一种用于分析水稻（Oryza Sativa L.）基因型和表型变异性的机器学习方法，以筛选抗旱品种

《Plant Gene》：A machine learning approach for analysis of genotypic and phenotypic variability in rice ( Oryza Sativa L.) for drought stress screening

水稻抗旱品种筛选及机器学习模型应用研究。通过分析耐旱品种Banglami与易感品种Ranjit的重组自交系种群，结合随机森林和支持向量机模型，发现开花期时长等表型特征与抗旱性显著相关，支持向量机模型分类准确率最优，为培育适应东部印度干旱气候的水稻品种提供新方法。

　　

Nabarun Roy | Bulbul Ahmed | Akanksha Yadav | Ankur Tripathi | Mritunjoy Barman | Pankaj Das | Uzma Manzoor

印度北方邦大诺伊达加尔戈蒂亚斯大学农业学院，邮编201308

摘要

Nutrition is fundamental to human health and environmental sustainability, with cereal crops serving as a vital component in feeding the global population. However, climate variability poses significant challenges to crop productivity, especially in developing countries. This study focuses on enhancing rice productivity in eastern India, a region prone to droughts, by employing statistical and machine learning methods. A population of Recombinant Inbred Lines (RILs) derived from drought-tolerant Banglami and drought-sensitive Ranjit rice varieties was evaluated under drought stress and non-stress conditions. Various phenotypic traits were analyzed, and machine learning models such as Random Forest (RF) and Support Vector Machines (SVM) were used to classify genotypes based on drought resistance. The results show that certain traits are closely related to yield under stress, with SVM outperforming other models in accurately identifying stress-resistant genotypes. This research provides insights for breeding improved rice varieties with enhanced stress tolerance, contributing to food security.

引言

植物材料

Recombinant Inbred Lines (RILs) were created by crossing the drought-tolerant Banglami variety with the drought-sensitive Ranjit variety. Banglami has a tall growth period and short, plump grains, while Ranjit is semi-dwarf with a longer growth period and slender grains.

表型分析

The phenotypic characteristics of RILs at the F6 and F7 stages were evaluated in 2018 and 2019. The average days to 50% flowering (DTF) was a key factor in determining drought tolerance. Under drought stress, the average DTF was 114.87 days.

讨论

The study highlights the need for drought-tolerant rice varieties, especially in India, where climate variability affects crop productivity. Genetic analysis revealed variability in phenotypes under stress conditions, indicating the potential for improved varieties.

结论

This research uses traditional and advanced methods to develop drought-resistant rice varieties, which is crucial for food security, especially in vulnerable regions. The findings have practical implications for breeding and food security.

CRediT作者贡献声明

Nabarun Roy: Writing, review & editing, data curation, conceptualization Bulbul Ahmed: Writing, review & editing, formal analysis, data curation Akanksha Yadav: Writing, review & editing, original draft Ankur Tripathi: Visualization, validation, supervision Mritunjoy Barman: Writing, original draft, validation, supervision Pankaj Das: Project administration, methodology, investigation Uzma Manzoor: Writing, review

利益冲突声明

The authors declare no commercial or financial conflicts that could affect the research results.

致谢

The authors thank the Department of Agricultural Biotechnology (ABT) at AAU, Jorhat, Assam, and the School of Agriculture (SOAG) at Galgotias University for financial and technical support.