Max Planck Institute of Molecular Plant Physiology, Potsdam-Golm, Germany.
Plant Biotechnol J. 2010 Oct;8(8):900-11. doi: 10.1111/j.1467-7652.2010.00516.x.
Biomarkers are used to predict phenotypical properties before these features become apparent and, therefore, are valuable tools for both fundamental and applied research. Diagnostic biomarkers have been discovered in medicine many decades ago and are now commonly applied. While this is routine in the field of medicine, it is of surprise that in agriculture this approach has never been investigated. Up to now, the prediction of phenotypes in plants was based on growing plants and assaying the organs of interest in a time intensive process. For the first time, we demonstrate in this study the application of metabolomics to predict agronomic important phenotypes of a crop plant that was grown in different environments. Our procedure consists of established techniques to screen untargeted for a large amount of metabolites in parallel, in combination with machine learning methods. By using this combination of metabolomics and biomathematical tools metabolites were identified that can be used as biomarkers to improve the prediction of traits. The predictive metabolites can be selected and used subsequently to develop fast, targeted and low-cost diagnostic biomarker assays that can be implemented in breeding programs or quality assessment analysis. The identified metabolic biomarkers allow for the prediction of crop product quality. Furthermore, marker-assisted selection can benefit from the discovery of metabolic biomarkers when other molecular markers come to its limitation. The described marker selection method was developed for potato tubers, but is generally applicable to any crop and trait as it functions independently of genomic information.
生物标志物可用于在表型特征变得明显之前预测其性质，因此是基础研究和应用研究的有价值的工具。医学领域几十年前就已经发现了诊断生物标志物，现在已经广泛应用。虽然这在医学领域是常规做法，但令人惊讶的是，农业领域从未对此进行过研究。到目前为止，植物表型的预测一直是基于种植植物并在耗时的过程中分析感兴趣的器官。在这项研究中，我们首次展示了代谢组学在不同环境下种植的作物中预测农艺重要表型的应用。我们的程序包括筛选大量非靶向代谢物的既定技术，同时结合机器学习方法。通过使用代谢组学和生物数学工具的这种组合，鉴定出可以用作生物标志物以改善性状预测的代谢物。可以选择预测代谢物，随后用于开发快速、靶向和低成本的诊断生物标志物检测，可用于育种计划或质量评估分析。鉴定出的代谢生物标志物可用于预测作物产品质量。此外，当其他分子标记物达到其局限性时，代谢生物标志物的发现可以使标记辅助选择受益。所描述的标记选择方法是为马铃薯块茎开发的，但通常适用于任何作物和性状，因为它独立于基因组信息而运作。
