Genetic Parameters Estimated in the Early Growth of Dimorphandra mollis Benth. Progenies

Abstract

The extractivism of Dimorphandra mollis Benth., which is a native tree from the Brazilian Cerrado biome, popularly known as fava d’anta, combined with the reduction in native vegetation area in the country over the years may result in a decrease in the specie’s natural populations. The objective of this study was to estimate the quantitative genetic parameters in nursery, hardening, and field phases, based on a progeny test. The experimental design adopted was randomized blocks (six blocks for the nursery and hardening phases, and four blocks for the field phase with 5 plants/plot and 72 mother trees), with evaluations of the collar diameter and seedling height at 30, 90, 150, 480, and 570 days after sowing, between the production and planting phases. Among the coefficients of variance, the phenotypic and additive ones showed the highest values. Heritabilities for height ranged from moderate to high (0.15 to 0.43), indicating good genetic control of the traits, high potential for selection, and possibility of genetic gains. The genetic divergence of the progenies resulted in division into five groups, confirming the existence of genetic variability among the evaluated progenies and the potential for conservation and breeding programs.

1. Introduction

Originally from the Cerrado biome of Brazil, Dimorphandra mollis Benth., commonly known as fava d’anta, is a tree of the legume family, with xerophytic habits, allowing it to occur in the different formations within the biome ranging from open grasslands with scattered shrubs and small trees to forested savanna formations. Its reproductive system is classified as facultative allogamous, and it has high intrapopulation and non-inbreeding genetic diversity [1,2]. Its fruits are exclusively obtained through the extractivism of specimens in their natural habitat, harvested at the ripening stage while still green due to the high content of bioflavonoids in this phase [3]. High bioflavonoid content is desired when extracting these beans, since the consumption of this plant compound has been suggested to present a wide range of health benefits [4]. Furthermore, bioflavonoids find widespread use in the pharmaceutical industry [5,6,7].

A lack of studies in the management techniques of any extractivism activity can affect a species’ population dynamics over time. In many cases, extraction can occur in a predatory or disorganized manner, without attention to the maintenance, reproduction of populations, and the development of individuals [8]. Considering the extractivism of D. mollis, ref. [9] has reported the risk of genetic erosion (considered the loss of genetic diversity between populations over time) that this activity represents for the species populations in the northern region of Minas Gerais state in Brazil, which is an important D. mollis extractivism region. In Piauí state, which also holds natural areas of D. mollis with extractivism practices, the genetic risk panorama of the species has not yet been assessed.

Habitat fragmentation can have serious consequences on forest health by reducing the effective population size, increasing inbreeding, loss of genetic diversity, and a discontinuous pattern of genetic variation, mainly due to limited gene flow [10]. The commercial interest in D. mollis bean pod extractivism is mainly due to the presence of the active ingredients in its fruits, especially bioflavonoids, among which rutin and quercetin stand out [3,11]. According to [12], in the 1990s, fava d’anta was responsible for 50% of the world’s rutin production, being extracted mainly from the States of Maranhão, Piauí and Minas Gerais. From 2011 to 2023, Brazil exported 1808.5 tons of rutin and derivatives, at an average price of USD 32.49 kg⁻¹, highlighting the importance of the species [13,14].

The reduction in natural populations may occur due to the predatory extractivism of D. mollis, and this effect can be potentialized by the increasing deforestation rates in the Cerrado Biome, specifically in the Cerrado-Caatinga transition zone which covers the present study area [15]. Investigations into the genetic basis of natural populations are crucial for the sustainable management of forest genetic resources, biodiversity, and adaptation to climate change [16]. Studies on the genetic characterization of D. mollis populations in the northern region of Minas Gerais revealed a high genetic diversity within populations [17]. Therefore, studies on phenotypic traits in progeny tests play a crucial role in predicting future genetic differentiation. The study and preservation of the genetic diversity of a tree species are highly relevant, aiding in the long-term survival and adaptability of the species, and presents a vital importance for those that are candidates for domestication. Genetic diversity is of utmost importance, especially facing projected climate changes in the upcoming decades, underscoring the imminent importance of genetic resource conservation [18].

