A Probabilistic Programming Approach to Protein Structure Superposition

TY - GEN

T1 - A Probabilistic Programming Approach to Protein Structure Superposition

AU - Moreta, Lys Sanz

AU - Al-Sibahi, Ahmad Salim

AU - Theobald, Douglas

AU - Bullock, William

AU - Rommes, Basile Nicolas

AU - Manoukian, Andreas

AU - Hamelryck, Thomas

PY - 2019/7

Y1 - 2019/7

N2 - Optimal superposition of protein structures is crucial for understanding their structure, function, dynamics and evolution. We investigate the use of probabilistic programming to superimpose protein structures guided by a Bayesian model. Our model THESEUS-PP is based on the THESEUS model, a probabilistic model of protein superposition based on rotation, translation and perturbation of an underlying, latent mean structure. The model was implemented in the deep probabilistic programming language Pyro. Unlike conventional methods that minimize the sum of the squared distances, THESEUS takes into account correlated atom positions and heteroscedasticity (i.e., atom positions can feature different variances). THESEUS performs maximum likelihood estimation using iterative expectation-maximization. In contrast, THESEUS-PP allows automated maximum a-posteriori (MAP)estimation using suitable priors over rotation, translation, variances and latent mean structure. The results indicate that probabilistic programming is a powerful new paradigm for the formulation of Bayesian probabilistic models concerning biomolecular structure. Specifically, we envision the use of the THESEUS-PP model as a suitable error model or likelihood in Bayesian protein structure prediction using deep probabilistic programming.

AB - Optimal superposition of protein structures is crucial for understanding their structure, function, dynamics and evolution. We investigate the use of probabilistic programming to superimpose protein structures guided by a Bayesian model. Our model THESEUS-PP is based on the THESEUS model, a probabilistic model of protein superposition based on rotation, translation and perturbation of an underlying, latent mean structure. The model was implemented in the deep probabilistic programming language Pyro. Unlike conventional methods that minimize the sum of the squared distances, THESEUS takes into account correlated atom positions and heteroscedasticity (i.e., atom positions can feature different variances). THESEUS performs maximum likelihood estimation using iterative expectation-maximization. In contrast, THESEUS-PP allows automated maximum a-posteriori (MAP)estimation using suitable priors over rotation, translation, variances and latent mean structure. The results indicate that probabilistic programming is a powerful new paradigm for the formulation of Bayesian probabilistic models concerning biomolecular structure. Specifically, we envision the use of the THESEUS-PP model as a suitable error model or likelihood in Bayesian protein structure prediction using deep probabilistic programming.

KW - Bayesian modelling

KW - deep probabilistic programming

KW - protein structure prediction

KW - protein superposition

UR - http://www.scopus.com/inward/record.url?scp=85071465138&partnerID=8YFLogxK

U2 - 10.1109/CIBCB.2019.8791469

DO - 10.1109/CIBCB.2019.8791469

M3 - Article in proceedings

BT - 2019 IEEE Conference on Computational Intelligence in Bioinformatics and Computational Biology, CIBCB 2019

A2 - Baruzzo, Giacomo

A2 - Daberdaku, Sebastian

A2 - Di Camillo, Barbara

A2 - Furini, Simone

A2 - Giordano, Emanuele Domenico

A2 - Nicosia, Giuseppe

PB - IEEE

T2 - 16th IEEE Conference on Computational Intelligence in Bioinformatics and Computational Biology, CIBCB 2019

Y2 - 9 July 2019 through 11 July 2019

ER -