Genetic Diversity of Local Rice Accessions Cultivated in Guinea Based on Agro-Morphological Traits and Identification of Sources of Tolerance to Iron Toxicity ()
1. Introduction
Rice is one of the most consumed cereals in the world [1] with a production of over 479.2 million tonnes [2]. Most of this production is in Asia.
Despite a large potential arable land area for rice cultivation of more than 200 million hectares, Africa produces only 29 million tonnes or 4% of world production [3] [4]. The main causes of low rice production are high production costs and low yields due to biotic factors including diseases (blast, helminthosporium, yellow mottle etc.) and weeds, and abiotic factors such as iron toxicity. In Guinea, rice is also the staple food of more than 12 million people. It is grown to nearly 520,000 ha [5]. However, like most irrigated and rainfed lowlands in West Africa, iron toxicity is a major constraint to rice production in lowlands bordered by iron-rich plateaus [6]. It reduces growth, tillering, and soil fertility [7] and consequently grain yields. Yield losses in rice due to iron toxicity are estimated to be between 12% and 100% in West Africa [6] [8] [9] [10]. Indeed, iron is widespread in West African lateritic soils in excessive amounts and thus causes nutritional disorders in rice. Critical levels of iron toxicity are above 500 mg Fe per kg−1 of dry weight leaves. However, the level of susceptibility to iron toxicity varies considerably among varieties [11] [12]. While the varieties of O. sativa showed growth problems at low external concentrations (less than 10 ppm Fe2+), cultivars of Oryza glaberrima subjected to higher doses (500 ppm Fe2+) for 3 days were not affected [13]. In Guinea, iron toxicity is much more prevalent in the lowlands of Lower Guinea and Forest Guinea. Despite the numerous works carried out by the Guinean Agricultural Research Institute in collaboration with Africarice to subdue the effects of this abiotic constraint through the creation of new varieties (ARICA6, ARICA18, NERICAL19, CK73) created by AfricaRice, IRRI and Kilissi that are iron-tolerant or iron-resistant, and through the use of alternative methods such as silica [14] or fertiliser, iron toxicity still hinders the development of rice production in Guinea. Indeed, farmers still face enormous difficulties related to the problem of large-scale dissemination of improved varieties, their adaptability, the availability of silica, the high price of fertiliser, and the lack of development of plains [1]. Therefore, a better response to farmers’ expectations is the exploitation of the genetic resources of their local cultivars better adapted to their environments to search for varieties resistant or tolerant to iron toxicity. Thus, the intensification and increase of rice production necessarily requires a better knowledge of the agro-morphological diversity of traditional varieties grown in rural areas. In particular, the level and structure of agromorphological diversity of local accessions must be assessed and genotypes that are potentially tolerant or resistant to iron toxicity must be identified.
2. Materials and Methods
2.1. Material
2.1.1. Study Sites
The experiments were carried out on two sites, Kilissi and Sérédou, indeed, most of the lowlands in these two localities are heavily affected by iron toxicity.
1) Kilissi site
The Kilissi Agricultural Research Station is located in the prefecture of Kindia, 135 km east of the capital Conakry, at an altitude of between 95 and 110 m and with geographical coordinates of 9˚59 north latitude and 12˚49 west longitude. The Prefecture of Kindia has a humid tropical climate characterised by two alternating seasons of equal duration. The dominant winds are the Monsoon and the Harmattan. The rainiest months in 2018 were June, July, and August. The maximum rainfall was recorded in August. The cumulative annual rainfall was 1656.7 mm for the last 10 months of the year. The field experiment was carried out under lowland conditions on an iron toxicity site with ultisol whose physical and chemical characteristics determined following the analysis of a composite sample are recorded in Table 1.
![]()
Table 1. Physical and chemical characteristics of Kilissi soil.
Source: Foulaya soil laboratory (LAB Sols F.), 2016.
2) Sérédou site
The Regional Agricultural Research Centre of Sérédou (CRRA-Sérédou) is located in Sérédou, one of the sub-prefectures of Macenta. It is located in the south of the prefecture, 38 km from the urban centre, at an altitude of 620 m and with geographical coordinates of 8˚32'37'' north latitude and 9˚28'22'' west longitude. The growing season begins in April and ends in December with a minimum temperature of 20.8˚C, an average of 25.8˚C and a maximum of 30.8˚C. A rainfall of 1992.4 mm for 179 rainy days was recorded during the trial. The average daily relative humidity and insolation were 77% and 6 hours, respectively. The field experiment was conducted in a lowland with hydromorphic clayey silt soil. The physico-chemical characteristics of the soil are recorded in Table 2.
