Hidden Genetic Diversity in Ethiopia’s Ancient Grain Tef Could Unlock Bigger Harvests

In the highlands of North Wollo, Ethiopia, a humble grass that has fed the region for millennia is revealing secrets that could reshape its future. Tef, the tiny-seeded cereal behind the beloved flatbread injera, has long been prized by farmers and consumers alike, yet its national average yield remains stubbornly low compared with other cereals. Now, a detailed field study conducted during the 2022 growing season at the Sirinka Agriculture Research Center has mapped the genetic variability hidden within sixty-four tef genotypes, offering breeders a practical roadmap for developing higher-yielding, more resilient varieties of this ancient crop.

The research, published in Innovations in Agriculture, evaluated sixty-four genotypes, most of them recombinant inbred lines derived from a cross between the breeding line DZ-Cr-387 and the cultivar Rosea, alongside a standard check variety called Mena and a local check from the Sirinka center itself. The trial was laid out in a simple lattice design with two replications, each plot measuring just one square meter, reflecting the small stature and dense sowing habits of the crop. Planting took place in the first week of July 2022, timed to the main cropping season, with recommended fertilization of 121 kilograms of NPS and 39 kilograms of urea per hectare and a seeding rate of 10 kilograms per hectare.

Across fifteen phonological and morph-agronomic traits, the analysis of variance uncovered pronounced differences among the genotypes, significant at both the 1 and 5 percent probability levels for most characters. Traits ranging from days to heading and physiological maturity to plant height, panicle length, culm length, peduncle length, number of spikelets per panicle, above-ground biomass, grain yield, harvest index, and lodging index all showed statistically meaningful separation. Only the number of fertile tillers per plant and main shoot panicle weight failed to distinguish the genotypes significantly, suggesting those characters are less useful for selection in this population.

The descriptive statistics revealed wide ranges for nearly every trait measured. Days to heading spanned 37 to 47 days, while physiological maturity ranged from 87 to 96.5 days. Plant height varied from 103.2 to 125.5 centimeters, panicle length from 40.9 to 53 centimeters, and culm length from 57.38 to 74.7 centimeters. The number of spikelets per panicle stretched from 387.5 to 735.4, and above-ground biomass ranged from 13,000 to 32,000 kilograms per hectare. Perhaps most striking for breeders, grain yield itself ranged from 2,095 to 4,205 kilograms per hectare, with a mean of 3,271 kilograms per hectare. Twenty-five of the sixty-four genotypes outperformed the standard check Mena, which yielded 3,380 kilograms per hectare, in terms of grain yield, a clear signal that the population harbors real potential for improvement.

To quantify how much of this variation is genetic rather than environmental, the researchers computed phenotypic and genotypic coefficients of variation, broad-sense heritability, and expected genetic advance for each trait. Lodging index showed the largest phenotypic coefficient of variation at 30.1 percent, while grain yield, biomass, and harvest index displayed moderate phenotypic variation in the 10 to 20 percent range. On the genotypic side, moderate coefficients of variation appeared for spikelets per panicle at 10.3 percent, grain yield at 12.3 percent, harvest index at 14.1 percent, and lodging index at 17.9 percent. Days to physiological maturity, by contrast, showed the lowest values, with a genotypic coefficient of variation of just 1.1 percent, indicating that maturity timing is tightly constrained and offers little room for selection.

Heritability estimates told a similarly nuanced story. Panicle length exhibited the highest broad-sense heritability at 74 percent, followed closely by culm length at 68.9 percent, meaning these structural traits are largely controlled by genetic factors and will respond predictably to selection. Moderate heritability values between 30 and 60 percent were recorded for plant height, days to heading, number of spikelets per panicle, above-ground biomass, grain yield, and harvest index, while traits such as grain filling period, days to maturity, tiller numbers, and main shoot panicle weight fell into the low category, their expression swamped by environmental noise. When heritability was paired with genetic advance as a percentage of the mean, harvest index at 22 percent and lodging index at 21.8 percent stood out as traits where selection could deliver substantial gains, whereas grain yield at 17.8 percent and spikelets per panicle at 16.4 percent showed moderate but still promising responses.

