Phylopower 0.1.0__tar.gz

phylopower-0.1.0/CITATION.cff

@@ -0,0 +1,22 @@
+cff-version: 1.2.0
+message: "If you use PhyloPower, please cite this software and the associated manuscript."
+title: "PhyloPower: manuscript-aligned power and sample-size estimation for meta-omics beta-diversity studies"
+type: software
+authors:
+  - family-names: Li
+    given-names: Yutong
+    affiliation: Jilin University
+version: 0.1.0
+date-released: 2026-07-24
+license: MIT
+repository-code: "https://github.com/Yutongli534/PhyloPower"
+keywords:
+  - microbiome
+  - metagenomics
+  - metaproteomics
+  - power analysis
+  - sample size
+  - beta diversity
+  - PERMANOVA
+  - phylogenetics
+doi: 10.5281/zenodo.21991013

phylopower-0.1.0/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 π-PhyloPower authors
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

phylopower-0.1.0/MANIFEST.in

@@ -0,0 +1,14 @@
+include README.md
+include LICENSE
+include CITATION.cff
+include REPRODUCIBILITY.md
+include phylopower/datagene/*.csv
+include phylopower/datagene/*.nwk
+include phylopower/datapro/sample_metadata.csv
+include phylopower/datapro/taxon_function_table.csv
+include phylopower/datapro/rooted_tree.nwk
+recursive-include phylopower/tutorials *.ipynb
+include requirements-lock.txt
+
+prune *.egg-info
+global-exclude *.py[cod]

phylopower-0.1.0/PKG-INFO

@@ -0,0 +1,300 @@
+Metadata-Version: 2.4
+Name: Phylopower
+Version: 0.1.0
+Summary: PhyloPower: manuscript-aligned power and minimum-sample-size estimation for meta-omics beta-diversity studies.
+Author: Yutong Li
+License-Expression: MIT
+Keywords: microbiome,metagenomics,metaproteomics,power-analysis,beta-diversity,PERMANOVA,phylogenetic,bootstrap,sample-size,PhyloFunc,Gemelli
+Classifier: Development Status :: 4 - Beta
+Classifier: Intended Audience :: Science/Research
+Classifier: Operating System :: POSIX :: Linux
+Classifier: Operating System :: MacOS
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.9
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Topic :: Scientific/Engineering :: Bio-Informatics
+Requires-Python: >=3.9
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Requires-Dist: numpy>=1.22
+Requires-Dist: pandas>=1.5
+Requires-Dist: scipy>=1.9
+Requires-Dist: scikit-bio>=0.6.3
+Requires-Dist: joblib>=1.2
+Requires-Dist: matplotlib>=3.7
+Requires-Dist: tqdm>=4.64
+Requires-Dist: psutil>=5.9
+Requires-Dist: phylofunc<3,>=2.0
+Requires-Dist: biopython>=1.80
+Provides-Extra: dev
+Requires-Dist: pytest>=7; extra == "dev"
+Requires-Dist: ruff>=0.4; extra == "dev"
+Requires-Dist: black>=24.0; extra == "dev"
+Dynamic: license-file
+
+# PhyloPower
+
+[![DOI](https://zenodo.org/badge/DOI/10.5281/zenodo.21991013.svg)](https://doi.org/10.5281/zenodo.21991013)
+![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)
+
+PhyloPower estimates the minimum balanced per-group sample size required for
+community-level meta-omics studies at a target statistical power and realized
+PERMANOVA effect size (ω²).
+
+## For reviewers
+
+The exact code version associated with the manuscript is permanently archived
+at Zenodo: https://doi.org/10.5281/zenodo.21991013. To verify the manuscript's
+claims, start from:
+
+- `tutorials/tutorial.ipynb` — end-to-end
+  walkthrough of both workflows on the bundled demo data;
+- `figures/README.md` — figure-to-script map: every
+  manuscript figure and supplementary figure has exactly one generating
+  script, with precomputed result caches so figures can be regenerated
+  without re-running the full simulations;
+- `REPRODUCIBILITY.md` — environment setup, seeds, and
+  Monte Carlo settings used for the manuscript;
+- `validation_datasets/` — processed external
+  validation cohorts with preparation summaries and checksums.
