter	CD	Measurement with graduated ruler
Number of leaves	NL	Counted manually
Leaf area size	LA	Measured with the automatic leaf area meter model Li-cor®, model: Li-3100C (Manufacturer LI-COR Biosciences, Sao Paulo, Brazil).
Fresh and dry shoot and root mass	FSM, FRM, DSM and DRM	Estimated on a precision scale.
Relative chlorophyll content	Chl	Determined using a portable chlorophyll meter (model SPAD-502, Minolta)
Root length, root volume, root surface projection, root surface area, and Mean root diameter	RL, RV, RSP, RA and MRD	Analyzed by EPSON STD 4800 root scanner and WinRHIZO Pro 2022a software.
Dickson Quality Index	DQI	Calculated by the equation [29]: DQI = [(RDM + SDM)/(Hei/SD + SDM/RDM)]

.

2.1.5. Statistical Analysis

Analysis of variance and mean tests were performed based on data obtained using the Scott–Knott methodology at 5% significance level. Based on the set of traits (16) and genotypes (33), the genetic distance matrix was estimated using the Standardized Euclidean Distance (ED), and genotype clustering was performed using the UPGMA (Unweighted Pair Group Arithmetic Mean) method. The cophenetic correlation coefficient between the graphical distance matrix and the original distance matrix was estimated. Principal component analysis (PCA) was performed, generating a Biplot scatter plot with all seedling traits. Furthermore, Pearson’s correlation among seedling traits was calculated. Statistical analyses were performed using the R (version 4.5.1) application [30].

2.2. Molecular Analysis

2.2.1. Sample Collection

For this study, 32 C. canephora genotypes grown in the state of Espírito Santo were selected. Leaves from these genotypes were collected from healthy plants at similar vegetative stage to ensure sample uniformity. Leaves were transported to the laboratory in hermetically sealed containers and stored at −20 °C until DNA extraction.

2.2.2. DNA Extraction

DNA extraction followed the “mini-prep” protocol of Doyle and Doyle (1990) [31], with modifications by Berilli et al. (2013) [32], where 200 mg of leaf tissue were macerated in liquid nitrogen and transferred to 2 mL Eppendorf tubes. Then, 800 μL of preheated extraction buffer containing 1% CTAB, 1.4 M NaCl, 20 mM EDTA, 100 mM Tris-HCl (pH 8.0), 1% PVP, and 0.1% 2-mercaptoethanol was added. Tubes were manually inverted for 5 min and incubated at 65 °C for 30 to 40 min in water bath, with manual agitation every 10 min.

After incubation, samples were centrifuged at 14,000 rpm for 5 min. The 600 μL supernatant was transferred to new Eppendorf tubes (1.5 mL), to which 600 μL of chloroform–isoamyl alcohol (24:1) was added. After further centrifugation at 14,000 rpm for 5 min, the supernatant was transferred to new tubes, and ice-cold 70% isopropanol was added, followed by overnight incubation at 4 °C. Samples were centrifuged at 14,000 rpm for 10 min, obtaining a pellet, which was washed twice, first with 300 μL of 70% ethanol and then with 300 μL of 95% ethanol. The pellet was resuspended in 200 μL of Tris-EDTA solution (10 mM Tris-HCl, 1 mM EDTA, pH 8.0) and incubated with RNAse (40 μg/mL) at 37 °C for 30 min. DNA concentrations were estimated using a 10% polyacrylamide gel and standardized to 10 ng/μL.

2.2.3. PCR Amplification

PCR amplification reactions followed the protocol of Zietkiewicz et al. (1994) [33], with modifications by Berilli et al. (2013) [32]. The final volume of 15 μL contained: 1.5 μL of 10X buffer (500 mM KCl, 100 mM Tris-HCl, pH 8.4, 1% Triton X-100), 0.6 μL of MgCl₂ (25 mM), 4.8 μL of dNTPs (0.2 mM), 0.3 μL of primer (0.5 mM), 0.2 μL of Taq DNA polymerase (5 U/μL), and 3 μL of genomic DNA (10 ng/μL), filling up with 4.6 μL ultrapure water. A negative control was included in each run to verify the absence of contamination.

