Overview

The maml repository is for the statistical analysis and machine learning of mass spectrometry data. The repository has three main workflows:

1D MS data
Individual 2D imaging data files
Multiple 2D imaging files

NOTE: This documentation is only for existing versions of the repository. It does not yet cover the new MSIUni class. This will follow soon.

1D MS data

See flowMSData.m for a simple analysis workflow.

2D imaging

Files that have been processed by the imzmlProc repository can be analysed using these workflows. All files will need to have annotations and be coregistered to an optical image. For help in these regards, see the documentation of the imzmlProc repository.

Single files

One file with annotated regions (ideally more than one type) can be opened and analysed using the functions outlined in flowMSImage.m. It starts with file loading and normalisation, and covers basic unsupervised learning and supervised classification.

Multiple files

The analysis of single imaging files is useful for checking on the data quality as well as the strength of classification models. To create more generalisable classification models, e.g. for unseen future data, the analysis of annotated regions from multiple files is required.

The workflow in flowMSImageMulti.m can be opened and worked through in consultation with the example provided.

Bugs, errors, etc…

Please document and bugs/errors in the Issues section, providing a minimum working example about how the error is created.

Additions

If there is anything you would like to have included in the repository, then there are two suggested pathways.

Leave a suggestion

For a particular feature, please create a new issue via the Issues page. Provide as much detail as you can.

Do it yourself

If you’d rather code any substantial changes yourself, please fork the repository and work on it yourself. When you’ve finished, please submit a pull request so your additions can be incorporated.