Sampling Conformational Ensembles of Highly Dynamic Proteins via Generative Deep Learning

Journal of Chemical Information and Modeling, Год журнала: 2025, Номер unknown

Опубликована: Фев. 21, 2025

Proteins are inherently dynamic, and their conformational ensembles play a crucial role in biological function. Large-scale motions may govern the protein structure–function relationship, numerous transient but stable conformations of intrinsically disordered proteins (IDPs) can Investigating to understand regulations disease-related aggregations IDPs is challenging, both experimentally computationally. In this paper, we first introduce deep learning-based model, termed Internal Coordinate Net (ICoN), which learns physical principles changes from molecular dynamics simulation data. Second, selected data points through interpolation learned latent space rapidly identify novel synthetic with sophisticated large-scale side chains backbone arrangements. Third, highly dynamic amyloid-β1–42 (Aβ42) monomer, our learning model provided comprehensive sampling Aβ42's landscape. Analysis these revealed clusters that could be used rationalize experimental findings. Additionally, method important interactions atomistic details not included training New showed distinct chain rearrangements probed by electron paramagnetic resonance amino acid substitution studies. This approach transferable for any available training. The work also demonstrated ability utilize natural conformation sampling.

Язык: Английский

---

Introduction to Machine Learning for Predictive Modeling I

Challenges and advances in computational chemistry and physics, Год журнала: 2025, Номер unknown, С. 3 - 30

Опубликована: Янв. 1, 2025

Язык: Английский

---

Sampling Conformational Ensembles of Highly Dynamic Proteins via Generative Deep Learning

bioRxiv (Cold Spring Harbor Laboratory), Год журнала: 2024, Номер unknown

Опубликована: Май 5, 2024

Proteins are inherently dynamic, and their conformational ensembles functionally important in biology. Large scale motions may govern protein structure function relationship, numerous transient but stable conformations of intrinsically disordered proteins (IDPs) can play a crucial role biological function. Investigating to understand regulations disease related aggregations IDPs is challenging both experimentally computationally. In this paper first an unsupervised deep learning based model, termed Internal Coordinate Net (ICoN), developed that learns the physical principles changes from molecular dynamics (MD) simulation data. Second, interpolating data points learned latent space selected rapidly identify novel synthetic with sophisticated large sidechains backbone arrangements. Third, highly dynamic amyloid beta 1 42 (Abeta42) monomer, our model provided comprehensive sampling Abeta42's landscape. Analysis these revealed clusters be used rationalize experimental findings. Additionally, method interactions atomistic details not included training New showed distinct sidechain rearrangements probed by EPR amino acid substitution studies. The proposed approach transferable for any available training. work also demonstrated ability utilize natural conformation sampling.

Язык: Английский

---

Sampling Conformational Ensembles of Highly Dynamic Proteins via Generative Deep Learning

Research Square (Research Square), Год журнала: 2024, Номер unknown

Опубликована: Июнь 28, 2024

Proteins are inherently dynamic, and their conformational ensembles functionally important in biology. Large-scale motions may govern protein structure–function relationship, numerous transient but stable conformations of intrinsically disordered proteins (IDPs) can play a crucial role biological function. Investigating to understand regulations disease-related aggregations IDPs is challenging both experimentally computationally. In this paper first an unsupervised deep learning-based model, termed Internal Coordinate Net (ICoN), developed that learns the physical principles changes from molecular dynamics (MD) simulation data. Second, interpolating data points learned latent space selected rapidly identify novel synthetic with sophisticated large-scale sidechains backbone arrangements. Third, highly dynamic amyloid-β1-42 (Aβ42) monomer, our learning model provided comprehensive sampling Aβ42’s landscape. Analysis these revealed clusters be used rationalize experimental findings. Additionally, method interactions atomistic details not included training New showed distinct sidechain rearrangements probed by EPR amino acid substitution studies. The proposed approach transferable for any available training. work also demonstrated ability utilize natural conformation sampling.

Язык: Английский

---

Evaluation and analysis of the effect of vocational and technical courses in higher education based on multiple regression analysis

Опубликована: Июнь 28, 2024

Язык: Английский

---