Cotton Genomics and Genetics 2025, Vol.16, No.5, 249-258 254 Figure 2 Description of SNPs in CottonSNP80K array. a: SNPs distributions on the 26 chromosomes of upland cotton. A01-A13 and D01-D13 in vertical axis are the serial number of 26 chromosomes; the horizontal axis shows chromosome length (Mb); the red region depicts SNP density (the number of SNPs per window). b: Distances between the SNPs. The vertical axis represents distances range (Kb) of SNPs. c: Distribution of genic and intergenic regions of selected SNPs (Adopted from Cai et al., 2017) 7 Integration of Genotyping with Breeding Decision Platforms 7.1 Integration with genomic selection (GS) models Today's high-throughput genotyping is not merely about providing a vast amount of SNP data; it has almost become a prerequisite for the operation of GS models. In the past, the prediction efficiency of breeding values was limited by data density and the operational threshold of model tools. Now, with platforms like AutoGP and IP4GS, the situation has changed (Li et al., 2023; Wu et al., 2025). They have user-friendly interfaces. Even breeders who don't know programming can model and predict by themselves. From genotype extraction, model parameter adjustment to the final phenotypic prediction, everything is packaged (Hickey et al., 2017; Yan et al., 2021). How to select the model, whether the prediction is accurate, and which parent is the best can all be quickly tested through these tools. Ultimately, they have transformed GS from merely an "expert-only" tool into a practical and operational productivity tool. 7.2 Joint analysis of high-throughput phenotyping and genotyping data Just looking at the genotype is not enough. In actual breeding, there are too many phenotypic and environmental interference factors. It is necessary to comprehensively consider the typing data, high-throughput phenotypic information and environmental background together in order to predict complex traits more accurately. Nowadays, some intelligent breeding strategies simply feed genomic, phenome and environmental data into the model together, and run the analysis in combination with machine learning or deep learning algorithms. The effect is actually more stable (Adunola et al., 2024). In terms of platforms, it's not necessary to set up your own server. Tools like Breedbase and IBP can basically help you organize, analyze, and even visualize your data (Varshney et al., 2016; Rasheed et al., 2017; Morales et al., 2020). Once the process is clear and the results are intuitive, the decision-making of the breeding plan will also be more confident. 7.3 The role of cloud computing and smart platforms in decision support Facing huge amounts of omics data, it is no longer realistic to rely on traditional desktop systems for analysis. At this point, the cloud platform becomes particularly crucial. For instance, the breeding management system (BMS)
RkJQdWJsaXNoZXIy MjQ4ODYzNA==