Limited proteolysis combined with mass spectrometry (LiP-MS) facilitates probing structural changes on a proteome-wide scale. This method leverages differences in the proteinase K accessibility of native protein structures to concurrently assess structural alterations for thousands of proteins in situ. Distinguishing different contributions to the LiP-MS signal, such as changes in protein abundance or chemical modifications, from structural protein alterations remains challenging. Here, we present the first comprehensive computational pipeline to infer structural alterations for LiP-MS data using a two-step approach. (1) We remove unwanted variations from the LiP signal that are not caused by protein structural effects and (2) infer the effects of variables of interest on the remaining signal. Using LiP-MS data from three species we demonstrate that this approach outperforms previously employed approaches. Our framework provides a uniquely powerful approach for deconvolving LiP-MS signals and separating protein structural changes from changes in protein abundance, post-translational modifications and alternative splicing. Our approach may also be applied to analyse other types of peptide-centric structural proteomics data, such as FPOP or molecular painting data.