Potato Agent: AI-Driven Data and Knowledge Exploration on an Agent-Ready Potato Multi-Omics Platform
Dong, Y.; Li, J.; Li, F.; Luo, J.; Jia, Y.; Li, D.; Wang, L.; Su, X.; Hu, J.; Shang, Y.; Huang, S.; Zhu, Y.; Jia, Y.
Show abstract
Potato is an important non-cereal food crop worldwide. However, the limited number of functionally validated genes remains a major bottleneck to favorable allele stacking and genome design breeding in potato. Rapid advances in AI agents offer a promising means to support crop breeding by translating natural-language questions into coordinated data analysis and knowledge retrieval. Their reliable use for potato breeding, however, is constrained by fragmented multi-omics resources that lack consistent curation and machine-accessible interfaces. Here, we constructed an agent-ready potato multi-omics database integrating genomic resources from 150 potato accessions, 259 bulk RNA-seq samples, and 14 spatial transcriptomic datasets into a pangenome, a tissue expression atlas, co-expression networks, and spatial expression maps accessible through open APIs. We developed 39 potato-specific Agent Skills for reproducible bioinformatics analysis and comprehensive data and knowledge exploration, enabling natural-language questions to be translated into standardized data-retrieval and analysis tasks. By integrating direct evidence from potato studies, functions of homologous genes in Arabidopsis, rice, and maize, and tissue expression patterns, we generated genome-wide functional predictions for 37,658 genes in the DM reference genome. We further developed Potato Agent as a multi-user, browser-based platform with isolated workspaces and online result preview, reducing the technical burden of agent deployment and providing direct access to integrated data, knowledge, and workflows. Case studies demonstrated its capabilities in reproducible bioinformatics analysis, agent-assisted identification of a tuber development regulator, scientific data visualization, and haplotype-aware promoter analysis and sgRNA design. Together, the agent-ready database and Potato Agent provide an integrated infrastructure for functional gene discovery and hybrid breeding in potato.
Matching journals
The top 8 journals account for 50% of the predicted probability mass.
Similar papers in this journal
- DeepPGDB: A Novel Paradigm for AI-Guided Interactive Plant Genomic Database 94%
- Artemisia Database: A Comprehensive Resource for Gene Expression and Functional Insights in Artemisia annua 91%
- EasyOmics-A graphical interface for population-scale omics data association, integration and visualization 90%
Similar papers in this journal
Similar papers in this journal
Similar papers in this journal
- Graph-based pangenomics maximizes genotyping density and reveals structural impacts on fungal resistance 91%
- Exceptional subgenome stability and functional divergence in allotetraploid teff, the primary cereal crop in Ethiopia 91%
- Purging of deleterious mutations during domestication in the predominant selfing crop soybean 91%
Similar papers in this journal
- High-throughput phenotyping with deep learning gives insight into the genetic architecture of flowering time in wheat 93%
- RWRtoolkit: multi-omic network analysis using random walks on multiplex networks in any species 92%
- Long-read and chromosome-scale assembly of the hexaploid wheat genome achieves high resolution for research and breeding 91%
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.