/msiPL

Python Implementation of the msiPL by Abdelmoula et al.

Primary LanguagePython

msiPL

Deep Learning based implementation for analysis of mass spectrometry imaging data

This readme file shows how to properly run the msiPL code

Paper: Walid Abdelmoula et al, msiPL: Non-linear Manifold and Peak Learning of Mass Spectrometry Imaging Data Using Artificial Neural Networks, bioRxiv, 2020

License: The Peak Learning software (msiPL) will be shared using the 3D Slicer Software License agreement.

Installations: Software and Libraries

We have implemented our machine learning model using the following software items:

1- Python(3.6.4)

2- Keras (2.1.5-tf) with a Tensorflow(1.8.0) backend.

3- Packages: numpy(1.14.2), sklearn(0.19.1), scipy(1.0.0), and h5py(2.7.1)

4- We implemented this model on Windows 10 PC workstation(Intel Xenon 3.3GHz, 512 GB RAM, 64-bit Windows, 2 GPUs NVIDIA TITAN Xp).

Demo

  • How to run the code?

    1- "msiPL_Run.py" is the main file that you should run first. The file should be running in a sequential manner, and we have provided required comments for instructions and guidance. In this file you will be able to:

      1.1. Load a dataset.
  1.2. Load the computational neural network architecture (VAE_BN).
  1.3. Train the model.
  1.3. Non-linear manifold learning and data visualization (non-linear dimensionality reduction)
  1.4. Evaluate the learning quality by estimation and reconstruction of the original data
  1.5. Peak Learning learning (Equation#4): to get a smaller list of informative peaks.
  1.6. Perform data clustering (GMM).
  1.7. Identify localized peaks within each cluster.

    2- "Computational_Model.py": implementation of the fully connected variational autoencoder, and regularized with batch normalization.

    3- "LearnPeaks.py": implementation of a function that identifies peaks of interest. It should be called after training the model, as instructed in "msiPL_Run.py".

    1. "msiPL_ForTesting.py": ultra-fast analysis on test data without any prior peak picking. You will need first to load the trained model from step#1 ("msiPL_Run.py").

If you used this implementation:

please cite the paper by Abdelmoula et al, msiPL