Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://kkrismer.github.io/seqgra/
seqgra: Synthetic rule-based biological sequence data generation for architecture evaluation and search
https://kkrismer.github.io/seqgra/
deep-learning interpretable-machine-learning neural-networks pytorch tensorflow
Last synced: 3 months ago
JSON representation
seqgra: Synthetic rule-based biological sequence data generation for architecture evaluation and search
- Host: GitHub
- URL: https://kkrismer.github.io/seqgra/
- Owner: kkrismer
- License: mit
- Created: 2019-09-17T14:26:55.000Z (about 5 years ago)
- Default Branch: master
- Last Pushed: 2022-05-30T18:35:51.000Z (over 2 years ago)
- Last Synced: 2024-07-18T10:01:31.590Z (4 months ago)
- Topics: deep-learning, interpretable-machine-learning, neural-networks, pytorch, tensorflow
- Language: Python
- Homepage: https://kkrismer.github.io/seqgra/
- Size: 8.87 MB
- Stars: 0
- Watchers: 6
- Forks: 2
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
- awesome-dl4g - seqgra - A deep learning pipeline that incorporates the rule-based simulation of biological sequence data and the training and evaluation of models (Software packages / Data wrangling)
README
# seqgra: Principled Selection of Neural Network Architectures for Genomics Prediction Tasks
[](https://opensource.org/licenses/MIT) [](https://doi.org/10.1093/bioinformatics/btac101) [](https://badge.fury.io/py/seqgra) [](https://travis-ci.com/kkrismer/seqgra)
https://kkrismer.github.io/seqgra/
## What is seqgra?
Sequence models based on deep neural networks have achieved state-of-the-art
performance on regulatory genomics prediction tasks, such as chromatin
accessibility and transcription factor binding. But despite their high
accuracy, their contributions to a mechanistic understanding of the biology
of regulatory elements is often hindered by the complexity of the predictive
model and thus poor interpretability of its decision boundaries. To address
this, we introduce seqgra, a deep learning pipeline that incorporates the
rule-based simulation of biological sequence data and the training and
evaluation of models, whose decision boundaries mirror the rules from the
simulation process. The method can be used to (1) generate data under the
assumption of a hypothesized model of genome regulation, (2) identify neural
network architectures capable of recovering the rules of said model, and (3)
analyze a model's predictive performance as a function of training set size,
noise level, and the complexity of the rules behind the simulated data.
## Installation
seqgra is a Python package that is part of [PyPI](https://pypi.org/),
the package repositories behind [pip](https://pip.pypa.io/en/stable/).
To install seqgra with pip, run:
```
pip install seqgra
```
To install seqgra directly from this repository, run:
```
git clone https://github.com/gifford-lab/seqgra
cd seqgra
pip install .
```
### System requirements
- Python 3.7 (or higher)
- *R 3.5 (or higher)*
- *R package `ggplot2` 3.3.0 (or higher)*
- *R package `gridExtra` 2.3 (or higher)*
- *R package `scales` 1.1.0 (or higher)*
The ``tensorflow`` package is only required if TensorFlow models are used
and will not be automatically installed by ``pip install seqgra``. Same is
true for packages ``torch`` and ``pytorch-ignite``, which are only
required if PyTorch models are used.
R is a soft dependency, in the sense that it is used to create a number
of plots (grammar-model-agreement plots,
grammar heatmaps, and motif similarity matrix plots) and if not available,
these plots will be skipped.
seqgra depends upon the Python package [lxml](https://lxml.de/), which in turn
depends on system libraries that are not always present. On a
Debian/Ubuntu machine you can satisfy those requirements using:
```
sudo apt-get install libxml2-dev libxslt-dev
```
## Usage
Check out the following help pages:
* [Usage examples](https://kkrismer.github.io/seqgra/examples.html): seqgra example analyses with data definitions and model definitions
* [Command line utilities](https://kkrismer.github.io/seqgra/cmd.html): argument descriptions for `seqgra`, `seqgras`, `seqgrae`, and `seqgraa` commands
* [Data definition](https://kkrismer.github.io/seqgra/dd.html): detailed description of the data definition language that is used to formalize grammars
* [Model definition](https://kkrismer.github.io/seqgra/md.html): detailed description of the model definition language that is used to describe neural network architectures and hyperparameters for the optimizer, the loss, and the training process
* [Simulators, Learners, Evaluators, Comparators](https://kkrismer.github.io/seqgra/slec.html): brief descriptions of the most important classes
* [seqgra API reference](https://kkrismer.github.io/seqgra/seqgra.html): detailed description of the seqgra API
* [Source code](https://github.com/gifford-lab/seqgra): seqgra source code repository on GitHub
## Citation
If you use seqgra in your work, please cite:
**seqgra: Principled Selection of Neural Network Architectures for Genomics Prediction Tasks**
Konstantin Krismer, Jennifer Hammelman, and David K. Gifford
Bioinformatics, Volume 38, Issue 9, 1 May 2022, Pages 2381–2388; DOI: https://doi.org/10.1093/bioinformatics/btac101
## Funding
We gratefully acknowledge funding from NIH grants 1R01HG008754 and
1R01NS109217.