The ISSR primers were selected after screening 32 primers with two DNA samples and a negative control, resulting in the choice of the following primers: UBC 818, UBC 834, UBC 841, UBC 842, UBC 845, UBC 848, UBC 849, UBC 854, UBC 855, UBC 825, UBC 857, UBC 859, UBC 868, UBC 875, UBC 879, UBC 889, UBC 895, and Echt5. Amplification was performed in an Eppendorf gradient thermocycler with the following program: initial denaturation at 94 °C for 5 min, followed by 40 cycles of 94 °C for 30 s, 52 °C for 40 s, and 72 °C for 1 min, with final extension at 72 °C for 5 min and maintenance at 25 °C.

2.2.4. Electrophoresis and Data Analysis

PCR products were stained with 3 μL of 6X “blue juice” dye and loaded onto a 7% polyacrylamide gel with 10X TAE buffer, submitted to 220 volts for 2 h and 30 min. Bands were visualized after staining with cybergreen for 30 min using a photoimaging device (Bio-Rad Laboratories, Hercules, CA, USA).

Band scoring was performed using binary visual analysis, where the presence of a clear and reproducible band was recorded as “1” and its absence as “0”. Only clear and well-defined bands were considered for scoring, while weak, diffuse, or blurred bands were excluded. Subsequently, descriptive analysis of data was performed with values for the total number of bands (TNB) and number of polymorphic bands (NPB).

2.2.5. Statistical Analysis

Genetic analysis was performed using the Jaccard similarity index [34], considering coincidence and discordance in the presence/absence of bands. The values obtained were used to construct a dendrogram using the UPGMA (Unweighted Pair Group Method with Arithmetic Mean), enabling the visualization of the genetic relationships among genotypes. The cutoff point for defining groups was estimated by the Mojena (1977) method [35], with k = 1.25 according to Milligan and Cooper (1985) [36], using the mean and standard deviation of the fusion levels to determine genetic similarity and divergence among individuals.

All these analyses were performed using the GENES software version 1900.2019.120 [37].

To evaluate the correlation between the genetic and morphological distance matrices of C. canephora genotypes, the Mantel test was applied [38]. For statistical analysis, the R application (version 4.5.1) was used [30].

3. Results

Analysis of variance revealed significant differences (p < 0.01) among C. canephora genotypes for all variables evaluated, demonstrating the existence of broad genetic diversity in the population under study (

Table 2. Mean square (MS), mean, and coefficient of variation (CV%) for different variables in C. canephora seedlings.

Variable	MS	Mean	CV (%)
Crown diameter (cm)	75.71 **	20.31	27.48
Stem diameter (mm)	3.30 **	10.78	8.99
Height (cm)	109.02 **	7.41	13.40
Number of leaves	59.62 **	9.60	33.59
Fresh shoot mass (g)	31.87 **	10.15	22.13
Fresh root mass (g)	8.59 **	4.10	29.51
Dry shoot mass (g)	4.15 **	3.01	25.18
Dry root mass (g)	1.07 **	1.27	31.42
Leaf area (cm2)	29,527.51 **	228.98	26.19
Dickson Quality Index	0.62 **	1.06	29.7
Chlorophyll	45.87 **	24.93	18.65
Root length (cm)	211,208 **	671.32	18.18
Root surface projection (cm2)	952.9 **	44.34	20.94
Root surface area (cm2)	9404.7 **	139.3	20.94
Mean root diameter (mm)	0.03 **	0.67	12.03
Root volume (cm3)	3.78 **	2.34	28.76

** Significant at 1% probability.

).

The coefficients of variation found in the present study varied among the characteristics evaluated, with the highest value found for number of leaves (33.59%) and the lowest for stem diameter (8.99%).

Table 3. Crown diameter (CD), stem diameter (SD), plant height (Hei), number of leaves (NL), fresh shoot mass (FSM) and fresh root mass (FRM), dry shoot mass (DSM) and dry root mass (DRM) of seedlings of the different C. canephora genotypes.

