Machine Learning Simulation Protocol for Protein Amide II Spectra

Sheng Ye

DOI: https://doi.org/10.5281/zenodo.10277699

2023-01-01

Abstract:Machine learning simulation data for protein amide II spectra.

What problem does this paper attempt to address?

Artificial Intelligence-based Amide-II Infrared Spectroscopy Simulation for Monitoring Protein Hydrogen Bonding Dynamics.

Sheng Ye,Kai Zhong,Yan Huang,Guozhen Zhang,Changyin Sun,Jun Jiang

DOI: https://doi.org/10.1021/jacs.3c12258

IF: 15

2024-01-01

Journal of the American Chemical Society

Abstract:The structurally sensitive amide II infrared (IR) bands of proteins provide valuable information about the hydrogen bonding of protein secondary structures, which is crucial for understanding protein dynamics and associated functions. However, deciphering protein structures from experimental amide II spectra relies on time-consuming quantum chemical calculations on tens of thousands of representative configurations in solvent water. Currently, the accurate simulation of amide II spectra for whole proteins remains a challenge. Here, we present a machine learning (ML)-based protocol designed to efficiently simulate the amide II IR spectra of various proteins with an accuracy comparable to experimental results. This protocol stands out as a cost-effective and efficient alternative for studying protein dynamics, including the identification of secondary structures and monitoring the dynamics of protein hydrogen bonding under different pH conditions and during protein folding process. Our method provides a valuable tool in the field of protein research, focusing on the study of dynamic properties of proteins, especially those related to hydrogen bonding, using amide II IR spectroscopy.
Machine Learning Spectroscopy Using a 2-Stage, Generalized Constituent Contribution Protocol

Jinming Fan,Chao Qian,Shaodong Zhou

DOI: https://doi.org/10.34133/research.0115

IF: 11

2023-01-01

Research

Abstract:A corrected group contribution (CGC)-molecule contribution (MC)-Bayesian neural network (BNN) protocol for accurate prediction of absorption spectra is presented. Upon combination of BNN with CGC methods, the full absorption spectra of various molecules are afforded accurately and efficiently-by using only a small dataset for training. Here, with a small training sample (<100), accurate prediction of maximum wavelength for single molecules is afforded with the first stage of the protocol; by contrast, previously reported machine learning (ML) methods require >1,000 samples to ensure the accuracy of prediction. Furthermore, with <500 samples, the mean square error in the prediction of full ultraviolet spectra reaches <2%; for comparison, ML models with molecular SMILES for training require a much larger dataset (>2,000) to achieve comparable accuracy. Moreover, by employing an MC method designed specifically for CGC that properly interprets the mixing rule, the spectra of mixtures are obtained with high accuracy. The logical origins of the good performance of the protocol are discussed in detail. Considering that such a constituent contribution protocol combines chemical principles and data-driven tools, most likely, it will be proven efficient to solve molecular-property-relevant problems in wider fields.
Machine Learning Spectroscopy Based on Group Contribution and Molecule Contribution Methods

Jinming Fan,Chao Qian,Shaodong Zhou

DOI: https://doi.org/10.21203/rs.3.rs-2139666/v1

2022-01-01

Abstract:Abstract A group contribution (GC) – molecule contribution (MC) – machine learning (ML) protocol for accurate prediction of absorption spectra is presented. Upon combination of ML with revised GC methods, both the maximum absorption wavelengths and the full spectra of various sorts of dyes are afforded accurately and efficiently – by using only a small data set for training. Further, by employing a MC method designed specifically for revGC and based on MC-interpretated mixing rule by, the spectra for mixtures are obtained with high accuracy. Considering that the revGC-MC-ML protocol combines chemical principles and data-driven tools, most likely, it will be proven efficient to solve molecular-property mattered problems in wider fields.
A Machine Learning Protocol for Predicting Protein Infrared Spectra.

