Global trends and collaborative networks in plant genetic engineering （1994-2024）

This dataset is derived from the Web of Science (WOS) Core Collection and systematically compiles English-language research articles and reviews related to transgenic plants published globally between January 1, 1986, and December 31, 2024. The data retrieval was completed on December 6, 2024, using a composite search strategy targeting records containing keywords such as "genetically modified," "plant," and "gene technology" in their titles and/or abstracts. The subject plants include, but are not limited to, major crops like rice, maize, cotton, soybean, and tomato. The final integrated dataset comprises 6,452 valid bibliographic records, with each entry containing complete metadata such as title, publication year, authors, affiliated countries, affiliated institutions, keywords, and citation counts.For data processing, the specialized bibliometric software CiteSpace 6.3.R1 and VOSviewer 1.6.20 were used for data cleaning, statistical analysis, and visualization. These tools facilitated the construction of multidimensional analytical maps, including country collaboration networks, institutional collaboration networks, author collaboration networks, and keyword co-occurrence networks. The dataset's temporal coverage spans from 1985 to 2024 with a yearly resolution. Spatially, it encompasses 78 countries and 603 research institutions, involving 29,857 authors.In the provided data tables, each row represents a single publication. Column labels correspond to fields such as the unique identifier, publication year, author list, country, institution, journal, keywords, and citation count. Citation counts are integer values without specific units of measurement. The overall data completeness is high, with no significant missing values identified. However, a potential limitation exists as the data source is restricted to the WOS database and English-language publications, which might lead to underrepresentation of research outputs from non-English speaking regions. Furthermore, during the keyword clustering process, automated algorithms might merge semantically similar but distinct terms (e.g., "CRISPR" and "gene editing"), which is a known source of semantic bias in bibliometric analyses.The primary dataset files are in Excel format, containing the raw literature list and pre-processed matrices for collaboration and keyword frequency analysis. These files are compatible with mainstream office software and statistical analysis tools.