Introduction

msAI - Software Tools to Create AI Models for Mass Spectrometry Data

Note: msAI is in development. Currently, all releases are development previews. Features are incomplete and may change.

Overview

msAI provides a library of python modules for…

  • Extracting data from mass spectrometry data files

    • Supported MS file types: mzML, TBD…

  • Importing sample metadata

    • Unify MS data and sample metadata in a single view

  • Partitioning data for training, testing, and validation

  • Quality analysis of MS runs / samples

    • Metadata verification

    • Anomaly detection

  • Plotting / visualization

  • AI / machine learning frameworks

    • TensorFlow

    • TBD…

  • Creating AI models

    • Neural network architectures - neuron layers, connections, activation functions

    • Model training - loss and optimization functions, hyperparameter tuning

    • Evaluation of models - measure accuracy, prevent overfitting

  • Model lifecycle management

    • Deploy, update, backup, restore

Installation

Install from the Python Package Index (PyPI):

pip install msAI