Sheng Ye,Kai Zhong,Jinxiao Zhang,Wei Hu,Jonathan D. Hirst,Guozhen Zhang,Shaul Mukamel,Jun Jiang

DOI: https://doi.org/10.1021/jacs.0c06530

IF: 15

2020-01-01

Journal of the American Chemical Society

Abstract:Infrared (IR) absorption provides important chemical fingerprints of biomolecules. Protein secondary structure determination from IR spectra is tedious since its theoretical interpretation requires repeated expensive quantum-mechanical calculations in a fluctuating environment. Herein we present a novel machine learning protocol that uses a few key structural descriptors to rapidly predict amide I IR spectra of various proteins and agrees well with experiment. Its transferability enabled us to distinguish protein secondary structures, probe atomic structure variations with temperature, and monitor protein folding. This approach offers a cost-effective tool to model the relationship between protein spectra and their biological/chemical properties.
A neural network protocol for electronic excitations of <i>N</i> -methylacetamide

Sheng Ye,Wei Hu,Xin Li,Jinxiao Zhang,Kai Zhong,Guozhen Zhang,Yi Luo,Shaul Mukamel,Jun Jiang

DOI: https://doi.org/10.1073/pnas.1821044116

IF: 11.1

2019-01-01

Proceedings of the National Academy of Sciences

Abstract:Significance UV absorption spectroscopy is an effective technique for characterizing protein structure. However its theoretical interpretation requires expensive first-principles simulations. We employ a neural network strategy to predict UV electronic spectra of peptide bonds. The protocol establishes structure–property relations and predicts ground-state dipole moments, as well as transition dipole moments. We establish machine learning as a useful spectroscopy simulation tool.
A neural network protocol for electronic excitations of N-methylacetamide.

Sheng Ye,Wei Hu,Xin Li,Jinxiao Zhang,Kai Zhong,Guozhen Zhang,Yi Luo,Shaul Mukamel,Jun Jiang

DOI: https://doi.org/10.1073/pnas.1821044116

2019-01-01

Abstract:UV absorption is widely used for characterizing proteins structures. The mapping of UV spectra to atomic structure of proteins relies on expensive theoretical simulations, circumventing the heavy computational cost which involves repeated quantum-mechanical simulations of excited-state properties of many fluctuating protein geometries, which has been a long-time challenge. Here we show that a neural network machine-learning technique can predict electronic absorption spectra of N-methylacetamide (NMA), which is a widely used model system for the peptide bond. Using ground-state geometric parameters and charge information as descriptors, we employed a neural network to predict transition energies, ground-state, and transition dipole moments of many molecular-dynamics conformations at different temperatures, in agreement with time-dependent density-functional theory calculations. The neural network simulations are nearly 3,000× faster than comparable quantum calculations. Machine learning should provide a cost-effective tool for simulating optical properties of proteins.
Machine Learning Molecular Dynamics for the Simulation of Infrared Spectra

Michael Gastegger,Jörg Behler,Philipp Marquetand

DOI: https://doi.org/10.48550/arXiv.1705.05907

IF: 2.552

2017-05-16

Chemical Physics

Abstract:Machine learning has emerged as an invaluable tool in many research areas. In the present work, we harness this power to predict highly accurate molecular infrared spectra with unprecedented computational efficiency. To account for vibrational anharmonic and dynamical effects -- typically neglected by conventional quantum chemistry approaches -- we base our machine learning strategy on ab initio molecular dynamics simulations. While these simulations are usually extremely time consuming even for small molecules, we overcome these limitations by leveraging the power of a variety of machine learning techniques, not only accelerating simulations by several orders of magnitude, but also greatly extending the size of systems that can be treated. To this end, we develop a molecular dipole moment model based on environment dependent neural network charges and combine it with the neural network potentials of Behler and Parrinello. Contrary to the prevalent big data philosophy, we are able to obtain very accurate machine learning models for the prediction of infrared spectra based on only a few hundreds of electronic structure reference points. This is made possible through the introduction of a fully automated sampling scheme and the use of molecular forces during neural network potential training. We demonstrate the power of our machine learning approach by applying it to model the infrared spectra of a methanol molecule, n-alkanes containing up to 200 atoms and the protonated alanine tripeptide, which at the same time represents the first application of machine learning techniques to simulate the dynamics of a peptide. In all these case studies we find excellent agreement between the infrared spectra predicted via machine learning models and the respective theoretical and experimental spectra.
Simulation Study of Chiral Two-Dimensional Ultraviolet Spectroscopy of the Protein Backbone

Darius Abramavicius,Jun Jiang,Benjamin M. Bulheller,Jonathan D. Hirst,Shaul Mukamel