Treat	CD		SD		Hei		NL		FSM		FRM		DSM		DRM
1	18.06	b	9.25	b	18.00	c	12.38	a	10.65	b	5.42	b	3.24	b	1.31	c
2	19.00	b	10.30	b	21.00	b	9.00	b	12.17	a	4.90	b	3.67	a	1.44	c
3	23.81	a	12.09	a	19.00	b	12.38	a	12.09	a	5.07	b	3.41	b	1.40	c
4	19.75	b	10.45	b	20.44	b	9.38	b	10.68	b	5.18	b	3.10	b	1.55	b
5	20.38	b	10.26	b	25.81	a	12.75	a	12.11	a	4.58	b	3.78	a	1.53	b
6	20.00	b	10.27	b	18.06	c	11.75	a	13.17	a	4.80	b	3.63	a	1.60	b
7	22.44	a	10.79	b	18.31	c	9.88	a	10.26	c	4.04	b	3.25	b	1.52	b
8	18.19	b	10.83	b	16.38	c	10.75	a	9.36	c	3.25	c	2.74	b	1.29	c
9	18.94	b	10.71	b	17.88	c	15.00	a	12.79	a	3.96	b	4.20	a	1.76	b
10	22.94	a	10.17	b	24.06	a	9.50	b	12.89	a	4.33	b	3.73	a	1.51	b
11	22.88	a	10.87	b	19.00	b	8.38	b	9.99	c	4.24	b	2.91	b	1.28	c
12	15.50	b	10.51	b	23.38	a	8.63	b	11.00	b	4.21	b	3.39	b	1.28	c
13	17.63	b	10.62	b	17.06	c	8.38	b	9.75	c	4.72	b	2.98	b	1.58	b
14	21.56	a	10.93	b	18.44	c	10.75	a	10.59	b	2.99	c	3.01	b	1.07	c
15	19.44	b	10.73	b	19.50	b	10.50	a	11.05	b	3.97	b	3.27	b	1.26	c
16	17.56	b	11.24	a	19.38	b	11.13	a	8.67	c	4.46	b	2.59	b	1.30	c
17	22.56	a	10.76	b	12.69	e	9.13	b	9.38	c	3.66	b	2.60	b	1.14	c
18	20.38	b	10.74	b	17.31	c	12.25	a	10.25	c	4.25	b	3.24	b	1.46	c
19	28.38	a	10.72	b	19.69	b	8.50	b	11.66	b	4.01	b	3.54	a	1.36	c
20	17.63	b	10.69	b	11.88	e	8.88	b	8.79	c	4.40	b	2.80	b	1.44	c
21	17.38	b	10.68	b	16.31	c	11.50	a	9.60	c	3.95	b	3.01	b	1.31	c
22	15.31	b	10.67	b	12.31	e	11.00	a	9.18	c	4.10	b	3.03	b	1.43	c
23	16.13	b	10.68	b	17.63	c	11.50	a	9.33	c	4.22	b	2.94	b	1.30	c
24	19.63	b	11.06	b	20.13	b	8.63	b	9.18	c	4.16	b	2.79	b	1.21	c
25	18.38	b	10.51	b	15.13	d	8.38	b	9.83	c	4.40	b	3.07	b	1.33	c
26	27.06	a	11.55	a	11.25	e	4.38	c	8.01	c	2.94	c	1.72	c	0.53	e
27	22.75	a	11.83	a	10.94	e	4.13	c	6.05	d	2.07	c	1.24	c	0.38	e
28	21.75	a	11.35	a	11.88	e	3.25	c	6.88	d	1.98	c	1.60	c	0.37	e
29	23.00	a	10.92	b	15.44	d	8.50	b	10.14	c	4.20	b	2.89	b	1.27	c
30	17.75	b	10.79	b	17.19	c	9.00	b	9.21	c	2.82	c	2.76	b	0.94	d
31	18.19	b	11.54	a	11.88	e	4.63	c	4.87	d	2.26	c	1.35	c	0.63	e
32	22.50	a	10.53	b	19.13	b	14.38	a	14.35	a	7.42	a	4.47	a	2.07	a
33	23.38	a	10.71	b	17.94	c	8.25	b	11.00	b	4.46	b	3.34	b	1.29	c

Means followed by the same letters in the columns do not differ from each other by the Scott–Knott test at 5% probability level.

and

Table 4. Leaf area (LA), Dickson quality index (DQI), chlorophyll index (Chl), root length (RL), root surface projection (RSP), root surface area (RA), mean root diameter (MRD), and root volume (RV) of seedlings of the different C. canephora genotypes.

