A Probabilistic Programming Approach to Protein Structure Superposition

Lys Sanz Moreta; Ahmad Salim Al-Sibahi; Douglas Theobald; William Bullock; Basile Nicolas Rommes; Andreas Manoukian; Thomas Hamelryck

doi:10.1109/CIBCB.2019.8791469

A Probabilistic Programming Approach to Protein Structure Superposition

Lys Sanz Moreta^*, Ahmad Salim Al-Sibahi, Douglas Theobald, William Bullock, Basile Nicolas Rommes, Andreas Manoukian, Thomas Hamelryck

^*Corresponding author for this work

1 Citation (Scopus)

Abstract

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.

Original language	English
Title of host publication	2019 IEEE Conference on Computational Intelligence in Bioinformatics and Computational Biology, CIBCB 2019
Editors	Giacomo Baruzzo, Sebastian Daberdaku, Barbara Di Camillo, Simone Furini, Emanuele Domenico Giordano, Giuseppe Nicosia
Number of pages	5
Publisher	IEEE
Publication date	Jul 2019
Article number	8791469
ISBN (Electronic)	9781728114620
DOIs	https://doi.org/10.1109/CIBCB.2019.8791469
Publication status	Published - Jul 2019
Event	16th IEEE Conference on Computational Intelligence in Bioinformatics and Computational Biology, CIBCB 2019 - Certosa di Pontignano, Siena, Italy Duration: 9 Jul 2019 → 11 Jul 2019

Conference

Conference	16th IEEE Conference on Computational Intelligence in Bioinformatics and Computational Biology, CIBCB 2019
Country/Territory	Italy
City	Certosa di Pontignano, Siena
Period	09/07/2019 → 11/07/2019
Sponsor	GlaxoSmithKline (GSK), IEEE, IEEE Computational Intelligence Society

Keywords

Bayesian modelling
deep probabilistic programming
protein structure prediction
protein superposition

Access to Document

10.1109/CIBCB.2019.8791469

Cite this

Moreta, L. S., Al-Sibahi, A. S., Theobald, D., Bullock, W., Rommes, B. N., Manoukian, A., & Hamelryck, T. (2019). A Probabilistic Programming Approach to Protein Structure Superposition. In G. Baruzzo, S. Daberdaku, B. Di Camillo, S. Furini, E. D. Giordano, & G. Nicosia (Eds.), 2019 IEEE Conference on Computational Intelligence in Bioinformatics and Computational Biology, CIBCB 2019 Article 8791469 IEEE. https://doi.org/10.1109/CIBCB.2019.8791469

A Probabilistic Programming Approach to Protein Structure Superposition. / Moreta, Lys Sanz; Al-Sibahi, Ahmad Salim; Theobald, Douglas et al.
2019 IEEE Conference on Computational Intelligence in Bioinformatics and Computational Biology, CIBCB 2019. ed. / Giacomo Baruzzo; Sebastian Daberdaku; Barbara Di Camillo; Simone Furini; Emanuele Domenico Giordano; Giuseppe Nicosia. IEEE, 2019. 8791469.

Research output: Chapter in Book/Report/Conference proceeding › Article in proceedings › Research › peer-review

Moreta, LS, Al-Sibahi, AS, Theobald, D, Bullock, W, Rommes, BN, Manoukian, A & Hamelryck, T 2019, A Probabilistic Programming Approach to Protein Structure Superposition. in G Baruzzo, S Daberdaku, B Di Camillo, S Furini, ED Giordano & G Nicosia (eds), 2019 IEEE Conference on Computational Intelligence in Bioinformatics and Computational Biology, CIBCB 2019., 8791469, IEEE, 16th IEEE Conference on Computational Intelligence in Bioinformatics and Computational Biology, CIBCB 2019, Certosa di Pontignano, Siena, Italy, 09/07/2019. https://doi.org/10.1109/CIBCB.2019.8791469

Moreta LS, Al-Sibahi AS, Theobald D, Bullock W, Rommes BN, Manoukian A et al. A Probabilistic Programming Approach to Protein Structure Superposition. In Baruzzo G, Daberdaku S, Di Camillo B, Furini S, Giordano ED, Nicosia G, editors, 2019 IEEE Conference on Computational Intelligence in Bioinformatics and Computational Biology, CIBCB 2019. IEEE. 2019. 8791469 doi: 10.1109/CIBCB.2019.8791469

Moreta, Lys Sanz ; Al-Sibahi, Ahmad Salim ; Theobald, Douglas et al. / A Probabilistic Programming Approach to Protein Structure Superposition. 2019 IEEE Conference on Computational Intelligence in Bioinformatics and Computational Biology, CIBCB 2019. editor / Giacomo Baruzzo ; Sebastian Daberdaku ; Barbara Di Camillo ; Simone Furini ; Emanuele Domenico Giordano ; Giuseppe Nicosia. IEEE, 2019.

@inproceedings{6a8fd548b91745fba285f5466294dddf,

title = "A Probabilistic Programming Approach to Protein Structure Superposition",

abstract = "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.",

keywords = "Bayesian modelling, deep probabilistic programming, protein structure prediction, protein superposition",

author = "Moreta, {Lys Sanz} and Al-Sibahi, {Ahmad Salim} and Douglas Theobald and William Bullock and Rommes, {Basile Nicolas} and Andreas Manoukian and Thomas Hamelryck",

year = "2019",

month = jul,

doi = "10.1109/CIBCB.2019.8791469",

language = "English",

editor = "Giacomo Baruzzo and Sebastian Daberdaku and {Di Camillo}, Barbara and Simone Furini and Giordano, {Emanuele Domenico} and Giuseppe Nicosia",

booktitle = "2019 IEEE Conference on Computational Intelligence in Bioinformatics and Computational Biology, CIBCB 2019",

publisher = "IEEE",

note = "16th IEEE Conference on Computational Intelligence in Bioinformatics and Computational Biology, CIBCB 2019 ; Conference date: 09-07-2019 Through 11-07-2019",

}

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 -

A Probabilistic Programming Approach to Protein Structure Superposition

Abstract

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this