Thus, studies of genetic variation within populations, through the analysis of quantitative traits in progeny tests using a mixed linear model and parameters estimated by the REML/BLUP method (restricted maximum likelihood/best linear unbiased prediction) can be used to predict the genetic superiority of individuals, aiding in the selection of those with the highest breeding values for further examination, as it has been applied in Parkia platycephala Benth. [19], Eugenia dysenterica DC. [20], and Roupala montana var. brasiliensis [21], and allows for determining the proportion of adaptive genetic variation that may respond to environmental changes or be exploited in forest breeding programs [22]. According to [23], the assessment of genetic divergence among progenies is a fundamental step for production or conservation, allowing for a more accurate direction of the strategy to be adopted to broaden the genetic base or generate gains.

Successful strategies for the conservation, management, and utilization of a species require a precise evaluation of the genetic variation within the population [24]. Therefore, the present study aims to assess the genetic variability within a native population of D. Mollis in the southern region of Piauí state, and to estimate genetic parameters for the main silvicultural polygenic traits. The test was implemented for ex situ conservation with the initial purpose of forming a base population with broad genetic variability. This population will later be utilized in a genetic improvement program.

2. Materials and Methods

2.1. Seed Collection Area

The collection of D. mollis seeds was carried out in the town of São Gonçalo do Gurguéia, Piauí, aiming to conserve genetic material and initiate a breeding program for the species. Seventy-two mother trees distributed across four populations, namely, populations 01 and 02 with one descendant each, in population 03 sixty-eight descendants and in population 04 two descendants (Figure 1), were selected for the experiment.

2.2. Seedling Production and Experimental Design

In December 2021, the fruits collected from D. mollis were processed, and seedling production was carried out in 150 m³ tubes filled with the commercial substrate and housed in the greenhouse of the Federal University of Piauí (UFPI), Bom Jesus Campus (9°04′46″ S and 44°19′38″ W). At 150 days of the hardening phase, the seedlings were transplanted into two-liter plastic bags. The region has a hot and humid climate, classified by Köppen as Aw, rainy tropical with a dry season in winter and average temperature of the hottest month greater than 22 °C [25]. During the nursery and hardening phases (Figure 2A,B), the experiment was set up in a completely randomized block design (CRBD), consisting of six blocks with 5 (plants/plot) and 72 mother trees, totaling 2160 seedlings.

In the field phase (Figure 2C,D), starting in January 2023, planting occurred at the Alvorada do Gurguéia Experimental Farm (FEAG/UFPI) (8°22′37″ S and 43°51′34″ W). The climate of the region is classified as semiarid, designated as Bsh, according to the Köppen climate classification [25]. A CRBD was also adopted with a reduction to four blocks with 5 (plants/plot) and 72 mother trees, totaling 1440 individuals.

Evaluations of the stem diameter in millimeters (mm) and height in centimeters (cm) were conducted at 30, 90, and 150 days in the nursery, and in the field at 480 and 570 days, using a digital caliper (Digimess-057856) and a graduated ruler.

2.3. Statistical Analysis and Genetic Parameter Estimates

Genetic parameter estimation was conducted using the R software version 4.1.2 [26] and lme4 package [27]. Variance components and genetic parameter estimates were obtained for the variables height (H in cm) and stem diameter (CD in mm) using the REML/BLUP method, with a mixed linear model (Equation (1)), considering open-pollinated progeny tests.

Yijk = µ + bi + tj + (tb)ij + eijk

(1)

where Yijk is the phenotypic value of the k-th individual of the j-th progeny in the i-th replication; µ is the fixed effect of the overall mean of the trait under analysis; bi is the fixed effect of the i-th replication; tj is the random effect of the j-th progeny; (tb)ij is the random effect of the interaction between the j-th progeny and the i-th replication; and eijk is the effect random of the relative experimental error of the k-th tree within the j-th progeny in the i-th replication.