+
+These companion files are distributed with the Zenodo archive and are not part
+of the PyPI package; `pip install` provides the Python package and the bundled
+demo datasets only.
+
+## Overview
+
+The manuscript-aligned implementation is `phylopower.cli`. It provides
+two raw-pool workflows:
+
+| Command | Synthetic generator | Distance |
+|---|---|---|
+| `gene` | PCAM | Gemelli phylogenetic RPCA |
+| `protein` | MDC-TF-MC | PhyloFunc |
+
+A step-by-step walkthrough of both workflows is provided in the tutorial
+notebook `tutorials/tutorial.ipynb`. Manuscript
+figure scripts and the figure-to-script map are documented in
+`figures/README.md`.
+
+`phylopower/cli.py` is a generated, self-contained runner. It embeds the
+lower-level core and all other project-local runtime modules, so it does not
+import an external `phylopower/core.py`.
+
+## Usage
+
+Run the workflows from this directory with `python -m phylopower`, or install
+the package first (`python -m pip install .`) and use the `phylopower`
+command. The bundled `phylopower/datagene/` and `phylopower/datapro/` demo
+datasets are found automatically; set `PHYLOPOWER_DATA_DIR` or provide input
+paths explicitly to use your own data.
+
+For the protein workflow, install the runtime dependencies with:
+
+```bash
+python -m pip install \
+  "numpy>=1.22" "pandas>=1.5" "scipy>=1.9" "scikit-bio>=0.6.3" \
+  "joblib>=1.2" "matplotlib>=3.7" "tqdm>=4.64" "psutil>=5.9" \
+  phylofunc "biopython>=1.80"
+```
+
+Both pandas 2.x and 3.x are supported; the workflows are verified against
+pandas 3.0 Copy-on-Write semantics (`DataFrame.to_numpy()` read-only views).
+
+The gene workflow must be launched from the compatible QIIME 2 environment
+that contains Gemelli and `biom-format`; selecting that environment only with
+`--qiime-env` is not a substitute for activating it because PCAM calls the
+Gemelli Python API in-process.
+
+```bash
+python -m phylopower --help
+
+python -m phylopower gene \
+  --table /data/feature_table.csv \
+  --tree /data/rooted_tree.nwk \
+  --taxonomy /data/taxonomy.csv \
+  --group /data/sample_metadata.csv \
+  --target-power 0.80 \
+  --out gene_result
+
+python -m phylopower protein \
+  --table /data/taxon_function_table.csv \
+  --tree /data/rooted_tree.nwk \
+  --group /data/sample_metadata.csv \
+  --target-power 0.80 \
+  --out protein_result
+```
+
+A full `protein` run with the default Monte Carlo settings targets
+publication-grade estimates and can take tens of minutes and several GB of
+memory. For a quick smoke test on the bundled demo data (reduced Monte Carlo
+effort, not suitable for publication estimates):
+
+```bash
+python -m phylopower protein \
+  --target-power 0.80 \
+  --min-n 2 --max-n 4 --coarse-step 2 \
+  --boot-number 10 --permutations 49 \
+  --pool-size-per-group 10 \
+  --mdctf-strength-candidates 6 --mdctf-preview-pool-size 60 \
+  --power-preview-boot-number 5 --power-preview-permutations 19 \
+  --out protein_quick
+```
+
+If `--target-omega2` is omitted, the observed ω² of the resolved pilot is used
+and recorded in `summary.json`.
+
+The effect grid is widened automatically when the target ω² exceeds the
+simulated range (`--pcam-scale-extend-max` for `gene`, `--mdctf-strength-max`
+for `protein`). If the target still falls outside the realized ω² range, the
+fitted power at the target would rest on curve extrapolation only: a warning
+is printed, the extrapolated fit is not allowed to qualify, no minimum n is
+reported, and `summary.json` records `"target_omega2_bracketed": false`
+(together with `low_omega_support_warning`). Widen the grid with the flags
+above and rerun in that case.