![]()
Table 2. Physico-chemical characteristics of the soil in Sérédou.
Source: Foulaya soil laboratory (LAB Sols F.), 2016.
2.1.2. Plant Material
Ninety local accessions collected from farmers in the prefectures of Kindia and Macenta and six imported improved varieties (AZUCENA, BOUAKE 189, CK 73, IR64, NERICAL 19 and AURYLUX 6) were used as plant material. The imported varieties were selected for their behaviour towards iron toxicity. These
improved varieties have already been tested in lowland areas with iron toxicity in the West African subregion, notably in Guinea, Benin, Burkina Faso, Nigeria, Ghana, and Côte d’Ivoire [15].
2.2. Methodology
2.2.1. Experimental Set-Up
Two trials were set up in July 2018 in Kilissi and Sérédou. At each site, the trial was conducted in an alpha lattice design with three replications. Each replication consists of eight blocks containing 12 elementary plots each. The area of an elementary plot was 5 m2 and the distances between replicates, blocks, and plot were, respectively, 1 m2, 0.5 cm, and 0.5 cm with a defence zone of 2 m2 on both sides of the trial. The total area of each trial was 2826.5 m2 with a length of 59.5 m and a width of 47.5 m.
2.2.2. Conduct of the Trials
The trials were conducted under lowland conditions. A nursery was set up on 24/07/2019 in Kilissi and on 25/07/2019 in Macenta (Figure 1). The preparation of the land following the installation of the nursery consisted of clearing, collection, and manual clearing followed by planting in both sites.
Twenty-one-day-old seedlings were transplanted to 5 m2 plots with two plants per plot with a spacing of 20 cm × 20 cm between the plants and between the rows, i.e. 5 m rows per elementary plot for a density of 625 plants per elementary plot (Figure 2).
![]()
Figure 1. Stages in the establishment of the rice nursery. (a) Rice sample; (b) Nursery seedlings; (c) Seedlings ready for transplanting.
![]()
Figure 2. Transplanting seedlings in the experimental plot. (a) Plot ready to be transplanted; (b) uprooting of plants; (c) Transplanting of rice; and (d) Transplanted plot.
Maintenance consisted of two manual weedings and one chemical weeding. Regarding fertilisation, 150 kg/ha of NPK fertiliser (15-15-15) was applied one day before transplanting, while 100 kg/ha of urea was applied in two sprays, 35 kg/ha at the beginning of tillering after the first weeding and 65 kg/ha at panicle initiation after the second weeding.
2.3. Data Collection
Observations were made on agro-morphological traits and iron toxicity scores on the leaves of rice plants. The quantitative variables were measured on 10 plants selected from the useful plot and the qualitative variables were observed on the whole plot and scores were noted after visual observations.
2.3.1. Quantitative Variables
To study the effect of the different treatments on the growth and development of rice, nine variables were measured. These are:
- the height of plants at maturity (HPM) measured on 10 plants in the working plot from the base of the plant to the tip of the highest panicle;
- the number of tufts per m2 (NT/m2) at the full tillering stage was determined by counting all plants per m2 in each test plot;
- the number of tillers per clump (NT/T) at 60 days after transplanting was determined by visual counting, which is the tillering ability of the variety. It is one of the most important indicators of good paddy grain yield;
- the number of panicles per m2 (NP/m2) at maturity, which gives information on the number of fertile tillers on 1 m2 per elementary plot and therefore affects yield;
- the total number of grains per panicle (NG/P) evaluated on one panicle per plant sampled with 1 m2 per variety;
- Grain yield (Rdt, kg/ha) after 85% - 90% maturity of panicles in the field and at 14% moisture. The yield per hectare is obtained in grams by weighing and extrapolated to kg/ha;
- the weight of 1000 seeds (P1000 Grn) was evaluated with an AIDE brand balance and precision Max 5000gx1G Max 11/b.O.10z by counting and weighing;
- the moisture content (TH%) was measured with a digital grain moisture meter, Precision Moistex SS 7 SATAKE, at a moisture content less than or equal to 14%;
- the sowing heading cycle (SSC), which represents the number of days between sowing and the beginning of the first rice plant to be headed (10%);
- the sowing-maturity cycle (SMC), which corresponds to the number of days between sowing and the maturity of the plant when at least 85% to 90% of the panicles have a straw colour.