The significance of the random and fixed effects was verified using the ranova and anova functions, respectively, implemented in the lme4 package [24]. The significance of the random effects (progeny and plot) was checked with the likelihood ratio test (LRT), using a Chi-square test with 1 degree of freedom. To assess the significance of the fixed effects (block), the F-test at a 5% probability level was applied.

The following variances were computed: genetic $\left({\widehat{\sigma }}_g^2\right)$; additive $({\widehat{\sigma }}_a^2=4\times {\widehat{\sigma }}_g^2)$; between progenies $\left({\widehat{\sigma }}_e^2\right)$; within progenies $\left({\widehat{\sigma }}_d^2\right);$and phenotypic $({\widehat{\sigma }}_p^2={\widehat{\sigma }}_a^2+{\widehat{\sigma }}_d^2+{\widehat{\sigma }}_e^2)$. From those variances, the values for narrow-sense heritability $({\widehat{h}}_a^2=\frac{{\widehat{\sigma }}_a^2}{{\widehat{\sigma }}_p^2})$, heritability within progenies ${(\widehat{h}}_d^2=\frac{0.75.{\widehat{\sigma }}_a^2}{{\widehat{\sigma }}_d^2}),$ and average heritability of progenies ${(\widehat{h}}_m^2=\frac{{\widehat{\sigma }}_a^2}{{\widehat{\sigma }}_a^2+\frac{{\widehat{\sigma }}_e^2}{J}+\frac{{\widehat{\sigma }}_d^2}{JK}})$were calculated, where J is the average number of replications and K is the average number of plants per plot. The following coefficients were computed: genetic variation (CVgp (%) =$\frac{\sqrt{0.25}\times {\widehat{\sigma }}_a^2}{\overline{m}}·100)$, additive genetic (CVgi (%) = $\frac{\surd ({\widehat{\sigma }}_a^2}{\overline{m}}·$100), and experimental (CVe (%) = $\frac{\surd {\widehat{\sigma }}_e^2}{\overline{m}}·$100), where $\overline{m}$ is the observed mean for the evaluated trait. In addition, the accuracy of progeny selection ($\widehat{r}$ =$\surd h_a^2$) and relative coefficient of variation = (b = $\frac{CVgp}{CVe}$) were also estimated. For all estimates, we followed [28].

Following the divergence analysis, a ranking of the progenies was carried out according to [29,30], who suggested the selection of the 30 best progenies, using the predicted genetic values and estimated gains through the REML/BLUP method [31]. The predicted breeding values (BLUPs) for the evaluated traits were obtained using the ranef function in the lme4 package. To obtain the ranking of progenies, the BLUP method was used to predict the genotypic values of each individual in relation to the total height (H, cm) and collar diameter (CD, mm), at the three measurement ages: 30 days (nursery), 150 days (hardening phase) and 570 days (field phase under rainfed conditions). A total of 72 progenies were evaluated, from which the best 30 progenies were identified at different development stages.

Data at 570 days were analyzed using multivariate statistics in the R environment, with the use of the MultivariateAnalysis package [32]. For cluster analysis, the UPGMA method [33] was employed, and the number of clusters in hierarchical methods was determined according to [34], based on the relative size of the fusion levels (average Euclidean distance) in the dendrogram. Ref. [34] proposed the selection of the number of groups at step j that first satisfies the inequality: αj > θk, where αj represents the value of the fusion level distance corresponding to step j (j = 1, 2,…, n); and θk represents the reference cutoff value, given by the following: the mean of the α values; the standard deviation of the α values; and k is a constant value (k = 1.25), fused for the definition of the optimal number of groups [35].

3. Results

The statistical analyses indicated that there were no significant differences among the progenies for the collar diameter at 480 and 570 days; however, there was a significant difference for the height variable in all evaluation periods (

Table 1. Significance of random effects (Chi-square–LRT test) and fixed effects (mean squares–F-test) of both models used in the analysis on height data (H) and collar diameter (CD) on the divergence analysis of seedling progenies of Dimorphandra mollis Benth. evaluated at 30, 90, 150, 480, and 570 days after the establishment of the progeny test.