Treat	LA		DQI		Chl		RL		RSP		RA		MRD		RV
1	244.24	c	1.02	b	23.64	b	823.75	a	54.19	b	170.25	b	0.66	c	2.82	b
2	281.45	b	1.10	b	24.95	b	806.74	a	50.96	b	160.10	b	0.63	d	2.55	b
3	287.13	b	1.18	b	24.00	b	771.08	a	53.08	b	166.76	b	0.68	c	2.90	b
4	241.41	c	1.17	b	24.49	b	765.86	a	51.92	b	163.11	b	0.68	c	2.81	b
5	285.38	b	1.05	b	22.58	b	861.79	a	53.67	b	168.60	b	0.62	d	2.64	b
6	308.17	b	1.31	a	27.31	a	920.50	a	55.09	b	173.05	b	0.60	d	2.63	b
7	226.04	c	1.25	a	24.95	b	795.56	a	49.29	b	154.86	b	0.61	d	2.42	c
8	233.24	c	1.11	b	28.33	a	738.36	b	44.14	c	138.66	c	0.60	d	2.08	c
9	270.37	b	1.48	a	22.04	b	807.96	a	49.60	b	155.81	b	0.61	d	2.40	c
10	311.83	b	1.09	b	23.80	b	724.32	b	53.00	b	166.51	b	0.74	b	3.10	b
11	220.63	c	1.03	b	23.95	b	733.08	b	43.92	c	137.97	c	0.60	d	2.07	c
12	213.84	c	0.96	b	23.88	b	682.05	b	44.07	c	138.46	c	0.65	c	2.25	c
13	226.75	c	1.31	a	27.55	a	720.44	b	48.83	b	153.41	b	0.68	c	2.63	b
14	252.09	c	0.91	b	22.25	b	681.49	b	41.02	c	128.85	c	0.60	d	1.95	c
15	263.81	c	1.03	b	22.59	b	667.81	b	44.65	c	140.28	c	0.67	c	2.35	c
16	223.84	c	1.04	b	25.36	b	684.99	b	45.72	c	143.62	c	0.67	c	2.40	c
17	222.13	c	1.08	b	25.29	b	665.01	b	43.16	c	135.59	c	0.65	c	2.24	c
18	239.42	c	1.23	b	25.31	b	863.43	a	51.88	b	162.97	b	0.60	d	2.45	c
19	288.22	b	1.12	b	24.93	b	603.77	b	39.88	c	125.28	c	0.67	c	2.09	c
20	170.05	d	1.37	a	23.68	b	761.82	a	50.96	b	160.11	b	0.67	c	2.70	b
21	213.61	c	1.12	b	24.05	b	622.53	b	41.40	c	130.06	c	0.66	c	2.18	c
22	184.46	d	1.36	a	31.76	a	674.07	b	44.80	c	140.73	c	0.66	c	2.35	c
23	200.01	d	1.09	b	24.29	b	650.38	b	43.44	c	136.48	c	0.67	c	2.31	c
24	226.11	c	0.97	b	21.38	b	522.09	c	37.30	d	117.18	e	0.72	b	2.10	c
25	192.36	d	1.16	b	27.76	a	709.23	b	48.42	b	152.11	b	0.68	c	2.64	b
26	145.29	d	0.53	d	29.53	a	285.69	d	21.84	e	68.62	e	0.80	a	1.37	d
27	107.29	e	0.38	d	23.55	b	281.43	d	18.94	e	59.51	e	0.70	b	1.02	d
28	107.41	e	0.36	d	24.46	b	243.56	d	17.41	e	54.68	e	0.75	b	1.00	d
29	255.94	c	1.14	b	26.23	b	518.72	c	37.11	d	116.59	d	0.72	b	2.12	c
30	187.58	d	0.82	c	24.29	b	547.52	c	32.50	d	102.11	d	0.59	d	1.52	d
31	87.01	e	0.62	c	21.13	b	543.97	c	31.10	d	97.69	d	0.57	d	1.40	d
32	366.00	a	1.65	a	28.94	a	822.34	a	70.99	a	223.03	a	0.86	a	4.85	a
33	273.13	b	1.10	b	24.38	b	652.04	b	48.93	b	153.73	b	0.75	b	2.98	b

Means followed by the same letters in the columns do not differ from each other by the Scott–Knott test at 5% probability level.

present data and the mean test for crown diameter (CD), stem diameter (SD), plant height (Hei), number of leaves (NL), fresh shoot mass (FSM) and fresh root mass (FRM), dry shoot mass (DSM) and dry root mass (DRM), leaf area (LA), Dickson quality index (DQI), chlorophyll index (Chl), root length (RL), root surface projection (RSP), root surface area (RA), mean root diameter (MRD), and root volume (RV) of seedlings of the 33 evaluated C. canephora genotypes.

