GONAD-ResNet

Table of Contents

• Description
• Getting started
  • Installing
  • Executing program
• Contributing
• License
• Authors
• Contact

---

Description

GONAD-ResNet is a neural network that identifies novel DSD-related genes by analyzing time-dependent single-cell gene expression data. It detects patterns similar to known DSD genes while filtering out irrelevant ones, narrowing thousands of candidates to a few. This accelerates gene discovery, improving diagnosis and research. we introduces “GONAD-ResNet,” a residual convolutional neural network designed to predict novel DSD-associated genes by learning complex patterns in time-dependent single-cell gene expression data. When applied to WES data from six patients (three XX, three XY) with DSD, GONAD-ResNet prioritized genes with expression patterns similar to known DSD genes while disregarding ubiquitous or irrelevant genes. This narrowed the list of potential candidates from around 1,000 to a few promising novel genes per patient. This innovative approach accelerates the discovery of new DSD-related genes, opening new research avenues and potentially improving patient outcomes.

Getting Started

Installing

conda

To run this pipeline, you need to have Conda installed on your machine. If you already have Conda installed, you can skip this step. Otherwise, you can install it by following the steps below:

1. Download:

   mkdir -p ~/miniconda3
   wget https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh -O ~/miniconda3/miniconda.sh

2. Run installer and initiate:

   bash ~/miniconda3/miniconda.sh -b -u -p ~/miniconda3
   rm -rf ~/miniconda3/miniconda.sh
   ~/miniconda3/bin/conda init bash
   ~/miniconda3/bin/conda init zsh

3. Refresh shell configuration:

   source ~/.bashrc

4. verify the installation:

   conda --version

Get Access to the repository:

1. Download the repository:

   Download the correspondent repository

   git clone https://github.com/hAzimi/GONAD-ResNet.git

   git checkout main

2. Recreate the environment locally:

   Once you have conda installed create the GONAD-ResNet environment by running:

   cd GONAD-ResNet
   conda env create -f environment.yml

   Verify that the environment is set up by doing:

   conda env list

   Then activate the GONAD-ResNet enviroment:

   conda activate GONAD-ResNet

Executing program

This repository contains a Streamlit web application that runs the gonad_resnet_webapp.py script.

To start the Streamlit web app, run the following command:

python -m streamlit run gonad_resnet_webapp.py

This will launch the application in your default web browser.

Dataset (required)

This project expects two AnnData .h5ad files (one per sex):

• Male dataset: download from Human developing gonads and save as human_main_male.h5ad
• Female dataset: download from Human developing gonads and save as human_main_female.h5ad

Place both files in a local folder of your choice (e.g. data/).

Update the dataset paths

Open create_inputs.py and update the paths inside data_loader() to point to where you saved the files:

# create_inputs.py
adataSC = anndata.read_h5ad('path_to_human_main_male.h5ad', backed=None)
adataSC = anndata.read_h5ad('path_to_human_main_female.h5ad', backed=None)

Contributing

We welcome contributions to improve this repository! If you would like to contribute, please follow these steps:

1. Fork the repository: Click the "Fork" button on the top right of the repository page.
2. Clone your fork:

   git clone https://github.com/hAzimi/GONAD-ResNet.git
   cd GONAD-ResNet

3. Create a new branch for your changes:

   git checkout -b feature-or-fix-name

4. Make your changes and commit:

   git add .
   git commit -m "Brief description of changes"

5. Push your branch to GitHub:

   git push origin feature-or-fix-name

6. Submit a Pull Request (PR): Open a PR against the main branch with a clear description of your changes.

Contribution Guidelines

• Ensure your code follows best practices and is well-documented.
• If modifying scripts, ensure that existing functionality remains intact.
• If your contribution is related to deep learning models or data analysis, please describe the changes and provide relevant results or comparisons.

For major changes, please open an issue first to discuss your ideas before implementation.

License

This project is licensed under the GNU General Public License v3.0 (GPLv3).

Authors

• Developed by Hamid Azimi
• Tested by Isabel Von Der DECKEN

Contact

For any issues or inquiries, please open an issue or contact

• Hamid Azimi
• Isabel Von Der DECKEN