Rice (Oryza sativa L.) is the most widely grown cereal in the world after wheat [1] . Nearly half the world’s population depends on rice as a staple food [2] . Rice has become a strategic crop for food security and political stability for a majority of African countries, particularly in West Africa [3] [4] . Despite efforts and political commitment, local productions are still unable to meet the food needs of local populations on the African continent [3] , as demand for rice has more than tripled from 1.9 to 5.8 million tons over the past two decades in sub-Saharan African countries [5] . In Burkina Faso, farming is essentially rain-fed, subsistence agriculture based on cereals, which account for more than 88% of the annual area under cultivation. Cereals constitute the staple diet of the population and, among them, rice ranks 4^th after sorghum, millet and maize in terms of cultivated area and food production. Rice production has risen thanks to increased acreage, following the lowlands and irrigated lands management, the promotion of rainfed rice varieties and subsidies for agricultural inputs and equipment. However, this increase in national rice production has been achieved at the same time as an increase in the import bill due to population growth and changing eating habits, mainly in urban areas. Annual per capita consumption of rice continues to rise, forcing the government of Burkina Faso to spend around 40 billion F CFA year⁻¹ [6] to cover the population’s rice needs. To reduce this import bill and improve food security, the second phase of Burkina’s Rice Development Strategy [7] has been drawn up to increase rice production to 3 million tons of grains rice by 2030 [8] to meet the country’s rice needs and increase stakeholders’ incomes through the competitiveness and sustainability of the national production. Agricultural industrialization is still in its infancy in Burkina Faso. However, national policy has launched an agricultural offensive aimed at achieving long-term food sovereignty by boosting the mechanization of farming operations, land and water conservation and agricultural product processing. Three agricultural growth poles (Bagré, Samendeni and Sourou) equipped with production, processing and marketing support infrastructures have been set up to make proven technologies available, with the prospect of promoting eight agricultural sectors, including rice ranked as the first strategic crop. Measures to support agricultural production have improved the performance of rice-growing in Burkina Faso, mainly by increasing the average yield per hectare from 2.7 tons to four tons of paddy grain ( http://www.fao.org/mafap/accueil-du-spaaa/fr/ ). However, this improvement in rice productivity in farmers’ fields is not enough to cover the country’s rice requirements. Indeed, national production covers less than 50% of consumption needs and forecast rice requirements in 2025 were estimated at 1.5 million tons, compared with production of 324,611 tons year⁻¹ in the period from 2010 to 2019 [8] . Food security in low-income countries such as Burkina Faso is severely compromised by low variety productivity and the vulnerability of agricultural production systems to the adverse effects of climate change [8] . Some initiatives should be considered to increase rice production by promoting efficient breeding techniques to ensure higher yields, resistance to pests and tolerance to abiotic stresses [9] . Following recent food crises, several West African countries have adopted strategies to increase rice production. These strategies included the large-scale use of improved crop varieties, better technical assistance for rice farmers and increasing the area under rice cultivation [10] .

Conventional breeding based on crossing genotypes of interest to select hybrids with the desired traits in the progeny is the most widely used method in national breeding programs. The crop variety development therefore involves scientists and the end-users (producers and the processors). The initial development is undertaken at research station for hybrid generation. The field evaluation of progenies based on participatory approach with stakeholders (producers and processors) select the best genotypes and the maker assisted selection (MAS) can be used by scientists for genotyping—final identification of the best materials. The end products (hybrids or composites) are tested by farmers and the best genotypes are registered as varieties in national seed catalog for seed production.

To provide seeds of improved varieties, induced mutagenesis is a breeding tool widely used by plant breeders to generate new crop genotypes that perform better in terms of yield and resilience to bio-aggressors and abiotic constraints [11] . It makes it possible to increase genetic variability and improve quantitative and qualitative crop traits in a much shorter time than conventional breeding [12] . The aim of the present study is to create genetic variability in two upland rice varieties with the prospect of selecting genotypes with desired agronomic performances.

Mutagenesis induced by gamma irradiation of seeds influenced the yield components of upland rice mutant lines at M4 generation. Significant variations were observed between mutant lines generated from the same rice variety for numbers of tillers and productive tillers, plant height, panicle length, paddy grain yield and days to 50% flowering. These significant differences result in genetic variability between mutant lines of the same parent variety. The days to 50% flowering of FKR45N or FKR47N mutant lines would be influenced by the effect of the coolness occurred during the plant growth phase coincided with the cool period of the year. This coolness may affect the development of rice seedlings by raising the days to 50% flowering.

The coefficient of variation was used to determine the variation between mutant lines generated from the same parent variety for each yield component. The value of this variation could be a reference for the selection of high-performance agronomic genotypes for more effective and efficient crop improvement [17] . The coefficient of variation can be used to classify variables. For example, if the coefficient of variation for the total number of tillers and the number of productive tillers is greater than 15%, this would explain a high level of variability between mutant lines for both traits. On the other hand, when the coefficient of variation is < 15%, its reveals little variation between mutant lines for the variables [18] .