Correlation analysis at both genotypic and phenotypic levels then illuminated which traits move together and which pull in opposite directions. Grain yield showed positive and highly significant correlations with the number of spikelets per panicle, plant height, and harvest index at both levels, and positive significant genotypic correlations with panicle length as well. Above-ground biomass also correlated positively and significantly with grain yield phenotypically. These relationships suggest that breeders selecting for taller plants with longer panicles, more spikelets per panicle, and greater biomass should simultaneously capture yield gains. Conversely, the lodging index displayed a negative and significant correlation with grain yield, reinforcing the long-standing observation that plants that fall over sacrifice both quantity and quality of harvest, a persistent bottleneck in tef production.

Other trait associations added further texture to the picture. Days to heading correlated positively and highly significantly with days to physiological maturity, plant height, panicle length, culm length, above-ground biomass, and grain yield, but negatively with harvest index and lodging index. Plant height showed strong positive genotypic correlations with culm length at 0.82 and panicle length at 0.72, while the number of total tillers per plant correlated at 0.92 with the number of fertile tillers per plant, indicating these paired traits are genetically intertwined. Peduncle length, interestingly, correlated negatively with the number of spikelets per panicle, hinting at a trade-off between stem architecture and reproductive capacity. Harvest index, a measure of how efficiently a plant converts total biomass into grain, was negatively correlated with days to heading and panicle length, suggesting earlier-maturing, more compact plants partition resources more efficiently into seed.

The broader context makes these findings especially consequential. Tef is an allotetraploid cereal with a chromosome number of 2n equals 4x equals 40, and Ethiopia is both its center of origin and its center of diversity, a living legacy of ancient farmer selection. The crop grows across 11 of Ethiopia’s 18 major agroecological zones and exhibits remarkable phenotypic plasticity, tolerating both moisture stress and waterlogging while commanding premium prices for its grain and straw. Yet the national average yield of roughly 1.85 tonnes per hectare pales against the 1.7 to 2.2 tonnes achieved by improved varieties on farmers’ fields and 2.2 to 2.8 tonnes on research fields, a gap driven by lodging susceptibility, moisture stress, weeds, seed shattering, and the limited genetic potential of widely grown cultivars.

The study concludes that while eleven of the fifteen evaluated traits showed low genotypic coefficients of variation, limiting direct selection gains, the combination of high heritability for panicle and culm length, substantial genetic advance for harvest index and lodging index, and strong positive correlations linking grain yield to spikelets per panicle, plant height, and harvest index provides a coherent selection strategy. The researchers recommend that these insights be integrated with advanced molecular techniques to pinpoint and validate key traits for breeding programs, and they emphasize that multi-location trials across diverse environments are imperative to estimate narrow-sense heritability and confirm the stability of these genetic signals. For a crop that anchors the diet of millions and is now attracting international attention, the genetic diversity documented in North Wollo may well be the raw material that finally lifts tef yields to match its cultural and nutritional importance.

Subject of Research: Genetic variability and trait correlations in tef genotypes in North Wollo, Ethiopia

Article Title: Genetic variability and associations among yield and yield related traits of Tef [ Eragrostis tef (Zucc.) Trotter] genotypes in North Wollo, Ethiopia

Article References: Wudu, E., & Mekonnen, F. (2025). Genetic variability and associations among yield and yield related traits of Tef [Eragrostis tef (Zucc.) Trotter] genotypes in North Wollo, Ethiopia. Innovations in Agriculture, 8, 1-9. https://doi.org/10.3897/ia.2025.148896

Image Credits: AI Generated

DOI: 10.3897/ia.2025.148896

Keywords: tef, Eragrostis tef, genetic variability, heritability, grain yield, Ethiopia, plant breeding, lodging index, harvest index, correlation analysis, North Wollo, recombinant inbred lines