zwbao/GAI_Paper

GAI Calculator

This is a Python script for calculating Gut Aging Index (GAI) based on OTU (Operational Taxonomic Units) data.

The datasets used in paper can be downloaded with download_data.sh BASH script.

Prerequisites

• Python 3.x
• pandas
• pycaret 2.3.5

Installation

1. Clone the repository:

git clone https://github.com/zwbao/GAI_paper.git

2. Install the required Python packages:

pip install pandas
pip install pycaret==2.3.5

Usage

1. Prepare the input files:

   • meta.tsv: The metadata file containing demographic information, with a column named 'id' as the index column.
   • otu.tsv: The OTU file containing operational taxonomic unit data, with a column named 'id' as the index column.

2. Run the script:

python gai_cal.py meta.tsv otu.tsv

3. Output files:
   • compare_models.tsv: A file containing the results of comparing different regression models.
   • tuned_best_model.tsv: A file containing the results of tuning the best model.
   • final_best_model_<date>.pkl: A file containing the finalized best model, where <date> represents the current date in the format YYYYMMDD.
   • adjust_values.tsv: A file containing average raw GAI values for different age ranges.
   • result.tsv: A file containing the final results, including age, raw GAI, adjusted value, and corrected GAI.

Using Pre-trained Models

To use the pre-trained models from the 'models' folder, follow these steps:

1. Load a model using PyCaret:

   from pycaret.classification import load_model
   model = load_model('path/to/model')

2. Prepare your data as specified in the meta.tsv and otu.tsv files, ensuring it matches the format the model expects.

3. Predict GAI using the loaded model:

   predictions = predict_model(model, data=unseen_data)

This allows for quick application of our research findings and replication of results using established models.

Contact

If you have any questions or issues, please feel free to contact me at zwbao1996@zju.edu.cn.