Principal component analysis revealed both positive and negative test values. Positive test values for the variables show that the means of variables in the same cluster are higher than the value of the overall mean. Negative test-values, suggest that the cluster averages for these variables are lower than the overall mean. Clusters 1 and 2 were characterized by low paddy grain yields, high plant height and negative test-values for the variables. On the other hand, the cluster 3, made up of FKR47N mutant lines, was the best-performing cluster, characterized by high paddy grain yields, taller plant and good tillering, with the number of productive tillers corresponding to 78.9% of the total number of tillers. Test-values for these variables were positive. These mutant lines could be selected for their high productivity. FKR47N mutant lines in cluster 2 show similar data of variables to those of cluster 3, with positive test-values for plant height and paddy grain yield, but negative test values for tiller number. Cluster 1 of FKR47N mutant lines was discriminated from cluster 3 by negative test values for total number of tillers, number of productive tillers, plant height and paddy grain yield. Cluster 1 mutant lines were characterized by their dwarfism and low paddy grain yield. However, these lines could be used for hybridization with other lines having agronomic traits of interest. According to [9] , the combination of hybridization and gamma-ray mutation has been used to generate new aromatic rice varieties coupled with high paddy grain yield. For example, some FKR47N and FKR45N mutant lines are agronomically efficient, with higher yield components [17] - [19] . Gamma irradiation created high-yielding mutant varieties in several plant species [20] [21] . From our results, the mutagenesis therefore has induced earlier flowering and higher paddy grain yield in 10 and nine mutant lines derived from the FKR45N and FKR47N parents, respectively.

The mutagenesis induced genetic variability in yield components of M2 and M3 lines generated from two cultivars [22] Indeed, the ionizing effects of gamma rays on the growth, morphology and yield of plant species including rice were highlighted [20] [21] whereas the gamma irradiation positively or negatively influenced some yield components of sorghum mutant lines derived from the Sariaso14 and Grinkan varieties [17] . The plant height of the M437P1S/300 mutant line was reduced by 20% compared to that of the FKR45N parent. On the other hand, the induced mutation may lead to an increase in wheat plant height [23] . Mutagenesis has induced either dwarf or tall plants, depending on the plant species or the type of mutagen used.

In addition to varietal productivity, the yield of irrigated rice varieties is higher than that of upland rice due to water control in irrigation conditions. This work has demonstrated that induced mutation using gamma ray has enhanced paddy grain yields in mutant lines M450P2S/350 (4 t ha⁻¹) and M422P3/300 (4.7 t ha⁻¹) respectively generated from varieties FKR45N and FKR47N, which are as productive as or more productive than some improved irrigated rice varieties. Indeed, the yields of these two upland rice mutants are similar to those of 16 irrigated varieties (4.5 t ha⁻¹) and higher than those of three others (3.3 t ha⁻¹), of which agronomic performances have been reported [4] . The yields of these two mutants are higher than those of popularized upland rice varieties. Indeed, an assessment of the contribution of upland rice to increased rice production in Burkina Faso showed that the yield of upland rice varieties varies between 1 t ha⁻¹ and 3 t ha⁻¹ for the majority of rice farmers, and the FKR45N variety is more productive than the FKR47N variety [24] . The superior yield of the M422P3/300 line of the FKR47N variety compared to that of the M450P2S/350 mutant line of the FKR45N variety confirms the effect of induced mutation on improving the productivity of a low-yielding genotype.

The results have shown that mutagenesis induced genetic variability could be a potential source for plant breeding, leading to early flowering and/or high yielding mutant lines. The mutant lines M450P1/350 and M451P1S/400 mutant lines from FKR45N variety and M430P1S/300 from FKR47N variety have shown therefore early flowering. These three mutant lines may be tolerant to post-flowering water stress due to occurrences of drought caused by early cessation of rainfall. The paddy grain yields of lines M450P2S/350 from FKR45N variety and M422P3/300 from FKR47N variety were improved by 120% and 20%, respectively. In addition, M422P3/300 line has high tillering and could, therefore, be recommended to livestock breeders for fodder.

To highlight the environmental stability and adaptability of the four mutant lines, a multi-location evaluation would be carried out in contrasting environments to confirm the agronomic performance observed. This would make it possible to study the genotype x environment interaction. Among rice production constraints, drought and Striga hermonthica infection are, respectively, the abiotic and abiotic constraints that cause major yield losses for rice growers. An evaluation of these mutant lines for their tolerance to water deficit and their resistance to S. hermonthica would, therefore, be considered for sustainable production under the strict rainfed conditions of local agro-ecologies.

The authors are grateful to the International Atomic Energy Agency (IAEA) (BFK5019, CRP-Striga D25005) for its assistance in the irradiation of rice seeds with gamma rays and its contribution to human and laboratory capacity building. They also thank the “Institut de l’Environnement et de Recherches Agricoles (INERA)”, Burkina Faso for facilitating the implementation of the field experiment at the Kamboinsé research Station.