DOI: https://doi.org/10.1021/ja101968g

IF: 15

2010-01-01

Journal of the American Chemical Society

Abstract:Amide n-pi* and pi-pi* excitations around 200 nm are prominent spectroscopic signatures of the protein backbone, which are routinely used in ultraviolet (UV) circular dichroism for structure characterization. Recently developed ultrafast laser sources may be used to extend these studies to two dimensions. We apply a new algorithm for modeling protein electronic transitions to simulate two-dimensional UV photon echo signals in this regime and to identify signatures of protein backbone secondary (and tertiary) structure. Simulated signals for a set of globular and fibrillar proteins and their specific regions reveal characteristic patterns of helical and sheet secondary structures. We investigate how these patterns vary and converge with the size of the structural motif. Specific chiral polarization configurations of the UV pulses are found to be sensitive to aspects of the protein structure. This information significantly augments that available from linear circular dichroism.
A Machine-Learning Protocol for Ultraviolet Protein-Backbone Absorption Spectroscopy under Environmental Fluctuations

Jinxiao Zhang,Sheng Ye,Kai Zhong,Yaolong Zhang,Yuanyuan Chong,Luyuan Zhao,Huiting Zhou,Sibei Guo,Guozhen Zhang,Bin Jiang,Shaul Mukamel,Jun Jiang

DOI: https://doi.org/10.1021/acs.jpcb.1c03296

IF: 3.466

2021-06-04

The Journal of Physical Chemistry B

Abstract:Ultraviolet (UV) absorption spectra are commonly used for characterizing the global structure of proteins. However, the theoretical interpretation of UV spectra is hindered by the large number of required expensive ab initio calculations of excited states spanning a huge conformation space. We present a machine-learning (ML) protocol for far-UV (FUV) spectra of proteins, which can predict FUV spectra of proteins with comparable accuracy to density functional theory (DFT) calculations but with 3–4 orders of magnitude reduced computational cost. It further shows excellent predictive power and transferability that can be used to probe structural mutations and protein folding pathways.The Supporting Information is available free of charge at <a class="ext-link" href="/doi/10.1021/acs.jpcb.1c03296?goto=supporting-info">https://pubs.acs.org/doi/10.1021/acs.jpcb.1c03296</a>.Additional computational details, figures, and tables (<a class="ext-link" href="/doi/suppl/10.1021/acs.jpcb.1c03296/suppl_file/jp1c03296_si_001.pdf">PDF</a>)This article has not yet been cited by other publications.

chemistry, physical
Two-Dimensional Electronic Correlation Spectroscopy of the N Pi* and Pi Pi* Protein Backbone Transitions: A Simulation Study

Zhenyu Li,Darius Abramavicius,Wei Zhuang,Shaul Mukamel

DOI: https://doi.org/10.1016/j.chemphys.2007.03.029

IF: 2.552

2007-01-01

Chemical Physics

Abstract:The two dimensional (2D) photon echo spectrum of the amide ultraviolet (UV) bands of proteins are simulated. Two effective exciton Hamiltonian parameter sets developed by Woody and Hirst, which predict similar CD spectra, may be distinguished by their very different 2DUV spectra. These differences are enhanced in specific configurations of pulse polarizations which provide chirality-induced signals.
Multipolar Force Fields for Amide-I Spectroscopy from Conformational Dynamics of the Alanine-Trimer

Padmabati Mondal,Pierre-André Cazade,Akshaya K. Das,Tristan Bereau,Markus Meuwly

DOI: https://doi.org/10.48550/arXiv.2106.10142

2021-06-18

Abstract:The dynamics and spectroscopy of N-methyl-acetamide (NMA) and trialanine in solution is characterized from molecular dynamics (MD) simulations using different energy functions, including a conventional point charge (PC)-based force field, one based on a multipolar (MTP) representation of the electrostatics, and a semiempirical DFT method. For the 1-d infrared spectra, the frequency splitting between the two amide-I groups is 10 cm$^{-1}$ from the PC, 13 cm$^{-1}$ from the MTP, and 47 cm$^{-1}$ from SCC-DFTB simulations, compared with 25 cm$^{-1}$ from experiment. The frequency trajectory required for determining the frequency fluctuation correlation function (FFCF) is determined from individual (INM) and full normal mode (FNM) analyses of the amide-I vibrations. The spectroscopy, time-zero magnitude of the FFCF $C(t=0)$, and the static component $\Delta_0^2$ from simulations using MTP and analysis based on FNM are all consistent with experiments for (Ala)$_3$. Contrary to that, for the analysis excluding mode-mode coupling (INM) the FFCF decays to zero too rapidly and for simulations with a PC-based force field the $\Delta_0^2$ is too small by a factor of two compared with experiments. Simulations with SCC-DFTB agree better with experiment for these observables than those from PC-based simulations. The conformational ensemble sampled from simulations using PCs is consistent with the literature , whereas that covered by the MTP-based simulations is dominated by P$_{\rm II}$ which agrees with and confirms recently reported, Bayesian-refined populations based on 1-dimensional infrared experiments. Full normal mode analysis together with a MTP representation provides a meaningful model to correctly describe the dynamics of hydrated trialanine.

