- Schneider Teixeira, Aline
- Balderas, Claudia
- Molina García, Antonio D.
Multivariate analysis employment has dramatically grown in the last years, for reducing big data sets’ inherent complexity, allowing decisive information extraction, not evident by simpler approximations. This growth is driven by the development of data analysis algorithms and computer routines, computing power increase, and by the availability of big datasets, sometimes derived from open data sources. Seed studies is one of the fields where multivariate analysis has proved its utility. Its application for seed behavior prediction under conservation conditions is especially interesting and relevant. Seed storability is a complex trait, ultimately determined by seed chemical composition and other genetic and physiological characteristics. Different compounds are related to seed desiccation and cold tolerance, longevity and successful recovery after storage. They include different size sugars and poly-alcohols, lipids, especially polyunsaturated fatty acids, prone to peroxidation, negatively correlated with seed vigor. Protein and other metabolites determined by various omics approaches can also relate to seed storage behavior. Some studies have utilized Principal Component Analysis (PCA), Linear Discriminant Analysis (LDA) and Partial Least Squares Regression (PLSR) to extract crucial information about sample variability, identifying groupings based on chemical composition or thermal stability. This study provides an overview of multivariate analysis and its contribution to plant genetic resources storage. FTIR spectra provide detailed insights into chemical composition and molecular interactions, meanwhile DSC data inform of water, proteins and lipids state. Various multivariate analysis and clustering techniques were examined, along with machine learning algorithms, such as Support Vector Machines (SVM), Neural Networks (NN), and Random Forests (RF). The methodologies explored show promising potential for providing valuable insights into seed conservation. Optimizing classification processes and enhancing quality control could improve germplasm bank practices. Moreover, the integration of advanced multivariate analysis and machine learning algorithms has the potential to achieve better seed management and long-term preservation strategies., Project SUST_CHIT_BIOPR (TED2021-129288B-C2) funded by MCIN/AEI/
10.13039/501100011033 and by the “European Union Next Generation EU/PRTR., Peer reviewed