Table 3 shows significant variation for all morphological characteristics analyzed. Genotype 32 stood out for presenting the highest values for fresh and dry shoot and root mass, with no statistical difference in fresh and dry shoot mass of genotypes 2, 5, 6, 9, and 10, indicating high growth potential and seedling vigor.

Regarding crown diameter (CD), the highest values were observed in genotypes 19, 26, 3, 33, 29, 10, 11, 27, 17, 32, 7, 28 and 14. These results suggest a greater light interception capacity, favoring vegetative development. Regarding stem diameter (SC), the genotypes that stood out were 3, 27, 26, 31, 28 and 16.

Regarding plant height, genotypes 5, 10, and 12 presented the highest values. The lowest values were recorded for genotypes 27, 26, 28, 31, 20, 22, 17 (<13 cm), indicating slower seedling development.

For the number of leaves (NL), genotypes 9 (15 leaves), 32, 5, 1, 3, 18, 6, 21, 23, 16, 22, 8, 14, 15 and 7 stood out with the highest values, which may be associated with the larger photosynthetic area and, consequently, higher biomass production.

The highest values of fresh mass (FSM) and dry shoot mass (FRM) were observed in genotypes 32, 6, 10, 9, 2, 5, and 3, indicating greater accumulation of aerial biomass in these materials. For fresh mass (DSM) and dry root mass (DRM), genotype 32 showed superior performance, suggesting a more vigorous root system.

The results presented in Table 4 reinforce the presence of genetic diversity among genotypes, especially regarding physiological and root characteristics. Genotype 32 once again stood out, presenting the largest leaf area (366.00 cm²), Dickson quality index (1.65), chlorophyll index (28.94), root surface projection (70.99 cm²), root surface area (223.03 cm²), mean root diameter (0.86 mm), and root volume (4.85 cm³). These results indicate superior performance in both shoot and root systems. It is noteworthy that genotype 32 corresponds to clone A1, a commercial material already genetically selected, which was used as a control in the present experiment.

Genotypes 26, 27, 28, and 31 presented the worst performances in almost all variables, with very low leaf area, Dickson quality index, root length, and root volume values. These results indicate significant limitations in seedling vigor and quality, suggesting the exclusion of these genotypes from selection programs.

Overall, the results demonstrate that the evaluated genotypes exhibit broad phenotypic and physiological diversity, with genotype 32 (A1) presenting the highest values for most traits analyzed. Other genotypes that presented similar values for the analyzed characteristics were genotypes 2, 3, 4, 5, 6, 10, and 33, the latter of which has also been registered and commercialized. On the other hand, the lowest values were found for genotypes 27 and 28, which indicates less vigorous seedlings.

Pearson’s correlation coefficient analysis for seedling of the 33 C. canephora genotypes revealed 80 significant correlations (p < 0.05, t-test), of which 67 were positive and 13 were negative (Figure 1). The highest correlations were found between RSP and RA (1.00), FSM and DSM (0.96), DRM and DQI (0.96), FSM and LA (0.95). The lowest correlations were found for DRM and MRD (−0.01), RA and MRD (−0.01), and RSP and MRD (−0.01).

Principal Component Analysis (PCA) is a recognized and effective method for dimensionality reduction, allowing the conversion of a large number of variables into a smaller set while maintaining most of the information contained in the original data. In the present study, PCA was performed to determine the total variance observed in the 33 C. canephora genotypes based on the morphophysiological variables of coffee seedlings.

The total contribution of principal components 1 and 2 was 75.5% (Figure 2), with PC1 accounting for 63.2% of the total variance and contributing the most to the observed diversity.

PC1 showed negative factor loadings for most traits, with the exception of SD and CD, while PC2 indicated positive factor loadings for SD, Chl, CD, MRD, FRM, RV, LA, FSM, and DQI (Figure 3).

For cluster analysis, variables with correlations greater than 90% were removed. Based on the morphophysiological data obtained by analyzing the remaining 11 variables using the UPGMA, the 33 C. canephora genotypes were grouped into four clusters (Figure 4). Cluster III (blue) was considered the largest, containing 18 genotypes, while clusters IV, II, and I contained eight, four, and three genotypes, respectively.