2.3.2. Qualitative Variables
Seven variables were assessed from sowing to maturity. They were assessed on all plants. They were:
- Panicle exertion (PE) which expresses the good or bad exit of the panicle from the plant. Thus, emergence was considered good if the distance between the base of the panicle leaf and the beginning of the secondary branching was greater than or equal to 10 cm.
- Plant vigour (PV), which was observed 21 and 42 days after sowing to differentiate robust green plants with active growth and good balance from stunted and stunted plants.
- The description of foliar symptoms of iron toxicity on the plant. For the observation of iron symptoms on rice leaves, the IRRI index scale was used for the assessment of iron toxicity [16] as recorded in Table 3. The generalized linear model [16] was used to test the effect of iron on yield. In addition, the assessment of iron toxicity symptoms (ITSS) was done by naked eye observation of the plants and the use of a 1 - 9 scale based on the International Rice Research Institute (IRRI) standard rice rating system: a score of 0 indicates normal growth and tillering, 9 indicates that almost all plants were dead or dying [17]. Plants were observed three times during the growth period (21DAR, 60DAR, and 90DAR) so that when iron toxicity symptoms were expressed on the foliage to monitor toxicity. Iron toxicity scores were recorded in all plots.
- Rice blast (RP) is the major disease of rice in South Saharan Africa (SSA). It is caused by Pyriculariosis oryzae, a fungus that attacks all organs of the plant: leaves, neck, nodes, panicle, rachis, and stems. Observations were made on the leaves that are most susceptible.
![]()
Table 3. IRRI scale for describing iron toxicity symptoms in the field (field damage to rice plants).
2.4. Data Analysis
The data collected were processed with the Excel spreadsheet and XLSTAT-Pro 7.1 software. The Excel spreadsheet was used to produce histograms of the distribution of the accessions for the qualitative variables, and XLSTAT-Pro 7.1 was used for the analysis of variance (ANOVA) based on a linear model to evaluate the level of variability of the material and to identify the discriminating characteristics and the Newman and Keuls test of separation of means at the threshold α = 5% to compare the performance of the varieties within each site and between sites.
3. Results
3.1. Kilissi Site
3.1.1. Variation of Qualitative Characteristics of Rice Accessions
At the level of the qualitative traits studied, a great variability was observed within the accessions. Thus, the score of iron toxicity symptoms varied from 0 to 9 and that of particularity from 0 to 5. For blast, most cultivars (62.5%) in the collection scored 0 and a minority (1.8%) scored 5. For iron toxicity, 20.48% of the plants scored 0 and a small proportion (5.9%) scored 9. For plant vigour, the score also varied from 0 to 5, with plants scoring 0 representing 1.73% and those scoring 5 representing 34.73% of the collection. Figure 3 shows the distribution diagrams of the accessions for each of the qualitative variables studied.
3.1.2. Variation of Quantitative Traits of Local Rice Accessions
1) Average performance of the local rice accessions evaluated
The characters studied discriminate significantly between accessions at the 1 and 5% thresholds, except for the height of plants at maturity (HPM), the number of panicles per plant (NP/P) and the number of seeds per panicle (NG/P), thus reflecting the existence of significant agro-morphological diversity within the collection (Table 4).
The coefficient of variation values ranged from 6.29% to 33.52% for moisture content and number of tillers per clump, respectively. The lowest values of the coefficient of variation were observed for “moisture content (6.29%) and panicle length (8.96%)”. The smallest significant difference was 0.411 for moisture content and 133.575 for yield.
The number of seeds per panicle varied from 46 for the least seed-producing variety to 155 for the most seed-producing. For yield, the lowest was 1143.66 kg/ha on average for the least productive accessions against 4475.33 kg/ha on average for the most productive accessions for an overall average of 2719.63 kg/ha. The number of days from sowing to flowering varied from 80 to 128 days.
![]()
Figure 3. Proportions of the modalities according to the qualitative traits. Legend: On the IRRI scale for iron toxicity: score 0: resistant plants, score 1: very tolerant, score 3: tolerant, score 5: sensitive, score 7: very susceptible (stunting), and score 9: very susceptible (death of plants).
