PathwayAge

We present PathwayAge, a biologically-informed, machine learning-based epigenetic clock that integrates pathway-level biological data to predict biological age and assess disease-related risks.

For business reasons, the core algorithm is currently available on a separate GitHub repository. If you are conducting research and would like access, please feel free to reach out via email pan.Li.1122@proton.me.

Repository Structure

• 1-PathwayAge Model/: This directory contains the code for building and evaluating the PathwayAge model using Methylome and Transcriptome data to predict biological age.
• 2-Aging Associated Pathways and Modules/: This directory provides instructions for identifying age-associated GO/KEGG pathways, and modules using a clustering approach.
• 3-Disease Risk Association /: This directory demonstrates that PathwavAge outperforms traditional clocks in predicting biological age acceleration and reveals distinct disease-specific aging patterns in aging-related diseases.
• 4-Disease-Specific pathways/: This directory contains analyses of disease-specific pathways across various conditions ,including sex-stratified differences within each disease.
• 5-reproducibility in Transcriptomics/: This directory contains replication analysis of PathwayAge using transcriptomics, along with validation of the reproducibility of the age-associated methylation pathways.

Installation

1. clone the repo:

git clone git@github.com:BioTransAI/pathwayAge.git

2. environments:

• Make sure the Anaconda or Miniconda is already installed.
  
• Move the file "pyBioTrans.yaml" under the path ~/miniconda3/bin/.
  

  mv pyBioTrans.yml ~/miniconda3/bin/
  cd ~/miniconda3/bin/
  conda env create -f pyBioTrans.yml
  conda activate pyBioTrans

Example Usage:

To predict biological age, please call the "PathwayAge" function.

• Supplying both the training dataset and the testing dataset:

  from pathwayAge import pathwayAge
  
  pathwayAge(
    # import Training data
    methylData = methylTrainData,
    # import testing data
    methylTestData = methylTestData,
    # name the file for the prediction results
    resultName = resultName,
    # Select one machine learning method
    predictionMode = predictionMode,
  )

• Using only the training dataset, the model will automatically perform nested cross validation to prevent data leakage into the testing data labels (Age) in stage1.

  from pathwayAge import pathwayAge
  
  pathwayAge(
    # import Training data
    methylData = methylTrainData,
    # name the file for the prediction results
    resultName = resultName,
    # select one machine learning method
    predictionMode = predictionMode,
  )