Fully cumstomizable robust Independent Components Analysis (ICA)
Project description
robustica
Fully customizable robust Independent Component Analysis (ICA).
Description
This package contains 3 modules:
-
RobustICA
Defines the most important class that allows to perform and customize robust independent component analysis.
-
InferComponents
Retrieves the number of components that explain a user-defined percentage of variance.
-
examples
Contains handy functions to quickly create or access example datasets.
A more user-friendly documentation can be found at https://crg-cnag.github.io/robustica/.
Requirements
In brackets, versions of packages used to develop robustica
.
numpy
(1.19.2)pandas
(1.1.2)scipy
(1.6.2)scikit-learn
(0.23.2)joblib
(1.0.1)tqdm
(4.59.0)- (optional)
scikit-learn-extra
(0.2.0): required only for clustering algorithms KMedoids and CommonNNClustering
Installation
[optional] scikit-learn-extra
incompatibility
To use the clustering algorithms KMedoids and CommonNNClustering, install a forked version first to avoid incompatibility with the newest numpy
(see #6 for more info on this).
pip install git+https://github.com/TimotheeMathieu/scikit-learn-extra
pip
pip install robustica
local (latest version)
git clone https://github.com/CRG-CNAG/robustica
cd robustica
pip install -e .
Usage
from robustica import RobustICA
from robustica.examples import make_sampledata
X = make_sampledata(ncol=300, nrow=2000, seed=123)
rica = RobustICA(n_components=10)
# note that by default, we use DBSCAN algorithm and the number of components can be smaller
# than the number of components defined.
S, A = rica.fit_transform(X)
# source matrix (nrow x n_components)
print(S.shape)
print(S)
(2000, 3)
[[ 0.00975714 0.00619138 0.00502649]
[-0.0021527 -0.0376857 0.0117938 ]
[ 0.00046302 0.01712561 0.00518039]
...
[ 0.00128344 -0.00767099 0.0047334 ]
[ 0.00644422 -0.00498327 0.01325542]
[ 0.0017873 -0.01739889 -0.00445954]]
# mixing matrix (ncol x n_components)
print(A.shape)
print(A)
(300, 3)
[[-1.79503194e-02 -1.05611924e+00 5.36688700e-01]
[ 1.03342514e-01 7.43471382e-02 4.90472157e-01]
[ 4.89753256e-01 -1.11300905e+00 -7.55809647e-01]
...
[ 4.30468472e-01 -4.87992838e-01 -7.77965512e-01]
[ 3.44078031e-02 4.09029805e-01 -7.29076312e-01]
[ 2.15557427e-02 2.89301273e-01 -2.96690459e-01]]
Tutorials
- Basic pipeline for exploratory analysis
- Using a custom clustering class
- Inferring the number of components
Contact
This project has been fully developed at the Centre for Genomic Regulation within the group of Design of Biological Systems
Please, report any issues that you experience through this repository's "Issues" or email:
License
robustica
is distributed under a BSD 3-Clause License (see LICENSE).
Citation
Anglada-Girotto, M., Miravet-Verde, S., Serrano, L., Head, S. A.. "robustica: customizable robust independent component analysis". BMC Bioinformatics 23, 519 (2022). DOI: https://doi.org/10.1186/s12859-022-05043-9
References
- Himberg, J., & Hyvarinen, A. "Icasso: software for investigating the reliability of ICA estimates by clustering and visualization". IEEE XIII Workshop on Neural Networks for Signal Processing (2003). DOI: https://doi.org/10.1109/NNSP.2003.1318025
- Sastry, Anand V., et al. "The Escherichia coli transcriptome mostly consists of independently regulated modules." Nature communications 10.1 (2019): 1-14. DOI: https://doi.org/10.1038/s41467-019-13483-w
- Kairov, U., Cantini, L., Greco, A. et al. Determining the optimal number of independent components for reproducible transcriptomic data analysis. BMC Genomics 18, 712 (2017). DOI: https://doi.org/10.1186/s12864-017-4112-9
Project details
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
File details
Details for the file robustica-0.1.4.tar.gz
.
File metadata
- Download URL: robustica-0.1.4.tar.gz
- Upload date:
- Size: 16.3 kB
- Tags: Source
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/5.1.1 CPython/3.10.8
File hashes
Algorithm | Hash digest | |
---|---|---|
SHA256 | 2ec0a10d8815a016c8319ffe4b460914044efd47cf4186e4ded2d3b96ca91aa9 |
|
MD5 | a0f3f8e8c782a713ae662f4640046c05 |
|
BLAKE2b-256 | 87174f0c142525f655729e5710b2763d8306a64b78efaa4488eab5750f45d18d |