A Bayesian semi-parametric model for thermal proteome profiling

Abstract

The thermal stability of proteins can be altered when they interact with small molecules, other biomolecules or are subject to post-translation modifications. Thus monitoring the thermal stability of proteins under various cellular perturbations can provide insights into protein function, as well as potentially determine drug targets and off-targets. Thermal proteome profiling is a highly multiplexed mass-spectrommetry method for monitoring the melting behaviour of thousands of proteins in a single experiment. In essence, thermal proteome profiling assumes that proteins denature upon heating and hence become insoluble. Thus, by tracking the relative solubility of proteins at sequentially increasing temperatures, one can report on the thermal stability of a protein. Standard thermodynamics predicts a sigmoidal relationship between temperature and relative solubility and this is the basis of current robust statistical procedures. However, current methods do not model deviations from this behaviour and they do not quantify uncertainty in the melting profiles. To overcome these challenges, we propose the application of Bayesian functional data analysis tools which allow complex temperature-solubility behaviours. Our methods have improved sensitivity over the state-of-the art, identify new drug-protein associations and have less restrictive assumptions than current approaches. Our methods allows for comprehensive analysis of proteins that deviate from the predicted sigmoid behaviour and we uncover potentially biphasic phenomena with a series of published datasets.

Hierarchical Gaussian process models explore the dark meltome of thermal proteome profiling experiments

“…Two caveats of NPARC addressed by Fang et al [17] However, the main limitation of NPARC remains the sigmoidal assumption. Indeed, as mentioned in several recent works [17][18][19], the use of NPARC limits the investigation of non-sigmoidal melting curves, which are likely to carry important biological information. This is a substantial limitation since it has been estimated [17] that up to 20% of the protein-level TPP datasets can show unconventional behaviours.…”

“…Indeed, as mentioned in several recent works [17][18][19], the use of NPARC limits the investigation of non-sigmoidal melting curves, which are likely to carry important biological information. This is a substantial limitation since it has been estimated [17] that up to 20% of the protein-level TPP datasets can show unconventional behaviours. Moreover, our exploration of the published phospho-peptide-level TPP-TR dataset [11] presented here suggests that about 44% of the phospho-peptides studied show non-sigmoidal behaviour (see Fig 5A).…”

Challenges and Opportunities for Bayesian Statistics in Proteomics

“…Again, these modeling choices can be evaluated using prior/posterior predictive checks. We refer to several discussion on choosing priors for Gaussian processes. ,,,, For applications of the Gaussian process to proteomics data, see Maboudi Afkham et al., Crook et al, Shin et al, and Fang et al…”

“…However, their nested model assumed that the conditional on the Bait cluster and the Prey clusters are independent, and their model assumed exchangeability (permutation leads to the same probability distribution) of the rows and columns. Fang et al 28 proposed a semiparametric model for thermal protein profiling after identifying proteins that deviate from classic sigmoid behavior. Semiparametric models combine interpretable parametric models with more flexible nonparametric models.…”

“…We refer to several discussion on choosing priors for Gaussian processes. 5 , 77 , 25 , 102 , 30 For applications of the Gaussian process to proteomics data, see Maboudi Afkham et al., 60 Crook et al, 21 Shin et al, 92 and Fang et al 28 …”