![]()
Table 4. Results of the analysis of variance of the 11 quantitative variables measured.
Legend: SSC: sowing cycle heading, HPM: plant height at maturity, NT/m2: number of clumps per m2, NT/T: number of tillers per clump, NP/m2: number of panicles per m2, NG/P: number of grains per panicle, LP: panicle length, Rdt (kg/ha): yield, P1000 Grns: 1000 seed weight, TH: moisture content, CSM: sowing maturity cycle, *: significant difference at 5%, **: significant difference at 1%, NS: non-significant difference, F: Fisher’s value, CV: coefficient of variation, LSD: smallest significant difference at 5% level.
2) Influence of the environmental factor on the variations of the quantitative traits
The results of the analysis of variance of the Kilissi site using the Newman and Keuls test of significance of the means presented in Table 5 reveal that the variables sowing epiposition cycle, 1000 grain weight, yield t/ha, and moisture content significantly discriminate the accessions. On the other hand, the variables height of plants at maturity, number of clumps per m2, number of tillers per clump, number of seeds per panicle, and panicle length did not show a significant difference between accessions.
3) Comparison of yields of local accessions and controls for the identification of the best accessions
The yields of the control varieties (Table 6) and the 15 best local cultivars identified (Table 7) at Kilissi show that some of the local accessions evaluated were affected by iron toxicity while others tolerated iron in the lowlands and gave good yields similar to or better than the controls. Yields of the control varieties ranged from 2164 kg/ha for the Bouaké189 variety to 4438 kg/ha for the CK73 variety. As for the local accessions identified, they obtained yields ranging from 3315.33 kg/ha for the Patem accession to 4475.33 kg/ha for the Gbai-bai accession with an overall average of 3898.36 kg/ha.
The comparison of the yields of the local cultivars with the controls is shown in Figure 4. Only three local cultivars have similar yields to the best controls (Orulux 6 and CK 73).
![]()
Table 5. Results of the Newman and Keuls test for significance of means in Kilissi site.
Legend: SSC: sowing cycle heading, HPM: plant height at maturity, NT/m2: number of tufts per m2, NT/T: number of tillers per tuft, NP/m2: number of panicles per m2, NG/P: number of grains per panicle, LP: panicle length, Rdt (kg/ha): yield, P1000 Grain: 1000 grain weight, TH: moisture content, SSC: sowing cycle maturity, *: Significant, **: highly significant, NS: not significant, Rep.: Replication.
![]()
Table 6. Performance of control varieties in terms of yields.
Legend: score 0: R: resistant; score 1: T: tolerant, score 3: S: susceptible.
![]()
Table 7. Performance of local accessions identified in terms of yield.
Legend: Means assigned the same letter in the same column or row are not significantly different at the 5% threshold by the SDPP calculation. Letters in the upper case refer to rows and those in the lower case to columns. S = Significant; NS = Non Significant, HS = Highly Significant, Rep. = Replication.
![]()
Figure 4. Variation in yields of local accessions identified in Kilissi.
3.2. Sérédou Site
3.2.1. Variation of Qualitative Traits of Local Rice Accessions
In Sérédou, a great variability was observed within the accessions for the qualitative traits (Figure 5). Thus, iron toxicity symptoms varied from 0 to 9 with 34.3% of the plants in the collection scoring 0 and a very small proportion (1.04%) scoring 9. For blast disease, the score varied from 0 to 5 with a majority of plants (65.63%) scoring 0 and a minority (1.29%) scoring 5. As for plant vigour, the score also varied from 0 to 5, with a large proportion of plants in the collection (36.45%) scoring 5 and a small proportion (6.25%) scoring 0.
![]()
Figure 5. Proportions of the modalities according to the qualitative characteristics. Legend: Using the IRRI scale for iron toxicity: score 0: resistant plants, score 1: very tolerant, score 3: tolerant, score 5: susceptible.
3.2.2. Variation of Quantitative Traits of Local Rice Accessions
1) Average performance of rice accessions evaluated in Sérédou
The analysis of variance (Table 8) shows that the parameters observed in this site significantly discriminate the accessions studied at the threshold of 1 and 5%. This result reveals the existence of an essential agro-morphological diversity within the collection for the characters studied.
