This directory contains all of the tutorials for selene.
Additionally, we have two tutorials that show how to apply trained models. Selene provides methods to run variant effect prediction and in silico mutagenesis, along with some visualization methods that we recommend running based on our Jupyter notebook tutorials.
- Comprehensive in silico mutagenesis tutorial:
- Tutorial with both the config file method and the non-config file method of running Selene. Also shows how to run variant effect prediction and visualize the difference scores. Contains an in silico mutagenesis example with known regulatory mutations:
We also have a tutorial demonstrating Selene's use to predict mean ribosomal load based on 5' UTR sequences:
regression_mpra_example. This is a good follow-up tutorial to the Getting Started tutorial if you are interested in training a regression model using Selene. It also shows how to run Selene with another model architecture.
The log statements printed in the training tutorials are from running the tutorials on a CUDA-enabled machine. Run times will be far longer (as described in the repository README) if you are running them only on CPU.
The process for adding a tutorial to selene is as follows:
- Create a subdirectory in the tutorials directory. The name of this subdirectory should be the name of the tutorial, formatted in snake-case.
- Write the tutorial in an ipython notebook in the subdirectory.
- Store all data for the tutorial in the subdirectory, and create a gzipped archive (i.e. a
*.tar.gzfile) with all the data required for the tutorial.
- Create a
*.nblinklink file in the
docs/source/tutorialsdirectory. This file will serve as a link to the tutorial's notebook file. Instructions for formatting this file can be found here.
- Add an entry for the tutorial to the list of tutorials in
make htmlfrom the