Effect	Height					CD
	30 Days	90 Days	150 Days	480 Days	570 Days	30 Days	90 Days	150 Days	480 Days	570 Days
Fixed {Block}	4.952 **	6.4938 **	29.753 **	13.709 **	10.414 **	4.5732 **	8.0143 **	53.287 **	17.683 **	13.493 **
Random {Progeny/Plot}	115.603 ** 16.004 **	36.797 ** 57.572 **	29.362 ** -	43.171 ** -	12.323 ** -	48.475 ** 39.071 **	5.987 * 89.708 **	16.951 ** -	2.0889 ns -	0.86691 ns -

^ns not significant; * significant at 5% probability; ** significant at 1% probability.

). The last measurements, 480 and 570 days, occurred at the peak of the dry season, contributing to the reduced growth rate and development of the progenies. Among the variances (

Table 2. Estimated genetic parameters for the seedlings of Dimorphandra mollis Benth. progenies, evaluated at 30, 90, 150, 480, and 570 days after the establishment of the progeny test, for height (H, cm) and collar diameter (CD, mm).

	Variance Component
	Nursery				Hardening		Field
Parameter	30 Days		90 Days		150 Days		480 Days		570 Days
	H (cm)	CD (mm)	H (cm)	CD (mm)	H (cm)	CD (mm)	H (cm)	CD (mm)	H (cm)	CD (mm)
$\mathrm{Additive}\mathrm{genetic}\mathrm{variance}\left({\widehat{\sigma }}_a^2\right)$	1.973	0.015	2.326	0.011	1.437	0.016	39.299	0.067	33.212	0.121
$\mathrm{Residual}\mathrm{variance}\left({\widehat{\sigma }}_e^2\right)$	2.384	0.032	5.485	0.074	8.346	0.131	89.872	0.999	106.360	1.605
$\mathrm{Variation}\mathrm{within}\mathrm{progenies}\left({\widehat{\sigma }}_d^2\right)$	0.180	0.004	0.894	0.016	-	-	-	-	-	-
$\mathrm{Phenotypic}\mathrm{variance}\left({\widehat{\sigma }}_p^2\right)$	4.537	0.051	8.705	0.101	9.783	0.148	129.171	1.066	139.572	1.726
$\mathrm{Narrow}-\mathrm{sense}\mathrm{heritability}{(\widehat{h}}_a^2)$	0.43 ± 0.08	0.30 ± 0.07	0.27 ± 0.06	0.11 ± 0.04	0.15 ± 0.05	0.11 ± 0.04	0.30 ± 0.08	0.15 ± 0.04	0.24 ± 0.07	0.07 ± 0.04
$\mathrm{Heritability}\mathrm{within}\mathrm{progenies}{(\widehat{h}}_d^2)$	0.89	0.73	0.66	0.35	-	-	-	-	-	-
$\mathrm{Average}\mathrm{heritability}\mathrm{of}\mathrm{progenies}{(\widehat{h}}_m^2)$	0.84	0.74	0.74	0.48	0.51	0.42	0.64	0.21	0.55	0.23
Coefficient of genetic variation (CVg p %)	10.62	3.79	7.00	2.39	5.02	2.62	11.05	3.30	10.05	3.96
Coefficient of additive genetic variation (CVgi %)	21.24	7.58	14.00	4.78	10.04	5.25	22.11	6.58	20.10	7.92
Coefficient of experimental variation (CVe %)	23.34	11.03	21.49	12.18	24.21	14.96	33.43	25.50	33.96	28.87
Relative coefficient of variation (b)	0.45	0.34	0.33	0.20	0.21	0.17	0.33	0.13	0.28	0.14
$\mathrm{Accuracy}\mathrm{in}\mathrm{progeny}\mathrm{selection}(\widehat{r}$%)	66.0	54.0	52.0	33.0	39.0	33.0	55.0	25.0	49.0	26.0
Mean	6.62	1.62	10.90	2.23	11.93	2.42	28.36	3.92	28.69	4.39

