A data-driven framework for identifying innovation pathways in rice abiotic stress research

Research on abiotic stress-related traits in rice—particularly salt tolerance, drought tolerance, and cadmium accumulation—is important for understanding environmental constraints on crop performance and supporting trait improvement. However, scientific innovation in this field still relies heavily on expert experience, and existing research data remain fragmented and dispersed across pathways, making it difficult to systematically integrate experimental technologies, biological entities, environmental factors, and phenotypic traits into traceable innovation pathways. To address this gap, this study develops a data-driven methodological framework that integrates multi-source data fusion, knowledge-base enhancement, multilayer heterogeneous-network construction, multidimensional innovation-feature measurement, and innovation-pathway identification. A curated literature dataset was first constructed by integrating full-text research articles with domain knowledge bases. A knowledge-base-enhanced entity–relation extraction model (KB-EREM), combining Bidirectional Encoder Representations from Transformers (BERT) with Lightweight Retrieval-Augmented Generation (LightRAG) was subsequently developed. A technology-data heterogeneous network was then built, comprising nine categories of scientific data and technical entities together with eight types of association relationships. On this basis, an innovation-feature measurement indicator system was established across five dimensions—data-drivenness, interdisciplinarity, connectivity, frontier orientation, and breakthrough potential—and the entropy-weighted Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) method was employed to quantify innovation characteristics along scientific research chains. An iterative depth-first search (DFS) algorithm was further used to identify innovation pathways within the heterogeneous network, thereby identifying potential innovation pathways in rice abiotic-stress research. The framework extracted 10,409 entity instances and 13,034 relation instances, with F1-scores of 87.58% for entity extraction and 75.03% for relation extraction. Empirical results show that innovation pathways in rice abiotic-stress research exhibit a characteristic “technological support → gene-related association → phenotypic realisation” structure, and the associations among molecular breeding technologies, environmental factors, and target traits occupy central positions in the network. Using salt-tolerance research as an example, pathways associated with SHMT3, HOG1, and OsVTC1-3 display high innovation-feature values, identifying these genes as literature-derived, computationally prioritised candidates associated with rice salt tolerance, whose biological roles require further experimental validation. The findings demonstrate that the proposed method effectively identifies traceable innovation pathways in rice abiotic-stress research, providing methodological support for literature-based candidate prioritisation, research-pathway tracing, and hypothesis generation in data-intensive crop research.

Authors

Institutions

Publication Details

Journal
Scientific Reports
Published
2026-11-19
DOI
https://doi.org/10.1038/s41598-026-79648-y
Primary Topic
scientometrics and bibliometrics research
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

A data-driven framework for identifying innovation pathways in rice abiotic stress research

Guojian Xian, Qiao Lin, Zhijie HU, Donghui Wu et al.
Scientific Reports
scientometrics and bibliometrics research
article

A data-driven framework for identifying innovation pathways in rice abiotic stress research

Guojian Xian, Qiao Lin, Zhijie HU, Donghui Wu, Lingshan Yang, Xuefu Zhang, Tan Sun
article en

Abstract

Research on abiotic stress-related traits in rice—particularly salt tolerance, drought tolerance, and cadmium accumulation—is important for understanding environmental constraints on crop performance and supporting trait improvement. However, scientific innovation in this field still relies heavily on expert experience, and existing research data remain fragmented and dispersed across pathways, making it difficult to systematically integrate experimental technologies, biological entities, environmental factors, and phenotypic traits into traceable innovation pathways. To address this gap, this study develops a data-driven methodological framework that integrates multi-source data fusion, knowledge-base enhancement, multilayer heterogeneous-network construction, multidimensional innovation-feature measurement, and innovation-pathway identification. A curated literature dataset was first constructed by integrating full-text research articles with domain knowledge bases. A knowledge-base-enhanced entity–relation extraction model (KB-EREM), combining Bidirectional Encoder Representations from Transformers (BERT) with Lightweight Retrieval-Augmented Generation (LightRAG) was subsequently developed. A technology-data heterogeneous network was then built, comprising nine categories of scientific data and technical entities together with eight types of association relationships. On this basis, an innovation-feature measurement indicator system was established across five dimensions—data-drivenness, interdisciplinarity, connectivity, frontier orientation, and breakthrough potential—and the entropy-weighted Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) method was employed to quantify innovation characteristics along scientific research chains. An iterative depth-first search (DFS) algorithm was further used to identify innovation pathways within the heterogeneous network, thereby identifying potential innovation pathways in rice abiotic-stress research. The framework extracted 10,409 entity instances and 13,034 relation instances, with F1-scores of 87.58% for entity extraction and 75.03% for relation extraction. Empirical results show that innovation pathways in rice abiotic-stress research exhibit a characteristic “technological support → gene-related association → phenotypic realisation” structure, and the associations among molecular breeding technologies, environmental factors, and target traits occupy central positions in the network. Using salt-tolerance research as an example, pathways associated with SHMT3, HOG1, and OsVTC1-3 display high innovation-feature values, identifying these genes as literature-derived, computationally prioritised candidates associated with rice salt tolerance, whose biological roles require further experimental validation. The findings demonstrate that the proposed method effectively identifies traceable innovation pathways in rice abiotic-stress research, providing methodological support for literature-based candidate prioritisation, research-pathway tracing, and hypothesis generation in data-intensive crop research.

Scientific Reports
Agricultural Information Institute (CN), Chinese Academy of Agricultural Sciences (CN), Ministry of Agriculture and Rural Affairs (CN)
scientometrics and bibliometrics research
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon