Plant molecular marker in AI era

Xiaoxu Li¹Zhengrong Hu²Wen Yu³He Xie⁴Pingjun Huang⁵Xinyao Zhang⁵Jia-Shuo Yang²Yangyang Li²Xiaonian Yang^5*Zhenchen Zhang^6*Wenxuan Pu^5*Zhiyuan Li^7*

• ¹Beijing Life Science Academy, Beijing, China
• ²Hunan Tobacco Research Institute, Changsha, China
• ³Institute of Tobacco Science, Fujian Provincial Tobacco Company, Fuzhou, China
• ⁴Yunnan Academy of Tobacco Agricultural Sciences, Kunming, China
• ⁵Technology Center, China Tobacco Hunan Industrial Co., Ltd.,, Changsha, China
• ⁶Guangdong Academy of Agricultural Sciences Crops Research Institute, Guangzhou, China
• ⁷Chinese Academy of Agricultural Sciences Institute of Tobacco Research, Qingdao, China

Plant molecular marker technologies have reshaped crop genetics and breeding by making it possible to analyse genome-wide variation with a precision that phenotype-based selection, even in experienced programmes, cannot reach in routine practice. This review summarises recent progress in marker platforms from classical RFLP and SSR systems to high-throughput SNP genotyping, with emphasis on KASP, multiple nucleotide polymorphism and multi-gene panel technologies, and on sequencing-based methods such as GBTS and Hyper-seq that often serve as an upstream discovery layer for targeted assays and databases. These platforms are increasingly integrated into practical workflows for marker-assisted and genomic selection, DNA fingerprinting, germplasm characterisation and plant variety protection, and multi-locus markers have become a central tool for high-resolution DUS testing and EDV determination that adds an independent layer of evidence to morphology-based assessments. Key challenges now include cross-platform standardisation, design of marker panels that balance cost with information content, interoperability of databases across institutions and countries, and the definition of molecular distance thresholds that are acceptable both biologically and in legal and regulatory settings. The review also considers the rapid integration of molecular marker data with artificial intelligence, including AI-driven marker discovery and panel optimisation, genomic prediction in multi-environment trials and the concept of an intelligent seed-industry operating system that links genotypic, phenotypic and environmental information in a coherent data framework. These developments collectively point to a shift from isolated marker assays towards platform-level, AI-supported infrastructures that can accelerate variety innovation and contribute to the modernisation and quality improvement of the seed industry.