Landscape and Emerging Directions of Single-Cell Genetic Variation Research Revealed by LLM-Assisted Analysis
Background Single-cell genetic variation research has expanded rapidly with advances in sequencing technologies and computational approaches and is increasingly important for investigating cellular heterogeneity, clonal evolution, and disease-associated genomic alterations. However, key methodological and experimental information is often dispersed across full-text publications, limiting systematic characterization of the field. To address this, we combined large-scale literature analysis with full-text structured information extraction assisted by the large language model (LLM) ChatGPT-4o to characterize technological, methodological, and application trends in single-cell genetic variation research. Results Information on sequencing technologies, variant types, disease applications, species, and variant detection tools was systematically extracted and organized. ChatGPT-4o-assisted extraction achieved a mean correctness of 82.50% and a mean completeness of 83.05% compared with manually curated annotations. Single-cell genetic variation research has expanded rapidly, accompanied by increasing international and institutional collaboration. scRNA-seq remains the predominant technology, while scATAC-seq, scDNA-seq, spatial transcriptomics, and multi-omics combinations are increasingly adopted. Cancer is the major application area, with tumor microenvironment, prognosis, and immunotherapy as prominent themes. SNVs remain the most extensively studied variant type, whereas combined SNV-CNV analyses are becoming more common, particularly in cancer studies. Bulk-oriented tools such as GATK remain widely used, while single-cell-specific methods are increasingly adopted to address low coverage, allelic dropout, and technical noise. Conclusions Single-cell genetic variation research is evolving toward integrated multi-omics, multi-variant analysis, and more specialized single-cell computational strategies. The increasing adoption of multi-omics technologies and single-cell-specific methods reflects the growing complexity and maturity of the field. Combining large-scale literature analysis with LLM-assisted full-text information extraction provides a systematic approach for capturing complex methodological and biological information and identifying emerging directions in single-cell genomics.
Authors
- Nianling Yang
- 易显富
- jingjing Zhang
- Xiaoran Bai
- Jiawen Mao
- Yongjian Wei
- Weiwei Gao
- Hechi Zou
- Hui Zhang
Institutions
- Tianjin Medical University (CN)
Publication Details
- Journal
- Research Square
- Published
- 2026-11-20
- DOI
- https://doi.org/10.21203/rs.3.rs-11657247/v1
- Primary Topic
- Single-cell and spatial transcriptomics
- Type
- preprint
Funders
- National Natural Science Foundation of China
- Tianjin Medical University