Optimal transport for time-course single cell data
Waddington-OT uses time-course data to infer how the probability distribution of cells in gene-expression space evolves over time, by using the mathematical approach of Optimal Transport (OT).
Waddington-OT depends on Python 3.
You can install all dependencies for wot with conda :
conda install cython h5py flask gunicorn numexpr numpy pandas scikit-learn scipy simplejson psutil conda install -c conda-forge pot
Or with pip :
pip install --user h5py docutils msgpack-python pip install --user cython pip install --user flask gunicorn numexpr numpy pandas scikit-learn scipy psutil pot
Install the wot package
pip install --user wot
Initializing an OT Model
wot uses an OTModel as its interface for computing transport maps.
You can initialize an OT Model in python with :
ot_model = wot.initialize_ot_model('matrix.txt', 'days.txt')
All Optimal Transport parameters can be customized when initializing the model. For instance, you could explicitely specify the defaults :
ot_model = wot.initialize_ot_model('matrix.txt', 'days.txt', tmap_prefix='tmaps', epsilon=.05, lambda1=10, lambda2=50, batch_size=50, tolerance=1e-2)
You can compute all transport maps with :
Loading Transport Maps
Once the transport maps have been created, you can operate on the transport maps using the TransportMapModel interface :
tmap_model = wot.model.TransportMapModel.from_directory('.')
All previously computed transport maps will be available.
If you want to keep the previously computed transport maps, simply initialize
a new model with a different prefix. Any model will only affect files that use
tmap_prefix, there is no interaction between models with different prefixes.
All data-processing functions are located in the
These include :
- ancestor census
- gene set scores
- gene regulatory networks (grn)
- local enrichment
- optimal transport validation
- trajectory trends
- convert matrix
- force layout
- wot server (interactive version of wot)
All of these are documented on wot's github pages website, with examples using simulated data to show how to use and plot the results of these commands.
The full documentation for wot is available on Github Pages : http://broadinstitute.github.io/wot
For more advanced usage, you may also browse the source code to read each function's documentation. Most of wot's internal functions have docstrings with a description of their parameters, output and examples on how to use them.
For more information about the internal functionning of wot, please refer to the Developer Notes
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
|Filename, size & hash SHA256 hash help||File type||Python version||Upload date|
|wot-0.3.5-py2.py3-none-any.whl (10.1 MB) Copy SHA256 hash SHA256||Wheel||py2.py3||Oct 17, 2018|
|wot-0.3.5.tar.gz (8.4 MB) Copy SHA256 hash SHA256||Source||None||Oct 17, 2018|