Triticeae Genomics and Genetics, 2025, Vol.16, No.3, 110-119 http://cropscipublisher.com/index.php/tgg 115 drought-resistant markers into the breeding process, making them one of the regular selection indicators. According to the current progress, the unified use of those QTLS and corresponding markers that have been repeatedly verified is very likely to make drought-resistant breeding both faster and more accurate. 6 Molecular Tools and Genomic Resources Enhancing QTL Discovery 6.1 Use of high-density SNP arrays and genotyping-by-sequencing (GBS) In the past, QTL positioning relied on low-density tagging, which was time-consuming, costly and not very efficient. But now the situation has changed. High-density SNP arrays and GBS enable researchers to genotype large populations more quickly and at a lower cost, and with much higher resolution. These methods can provide dense markers across the entire genome, which helps to more accurately lock onto QTLS, especially those rare or harmful variations that are easily missed by traditional methods (Borevitz and Chory, 2004). Of course, typing alone is not enough. After combining genome-wide variation data with SNP markers, the detection ability will be stronger, especially in complex field environments (Macleod et al., 2016). In addition, tools such as QTLseqr and FastQTL have also saved a lot of trouble in data processing and are convenient and efficient for batch analysis (Ongen et al., 2015; Mansfeld and Grumet, 2017). 6.2 Genome-wide association studies (GWAS) complementing QTL mapping When many people mention QTL, they only think of location maps, but in fact, GWAS has long been a main tool. Its approach is different-instead of relying on population construction, it uses existing natural variation resources to find loci related to traits (Zhang et al., 2022). The strength of GWAS lies in "casting a net" on a whole-genome scale. When used in combination with QTL data and annotation information, the effect will be better (Huang et al., 2022). Nowadays, there are also many online tools assisting this type of analysis, such as ezQTL and QTLbase2. They not only make the results more intuitive but also superimpose and compare the results of GWAS and QTL, making it convenient to identify those loci with true biological significance. 6.3 Integration of transcriptomics and gene annotation in QTL fine-mapping To understand the mechanism behind QTL, merely relying on position is far from enough. At this point, transcriptome data, gene expression and various annotation information all need to be brought in for analysis together. Data like eQTL (expression QTL), in combination with transcriptional information and variant annotations, can significantly improve localization accuracy and also help infer functional mechanisms (Wen, 2016). Especially for those variations that are not very obvious in location but have important functions, it is even more necessary to rely on these integrations. Some new methods have emerged, such as BayesRC or EPISPOT. These models incorporate biological background knowledge, omics data, and regulatory characteristics into the analysis, which are very helpful for identifying specific variations (Ruffieux et al., 2020). Furthermore, databases such as QTLbase2 and QTLtools have been able to support the exploration of QTL under various biological conditions and molecular levels, and the integration of resources is becoming increasingly in place (Delaneau et al., 2016). 7 Future Directions in Drought-Resilient Triticeae Breeding 7.1 Potential of genomic selection (GS) and machine learning in predicting drought performance Traditional breeding methods are not ineffective, but when it comes to the complex trait of drought resistance, their efficiency often causes concern. Nowadays, genomic selection (GS) is being adopted by an increasing number of breeding projects. One of its advantages is that it can make predictions based on the markers of the entire genome, without having to wait for field trials to know the results. This is indeed quite effective in saving time and accelerating the breeding process. Moreover, once the GS model combines high-throughput phenotypic and environmental data, the accuracy of prediction is usually higher than that of traditional methods (Mwadzingeni et al., 2016). Of course, GS alone is not enough. In recent years, the integration of machine learning has further advanced this type of prediction. It can integrate data from omics, environment, and even planting management, simulate the interaction between genotypes and the environment, and help breeders more accurately select materials suitable for arid regions (Cooper and Messina, 2022). Some projects have already started to test the waters with these tools on wheat and barley, and the prospects seem quite promising (Caccialupi et al., 2023).
RkJQdWJsaXNoZXIy MjQ4ODYzNA==