The number of days for the sowing epiposition cycle varied from 81.66 days for short-cycle accessions to 130 days for long-cycle accessions. As for the number of grains per panicle, it varied from 99.33 grains for the least seed-productive accessions to 165.33 grains for the most seed-productive accessions. For grain yield, the lowest was 1555.66 kg/ha for the least productive accessions against 5886 kg/ha for the most productive ones with an average of 3349 kg/ha. The sowing maturity cycle varied from 103.33 days for the short-cycle varieties to 165 days for the long-cycle accessions. These variations reflect a very large difference between accessions. As for the coefficient of variation, the extreme values are 18.79% and 35.97% for the parameters observed. The smallest significant difference varies between 0.36 and 6.25.
2) Influence of the environmental factor on the variations of the quantitative traits
The results of the analysis of variance of the Sérédou site with the Newman and Keuls test of significance of means presented in Table 9 reveal that the variables significantly discriminate the accessions. The coefficient of variation varied from 19.79 for moisture content to 35.97% for grain yield.
3) Comparison of yields of local accessions and control and identification of the best accessions
The performance of the control varieties and the 15 best local accessions identified in the site are recorded in Table 10 and Table 11, respectively. The control varieties obtained yields ranging from 1554.67 kg/ha for variety CK73 to 3587.67 for the variety ORYLUX 6. As for the 15 local accessions identified, the yield obtained varied from 2836.33 kg/ha (Yamalopou accession) to 5886.3 kg/ha (Kagbowoletaie accession) with an overall average of 4535.7 kg/ha.
Toxicity scores ranged from 0 for resistant varieties to 5 for susceptible varieties. Figure 6 shows the yield variations of the local accessions and control varieties. These local cultivars had similar or much higher grain yields than the controls used. Except for the four local cultivars, all other 11 accessions showed a higher yield than the best control Orulux 6.
![]()
Table 8. Results of the analysis of variance of the 11 quantitative variables measured.
Legend: SSC: sowing cycle heading. HPM: plant height at maturity. NT/m2: number of clumps per m2. NT/T: number of tillers per clump. NP/m2: number of panicles per m2. NG/P: number of seeds per panicle. LP: panicle length. Rdt (kg/ha): yield. P1000 Grain: 1000 seed weight. TH: moisture content. CSM: sowing maturity cycle. %. CV: coefficient of variation. ***: very highly significant (pr < 0.0001). LSD: smallest significant difference at 5% level.
![]()
Table 9. Results of the Newman and Keuls test for significance of means in Kilissi site.
Legend: SSC: sowing cycle heading HPM: plant height at maturity. NT/m2: number of tufts per m2. NT/T: number of tillers per tuft. NP/m2: number of panicles per m2. NG/P: number of grains per panicle. LP: panicle length. Rdt (kg/ha): yield. P1000 Grn: weight of 1000 grains. TH: moisture content. CSM: sowing maturity cycle. CV: coefficient of variation. **: highly significant (Pr ≤ 0.001).
![]()
Table 10. Performance of control varieties in terms of yield.
Legend: Score 1: T: tolerant; Score 5: S: susceptible, Sign.: Significance, IT: Iron toxicity score.
![]()
Table 11. Performance of local accessions identified in terms of yield.
Legend: Means with the same letter in the same column or row are not significantly different at the 5% threshold by the SDPP calculation. Letters in the upper case refer to rows and letters in the lower case refer to columns. S = Significant; NS = Non Significant; HS = Highly Significant, Rep. = Replication.
3.2.3. Selection of the Best Performing Accessions in Both Sites
Of the 30 best local cultivars identified in the Kilissi and Sérédou sites, the 12 best performing and stable accessions in terms of yield, tolerance, or resistance to iron toxicity were selected. Indeed, these accessions listed in Table 12 have
![]()
Figure 6. Yield variation of local rice accessions in the Sérédou site.
![]()
Table 12. Local accessions selected in both sites.
average yields in both sites ranging from 3169.66 kg/ha (Coyady accession) to 4662.33 kg/ha (Komassataie accession). Three cultivars were found to be resistant to iron toxicity (score 0) and nine are tolerant to abiotic stress with scores between 1 and 1.5. Indeed, these accessions were the most stable in terms of yield in both sites.
