Latest news

v0.9.1 is released with a lot of improvements!

Please checkout the tutorials and documentations at 📄cma2015.github.io/DEM.

• The DEM is implemented in the Python package biodem, which comprises 4 modules: data preprocessing, dual-extraction modeling, phenotypic prediction, and functional gene mining.
• For more details, please check out our publication. 🖱️Click to copy citation

Installation

System requirements

• Python 3.10 / 3.11 / 3.12.
• Optional: Hardware accelerator supporting PyTorch.

Recommended: NVIDIA graphics card with 12GB memory or larger.

Install biodem

Conda / Mamba is recommended for installation.

1. Create a conda environment:

   mamba create -n dem python=3.11
   mamba activate dem
   
   # Install PyTorch with CUDA support
   mamba install pytorch torchvision torchaudio pytorch-cuda=12.4 -c pytorch -c nvidia
   

2. Install biodem package from PyPI

   pip install biodem
   

Usage

Please checkout the documentations at 📄cma2015.github.io/DEM.

biodem comprises 4 functional modules:

1. Data preprocessing

Nested cross-validation is recommended for data preprocessing.

• Steps:
  1. Split data into nested cross-validation sets.
  2. Imputation & standardization.
  3. Feature selection using the variance threshold filter and Random Forests.
  4. SNP2Gene transformation.

2. Dual-extraction modeling

• It takes preprocessed multi-omics data and phenotypic data as inputs. DEM is capable of performing both classification and regression tasks.

3. Phenotypic prediction

• It loads the trained DEM model checkpoint and performs phenotypic prediction.

4. Functional gene mining

• It performs functional gene mining based on the trained DEM model through feature ranking by permutation.

Citation

Please cite our paper if you use this package:

@article{renDualextractionModelingMultimodal2024a,
  title = {Dual-Extraction Modeling: {{A}} Multi-Modal Deep-Learning Architecture for Phenotypic Prediction and Functional Gene Mining of Complex Traits},
  shorttitle = {Dual-Extraction Modeling},
  author = {Ren, Yanlin and Wu, Chenhua and Zhou, He and Hu, Xiaona and Miao, Zhenyan},
  year = {2024},
  month = sep,
  journal = {Plant Communications},
  volume = {5},
  number = {9},
  pages = {101002},
  issn = {25903462},
  doi = {10.1016/j.xplc.2024.101002},
  langid = {english}
}

Asking for help

If you have any questions, please contact us via GitHub issues or email us.