), for the additive (${\widehat{\sigma }}_a^2$) and within progeny variances ${(\widehat{h}}_d^2)$, $\left({\widehat{\sigma }}_a^2\right)$ showed higher values, with values ranging from 0.01 to 39.29, and at 480 days, for the height character, it presented the highest value among the evaluations. The narrow-sense heritability ${(\widehat{h}}_a^2)$ presented a range of 0.07 to 0.43, while the heritability within progenies ${(\widehat{h}}_d^2)$ ranged from 0.35 to 0.89, and the average heritability ${(\widehat{h}}_m^2)$ ranged from 0.21 to 0.84. All calculated heritabilities showed a reduction as the evaluation age increased. During development in the nursery, the average heritability behaved similar between the progenies; however, in the hardening and field phases, the progenies showed a reduction in heritability, which could be due to the genotype–environment interaction or even genetic inheritance from the related mother trees (epigenetics).

The genetic variation coefficient results (CV_gp %) ranged from 2.39 to 11.05%, indicating variation among the progenies and suggesting the possibility of genetic gains in both evaluated parameters (height and collar stem diameter). The values of additive genetic coefficient of variation (CV_gi %) ranged from 4.78 to 22.11%. The experimental coefficient of variation (CV_e %) ranged from 11.03 to 33.96%. The relative coefficient of variation (b) and accuracy in progeny selection ($\widehat{r}$) both indicated that it is possible to obtain gains in the selection of genetic material.

The top 30 progenies for the ages of 30, 150, and 570 days are presented in

Table 3. Progeny ranking via the BLUP/REML of estimated gains in height (H, cm) and collar diameter (CD, mm) of the 30 best progenies of Dimorphandra mollis Benth at three different development phases: nursery (30 days), hardening (150 days), and field (570 days after sowing). Light- and dark-shaded values indicate progenies that were identified with superior H and CD growth in all three development phases.

	Nursery (30 Days)				Hardening (150 Days)				Field (570 Days)
Progeny	H (cm)	Progeny	CD (mm)	Progeny	H (cm)	Progeny	CD (mm)	Progeny	H (cm)	Progeny	CD (mm)
68	2.19	15	0.12	25	1.05	45	0.11	44	6.51	44	0.16
15	1.26	33	0.11	55	1.03	25	0.11	28	3.67	33	0.12
28	1.22	31	0.09	9	1.01	33	0.10	9	3.28	59	0.12
19	1.09	71	0.09	68	0.92	51	0.07	45	3.14	65	0.11
55	0.77	28	0.07	19	0.87	70	0.06	47	2.38	45	0.11
26	0.73	19	0.06	7	0.70	48	0.06	51	2.35	51	0.10
33	0.70	26	0.06	67	0.50	19	0.06	27	2.35	28	0.10
31	0.66	57	0.06	51	0.48	65	0.05	33	2.18	72	0.09
25	0.65	41	0.06	32	0.43	55	0.04	1	2.02	9	0.09
10	0.64	5	0.06	20	0.41	62	0.04	40	1.97	35	0.08
22	0.60	51	0.05	30	0.39	32	0.04	60	1.84	24	0.08
71	0.58	65	0.05	60	0.39	38	0.04	72	1.80	17	0.07
40	0.57	47	0.04	22	0.39	30	0.04	62	1.45	47	0.07
47	0.51	38	0.04	33	0.36	68	0.04	2	1.43	41	0.07
23	0.51	39	0.04	27	0.34	41	0.03	3	1.37	61	0.06
9	0.46	34	0.04	5	0.28	60	0.03	68	1.32	30	0.05
60	0.46	56	0.03	10	0.28	71	0.03	30	1.17	69	0.05
70	0.39	53	0.03	66	0.27	64	0.03	25	1.10	7	0.05
30	0.38	32	0.03	48	0.23	34	0.03	26	1.03	70	0.04
5	0.36	6	0.03	43	0.19	44	0.03	7	0.98	60	0.04
56	0.32	10	0.02	58	0.19	4	0.03	46	0.97	46	0.03
14	0.27	27	0.02	59	0.17	49	0.03	69	0.85	71	0.03
67	0.23	43	0.02	47	0.16	27	0.03	17	0.74	14	0.03
4	0.21	23	0.01	12	0.16	5	0.02	41	0.67	67	0.02
6	0.19	3	0.01	16	0.15	21	0.02	67	0.64	2	0.02
65	0.19	22	0.01	50	0.12	23	0.02	42	0.61	6	0.02
41	0.17	4	0.01	61	0.11	43	0.02	52	0.56	11	0.02
16	0.16	72	0.01	35	0.11	20	0.02	59	0.43	27	0.01
27	0.15	40	0.01	18	0.08	52	0.01	4	0.38	66	0.01
48	0.13	7	0.01	41	0.08	39	0.01	6	0.27	53	0.01

