Reproducibility package for: Integrative single-cell transcriptomic and proteomic modeling maps host protease sources in the dentin–pulp complex and estimates the coverage of combined MMP–cathepsin inhibition

Reproducibility package — Integrative single-cell and proteomic modeling of the dentin–pulp protease environment This deposit contains all data files and the complete analysis script required to reproduce every figure and quantitative result of the associated study, including all sensitivity and robustness re-analyses. The workflow reconstructs a cell-resolved protease–substrate degradation network of the dentin–pulp complex from publicly deposited single-cell transcriptomic and proteomic datasets and evaluates in-silico protease-inhibition scenarios. Executing the provided script regenerates the main figures, the sensitivity figure, and the revision figures, together with the underlying result tables. Contents Supplementary_reproduction_code.py — single self-contained analysis script (Modules 1–7 and Re-analyses A–E). requirements.txt and environment.yml — pinned software dependencies (Python 3.11). DATA_MANIFEST.csv — every data file with its size, SHA-256 checksum, source accession, and description. LICENSE — MIT License for the analysis code. README.md — full setup and execution instructions. data/ — all input and intermediate files read by the script (raw single-cell matrices, proteomic tables, curated model inputs, intermediate result tables, and optional pre-computed single-cell checkpoints). How to reproduce Create the software environment with conda env create -f environment.yml (or pip install -r requirements.txt in a Python 3.11 environment). Keep the data/ folder next to the script, or set the DENTIN_DATA_DIR environment variable to its location. Run python Supplementary_reproduction_code.py. All file paths are resolved relative to a single data directory, so no code editing is required. Random seeds are fixed (Monte Carlo seed = 42; 2,000 iterations), so uncertainty intervals and permutation-null results are fully reproducible. Internet access is required only for one step (Module 5 queries the public UniProt REST API to map protein accessions to gene symbols); all other modules run offline. Data provenance The primary datasets remain available from their original public repositories and should be cited by accession: Gene Expression Omnibus (GSE164157, GSE185222, GSE251953) and PRIDE/ProteomeXchange (PXD070849, PXD002264). The copies included here are provided solely to enable one-step reproduction; integrity can be verified against the SHA-256 checksums in DATA_MANIFEST.csv. License and citation The analysis code is released under the MIT License; redistributed third-party datasets remain under the terms of their source repositories. When reusing these materials, please cite both this record (by its DOI) and the associated article.