Ethiopian Sesame Trial Finds a Champion Variety That Wins Almost Everywhere

Sesame may be one of the oldest oilseed crops on Earth, cultivated for more than five thousand years in the Near East and Africa, but a new study from western Ethiopia shows that the ancient plant still has plenty of surprises locked in its genes. Researchers at the Bako Agricultural Research Center have spent three growing seasons putting twelve sesame genotypes through their paces across six distinct environments in western Oromia, and the results point to one clear winner: a breeding line called BKC-047 that out-yielded the best commercial check variety by a striking 28 percent while staying remarkably consistent no matter where it was planted.

The stakes are higher than they might appear. Sesame seeds carry the highest oil content of any major oilseed crop, roughly 60 percent, compared with about 20 percent for soybean, 40 percent for rapeseed and 45 percent for sunflower. Ethiopia ranks among the top producers globally and is the fifth largest sesame exporter worldwide, behind only Sudan, Nigeria, India and Tanzania. Yet national productivity sits at a modest 0.74 tonnes per hectare across roughly 270,000 hectares, held back by a shortage of varieties that can cope with the country’s wildly variable rain-fed conditions.

The core problem the researchers tackled is known as genotype by environment interaction, or GEI. When this phenomenon is strong, a variety that thrives in one location or year may flop in another, masking its true genetic potential and complicating the breeder’s task of selecting lines that will perform reliably for farmers. To untangle these effects, the team grew four advanced crossing-derived lines, seven locally collected genotypes and two standard checks in a randomized complete block design with three replications at three locations, Bako, Uke and Cheweka, over the 2022 to 2024 seasons, treating each location-year combination as a separate environment.

Because the trial was unbalanced, with not every location tested in every year, the team fitted a linear mixed model using restricted maximum likelihood, or REML, which handles missing observations gracefully. Genotype was treated as a fixed effect while environment, the genotype-by-environment interaction term and replications nested within environments were treated as random. Adjusted genotype means from this model then fed into two complementary statistical frameworks: the Additive Main effects and Multiplicative Interaction model, known as AMMI, and the Genotype plus Genotype-by-Environment biplot, or GGE biplot, which together have become the gold standard for dissecting multi-environment trial data.

The combined analysis of variance delivered an unambiguous verdict. Genotypes, environments and their interaction were all highly significant for seed yield and most agronomic traits, confirming real genetic variability among the tested lines. But the AMMI analysis revealed something more striking: the genotype-by-environment interaction alone explained 40.96 percent of the total variation in yield, dwarfing the environment’s 30.15 percent share and the genotype’s 18.36 percent. In other words, how a sesame line responded to changing conditions mattered more than either its inherent genetics or the environment itself, a pattern that makes multi-environment testing essential before any variety release.

The first two interaction principal component axes captured the bulk of this interaction structure, with IPCA1 accounting for 45.55 percent and IPCA2 for 29.48 percent of the GEI variation, a combined 75.03 percent. On the AMMI 1 biplot, genotypes G2, G6, G7, G8 and G9 all sat to the right of the grand mean, marking them as high yielders. Crucially, G2, the BKC-047 line, paired the highest mean yield of 823 kilograms per hectare with an IPCA1 score close to zero, the statistical signature of a genotype that barely interacts with its environment and therefore performs steadily everywhere.

The GGE biplot analysis reinforced the picture from a different angle. In the mean-versus-stability view, genotype stability is judged by the length of each genotype’s projection from the average environment axis, with shorter vectors indicating greater stability. BKC-047 again emerged as the strongest combination of high yield and stability, and in the ranking biplot it landed closest to the concentric circles defining the ideal genotype, sitting near the center of the plot where breeders dream of finding their candidates. By contrast, G4 and G11 were flagged as poor combinations of low yield and instability.

Not every promising line was a universal performer. The which-won-where analysis divided the six environments into three mega-environments, with vertex genotypes G3, G6, G8, G11 and G4 each dominating particular sectors. BK-005, designated G6, won in environments E1, E4 and E6 with a yield of 813 kilograms per hectare, only 26 percent above the check, while G8 took E3 and E5 and G3 won E2. But the strong interaction effects revealed in the biplots showed that these winners possessed specific rather than broad adaptation, shining in favored niches while remaining environment-dependent elsewhere. The environment analysis itself proved informative too: Cheweka 2023 and Uke 2024 were both highly discriminating and representative, making them the best testing sites, while closely associated environment pairs suggested the network could be trimmed without losing information.

The trial also tracked bacterial blight, a serious disease caused by Xanthomonas campestris pv. sesami that thrives in high-rainfall areas and can strip plants of leaves, flowers and capsules. Severity was scored on a one-to-nine scale under natural field conditions at flowering and fruiting stages. Here again BKC-047 led the field with the lowest severity score of 1.83, firmly in the resistant range, followed by G9 at 2.00 and G3 at 2.33, while G1 fared worst at 6.67. The combination of top yield and blight tolerance makes BKC-047 an unusually complete package for smallholder farmers who cannot afford fungicides.

The researchers are candid about the study’s limits: only six environments in western Oromia were sampled, the dataset was unbalanced across years, and three seasons cannot capture the full sweep of climatic variability. Even so, the recommendation is clear. BKC-047, with its broad adaptation, top yield and disease tolerance, moves forward into variety verification trials targeting wide cultivation across western Oromia and similar agro-ecologies, while BK-005 will be verified specifically in the environments where it excelled. If verification succeeds, Ethiopian sesame farmers may soon plant a variety bred not just to yield well somewhere, but to yield well almost everywhere.

Subject of Research: Genotype by environment interaction and yield stability analysis of sesame genotypes in western Oromia, Ethiopia

Article Title: Genotype by environment interaction and AMMI analysis of sesame (Sesamum indicum L.) genotypes at western Oromia of Ethiopia

Article References: Takele, F., Dabesa, A., & Samuel, D. (2026). Genotype by environment interaction and AMMI analysis of sesame (Sesamum indicum L.) genotypes at western Oromia of Ethiopia. BMC Agriculture, 2(1), Article 30. https://doi.org/10.1186/s44399-026-00056-8

Image Credits: AI Generated

DOI: 10.1186/s44399-026-00056-8

Keywords: sesame, genotype by environment interaction, AMMI analysis, GGE biplot, yield stability, Ethiopia, plant breeding, bacterial blight, multi-environment trials, oilseed crops, agricultural genetics, variety release