permutation tests #43
Conversation
- saved processed dataset to example_data - removed data cleaning code from notebook - added descriptions and permutation explanation in markdown
enryH
left a comment
enryH
left a comment
There was a problem hiding this comment.
I will test it now.
| @@ -0,0 +1,53 @@ | |||
| # preprocessing | |||
There was a problem hiding this comment.
I would rename this to something similar to compositional.py - I would like to not have it specific to microbiome at best, but we can group it here for now. Maybe you could move the file as compositional.py to transform/compositional.py?
| @@ -0,0 +1,25 @@ | |||
| eff_id,inf_id,sampling_location,sampling_read,eff_abundance,inf_abundance | |||
There was a problem hiding this comment.
could you add a README.md to the folder briefly putting a link where this originates from? And maybe if it is a subset, a small preprocessing script?
There was a problem hiding this comment.
Pull Request Overview
This PR adds comprehensive permutation testing functionality to the acore library, including implementations for paired and independent sample tests, chi-squared tests, and Jaccard similarity calculations. The changes also include documentation, examples, and minor code quality improvements.
- Implements three main permutation test functions:
paired_permutation,indep_permutation, andchi2_permutation - Adds Jaccard similarity calculations for graph comparison
- Includes comprehensive test suite and tutorial documentation with real-world metagenomics example
Reviewed Changes
Copilot reviewed 15 out of 15 changed files in this pull request and generated 17 comments.
Show a summary per file
| File | Description |
|---|---|
| src/acore/permutation_test/init.py | Core permutation test implementations with support for multiple metrics |
| src/acore/permutation_test/internal_functions.py | Helper functions for permutation and contingency table operations |
| src/acore/permutation_test/jaccard.py | Jaccard similarity calculation functions for set comparison |
| src/acore/types/permutation_test.py | Pydantic model for permutation test results |
| tests/test_permutation_test.py | Comprehensive test suite for permutation test functions |
| docs/api_examples/permutation_testing.py | Tutorial demonstrating permutation testing on metagenomics data |
| docs/api_examples/permutation_testing.ipynb | Jupyter notebook version of the tutorial |
| docs/index.rst | Updated documentation index to include permutation testing tutorial |
| src/acore/microbiome/internal_functions.py | New CLR transformation functions for compositional data |
| example_data/mgnify/Ju2018_GO0017001_enf_inf_paired.csv | Example dataset for tutorial |
| src/acore/multiple_testing/init.py | Code style fix: changed membership test to use not in operator |
| src/acore/enrichment_analysis/init.py | Code style fix: changed membership tests to use not in operator |
| src/acore/decomposition/umap.py | Added blank line for formatting consistency |
| src/acore/io/uniprot/uniprot.py | Removed extra blank lines |
| CONTRIBUTING.rst | Fixed installation command quoting and corrected PR template path |
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
| # ## Data preparation details | ||
| # | ||
| # ### Downloading | ||
| # The analysed samples were downloaded via the [MGnify API](https://www.ebi.ac.uk/metagenomics/api/docs/). The inffluent (INF) and effluent (EFFF) datasets have paired samples and we also needed to download the sample metadata (also available via Mgnify API) to assign the correct pairing. |
There was a problem hiding this comment.
Multiple spelling errors: "inffluent" should be "influent" and "EFFF" should be "EFF".
| "## Data preparation details\n", | ||
| "\n", | ||
| "### Downloading\n", | ||
| "The analysed samples were downloaded via the [MGnify API](https://www.ebi.ac.uk/metagenomics/api/docs/). The inffluent (INF) and effluent (EFFF) datasets have paired samples and we also needed to download the sample metadata (also available via Mgnify API) to assign the correct pairing.\n", |
There was a problem hiding this comment.
Multiple spelling errors: "inffluent" should be "influent" and "EFFF" should be "EFF".
| def random_choice_equal(): | ||
| return np.random.choice( | ||
| ["A", "B", "C", "D"], size=100, replace=True, p=[0.25, 0.25, 0.25, 0.25] | ||
| ) | ||
|
|
||
|
|
||
| @pytest.fixture | ||
| def random_choice_unequal(): | ||
| return np.random.choice( | ||
| ["A", "B", "C", "D"], size=100, replace=True, p=[0.1, 0.25, 0.05, 0.6] | ||
| ) |
There was a problem hiding this comment.
Test fixtures that use random number generation without a fixed seed can lead to non-reproducible test failures. Consider using np.random.default_rng(seed=...) or np.random.seed() to ensure reproducibility.
| from acore import permutation_test as pt | ||
| import pytest | ||
| import numpy as np | ||
| from scipy.stats import ttest_rel, ttest_ind, mannwhitneyu, wilcoxon |
There was a problem hiding this comment.
Import of 'ttest_ind' is not used.
Import of 'mannwhitneyu' is not used.
Import of 'wilcoxon' is not used.
| from scipy.stats import ttest_rel, ttest_ind, mannwhitneyu, wilcoxon | |
| from scipy.stats import ttest_rel |
|
@angelphanth Definitely 'squash and merge` this PR. |
Co-authored-by: Henry Webel <heweb@dtu.dk>
Co-authored-by: Henry Webel <heweb@dtu.dk>
mainly typos Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
|
@angelphanth I merge main, where some documentation files were updated (see #47) |
Description
Add a subpackage (as new model) to core.
Tasks Checklist
__init__.pyin the new folder, so thatthey are available when the subpackage is imported.
Create Pandera schema in a file with subpackage name in thesrc/acore/typesfolder.Optimal is to have only one output schema of results per subpackage or module.
datafolder, or reuse an existing one for testingdocs/api_examplesfolder with that dataindex.rstfile in thedocsfolder with the new example/testfolder with the name of the subpackage or moduleusing pytest or unittests to test your new functionality.