Implementation of a Generalized Additive Model (GAM) for Soybean Maturity Prediction in African Environments

Time to maturity (TTM) is an important trait in soybean breeding programs. However, soybeans are a relatively new crop Africa. As such, TTM information for not yet as well defined other major producing areas. Multi-environment trials (METs) allow breeders analyze performance across diverse conditions, but also pose statistical challenges (e.g., unbalanced data). Modern methods, e.g., generalized additive models (GAMs), can flexibly smooth range of responses while retaining observations that could be lost under approaches. We leveraged 5 years data from MET program Africa identify the best geographical and seasonal variables explain site genotypic differences TTM. Using cycle features minimum temperature, daylength) along with trial geolocation (longitude, latitude), GAM predicted within 10 days average observed (RMSE = 10.3; x 109 post-planting). Furthermore, we found significant between cultivars (p < 0.05) sensitivity temperature daylength. Our results show potential advance design systems enhance planting decisions