From the morphological analyses, it was possible to verify differences among genotypes and confirm the existence of genetic diversity. However, for a deeper understanding of this diversity, it is important to use more in-depth analyses, such as molecular analysis.

For molecular analysis, 18 ISSR primers (

Table 5. Primer sequence, total number of bands (TNB), and number of polymorphic bands (NPB) used to characterize the C. canephora diversity.

Primers	Sequence	TNB	NPB
UBC 819	GTGTGTGTGTGTGTGTA	5	5
UBC 834	GAGAGAGAGAGAGYT	7	6
UBC 841	GAGAGAGAGAGAGAGAYC	5	5
UBC 842	GAGAGAGAGAGAGAGAYG	5	5
UBC 845	CTCTCTCTCTCTCTCTRG	4	4
UBC 848	CACACACACACACACARG	3	3
UBC 849	GTGTGTGTGTGTGTGTYA	5	4
UBC 854	TCTCTCTCTCTCTCTCRG	5	5
UBC 855	ACACACACACACACACYT	4	4
UBC 825	ACACACACACACACACT	13	13
UBC 857	ACACACACACACACACYG	5	5
UBC 859	TGTGTGTGTGTGTGTGRC	6	6
UBC 868	GAAGAAGAAGAAGAAGAA	3	2
UBC 875	ACACACACACACACACYG	10	7
UBC 879	CTTCACTTCACTTCA	10	7
UBC 889	DBDACACACACACACAC	10	8
UBC 895	AGGTCGCGGCCGCNNNNNNAT	10	8
Echt 5	AGACAGACGC	7	5

) were tested on the 32 C. canephora accessions that presented amplifications for these accessions; one was excluded due to absent DNA.

Molecular analysis using 18 ISSR primers yielded a total of 117 amplified bands, of which 102 (87.2%) were polymorphic, indicating genetic diversity among the 32 C. canephora genotypes evaluated (Table 5).

Of the 18 primers used, UBC 825, UBC 895, UBC 889, UBC 879, and UBC 875 performed better for the genotypes under study, with greater number of amplified bands. On the other hand, primer UBC 868 presented the lowest number of total (3) and polymorphic (2) bands, indicating lower genetic discrimination capacity. However, even primers with the lowest number of bands contributed to the composition of the genetic profile of genotypes, being useful for complementing the analyses.

Molecular analysis using ISSR markers allowed the formation of 16 distinct groups among the 32 C. canephora genotypes, with cutoff of 50% (Figure 5). The largest group formed was group XII, composed of four genotypes (18, 19, 17, and 16).

The Mantel test was employed to assess the correlation between the genetic distance matrix, obtained through ISSR markers, and the morphophysiological distance matrix, derived from the characteristics evaluated in the C. canephora seedlings. The results showed a significant positive correlation (r = 0.228; p = 0.018) between the two distance matrices, indicating a statistically significant association, albeit of low to moderate magnitude, between the molecular genetic diversity and the morphophysiological variability of the studied genotypes’ seedlings.

4. Discussion

There was genetic diversity among the 33 C. canephora genotypes analyzed. This diversity is essential for strengthening breeding programs and ensuring sustainable coffee production [21]. Studies have shown that genetic diversity among C. canephora genotypes can significantly influence traits such as photosynthetic efficiency, nutrient uptake, and stress tolerance, which are vital for adaptation to different environmental conditions [20,26,39]. Furthermore, the identification of distinct genetic groups allows for targeted breeding strategies, which can improve yield and resistance to pests and diseases [21,26], while contributing to the preservation of genetic resources [40].

The genetic diversity observed in the seedling phase offers opportunities for the early selection of more vigorous and adapted plants, boosting crop productivity and sustainability [41] and accelerating the breeding program toward the release of new clones.

The results found by the mean test analysis indicate that, in addition to genotype 32, a commercial clone, other genotypes such as 2, 3, 4, 5, 6, 10, and 33 also stood out for presenting similar and high values in several traits desirable for the development of vigorous seedlings with good root structure and efficient photosynthetic capacity. The similarity in the values of these traits among these genotypes suggests that they also have the potential to be used in breeding programs and in the production of high-quality seedlings. The use of superior genotypes in the production of quality C. canephora seedlings is important due to their impact on overall plant growth and performance [7]. Furthermore, integrating these superior genotypes into seedling production systems can lead to better field establishment and success rates in coffee plantations [41].

