Organ Age Predictor

An organ age prediction tool based on proteomic data.

The predictor uses Olink NPX proteomic measurements together with sample metadata to generate predictions.

Repository Contents

├── README.md
├── tutorial.ipynb
├── pre.cp39-win_amd64.pyd 
└── model/ 
    └── model.bin (an empty file)

The real model file are distributed separately through the GitHub Releases page.

Download the following file from the latest release:

model.bin

After downloading, organize the files as follows:

your_project/
├── tutorial.ipynb
├── pre.cp39-win_amd64.pyd
├── metadata.csv
├── protein.csv
└── model/
    └── model.bin

System Requirements

The provided compiled module:

pre.cp39-win_amd64.pyd

is built for:

• Windows 64-bit
• CPython 3.9

Python 3.9 is therefore recommended when using this release.

Install the required Python packages:

pip install pandas numpy cryptography

Input Data

Two input files are required:

metadata.csv
protein.csv

1. Metadata

metadata.csv should contain sample metadata.

• Sample IDs should be used as the row index.
• Chronological age should be provided in a column named Age.
• Age values should be numeric.

Example:

Sample ID	Age
Sample_001	52
Sample_002	61
Sample_003	47

2. Proteomic Data

protein.csv should contain the proteomic measurements.

• Sample IDs should be used as the row index.
• Protein names should be used as column names.
• Values should correspond to Olink NPX measurements.

Example:

Sample ID	Protein_A	Protein_B	Protein_C
Sample_001	8.21	4.72	6.15
Sample_002	7.95	4.91	5.88
Sample_003	8.43	5.02	6.34

The sample IDs in the metadata and proteomic data should correspond to the same samples.

Notes

Python Version

The provided .pyd file is compiled specifically for CPython 3.9 on 64-bit Windows.

Using another Python version, such as Python 3.10, 3.11, or 3.12, may result in an import error.

You can check your Python version with:

python --version

A compatible environment should report Python 3.9.x.

Platform Compatibility

The current binary release is intended for Windows 64-bit systems.

The provided .pyd file cannot be imported directly on Linux or macOS.