The study was conducted from May to August 2025 within the Plant Health and Quality Control Service (SCPQ) of the Directorate for Plant Protection and Conditioning (DPVC), under the Ministry of Agriculture, Animal Resources, and Fisheries (MARAH), Ouagadougou, Burkina Faso.

The three municipalities of Komsilga, Koubri, and Saaba are located in Kadiogo province (Central region), 20 - 30 km from the capital Ouagadougou. They form a peri-urban agricultural corridor combining high population density with significant rainfed crop production potential, especially maize. Komsilga and Saaba have populations exceeding 100,000 and 80,000 inhabitants respectively; Koubri, the most rural of the three (~555 km², 26 villages), hosts 13,067 households whose livelihoods depend primarily on farming and livestock.

Maize seeds were sampled from municipal warehouses using stratified random sampling. For each variety, a seed probe was inserted at the four corners and centre of a minimum of five 100 kg bags to obtain a representative 500 g composite sample. Ten composite samples (50 primary draws in total) were collected, hermetically sealed in labelled plastic bags, and transported to the laboratory. To ensure uniform geographic coverage, one warehouse was selected from each municipality via spatial stratification. Subsequently, bag samples were obtained through simple random sampling, whereby an equal and independent probability of selection was afforded to every unit within the designated facilities. The four certified improved varieties under study were: BARKA, SR21, and KEJ (present in all three municipalities) and FBC6 (Komsilga only). Microbiological sampling was conducted following one month of municipal storage, subsequent to initial agronomic validations.

2.3.1. Culture Medium Preparation and Inoculation

Sabouraud chloramphenicol agar (Liofilchem, Italy) was prepared at 65.5 g/L distilled water, sterilised by autoclaving (121˚C, 15 min), and poured into Petri dishes. Twenty-five grams of each composite sample were homogenised in 225 mL sterile buffered peptone water (BPW) to obtain a 10⁻¹ stock solution. Serial decimal dilutions up to 10⁻⁵ were prepared. One hundred microlitres of each dilution were surface-plated in duplicate on the selective medium and incubated at 37˚C for five days.

2.3.2. Colony Enumeration and Identification

Only plates containing 15 - 300 colonies were retained for enumeration. Results were expressed in colony-forming units per gram (CFU/g) using the standard formula: CFU/g = ΣC/(V × [n₁ + 0.1 × n₂] × d), where C is the total colony count on selected plates, V the volume plated (0.1 mL), n₁ and n₂ the number of plates retained at the first and second dilution respectively, and d the dilution factor of the first selected dilution.

Fungal identification was performed at genus level by macroscopic observation (colony colour, texture, pigmentation, reverse) and microscopic examination between slide and cover slip (mycelia, conidiophores, conidia, sporangia) according to [10].

For each variety, 100 seeds were placed on moistened blotting paper, covered with a second pre-soaked sheet, and rolled into cylinders (four replicates per variety: R1 - R4). Rolls were placed vertically in a germination chamber at 25˚C. The first count (normal seedlings, abnormal seedlings, ungerminated seeds) was performed at day 3 (D3); ungerminated seeds were kept in place until the final count at day 7 (D7). The germination rate (GR) was calculated as the number of normal seedlings relative to total seeds tested, expressed as a percentage.

Data were entered in Microsoft Excel 16 and analysed with Jamovi 2.3. Pearson’s correlation coefficient (r) was used to assess the linear relationship between fungal load and germination rate. The significance threshold was set at p < 0.05.

Table 1. Fungal load (CFU/g) of subsidized maize seeds by municipality and variety.

CFU/g: colony-forming units per gram.

Fungal loads varied from 2.3 × 10² to 1.61 × 10³ CFU/g depending on variety and municipality (Table 1). The highest load was recorded in KEJ from Koubri (1.61 × 10³ CFU/g), followed by BARKA from Koubri (7.7 × 10² CFU/g). The lowest loads were observed in SR21 from Saaba and KEJ from Komsilga (both 2.3 × 10² CFU/g). Marked inter-site variation was observed for the KEJ variety, whose fungal load rose seven-fold between Komsilga and Koubri, suggesting a strong influence of local storage conditions.