+
+### Monte Carlo settings
+
+The default settings are `--boot-number 200` bootstrap iterations per
+effect level and `--permutations 199` permutations per PERMANOVA; scikit-bio
+applies the `+1` correction internally, so p-value denominators are 200.
+Publication-grade estimates should use substantially larger counts — see
+`REPRODUCIBILITY.md`. All stochastic steps, including
+the PERMANOVA permutation streams and tree perturbation, are deterministically
+derived from `--random-seed`, so a fixed seed reproduces a run bit-for-bit.
+
+To inspect the exact embedded-module hashes:
+
+```bash
+python phylopower/cli.py --standalone-info
+```
+
+The maintainable build sources are `phylopower/_cli_source.py` and
+`phylopower/_core_source.py`; both are included under `phylopower/`. They
+are used only to regenerate the self-contained runners; the generated
+`phylopower/cli.py` does not read them at runtime. The regeneration script
+(`scripts/build_standalone.py`) is maintained in the GitHub repository.
+
+### Optional controls
+
+- **Parallelism:** `--scenario-n-jobs` sets the number of independent effect
+  scenarios processed concurrently, and `--bootstrap-n-jobs` sets the joblib
+  workers used inside the bootstrap/PERMANOVA loops (`0` = automatic).
+- **Tree perturbation:** `--tree-jitter-sigma` and `--tree-nni-prob` set the
+  perturbation magnitudes; `--tree-support-threshold` restricts NNI to nodes
+  below a given support value; `--tree-set trees.nwk` estimates both
+  magnitudes from a set of bootstrap or posterior trees, with explicit values
+  taking precedence. All tree-perturbation options are off by default.
+- **Pilot stability diagnostic:** `phylopower diagnose --modality
+  {gene,protein} --table ... --tree ... --group ... [--taxonomy ...]
+  --out diagnosis/` subsamples the pilot, reruns the workflow with a reduced
+  effect grid, and reports the median and interquartile range of the resulting
+  minimum n.
+
+### Versions used for the manuscript
+
+The reported analyses used Python 3.10.14, QIIME 2 2024.10.1 (environment
+`qiime2-metagenome-2024.10`), Gemelli 0.0.12, PhyloFunc 2.0, scikit-bio 0.6.3,
+biom-format 2.1.15, NumPy 1.26.4, pandas 2.2.2, SciPy 1.10.1, Biopython 1.83,
+joblib 1.4.2, Matplotlib 3.8.4, tqdm 4.67.3, and psutil 6.1.0. The exact pins
+are provided in `requirements-lock.txt`, which is shipped with the source
+distribution and the archived release.
+
+## Package installation
+
+From this source directory:
+
+```bash
+python -m pip install .
+phylopower --help
+```
+
+The installed `phylopower` command and `python -m phylopower` both invoke
+`phylopower.cli`.
+
+The tutorial notebook is installed with the package at
+`phylopower/tutorials/tutorial.ipynb`; it is also provided as
+`tutorials/tutorial.ipynb` in the archived release.
+
+
+The protein workflow requires the dependencies declared in
+`pyproject.toml`. The gene workflow additionally requires a QIIME 2
+environment containing compatible versions of Gemelli, `biom-format`, and
+scikit-bio. The manuscript development environment was named
+`qiime2-metagenome-2024.10`. Activate that environment before installing and
+running the gene workflow:
+
+```bash
+conda activate qiime2-metagenome-2024.10
+python -m pip install .
+phylopower gene --target-power 0.80 --out gene_result
+```
+
+## Python API
+
+```python
+from phylopower import (
+    compute_gene_min_sample_size,
+    compute_protein_min_sample_size,
+    demo_path,
+)
+
+result = compute_protein_min_sample_size(
+    table=demo_path("datapro", "taxon_function_table.csv"),
+    tree=demo_path("datapro", "rooted_tree.nwk"),
+    group=demo_path("datapro", "sample_metadata.csv"),
+    target_power=0.80,
+    target_omega2=0.05,
+    out="protein_result",
+)
+```
+
+The compatibility aliases `compute_taxon` and `compute_taxon_function` point
+to these manuscript-aligned functions; they do not call the older workflows
+in `core.py`.