. From this ranking, it was found that ten individuals had gains in height in all phases of evaluation, namely 9, 25, 27, 30, 33, 41, 47, 60, 67, and 68; and for the collar diameter, only three individuals: 27, 33 and 41.

Using the Unweighted Pair Group Method with Arithmetic Mean (UPGMA) clustering method with an average Euclidean distance at the level of progenies, it was possible to identify the cutoff point using the Mojena method on the dendrogram (Figure 3) and divide the progenies into five groups, with Group I formed by a single progeny, Group II by 10 progenies, Group III by 4 progenies, Group IV by 20 progenies, and Group V by 37 progenies. Progenies within the same group exhibit genetic similarity to each other. Among the five groups of genetic dissimilarity found, Groups III, IV, and V stand out for their greater genetic potential for the traits considered in the research.

4. Discussion

The progeny test for D. mollis did not show a significant difference in the field phase possibly due to factors such as adaptation to the environment, water availability, competition or even the development of the seedling’s root system during establishment in the field. According to [36,37], the diameter of a tree stem can be sensitive to different water levels, which can be modulated by turgescence pressure or the osmotic potential of the plant, and mainly by the water potential of the xylem. The diameter of the stem may also vary according to the water content of the stem, even within a period of one day [37].

Estimates of genetic parameters can vary according to the population, environment, trait, age, and method used, until the plant reaches its full development in the field. During this period, there may be greater or lesser environmental influence on the expression of traits related to individual growth [38,39]. The existence of genetic variability and the possibility of obtaining gains through selection indicates that the experimental design applied in this study ensured environmental control in the progeny test; therefore, the parameter estimates obtained through the deviance analysis are reliable [40,41].

Heritability is a frequently estimated parameter as it indicates a strong correlation with the success of genetic improvement through selection. Ref. [42] classified genetic heritability values ranging from 0.1 to 15% as low, 15 to 50% as moderate, and over 50% as high. Heritability is highly important for genetic selection, as alleles and their effects segregate into subsequent generations [43]. Average heritability ${(\widehat{h}}_m^2)$ for height and collar diameter varied from moderate to high in all phases of the present study (Table 2). These heritability values indicate genetic variability that is inheritable within the population or the relative amount of genetic variance useful for improving offspring through sexual propagation [30,44].

According to the classification proposed by [45], the narrow-sense heritability ${(\widehat{h}}_a^2)$for height at 30 days (0.43) can be considered moderate, and at 90, 480, and 570 days (0.27, 0.30 and 0.24, respectively) high, while at 150 days (0.15), the value can be classified as low. Regarding collar diameter, at 30, 90, 150, 480, and 570 days the heritability values can be classified as low. Within-progeny heritability ${(\widehat{h}}_d^2)$ values ranged from moderate to very high, varying between 0.35 and 0.89, while the mean heritability ${(\widehat{h}}_m^2)$ values ranged from high to very high, with variation between 0.21 and 0.84. This indicates an interaction between additive and environmental effects throughout the evaluated progeny growth period [41].