Microbiological analyses, conducted in triplicate (n = 3) to ensure precision, revealed four pathogenic fungal genera across all samples (Table 2); the reliability of these estimates is supported by low dispersion measures, with standard deviations for fungal counts and germination rates remaining within a narrow margin of the mean. Aspergillus spp. was the most prevalent genus, isolated from every sample, while Penicillium spp. appeared in seven out of ten; Rhizopus spp. was restricted to BARKA (Koubri and Saaba), and Fusarium spp. was found exclusively in SR21 from Saaba.

Table 2.Fungal pathogens identified by municipality and variety.

Aspergillus spp.: colonies granular, yellowish-green or black, fast-growing; smooth, long conidiophores with a globular, biseriately-arranged terminal vesicle; globular conidia.

Fusarium spp.: colonies woolly, whitish to pink; septate mycelium; abundant sickle-shaped macroconidia.

Penicillium spp.: colonies powdery green, rapidly spreading; chains of globular conidia borne on branched, non-septate conidiophores.

Rhizopus spp.: colonies cottony, brown-black at maturity; siphoned hyphae; globular sporangia.

Overall germination rates were satisfactory, ranging from 81% to 96% (Table 3). SR21 achieved the highest germination in two municipalities (96% in Komsilga; 95% in Saaba). Conversely, KEJ from Koubri recorded the lowest germination (81%), coinciding with the study’s highest fungal load (1.61 × 10³ CFU/g). BARKA showed consistent germination of 88% - 90% across all three sites, whereas FBC6 (Komsilga only) reached 91%.

Table 3.Germination rates (%) of subsidized maize varieties by municipality.

Statistical analysis revealed a very strong, highly significant negative linear correlation between fungal load and germination rate (r = −0.949; p = 0.001) (Table 4). While the correlation is highly significant, these results should be interpreted with caution given the limited number of lots sampled. A Pearson coefficient this close to −1 indicates that virtually all variance in germination rate is explained by variation in fungal load. The p-value (0.001) is well below the 0.05 threshold, confirming that the relationship is not due to chance.

Table 4.Results of Pearson’s correlation analysis between fungal load and germination rate.

***p < 0.001; r: Pearson’s correlation coefficient.

The four genera identified (Aspergillus, Penicillium, Fusarium, and Rhizopus) are consistent with those consistently reported in the literature as the dominant fungal contaminants of stored maize in sub-Saharan Africa. In a 2025 historical review of Nigerian maize contamination data, [4] found that Aspergillus spp. accounted for 37.3% of all isolates, followed by Fusarium spp. (23.1%), Penicillium spp. (13.8%), and Rhizopus spp. (5.4%) a ranking that mirrors our findings.

At the national level, our results align with those of [7], who reported a predominance of Aspergillus section Flavi (75%) in Burkinabè maize seeds, and with a 2024 nationwide study that identified Fusarium verticillioides, Penicillium spp., and Aspergillus spp. as the dominant taxa across 68 samples from major maize-growing areas of Burkina Faso [8]. Comparable assemblages were reported by [11] in Côte d’Ivoire, [12] in Ethiopia, and [13] in Pakistan, underlining the cosmopolitan character of these storage pathogens.

The presence of Fusarium spp. is particularly concerning: F.verticillioides, the principal fumonisin producer, causes maize ear rot responsible for yield losses of 13% - 70% in sub-Saharan Africa [14]. Aspergillus flavus is the chief producer of aflatoxin B1—a Group 1 human carcinogen (IARC) and is also known to directly impair seedling coleoptile elongation [15]. The reported mean aflatoxin concentration of 517 µg/kg in Burkinabè maize [9] over 100 times the EU regulatory limit of 4 µg/kg for direct human consumption underscores the severity of the public health dimension beyond agronomic concerns.