+
+## Input files
+
+### Gene workflow
+
+- feature table: CSV with a `Taxon` column and one column per sample;
+- rooted phylogenetic tree: Newick;
+- taxonomy: CSV containing `Feature ID` and `Taxon`;
+- group map: CSV containing `sample_id` and `group_name`.
+
+### Protein workflow
+
+- Taxon–Function table: CSV with `Taxon`, `Function`, and one column per
+  sample;
+- rooted phylogenetic tree: Newick;
+- group map: CSV containing `sample_id` and `group_name`.
+
+The raw-pool implementation supports two or more groups (the per-group sample
+size recommendation assumes a balanced design with the same n per group).
+
+## Outputs
+
+Each run writes:
+
+- `summary.json`;
+- `power_by_sample_size.csv`;
+- `scenario_metrics_by_sample_size.csv` (per-scenario power with Monte Carlo
+  standard error and 95% Wilson interval columns);
+- `sample_size_decision.png`.
+
+The Python API returns the same summary and two result tables in a dictionary.
+
+## Reproducibility
+
+See `REPRODUCIBILITY.md`.
+
+The bundled datasets are convenient execution examples. Before a public
+Zenodo release, the depositors should verify and document their original
+accessions, citations, and redistribution terms.
+
+## Citation and license
+
+Citation metadata are provided in `CITATION.cff`. PhyloPower
+is distributed under the MIT License; see `LICENSE`.

phylopower-0.1.0/README.md

@@ -0,0 +1,264 @@
+# PhyloPower
+
+[![DOI](https://zenodo.org/badge/DOI/10.5281/zenodo.21991013.svg)](https://doi.org/10.5281/zenodo.21991013)
+![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)
+
+PhyloPower estimates the minimum balanced per-group sample size required for
+community-level meta-omics studies at a target statistical power and realized
+PERMANOVA effect size (ω²).
+
+## For reviewers
+
+The exact code version associated with the manuscript is permanently archived
+at Zenodo: https://doi.org/10.5281/zenodo.21991013. To verify the manuscript's
+claims, start from:
+
+- `tutorials/tutorial.ipynb` — end-to-end
+  walkthrough of both workflows on the bundled demo data;
+- `figures/README.md` — figure-to-script map: every
+  manuscript figure and supplementary figure has exactly one generating
+  script, with precomputed result caches so figures can be regenerated
+  without re-running the full simulations;
+- `REPRODUCIBILITY.md` — environment setup, seeds, and
+  Monte Carlo settings used for the manuscript;
+- `validation_datasets/` — processed external
+  validation cohorts with preparation summaries and checksums.
+
+These companion files are distributed with the Zenodo archive and are not part
+of the PyPI package; `pip install` provides the Python package and the bundled
+demo datasets only.
+
+## Overview
+
+The manuscript-aligned implementation is `phylopower.cli`. It provides
+two raw-pool workflows:
+
+| Command | Synthetic generator | Distance |
+|---|---|---|
+| `gene` | PCAM | Gemelli phylogenetic RPCA |
+| `protein` | MDC-TF-MC | PhyloFunc |
+
+A step-by-step walkthrough of both workflows is provided in the tutorial
+notebook `tutorials/tutorial.ipynb`. Manuscript
+figure scripts and the figure-to-script map are documented in
+`figures/README.md`.
+
+`phylopower/cli.py` is a generated, self-contained runner. It embeds the
+lower-level core and all other project-local runtime modules, so it does not
+import an external `phylopower/core.py`.
+
+## Usage
+
+Run the workflows from this directory with `python -m phylopower`, or install
+the package first (`python -m pip install .`) and use the `phylopower`
+command. The bundled `phylopower/datagene/` and `phylopower/datapro/` demo
+datasets are found automatically; set `PHYLOPOWER_DATA_DIR` or provide input
+paths explicitly to use your own data.