The observed values for heritability in this study are consistent with those reported in the literature for native forest species of the Cerrado biome. Ref. [46] estimated the mean heritability ${(\widehat{h}}_m^2)$ values for height and diameter at the breast height of 0.38 and 0.32, respectively, for the seedlings of Dipteryx alata at 900 days after sowing. Ref. [20] evaluated progenies of Eugenia dysenterica DC. at 360 days after sowing and found ${\widehat{h}}_m^2$ values for the height and collar diameter of 0.68 and 0.60, respectively. Ref. [19] assessed the species Parkia platycephala Benth. in the nursery phase and observed ${\widehat{h}}_m^2$ values for the height and collar diameter of 0.27 and 0.51, respectively. However, in the field phase, the authors recorded lower ${\widehat{h}}_m^2$values (0.11 for height and 0.20 for collar diameter), indicating that the progenies in the present study, with ${\widehat{h}}_m^2$ values for the heights at 30, 90, 150, 480, and 570 days of 0.84, 0.74, 0.51, 0.64, and 0.55, and the collar diameters of 0.74, 0.48, 0.42, 0.21, and 0.23, experienced a positive genotype–environment interaction throughout the evaluated plant development phases. Ref. [47] evaluated progenies of A. aculeata at 4 to 5 years of age and reported ${\widehat{h}}_m^2$ values of 0.63 for height and 0.57 for collar diameter.

The experimental coefficient of variation CVe (%) estimated in the present study, ranging from 11.03 to 33.96%, presented lower values than the ones found in the literature for native forest species, and this is an interesting fact because, the smaller the environmental variation, the greater the experimental control, demonstrating that the current experiment had low environmental influence. Ref. [21] estimated CVe values for the seedlings of Roupala montana var. brasiliensis at 180 days ranging from 25.55 to 57.14%. Ref. [48] evaluated the seedlings of Dipteryx alata Vogel. at 360 days and found CVe values ranging from 39.86 to 52.36%. Ref. [49] assessed height in Jacaranda cuspidifolia (Mart.) progenies at 360 and 720 days after sowing, and found CVe values of 25.8% and 31.7%, respectively. Ref. [20] evaluated progenies of Eugenia dysenterica DC. at 360 days and found CVe values for the diameter and height of 20.57% and 40.63%, respectively. According to [46], high-amplitude values of the experimental coefficient of variation are expected for field tests in native species.

An important indicator of genetic variability to be explored for breeding purposes is the coefficient of genetic variation of the trait (CVgp %), which is directly proportional to the genetic variance, and should be equal to or greater than the coefficient of experimental variation [47,50]. In this study, CVgp ranged from 2.39% (low) for diameter at 90 days to 11.05% (high) for height at 480 days. This study demonstrates that phenotype expression for height was influenced by the genotype of the progenies throughout the study. According to the classification proposed by [45], CVgp values can be considered high when they exceed 10.97% and 13.93% for height and diameter, respectively.

Ref. [46] found CVgp values ranging from 11.3 to 11.9% for Dipteryx alata at 900 days, which can be considered high (>7%, a value commonly used as baseline for comparison) [40]. While [51] evaluated the same species and reported higher CVgp values, between 17.1 and 31.1%., [20] studied progenies of Eugenia dysenterica at 360 days and found CVgp values for height ranging from 14.16 to 21.29% and for stem diameter from 6.20 to 7.55%.

Consistent with the values reported by [45], the results found in the present study for CVgi (%) in the nursery phase for height were considered high, for collar diameter low-to-moderate, and the field phase for height and diameter was classified as very high. Ref. [52] found CVgi (%) values of 27.19 and 15.06% for height and diameter, respectively, in Ilex paraguariensis progenies at 150 days, suggesting that such values may be suitable for a breeding program. Ref. [19] assessed Parkia platycephala and reported CVgi values in the nursery phase for height and stem diameter of 9.01 and 29.09%, respectively, and in the field, values of 4.58 and 31.59%, respectively.

