Inferring differential subcellular localisation in comparative spatial proteomics using BANDLE

Oliver M. Crook,Colin T. R. Davies,Lisa M. Breckels,Josie A. Christopher,Laurent Gatto,Paul D. W. Kirk,Kathryn S. Lilley

DOI: https://doi.org/10.1038/s41467-022-33570-9

IF: 16.6

2022-10-10

Nature Communications

Abstract:Abstract The steady-state localisation of proteins provides vital insight into their function. These localisations are context specific with proteins translocating between different subcellular niches upon perturbation of the subcellular environment. Differential localisation, that is a change in the steady-state subcellular location of a protein, provides a step towards mechanistic insight of subcellular protein dynamics. High-accuracy high-throughput mass spectrometry-based methods now exist to map the steady-state localisation and re-localisation of proteins. Here, we describe a principled Bayesian approach, BANDLE, that uses these data to compute the probability that a protein differentially localises upon cellular perturbation. Extensive simulation studies demonstrate that BANDLE reduces the number of both type I and type II errors compared to existing approaches. Application of BANDLE to several datasets recovers well-studied translocations. In an application to cytomegalovirus infection, we obtain insights into the rewiring of the host proteome. Integration of other high-throughput datasets allows us to provide the functional context of these data.

multidisciplinary sciences