On the other hand, genotypes 27 and 28 presented the lowest values for the variables analyzed, indicating lower seedling vigor. This can be attributed to genetic factors that negatively influence plant development, resulting in lower resource use efficiency and reduced growth capacity [8]. Analyzing seedlings of conilon coffee genotypes, Lani et al. (2005) [6] observed that high-quality genotypes demonstrated an increase of up to 20% in production compared to lower-quality seedling genotypes. Furthermore, inadequate seedling quality can increase competition with weeds, which negatively affects nutrient and water availability, affecting plant growth and root development [42]. However, despite being less vigorous, these genotypes are important for the genetic diversity of the species and may present other favorable characteristics for cultivation that were not addressed in this study, such as productivity and resistance to biotic and abiotic factors [26].

Regarding the correlation between variables collected for coffee seedlings, high correlation was observed between dry root mass and DQI, as well as between fresh shoot mass and leaf area, revealing strong interdependence between traits. Such positive correlations indicate that genotypes with more vigorous growth tend to develop a more robust root structure and accumulate greater amount of biomass, which may be indicative of greater vigor and potential for adaptation to field conditions, since the root system is directly related to plant height and yield in C. canephora [9].

On the other hand, lower correlations, with values close to zero, suggest low interdependence between traits, indicating, for example, that root diameter is not a determining factor for root biomass accumulation. This lack of significance may indicate that other factors, such as genetic or environmental factors, are more strongly influencing these traits. Genetic mapping studies highlight the complexity of trait interactions, where quantitative trait loci (QTL) for plant yield and height have been identified, but the intricate relationships between various traits do not always produce strong correlations [43,44].

In principal component analysis, PC1 showed negative factor loadings for most of the traits analyzed, except for stem diameter and the Spad index, which measures leaf chlorophyll. This suggests that PC1 is inversely related to most morphophysiological traits of seedlings, as negative loadings indicate that, as one variable increases, another decreases [45]. In practical terms, genotypes with higher values on PC1 tend to have lower values in these variables, which may indicate negative association between vigor and these specific variables. The exception of stem diameter and the Spad index (chlorophyll) may suggest that, regardless of the other variables, stem development and chlorophyll content are limited or even increased in genotypes that express these traits.

PC2 showed positive factor loadings for stem diameter and the Spad index, crown diameter, root diameter, fresh root mass, root volume, leaf area, fresh root and shoot mass and the Dickson Quality Index (DQI). This indicates that PC2 is directly related to these traits, indicating that genotypes with high PC2 values tend to exhibit more vigorous development regarding these traits. The positive value with the DQI, which is a widely used index to evaluate seedling quality, indicates that these genotypes have high quality potential [7].

The application of PCA-based morphophysiological analysis in the breeding and selection of C. canephora varieties can significantly increase the efficiency and accuracy of genotype selection. By using principal component analysis (PCA), it is possible to reduce the dimensionality of complex datasets, enabling the identification of key traits that contribute to the quality and adaptability of coffee varieties, similar to the methods used in the selection of C. arabica genotypes based on sensory and physical characteristics [46].

Cluster analysis is important because it clearly and efficiently reveals the genetic divergence among the analyzed materials. In this sense, cluster analyses revealed that the genotypes have genetic diversity, being grouped into four clusters using the UPGMA.

Group III stands out as the largest group, containing 18 of the 33 genotypes, indicating genetic similarity among most genotypes analyzed. This concentration of genotypes in this group suggests that they have similar morphophysiological characteristics, possibly resulting from a common genetic base [47]. This dense clustering may be advantageous for breeding programs seeking uniformity and stability in specific traits, such as vigorous growth, resistance to biotic or abiotic stresses, and productivity [48].

On the other hand, groups IV, II, and I, containing eight, four and three genotypes, respectively, stand out for their genetic diversity in relation to the larger group III. The fact that these smaller groups contain few genotypes indicates that these individuals present unique or less common morphophysiological characteristics, making them genetically different from the larger majority group. This type of diversity is crucial for breeding programs, as these unique genotypes may present rare alleles or genetic characteristics that provide them specific adaptive or agronomic advantages, such as greater resistance, better adaptation to extreme climatic conditions, or superior sensory qualities [9,39,40,49]. According to Senra et al. (2022) [10], obtaining genetic diversity among coffee genotypes in the initial phase of experimentation is important to determine selection and recombination strategies for controlled crosses.