The fungal loads recorded in this study (2.3 × 10² to 1.61 × 10³ CFU/g) are within the ranges reported for cereal seeds stored under sub-optimal conditions in West Africa. The pronounced inter-site variability observed for KEJ a seven-fold difference between Komsilga and Koubri suggests that local storage conditions (warehouse relative humidity, ventilation, ambient temperature, and prior storage duration) exert a stronger influence on contamination levels than variety genetics.

Storage fungi, particularly Aspergillus and Penicillium, develop preferentially at relative humidity above 70% and temperatures of 25˚C - 40˚C [16]. In the Sahelian climate of the Kadiogo region where ambient temperatures regularly exceed 35˚C during the dry season non-air-conditioned municipal warehouses routinely provide such conditions. Grain moisture content at intake is another critical parameter: if seeds are stored above the recommended 12% - 13% moisture threshold without prior drying, fungal proliferation is virtually inevitable. The absence of systematic humidity monitoring in the sampled warehouses represents a major structural gap in the current seed subsidy supply chain.

The highly significant negative correlation between fungal load and germination rate (r = −0.949; p = 0.001) is the central finding of this study. It demonstrates that fungal contamination is a major limiting factor for seed viability under the storage conditions observed. The KEJ/Koubri sample provides the clearest illustration: the highest fungal load (1.61 × 10³ CFU/g) coincided with the lowest germination rate (81%), the only sample falling below the 85% minimum germination threshold commonly adopted by West African seed certification standards.

Our results are consistent with those of [14], who found that maize seeds infected by Aspergillus niger, Penicillium spp., and Fusarium oxysporum showed germination reductions of 10% - 30% compared to healthy seeds. They also corroborate the findings of [13], who documented an inverse relationship between storage fungal loads and germination capacity in Ethiopian maize, and of [17], who reported significant seedling emergence failures in contaminated Burkinabè maize lots. At the mechanistic level, fungal pathogens impair germination primarily by 1) producing enzymes (cellulases, amylases, lipases) that degrade seed storage compounds and embryo membranes, 2) secreting mycotoxins that disrupt metabolic processes within the embryo, and 3) competing for oxygen and nutrients within the seed coat [6][16].

It should be noted that, although most germination rates observed (81% - 96%) remained above the commonly cited 80% threshold for certified seeds, the strong negative correlation trajectory indicates that any further deterioration of storage conditions or prolonged warehouse tenure could drive multiple varieties below acceptable germination thresholds. Moreover, since mycotoxin content was not quantified in the present study, the risk to end-users (farmers, livestock, and consumers) from sub-lethal but chronic toxin exposure remains an open and urgent research question.

These findings carry direct implications for the design and governance of Burkina Faso’s seed subsidy program. The current absence of systematic pre-distribution microbiological screening represents a critical gap, given that contaminated seeds can negate the investment made by both the state and smallholder farmers. Several West African countries including Ghana and Nigeria have begun integrating fungal and mycotoxin screening into their national seed certification protocols with measurable gains in germination success rates.

Fungal contamination of subsidized seeds also has broader food security implications. While these consistent laboratory results confirm a high fungal presence, it is important to note that mycotoxins were not measured in these samples, so the public-health implications remain indirect, serving as a proxy for risk rather than a confirmed chemical assessment. As highlighted by [9], the estimated probable daily intake of aflatoxins from maize in Burkina Faso (29 - 432 ng/kg body weight/day) generates a projected liver cancer burden of up to 28 cases per 100,000 persons per year. Poor seed germination that reduces crop yields simultaneously reduces household food availability and income compounding the mycotoxin-related health burden. In a context where agricultural subsidies are intended to strengthen food sovereignty, ensuring the microbiological integrity of subsidized inputs is therefore both an agronomic and a public health imperative.