Back

EpiKG2DAG: a Framework for Automated DAG Construction from Biomedical Text

DU, J.; Deng, G.

2026-08-11 health informatics
10.64898/2026.08.09.26360023 medRxiv
Show abstract

While Directed Acyclic Graphs (DAGs) are essential for causal inference, their construction often relies on expert heuristics, which bypasses systematic evidence synthesis and creates a critical "evidence retrieval gap" in causal modeling. This study introduces EpiKG2DAG, a framework that supports evidence-anchored candidate DAG generation by transforming unstructured biomedical abstracts into structured epidemiological associations. We utilized DeepSeek-V3 to extract exposure-outcome association triplets from 189,266 abstracts and employed SapBERT for semantic normalization against UMLS concepts. The resulting Epidemiological Knowledge Graph (EpiKG) enables the automated identification of candidate confounders, mediators, and colliders based on graph-theoretic motifs and literature-derived evidence. A case study on COVID-19 and AKI demonstrates that the framework uncovers non-obvious confounders, such as air pollution, while ensuring evidence traceability. This work contributes to the field by mitigating the knowledge-acquisition bottleneck and providing a transparent, reproducible foundation for evidence-based causal modeling.

Matching journals

The top 7 journals account for 50% of the predicted probability mass.

1
Journal of Biomedical Informatics
○ 47 papers in training set
Top 0.1%
11.7%
Similar papers in this journal
2
Bioinformatics
◐ 1204 papers in training set
Top 2%
10.8%
Similar papers in this journal
3
BMC Medical Informatics and Decision Making
○ 43 papers in training set
Top 0.2%
7.2%
Similar papers in this journal
4
Journal of the American Medical Informatics Association
◐ 71 papers in training set
Top 0.5%
6.6%
Similar papers in this journal
5
JAMIA Open
◐ 42 papers in training set
Top 0.3%
6.2%
Similar papers in this journal
6
npj Digital Medicine
○ 118 papers in training set
Top 1%
4.8%
7
Patterns
○ 78 papers in training set
Top 0.3%
4.8%
50% of probability mass above
8
BMC Medical Research Methodology
○ 47 papers in training set
Top 0.2%
4.3%
9
Nature Communications
○ 5641 papers in training set
Top 33%
4.0%
10
PLOS ONE
● 5266 papers in training set
Top 33%
4.0%
11
JMIR Medical Informatics
◐ 18 papers in training set
Top 0.2%
3.2%
12
Scientific Reports
○ 3612 papers in training set
Top 34%
3.2%
13
Database
◐ 61 papers in training set
Top 0.4%
1.7%
14
Artificial Intelligence in Medicine
○ 17 papers in training set
Top 0.4%
1.5%
15
European Journal of Epidemiology
○ 43 papers in training set
Top 0.5%
1.3%
16
GigaScience
◐ 212 papers in training set
Top 3%
1.3%
17
Computers in Biology and Medicine
○ 128 papers in training set
Top 4%
1.1%
18
iScience
○ 1154 papers in training set
Top 28%
1.1%
19
Journal of Medical Internet Research
◐ 87 papers in training set
Top 2%
1.0%
20
GENETICS
◐ 483 papers in training set
Top 4%
1.0%
21
Scientific Data
○ 209 papers in training set
Top 3%
0.8%
22
Science Advances
● 1243 papers in training set
Top 31%
0.8%
23
BMC Bioinformatics
○ 457 papers in training set
Top 6%
0.8%
24
Genome Biology
○ 637 papers in training set
Top 9%
0.8%
25
IEEE Journal of Biomedical and Health Informatics
● 37 papers in training set
Top 1%
0.8%
26
Communications Medicine
○ 113 papers in training set
Top 5%
0.6%
27
Frontiers in Bioinformatics
○ 49 papers in training set
Top 2%
0.6%

"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.