The 18 ISSR primers used were efficient in observing genetic diversity among genotypes, allowing for the verification of similarity and dissimilarity among genotypes through the bands presented in the analysis. Other authors also used ISSR markers when studying C. canephora [14,15,16,17,18].

Through molecular analysis of the 32 genotypes evaluated, it was possible to find 16 distinct groups, revealing the presence of genetic diversity among genotypes. Analyses using molecular markers performed by other authors also showed broad genetic diversity in C. canephora, highlighting the relevance of this information for germplasm hybridization and the development of clonal varieties [50,51,52]. The genetic diversity revealed by ISSR markers confirms the allogamous nature of the species and is in line with other studies that also found high diversity in C. canephora populations in Espírito Santo [26,40], indicating that the local genetic base is rich for breeding programs.

The most genetically close genotypes were genotypes 31 and 32, which share DNA fragments similar to genotype 28, all belonging to group I. This genetic proximity suggests a common genetic structure but highlights that the individuals are not identical and may therefore be genetically close, with few differences between them.

The largest group identified was Group XII, composed of four genotypes (18, 19, 17, and 16). The presence of these genotypes in a single group suggests genetic similarity among them. This group may represent a set of genotypes that share a common genetic base or that have been subjected to similar selection, resulting in convergent genetic traits. The identification of a homogeneous group is useful for breeding programs, as it allows the selection of genotypes with desirable agronomic traits, ensuring consistent traits and maximizing hybrid vigor [53].

Particularly interesting are the groups that contain only one genotype (Groups II, III, VII, X, and XVI). These isolated genotypes possibly contain unique or differentiated genetic characteristic traits not shared with other genotypes in the study. These genotypes are important for the conservation of genetic resources, as they may contain valuable traits that, if preserved and incorporated into breeding programs, could improve the characteristics of C. canephora [54].

The correlation analysis between genetic and morphological diversity, assessed by the Mantel test, reinforces this interpretation by indicating that part of the observed phenotypic variability is associated with genetic divergence between genotypes. The Mantel statistic of 0.228 suggests that approximately 22.8% of the variation observed in the morphological characteristics can be explained by the genetic diversity detected by the ISSR markers. This value, despite being considered low to moderate according to correlation classification [55], is statistically significant (p < 0.05).

These results are consistent with previous studies on coffee that utilized the Mantel test. Ferrão et al. (2013) [56] observed significant correlations between molecular markers and morphological traits in C. canephora, demonstrating the effectiveness of this approach for coffee germplasm characterization. Similarly, Leroy et al. (2014) [57] found significant correlations between different types of molecular markers when studying core collections of C. canephora.

The low to moderate correlation (r = 0.228) can be attributed to several factors. ISSR markers are considered neutral and randomly distributed in the genome, meaning they may not be closely linked to the quantitative trait loci (QTLs) controlling the measured morphological characteristics [58]. Furthermore, morphophysiological characteristics are often controlled by multiple genes and are typically influenced by environmental factors, thus reducing the direct correlation with neutral markers [59].

Therefore, the significant correlation found in the present study validates the complementary use of ISSR markers and morphological traits for characterizing genetic diversity in C. canephora. This result has important practical implications for pre-breeding programs, as it suggests that selection based on morphological characteristics may, at least partially, reflect the molecular genetic diversity of the genotypes, facilitating the identification of genetically diverse material for targeted crosses.

5. Conclusions

There is genetic diversity among C. canephora genotypes under study, indicating potential for use in breeding programs and for the conservation of the species’ genetic resources.

Genotypes 2, 3, 4, 5, 6, and 10, along with the genotypes used as controls (32 and 33), stood out as those that produced the most vigorous seedlings for the traits evaluated, which may contribute to good establishment of seedlings in the field and overall plant performance.

The results obtained through the Mantel test demonstrated a positive and significant correlation (r = 0.228; p = 0.018) between the genetic and morphophysiological distances of the C. canephora genotypes, evidencing that the morphophysiological variability observed in the seedlings reflects, in part, the genetic diversity detected by the ISSR markers.

Finally, the finding that the morphophysiological and molecular groupings differ demonstrates the importance of carrying out studies with field evaluations in adult plants for a deeper understanding of the diversity and potential of these genotypes.