Tree Genetics and Molecular Breeding 2025, Vol.15, No.5, 192-201 http://genbreedpublisher.com/index.php/tgmb 198 8.3 Limited field-level validation of laboratory findings Most research on theobromine is conducted in laboratories, mainly at the cellular, tissue or organ level. The drawback is the lack of validation in the field environment (Koyama et al., 2003; Zheng et al., 2004). The metabolic and gene expression patterns in the laboratory often differ from those in the field environment. Especially when facing environmental stress, pests and diseases or agronomic management, the differences are more obvious (Gallego et al., 2021). In addition, post-harvest treatments, such as fermentation and baking, can also significantly affect the content of theobromine and related alkaloids, but the molecular mechanism is not fully understood (Febrianto and Zhu, 2022; Cortez et al., 2023). Therefore, for laboratory achievements to be truly applied to field production, more research is still needed. 9 Future Perspectives 9.1 Multi-omics integration: metabolomics, genomics, transcriptomics, proteomics Multi-omics integration has become an important method for studying complex biological processes. Analyzing the genome, transcriptome, proteome and metabolome together can provide a more comprehensive understanding of the synthesis process and regulatory network of theobromine. This method can link genes and phenotypes, and also identify key regulatory factors and metabolic pathways, providing a basis for functional gene research and molecular marker development (Picard et al., 2021; Wörheide et al., 2021; Ramos-López et al., 2022). In the study of cocoa processing and flavor, multi-omics analysis also revealed metabolite differences in different regions, varieties and processing stages, and more molecules related to flavor were understood through proteomic and peptidomic studies (Agyirifo et al., 2019; Herrera-Rocha et al., 2023). In the future, with the development of high-throughput technologies and data analysis tools, multi-omics integration will play a greater role in cocoa quality improvement and functional component development (Mahmood et al., 2022; Shankar and Sharma, 2022; Cembrowska-Lech et al., 2023). 9.2 AI-driven metabolite prediction and pathway modeling Artificial intelligence (AI) and machine learning (ML) provide new means for the integration of multi-omics data and the analysis of complex networks. AI can help predict the relationship among genotypes, environment and phenotypes through deep learning, knowledge graphs and generative models, and can also discover new molecular targets and biomarkers (Picard et al., 2021; Mahmood et al., 2022; Yan and Wang, 2022). In metabolite prediction, AI can handle complex big data and also improve the training effect of models through synthetic data, thereby more accurately identifying new metabolic pathways and key enzymes (Shankar and Sharma, 2022; Wu and Xie, 2024). These technologies will contribute to more precise regulation of theobromine and related flavors and nutritional components (Yan and Wang, 2022; Cembrowska-Lech et al., 2023). 9.3 Toward precision breeding for optimized flavor and nutritional value Combining multi-omics with AI provides a new direction for the precise breeding of Theobroma cacao. By integrating genomic, transcriptomic, proteomic and metabolomic data and combining them with AI models, complex traits such as flavor and nutrition can be predicted more quickly and accurately (Shankar and Sharma, 2022; Yan and Wang, 2022). Precision breeding can accelerate the selection and breeding of superior varieties, and also improve the flavor and nutrition of cocoa in a targeted manner according to different market demands, achieving optimization from genes to products (Mahmood et al., 2022). In the future, with the accumulation of more data and the improvement of AI algorithms, intelligent breeding relying on molecular markers and phenotypic prediction will become an important driving force for promoting the high-quality development of Theobroma cacao industry (Yan and Wang, 2022; Cembrowska-Lech et al., 2023). Acknowledgments The authors appreciate the comments from two anonymous peer reviewers on the manuscript of this study. The authors also thank the group members for helping to organize the research data. Conflict of Interest Disclosure The authors affirm that this research was conducted without any commercial or financial relationships that could be construed as a potential conflict of interest.
RkJQdWJsaXNoZXIy MjQ4ODYzNA==