Abstract
We present an update of the MaxQuant software for isobaric labeling data and evaluate its performance on benchmark data sets. Impurity correction factors can be applied to labels mixing C- and N-type reporter ions such as TMT Pro. Application to a single-cell multispecies mixture benchmark shows the high accuracy of the impurity-corrected results. TMT data recorded with FAIMS separation can be analyzed directly in MaxQuant without splitting the raw data into separate files per FAIMS voltage. Weighted median normalization is applied to several data sets, including large-scale human body atlas data. In the benchmark data sets, the weighted median normalization either removes or strongly reduces the batch effects between different TMT plexes and results in clustering by biology. In data sets including reference channels, we find that weighted median normalization performs as well or better when the reference channels are ignored and only the sample channel intensities are used, suggesting that the measurement of reference channels is unnecessary when using weighted median normalization in MaxQuant. We demonstrate that MaxQuant including the weighted median normalization performs well on multinotch MS3 data, as well as on phosphorylation data. MaxQuant is freely available for any purpose and can be downloaded from https://www.maxquant.org/.