Code and data for Auditing Protein–Protein Interaction Benchmarks by Relation Pattern and Entity Boundary (APBC 2026).
conda env create -f environments/PPI-stanza.yml
conda activate PPI-stanza
python -m src check
Run all cells of reproduce.ipynb with the PPI-stanza kernel, or:
python -m src reproduce statistics --output DIR
python -m src reproduce full --output DIR
full needs the parser models. Seeds are in src/reproduction/experiment.json.
python -c "import stanza; stanza.download('en', package='craft', processors='tokenize,pos,lemma,depparse')"
python -c "import stanza; stanza.download('en', package=None, processors={'coref': 'default'})"
python -m src sentence "PROTEIN0 binds PROTEIN1."
Versions are in environments/VERSIONS.md.
| Location | Contents |
|---|---|
reproduce.ipynb | Reproduction notebook |
src/ | Reader; paper analyses in src/reproduction/ |
data/ | Corpora, lexicons and rules |
supplements/ | Supplement PDF, readable result tables, ledgers and anonymous ratings |
environments/ | Conda environment and versions |
@inproceedings{kan2026auditing,
title = {Auditing Protein--Protein Interaction Benchmarks by Relation Pattern and Entity Boundary},
author = {Kan, Li-Yu and Hsu, Wen-Lian},
booktitle = {Asia Pacific Bioinformatics Conference (APBC 2026)},
year = {2026}
}
The codebase is released under the MIT License (LICENSE). The corpora in
data/datasets/ are not covered by this license and remain subject to their original
terms. Please refer to each dataset's README for available licensing information.
Code and data for Auditing Protein–Protein Interaction Benchmarks by Relation Pattern and Entity Boundary (APBC 2026).
conda env create -f environments/PPI-stanza.yml
conda activate PPI-stanza
python -m src check
Run all cells of reproduce.ipynb with the PPI-stanza kernel, or:
python -m src reproduce statistics --output DIR
python -m src reproduce full --output DIR
full needs the parser models. Seeds are in src/reproduction/experiment.json.
python -c "import stanza; stanza.download('en', package='craft', processors='tokenize,pos,lemma,depparse')"
python -c "import stanza; stanza.download('en', package=None, processors={'coref': 'default'})"
python -m src sentence "PROTEIN0 binds PROTEIN1."
Versions are in environments/VERSIONS.md.
| Location | Contents |
|---|---|
reproduce.ipynb | Reproduction notebook |
src/ | Reader; paper analyses in src/reproduction/ |
data/ | Corpora, lexicons and rules |
supplements/ | Supplement PDF, readable result tables, ledgers and anonymous ratings |
environments/ | Conda environment and versions |
@inproceedings{kan2026auditing,
title = {Auditing Protein--Protein Interaction Benchmarks by Relation Pattern and Entity Boundary},
author = {Kan, Li-Yu and Hsu, Wen-Lian},
booktitle = {Asia Pacific Bioinformatics Conference (APBC 2026)},
year = {2026}
}
The codebase is released under the MIT License (LICENSE). The corpora in
data/datasets/ are not covered by this license and remain subject to their original
terms. Please refer to each dataset's README for available licensing information.