Pre- and Post-Processing Workflow for Affinity Purification Mass Spectrometry Data

The
reliable detection of protein–protein interactions by
affinity purification mass spectrometry (AP-MS) is crucial for the
understanding of biological processes. Quantitative information can
be used to separate truly interacting proteins from false-positives
by contrasting counts of proteins binding to specific baits with counts
of negative controls. Several approaches have been proposed for computing
scores for potential interaction proteins, for example, the commonly
used SAINT software. However, it remains a subjective decision where
to set the cutoff score for candidate selection; furthermore, no precise
control for the expected number of false-positives is provided. In
related fields, successful data analysis strongly relies on statistical
pre- and post-processing steps, which, so far, have played only a
minor role in AP-MS data analysis. We introduce a complete workflow,
embedding either the scoring method SAINT or alternatively a two-stage
Poisson model into a pre- and post-processing framework. To this end,
we investigate different normalization methods and apply a statistical
filter adjusted to AP-MS data. Furthermore, we propose permutation
and adjustment procedures, which allow the replacement of scores by
statistical p values. The performance of the workflow
is assessed on simulations as well as on a study focusing on interactions
with the T3SS in Salmonella Typhimurium. Preprocessing
methods significantly increase the number of detected truly interacting
proteins, while a constant false-discovery rate is maintained. The
software solution is freely available.