Efficient Generative Modelling of Protein Structure Fragments using a Deep Markov Model

Research output: Chapter in Book/Report/Conference proceedingArticle in proceedingsResearchpeer-review

Standard

Efficient Generative Modelling of Protein Structure Fragments using a Deep Markov Model. / Thygesen, Christian B.; Al-Sibahi, Ahmad Salim; Steenmanns, Christian S.; Sanz Moreta, Lys; Sørensen, Anders B.; Hamelryck, Thomas Wim.

International Conference on Machine Learning, 18-24 July 2021, Virtual. PMLR, 2021. p. 10258-10267 (Proceedings of Machine Learning Research, Vol. 139).

Research output: Chapter in Book/Report/Conference proceedingArticle in proceedingsResearchpeer-review

Harvard

Thygesen, CB, Al-Sibahi, AS, Steenmanns, CS, Sanz Moreta, L, Sørensen, AB & Hamelryck, TW 2021, Efficient Generative Modelling of Protein Structure Fragments using a Deep Markov Model. in International Conference on Machine Learning, 18-24 July 2021, Virtual. PMLR, Proceedings of Machine Learning Research, vol. 139, pp. 10258-10267, 38th International Conference on Machine Learning, 18/07/2021. <https://proceedings.mlr.press/v139/thygesen21a.html>

APA

Thygesen, C. B., Al-Sibahi, A. S., Steenmanns, C. S., Sanz Moreta, L., Sørensen, A. B., & Hamelryck, T. W. (2021). Efficient Generative Modelling of Protein Structure Fragments using a Deep Markov Model. In International Conference on Machine Learning, 18-24 July 2021, Virtual (pp. 10258-10267). PMLR. Proceedings of Machine Learning Research Vol. 139 https://proceedings.mlr.press/v139/thygesen21a.html

Vancouver

Thygesen CB, Al-Sibahi AS, Steenmanns CS, Sanz Moreta L, Sørensen AB, Hamelryck TW. Efficient Generative Modelling of Protein Structure Fragments using a Deep Markov Model. In International Conference on Machine Learning, 18-24 July 2021, Virtual. PMLR. 2021. p. 10258-10267. (Proceedings of Machine Learning Research, Vol. 139).

Author

Thygesen, Christian B. ; Al-Sibahi, Ahmad Salim ; Steenmanns, Christian S. ; Sanz Moreta, Lys ; Sørensen, Anders B. ; Hamelryck, Thomas Wim. / Efficient Generative Modelling of Protein Structure Fragments using a Deep Markov Model. International Conference on Machine Learning, 18-24 July 2021, Virtual. PMLR, 2021. pp. 10258-10267 (Proceedings of Machine Learning Research, Vol. 139).

Bibtex

@inproceedings{8a5efeb8a3bf48a3bd471577e4464536,
title = "Efficient Generative Modelling of Protein Structure Fragments using a Deep Markov Model",
abstract = "Fragment libraries are often used in protein structure prediction, simulation and design as a means to significantly reduce the vast conformational search space. Current state-of-the-art methods for fragment library generation do not properly account for aleatory and epistemic uncertainty, respectively due to the dynamic nature of proteins and experimental errors in protein structures. Additionally, they typically rely on information that is not generally or readily available, such as homologous sequences, related protein structures and other complementary information. To address these issues, we developed BIFROST, a novel take on the fragment library problem based on a Deep Markov Model architecture combined with directional statistics for angular degrees of freedom, implemented in the deep probabilistic programming language Pyro. BIFROST is a probabilistic, generative model of the protein backbone dihedral angles conditioned solely on the amino acid sequence. BIFROST generates fragment libraries with a quality on par with current state-of-the-art methods at a fraction of the run-time, while requiring considerably less information and allowing efficient evaluation of probabilities.",
author = "Thygesen, {Christian B.} and Al-Sibahi, {Ahmad Salim} and Steenmanns, {Christian S.} and {Sanz Moreta}, Lys and S{\o}rensen, {Anders B.} and Hamelryck, {Thomas Wim}",
year = "2021",
language = "English",
series = "Proceedings of Machine Learning Research",
pages = "10258--10267",
booktitle = "International Conference on Machine Learning, 18-24 July 2021, Virtual",
publisher = "PMLR",
note = "38th International Conference on Machine Learning ; Conference date: 18-07-2021 Through 24-07-2021",

}

RIS

TY - GEN

T1 - Efficient Generative Modelling of Protein Structure Fragments using a Deep Markov Model

AU - Thygesen, Christian B.

AU - Al-Sibahi, Ahmad Salim

AU - Steenmanns, Christian S.

AU - Sanz Moreta, Lys

AU - Sørensen, Anders B.

AU - Hamelryck, Thomas Wim

PY - 2021

Y1 - 2021

N2 - Fragment libraries are often used in protein structure prediction, simulation and design as a means to significantly reduce the vast conformational search space. Current state-of-the-art methods for fragment library generation do not properly account for aleatory and epistemic uncertainty, respectively due to the dynamic nature of proteins and experimental errors in protein structures. Additionally, they typically rely on information that is not generally or readily available, such as homologous sequences, related protein structures and other complementary information. To address these issues, we developed BIFROST, a novel take on the fragment library problem based on a Deep Markov Model architecture combined with directional statistics for angular degrees of freedom, implemented in the deep probabilistic programming language Pyro. BIFROST is a probabilistic, generative model of the protein backbone dihedral angles conditioned solely on the amino acid sequence. BIFROST generates fragment libraries with a quality on par with current state-of-the-art methods at a fraction of the run-time, while requiring considerably less information and allowing efficient evaluation of probabilities.

AB - Fragment libraries are often used in protein structure prediction, simulation and design as a means to significantly reduce the vast conformational search space. Current state-of-the-art methods for fragment library generation do not properly account for aleatory and epistemic uncertainty, respectively due to the dynamic nature of proteins and experimental errors in protein structures. Additionally, they typically rely on information that is not generally or readily available, such as homologous sequences, related protein structures and other complementary information. To address these issues, we developed BIFROST, a novel take on the fragment library problem based on a Deep Markov Model architecture combined with directional statistics for angular degrees of freedom, implemented in the deep probabilistic programming language Pyro. BIFROST is a probabilistic, generative model of the protein backbone dihedral angles conditioned solely on the amino acid sequence. BIFROST generates fragment libraries with a quality on par with current state-of-the-art methods at a fraction of the run-time, while requiring considerably less information and allowing efficient evaluation of probabilities.

M3 - Article in proceedings

T3 - Proceedings of Machine Learning Research

SP - 10258

EP - 10267

BT - International Conference on Machine Learning, 18-24 July 2021, Virtual

PB - PMLR

T2 - 38th International Conference on Machine Learning

Y2 - 18 July 2021 through 24 July 2021

ER -

ID: 300919805