A Python package for processing and analysing the output of proteomics software MaxQuant.

PyMaxQuant is available via the Python package index at PyPi and can be installed in the usual way with:

pip install pymaxquant

Once installed the package is available for import using:

import maxquant

The package is organised into multiple submodules for different purposes, eg.

1. io for reading and writing both MaxQuant and Perseus format files (input/output)
2. filters for filtering data by quality and features
3. process incorporating experimental design, labels to index, expand-side-table (Perseus) and more
4. normalization for performing normalisation methods, e.g. remove column median
5. annotations adding annotation metadata for quantified proteins
6. analysis performing simple analyses, e.g. column correlations
7. plots standard plot outputs for overviews of data

The goal is to provide a simple scripting approach to replicate many of the common steps for interacting with the output of MaxQuant. Many of the steps implemented are based on similar steps used in the MaxQuant sister software Perseus. While currently Perseus has more features, it has stability issues with the larger datasets we are currently using. Having the processing steps implemented in Python allows for simple processing workflow scripts to be created and re-used.

An example Phosphoproteomic label-free-quantification workflow would be as follows:

import maxquant
df = maxquant.io.read_maxquant('Phospho (STY)Sites.txt')

df = maxquant.filter.filter_localization_probability(df)

df = maxquant.filter.remove_reverse(df)
df = maxquant.filter.remove_only_identified_by_site(df)
df = maxquant.filter.remove_potential_contaminants(df)

# Use standard Pandas dataframe manipulations to set an index
df.set_index('Proteins', inplace=True)
df = df.filter(regex='Intensity ')

df = df.process.expand_side_table(df)

# Remove the multiplicity column
df = df.filter(regex='Intensity ')

df = maxquant.process.apply_experimental_design(df, 'experimentalDesignTable.txt')

# The result of this step will be a multilevel index Class, Replicate
# built by matching sample labels using regex
indices = [
    ('Class': '^(.*)_',
    ('Replicate': '_(\d)', 
]
df = maxquant.process.build_index_from_labels(df, indices)

Provided functions are based on our current requirements, but will be expanded in future.

PyMaxQuant is open source software and available under the BSD 2-clause (Simplified) license.