Genomics and Applied Biology 2024, Vol.15, No.6, 307-319 http://bioscipublisher.com/index.php/gab 313 In conclusion, each model has its strengths and weaknesses, and no single model is universally optimal. GBLUP is efficient for highly polygenic traits, while Bayesian models excel when specific markers have substantial effects. Machine learning models may provide an edge in cases of complex gene interactions. Hybrid approaches, combining different models, can sometimes achieve better accuracy, capturing a broader range of genetic effects. 5.3 Adaptability of prediction models under different environmental conditions One of the challenges in applying genomic prediction models in yam breeding is the adaptability of these models under diverse environmental conditions. Yam is cultivated in a variety of climatic regions, and traits like yield and disease resistance can vary significantly with environmental conditions. Prediction models must be capable of accounting for genotype-by-environment (GxE) interactions to be effective across different regions. GBLUP, for instance, can be extended with environmental interaction terms to model GxE effects, allowing predictions to consider both the genetic and environmental components. This is especially beneficial in crops like yam, where performance can be highly influenced by local climate and soil conditions (Crossa et al., 2017). Models incorporating environmental covariates have shown improved accuracy in predicting trait performance across different environments. For instance, GBLUP models with interaction terms are used in crops like barley and wheat, where they have been shown to perform better than standard models in multi-environment trials. This adaptability is crucial in yam breeding, as traits such as drought resistance, tuber yield, and disease resistance are influenced by climatic variability. Researchers have also explored the use of multivariate models and mixed-model approaches to better capture GxE interactions, which can significantly improve prediction accuracy in yam (Ge et al., 2020). Additionally, recent advances in high-throughput phenotyping and environmental monitoring provide data that can be incorporated into genomic prediction models. By integrating phenotypic and environmental data, machine learning models, like deep learning, can capture complex interactions that simpler models may miss. For example, CNNs and MLPs, when trained with environmental covariates, show promising accuracy improvements for yield predictions in crops with variable growing conditions (Sandhu et al., 2021). This approach holds potential for yam breeding, enabling breeders to make predictions that are more accurate and tailored to specific environments. Overall, the adaptability of genomic prediction models under diverse conditions enhances their utility, making them valuable tools for yam breeders working across various climates. 6 Strategies to Shorten the Yam Breeding Cycle 6.1 Application of speed breeding techniques Speed breeding (SB) is an innovative technique designed to accelerate plant growth and development through the use of controlled environments, including extended photoperiods, greenhouses, and precise temperature and humidity settings. By adjusting these factors, SB allows multiple plant generations to complete their life cycles within a single year, significantly advancing the breeding process. Originally demonstrated in wheat, SB has shown potential for use in long-generation crops like yam, where it can reduce the time required for seed maturation and full plant development, ultimately speeding up the breeding cycle and improving genetic gains (Watson et al., 2019). In yam breeding, speed breeding techniques could be applied by maintaining plants under artificial light for extended periods, simulating continuous day conditions to maximize photosynthesis and encourage rapid growth. Using greenhouse facilities, yam plants could undergo more frequent seed-to-seed cycles, reducing the time between generations. These methods allow breeders to focus on critical early traits such as seedling vigor, disease resistance, and drought tolerance under controlled conditions before field evaluation. Mondo et al. (2021) demonstrated that changes in calcium nitrate concentration significantly affect pollen germination rates in these two yam varieties, particularly with a notable increase in pollen germination at optimal concentrations. This finding supports the application of appropriate calcium nitrate concentrations in the breeding process to enhance yam pollen viability, thereby increasing the likelihood of successful pollination and fruit set (Mondo et al., 2021) (Figure 2). For example, studies in wheat have shown that by reducing the photoperiod, plants can go through several generations in one year, significantly expediting breeding cycles (Begna, 2022).
RkJQdWJsaXNoZXIy MjQ4ODYzMg==