Montesinos-López Abelardo, Crespo-Herrera Leonardo, Dreisigacker Susanna, Gerard Guillermo, Vitale Paolo, Saint Pierre Carolina, Govindan Velu, Tarekegn Zerihun Tadesse, Flores Moisés Chavira, Pérez-Rodríguez Paulino, Ramos-Pulido Sofía, Lillemo Morten, Li Huihui, Montesinos-López Osval A, Crossa Jose
Departamento de Matemáticas, Centro Universitario de Ciencias Exactas e Ingenierías (CUCEI), Universidad de Guadalajara, Guadalajara, Jalisco, Mexico.
International Maize and Wheat Improvement Center (CIMMYT), Texcoco, Estado. de México, Mexico.
Front Plant Sci. 2024 Mar 4;15:1324090. doi: 10.3389/fpls.2024.1324090. eCollection 2024.
In the field of plant breeding, various machine learning models have been developed and studied to evaluate the genomic prediction (GP) accuracy of unseen phenotypes. Deep learning has shown promise. However, most studies on deep learning in plant breeding have been limited to small datasets, and only a few have explored its application in moderate-sized datasets. In this study, we aimed to address this limitation by utilizing a moderately large dataset. We examined the performance of a deep learning (DL) model and compared it with the widely used and powerful best linear unbiased prediction (GBLUP) model. The goal was to assess the GP accuracy in the context of a five-fold cross-validation strategy and when predicting complete environments using the DL model. The results revealed the DL model outperformed the GBLUP model in terms of GP accuracy for two out of the five included traits in the five-fold cross-validation strategy, with similar results in the other traits. This indicates the superiority of the DL model in predicting these specific traits. Furthermore, when predicting complete environments using the leave-one-environment-out (LOEO) approach, the DL model demonstrated competitive performance. It is worth noting that the DL model employed in this study extends a previously proposed multi-modal DL model, which had been primarily applied to image data but with small datasets. By utilizing a moderately large dataset, we were able to evaluate the performance and potential of the DL model in a context with more information and challenging scenario in plant breeding.
在植物育种领域，已经开发并研究了各种机器学习模型，以评估未见表型的基因组预测（GP）准确性。深度学习已显示出前景。然而，大多数关于植物育种中深度学习的研究都局限于小数据集，只有少数研究探索了其在中等规模数据集上的应用。在本研究中，我们旨在通过使用一个适度大的数据集来解决这一局限性。我们检验了一个深度学习（DL）模型的性能，并将其与广泛使用且强大的最佳线性无偏预测（GBLUP）模型进行比较。目标是在五折交叉验证策略的背景下以及使用DL模型预测完整环境时评估GP准确性。结果显示，在五折交叉验证策略中包含的五个性状中的两个性状方面，DL模型在GP准确性上优于GBLUP模型，其他性状的结果相似。这表明DL模型在预测这些特定性状方面具有优越性。此外，当使用留一环境法（LOEO）预测完整环境时，DL模型表现出具有竞争力的性能。值得注意的是，本研究中使用的DL模型扩展了先前提出的多模态DL模型，该模型主要应用于图像数据，但使用的是小数据集。通过使用一个适度大的数据集，我们能够在植物育种中更具信息和挑战性的场景下评估DL模型的性能和潜力。