4. Discussion
The minimum and maximum values showed important deviations. The high coefficients of variation for most of the characters and the presence of several modalities for some qualitative characters reflect the existence of a great morphological variability within the studied accessions. The discrimination of these accessions especially by the variables such sowing-maturity cycle, number of grains per panicle, number of clumps per m2, 1000 grain weight, and yield shows that these could be used as important agronomic traits to effectively differentiate between the local accessions evaluated. The results obtained are similar to those reported by [10] [18]. In the collection, regarding iron toxicity symptoms, the plants reacted differently and showed typical symptoms of iron toxicity for both sensitive and tolerant accessions. No symptoms were reported on the resistant ones. These symptoms appeared on the leaf blades in the form of brown spots characteristic of bronzing, which spread to the plots containing the sensitive and tolerant varieties. In these plots affected by iron toxicity, a decrease in leaf biomass and leaf water content was also observed. According to [19], leaf and root dry weight are the most sensitive parameters to iron toxicity in the early stages of plant development. In our study, the amount of biomass has decreased the dry weight of the leaves will necessarily decrease and this will lead to a strong reduction in the growth of the plants in the plots. In the presence of excess iron, the iron concentration in the rice plant increases considerably. As observed earlier, iron is distributed differently among the organs accumulation mainly in the roots. Although the iron concentration in the roots is the highest, the iron content in the sheaths and blades of treated plants remains relatively high. It is accepted that the critical threshold for iron toxicity in Asian rice species is 700 mg·kg−1 D.M. [20]. The iron concentration measured in the plots where our plants were grown reaches up to 2300 mg·kg−1 M.S, indicating that this threshold is largely exceeded and that the stress conditions have been met. Among the local varieties evaluated, some are susceptible and others tolerant or resistant and showed more pronounced toxicity symptoms for the morphological parameters studied. The different responses of the varieties can be explained either by the implementation of different resistance mechanisms by the plants or by different levels of resistance for the same resistance mechanism [21]. In both sites (Kilissi and Sérédou), 30 accessions were identified including 15 accessions per site based on their iron resistance, yields, and the choice made by farmers according to certain criteria such as size, grain colour, senescence and tillering. Of these 30 accessions, 12 were selected for their stability in terms of yield and low sensitivity to iron toxicity. These selected accessions had higher yields than the control varieties. Rice plots with many more symptoms often have low yields, whereas plots with fewer symptoms or no symptoms at all have higher yields. Yields are variable depending on the degree of toxicity of the varieties tested. All varieties had significantly lower yields under iron toxicity stress. The iron toxicity constraint induced a high yield loss ranging from 15% to 25% of the potential yield. [18] found a yield loss of 43%, [22] reported a loss of 59.5%. According to [23], the yield loss of rice varies from 12% to 100% in West Africa under iron toxicity conditions. Previous studies realized by [24] on the genetic diversity of 204 accessions of O. glabberima cultivated under iron toxicity also reported different tolerance levels to iron toxicity stress.
5. Conclusion
The agro-morphological characterisation has contributed to a better knowledge of the genetic resources. The traditional varieties from the two study areas (Kindia and Macenta) represent a fairly large genetic potential. From the results obtained, it appears that there is a great variability within the accessions and a high potential of cultivars tolerant or resistant to iron toxicity. Thus, 30 local accessions with similar or higher grain yields than the controls were identified in the two sites, of which 12 proved stable in the two environments (Kilissi and Sérédou). In view of the problems posed by iron toxicity in lowland rice, the 12 local accessions selected could serve as a basic material for rice breeding for resistance or tolerance to this abiotic stress. In addition, genetic characterisation of these cultivars using molecular markers and identification of QTLs associated with iron resistance could complement and strengthen the results of this study.
Data Availability
The data used to support the findings of this work are available upon request from the corresponding author.
Authors’ Contributions
This work was carried out in collaboration between both authors. MLB & MS wrote the protocol, wrote the first version of the manuscript, MLB & NS, performed the statistical analysis., NS & MHO reviewed the experimental protocol and corrected all versions of the manuscript, SC ensured monitoring work at the Guinea Agronomic Research Research Institute and correcting the first versions of the manuscript. NS, MHO, KRN, PBK & MS provided guidance, protocol validation and final manuscript version.
Acknowledgements
We thank all members of the Genetics and Plant Breeding Team (EGAP) of the Biosciences Laboratory of the Joseph KI-ZERBO University for the corrections of this manuscript and the IRAG for the institutional support and accompaniment during the experiments. We also thank the WAAPP (PPAAO project) for the financial support of this work.