+
+For the protein workflow, install the runtime dependencies with:
+
+```bash
+python -m pip install \
+  "numpy>=1.22" "pandas>=1.5" "scipy>=1.9" "scikit-bio>=0.6.3" \
+  "joblib>=1.2" "matplotlib>=3.7" "tqdm>=4.64" "psutil>=5.9" \
+  phylofunc "biopython>=1.80"
+```
+
+Both pandas 2.x and 3.x are supported; the workflows are verified against
+pandas 3.0 Copy-on-Write semantics (`DataFrame.to_numpy()` read-only views).
+
+The gene workflow must be launched from the compatible QIIME 2 environment
+that contains Gemelli and `biom-format`; selecting that environment only with
+`--qiime-env` is not a substitute for activating it because PCAM calls the
+Gemelli Python API in-process.
+
+```bash
+python -m phylopower --help
+
+python -m phylopower gene \
+  --table /data/feature_table.csv \
+  --tree /data/rooted_tree.nwk \
+  --taxonomy /data/taxonomy.csv \
+  --group /data/sample_metadata.csv \
+  --target-power 0.80 \
+  --out gene_result
+
+python -m phylopower protein \
+  --table /data/taxon_function_table.csv \
+  --tree /data/rooted_tree.nwk \
+  --group /data/sample_metadata.csv \
+  --target-power 0.80 \
+  --out protein_result
+```
+
+A full `protein` run with the default Monte Carlo settings targets
+publication-grade estimates and can take tens of minutes and several GB of
+memory. For a quick smoke test on the bundled demo data (reduced Monte Carlo
+effort, not suitable for publication estimates):
+
+```bash
+python -m phylopower protein \
+  --target-power 0.80 \
+  --min-n 2 --max-n 4 --coarse-step 2 \
+  --boot-number 10 --permutations 49 \
+  --pool-size-per-group 10 \
+  --mdctf-strength-candidates 6 --mdctf-preview-pool-size 60 \
+  --power-preview-boot-number 5 --power-preview-permutations 19 \
+  --out protein_quick
+```
+
+If `--target-omega2` is omitted, the observed ω² of the resolved pilot is used
+and recorded in `summary.json`.
+
+The effect grid is widened automatically when the target ω² exceeds the
+simulated range (`--pcam-scale-extend-max` for `gene`, `--mdctf-strength-max`
+for `protein`). If the target still falls outside the realized ω² range, the
+fitted power at the target would rest on curve extrapolation only: a warning
+is printed, the extrapolated fit is not allowed to qualify, no minimum n is
+reported, and `summary.json` records `"target_omega2_bracketed": false`
+(together with `low_omega_support_warning`). Widen the grid with the flags
+above and rerun in that case.
+
+### Monte Carlo settings
+
+The default settings are `--boot-number 200` bootstrap iterations per
+effect level and `--permutations 199` permutations per PERMANOVA; scikit-bio
+applies the `+1` correction internally, so p-value denominators are 200.
+Publication-grade estimates should use substantially larger counts — see
+`REPRODUCIBILITY.md`. All stochastic steps, including
+the PERMANOVA permutation streams and tree perturbation, are deterministically
+derived from `--random-seed`, so a fixed seed reproduces a run bit-for-bit.
+
+To inspect the exact embedded-module hashes:
+
+```bash
+python phylopower/cli.py --standalone-info
+```
+
+The maintainable build sources are `phylopower/_cli_source.py` and
+`phylopower/_core_source.py`; both are included under `phylopower/`. They
+are used only to regenerate the self-contained runners; the generated
+`phylopower/cli.py` does not read them at runtime. The regeneration script
+(`scripts/build_standalone.py`) is maintained in the GitHub repository.
+
+### Optional controls
+
+- **Parallelism:** `--scenario-n-jobs` sets the number of independent effect
+  scenarios processed concurrently, and `--bootstrap-n-jobs` sets the joblib
+  workers used inside the bootstrap/PERMANOVA loops (`0` = automatic).
+- **Tree perturbation:** `--tree-jitter-sigma` and `--tree-nni-prob` set the
+  perturbation magnitudes; `--tree-support-threshold` restricts NNI to nodes
+  below a given support value; `--tree-set trees.nwk` estimates both
+  magnitudes from a set of bootstrap or posterior trees, with explicit values
+  taking precedence. All tree-perturbation options are off by default.