According to [30], accuracy values above 75% are considered optimal. Thus, the values found for height and stem diameter ranged from 25 to 66%, demonstrating good precision in accessing genetic variation from the observed phenotypic variation in the evaluated traits. This indicates that the selection to be performed will provide a greater guarantee of success, reinforcing that the experimental design and the number of repetitions in the experiment were efficient in controlling environmental effects.

According to [53], relative variation coefficient values close to unity (1) indicate a strong relationship between genotypic and environmental effects. The values found in the present study for height (0.21 to 0.45) and collar diameter (0.13 to 0.34) indicate a moderate relationship between the genotypic and environmental effects. Ref. [47] evaluated A. aculeata progenies at 4 to 5 years of age and found values of 0.79 and 0.57 for height and stem diameter, respectively.

Ranking individuals based on their genetic values is highly important for selecting materials destined for seed orchards and use in the seed propagation of elite genotypes. Predicted genetic values relative to all candidate individuals allow for the establishment of the best strategy to increase breeding efficiency [54]. The existence of the differences between the progenies for the traits evaluated over time, high values of variances, heritability, coefficients of variation and accuracy, support results that there was genetic variation among the individuals evaluated, as well as between the progenies, guaranteeing reliability and precision in selection [55]. Evaluating genetic divergence among different progenies through quantitative genetics studies is a fundamental step for programs that use genetic variability for production and conservation purpose. It allows for a more precise direction to be taken in the strategy to be adopted, such as expanding the available genetic base to achieve genetic gains [23,56].

Elements selected among and within dissimilarity groups represent an excellent choice as parents in induced crosses for planned recombination of future cycles of recurrent selection, contributing to increasing the intensity of genetic variability in breeding programs and reducing the risks of a narrow genetic base [53]. Therefore, combinations obtained among the evaluated individuals contribute to a greater concentration of favorable alleles, which can release a wider range of genetic variability and exploit heterosis, also enabling the emergence of desirable transgressive segregants in the descendant progenies.

The meta-analyses of hundreds of datasets have shown that the diversity of plants and animals is decreasing due to habitat degradation, population loss, over harvesting, invasive species, and increased extreme weather events [57]. Many Brazilian native species with potential for timber and non-timber use still need to be studied in terms of genetic gains with selection, conservation, and management. The lack of research makes it difficult to assess the genetic potential of native species in forest breeding programs. Thus, the present study provides initial results, demonstrating important implications regarding the genetic structure of the D. mollis population in Piauí, highlighting the importance of genetic variability in ensuring the long-term survival and conservation of populations.

Given the social and economic importance of the commercialization of fava d’anta fruit in the Cerrado, any progeny selected based on height and diameter growth superiority alone must be confirmed by future monitoring of the progenies’ fruit quality and producing capacity. However, we believe that height and diameter growth are adequate proxy indicators of superior producing trees, given that a larger fruit production can be expected from larger trees, as well as by the fact that larger diameter D. mollis trees have a greater probability of presenting fruit yield compared to smaller ones [58]. Besides fruit production capacity, another important trait to consider when selecting a superior progeny is its ability to support prolonged dry periods that are common in the study area. Thus, we recommend that the findings reported here be confirmed by continuous monitoring of the progeny test.

5. Conclusions

There was considerable genetic variation among the 72 progenies of Dimorphandra mollis Benth. evaluated for height and collar diameter. The tested genetic material, according to the estimated coefficients, has potential for exploration in breeding, management, and genetic conservation programs for the species.

The progenies belonging to groups III, IV, and V stand out as promising individuals with genetic potential for the traits considered in the research. Future combinations between these individuals can generate descendants with high genetic variability and desired superior characteristics.

The progeny test was implemented at the Alvorada do Gurguéia Experimental Farm (FEAG/UFPI) and will serve to obtain relevant information on future aspects regarding the species.