+- **Pilot stability diagnostic:** `phylopower diagnose --modality
+  {gene,protein} --table ... --tree ... --group ... [--taxonomy ...]
+  --out diagnosis/` subsamples the pilot, reruns the workflow with a reduced
+  effect grid, and reports the median and interquartile range of the resulting
+  minimum n.
+
+### Versions used for the manuscript
+
+The reported analyses used Python 3.10.14, QIIME 2 2024.10.1 (environment
+`qiime2-metagenome-2024.10`), Gemelli 0.0.12, PhyloFunc 2.0, scikit-bio 0.6.3,
+biom-format 2.1.15, NumPy 1.26.4, pandas 2.2.2, SciPy 1.10.1, Biopython 1.83,
+joblib 1.4.2, Matplotlib 3.8.4, tqdm 4.67.3, and psutil 6.1.0. The exact pins
+are provided in `requirements-lock.txt`, which is shipped with the source
+distribution and the archived release.
+
+## Package installation
+
+From this source directory:
+
+```bash
+python -m pip install .
+phylopower --help
+```
+
+The installed `phylopower` command and `python -m phylopower` both invoke
+`phylopower.cli`.
+
+The tutorial notebook is installed with the package at
+`phylopower/tutorials/tutorial.ipynb`; it is also provided as
+`tutorials/tutorial.ipynb` in the archived release.
+
+
+The protein workflow requires the dependencies declared in
+`pyproject.toml`. The gene workflow additionally requires a QIIME 2
+environment containing compatible versions of Gemelli, `biom-format`, and
+scikit-bio. The manuscript development environment was named
+`qiime2-metagenome-2024.10`. Activate that environment before installing and
+running the gene workflow:
+
+```bash
+conda activate qiime2-metagenome-2024.10
+python -m pip install .
+phylopower gene --target-power 0.80 --out gene_result
+```
+
+## Python API
+
+```python
+from phylopower import (
+    compute_gene_min_sample_size,
+    compute_protein_min_sample_size,
+    demo_path,
+)
+
+result = compute_protein_min_sample_size(
+    table=demo_path("datapro", "taxon_function_table.csv"),
+    tree=demo_path("datapro", "rooted_tree.nwk"),
+    group=demo_path("datapro", "sample_metadata.csv"),
+    target_power=0.80,
+    target_omega2=0.05,
+    out="protein_result",
+)
+```
+
+The compatibility aliases `compute_taxon` and `compute_taxon_function` point
+to these manuscript-aligned functions; they do not call the older workflows
+in `core.py`.
+
+## Input files
+
+### Gene workflow
+
+- feature table: CSV with a `Taxon` column and one column per sample;
+- rooted phylogenetic tree: Newick;
+- taxonomy: CSV containing `Feature ID` and `Taxon`;
+- group map: CSV containing `sample_id` and `group_name`.
+
+### Protein workflow
+
+- Taxon–Function table: CSV with `Taxon`, `Function`, and one column per
+  sample;
+- rooted phylogenetic tree: Newick;
+- group map: CSV containing `sample_id` and `group_name`.
+
+The raw-pool implementation supports two or more groups (the per-group sample
+size recommendation assumes a balanced design with the same n per group).
+
+## Outputs
+
+Each run writes:
+
+- `summary.json`;
+- `power_by_sample_size.csv`;
+- `scenario_metrics_by_sample_size.csv` (per-scenario power with Monte Carlo
+  standard error and 95% Wilson interval columns);
+- `sample_size_decision.png`.
+
+The Python API returns the same summary and two result tables in a dictionary.
+
+## Reproducibility
+
+See `REPRODUCIBILITY.md`.
+
+The bundled datasets are convenient execution examples. Before a public
+Zenodo release, the depositors should verify and document their original
+accessions, citations, and redistribution terms.
+
+## Citation and license
+
+Citation metadata are provided in `CITATION.cff`. PhyloPower
+is distributed under the MIT License; see `LICENSE`.