Biological Physics
HSQC Spectra Simulation and Matching for Molecular Identification

Martin Priessner,Richard J. Lewis,Magnus J. Johansson,Jonathan M. Goodman,Jon Paul Janet,Anna Tomberg

DOI: https://doi.org/10.1021/acs.jcim.3c01735

IF: 6.162

2024-03-27

Journal of Chemical Information and Modeling

Abstract:In the pursuit of improved compound identification and database search tasks, this study explores heteronuclear single quantum coherence (HSQC) spectra simulation and matching methodologies. HSQC spectra serve as unique molecular fingerprints, enabling a valuable balance of data collection time and information richness. We conducted a comprehensive evaluation of the following four HSQC simulation techniques: ACD/Labs (ACD), MestReNova (MNova), Gaussian NMR calculations (DFT), and a graph-based neural network (ML). For the latter two techniques, we developed a reconstruction logic to combine proton and carbon 1D spectra into HSQC spectra. The methodology involved the implementation of three peak-matching strategies (minimum-sum, Euclidean-distance, and Hungarian distance) combined with three padding strategies (zero-padding, peak-truncated, and nearest-neighbor double assignment). We found that coupling these strategies with a robust simulation technique facilitates the accurate identification of correct molecules from similar analogues (regio- and stereoisomers) and allows for fast and accurate large database searches. Furthermore, we demonstrated the efficacy of the best-performing methodology by rectifying the structures of a set of previously misidentified molecules. This research indicates that effective HSQC spectral simulation and matching methodologies significantly facilitate molecular structure elucidation. Furthermore, we offer a Google Colab notebook for researchers to use our methods on their own data (https://github.com/AstraZeneca/hsqc_structure_elucidation.git).

chemistry, multidisciplinary, medicinal,computer science, interdisciplinary applications, information systems
Accurate Machine Learning Prediction of Protein Circular Dichroism Spectra with Embedded Density Descriptors

Luyuan Zhao,Jinxiao Zhang,Yaolong Zhang,Sheng Ye,Guozhen Zhang,Xin Chen,Bin Jiang,Jun Jiang

DOI: https://doi.org/10.1021/jacsau.1c00449

IF: 8

2021-01-01

JACS Au

Abstract:A data-driven approach to simulate circular dichroism (CD) spectra is appealing for fast protein secondary structure determination, yet the challenge of predicting electric and magnetic transition dipole moments poses a substantial barrier for the goal. To address this problem, we designed a new machine learning (ML) protocol in which ordinary pure geometry-based descriptors are replaced with alternative embedded density descriptors and electric and magnetic transition dipole moments are successfully predicted with an accuracy comparable to first-principle calculation. The ML model is able to not only simulate protein CD spectra nearly 4 orders of magnitude faster than conventional first-principle simulation but also obtain CD spectra in good agreement with experiments. Finally, we predicted a series of CD spectra of the Trp-cage protein associated with continuous changes of protein configuration along its folding path, showing the potential of our ML model for supporting real-time CD spectroscopy study of protein dynamics.
Geometry and Excitation Energy Fluctuations of NMA in Aqueous Solution with CHARMM, AMBER, OPLS, and GROMOS Force Fields: Implications for Protein Ultraviolet Spectra Simulation.

Zhenyu Li,Haibo Yu,Wei Zhuang,Shaul Mukamel

DOI: https://doi.org/10.1016/j.cplett.2007.12.022

IF: 2.719

2007-01-01

Chemical Physics Letters

Abstract:Molecular dynamics (MD) simulations are performed for N-methylamide (NMA) in water at 300 K with different force fields. Compared to the three all-atom force fields (CHARMM22, AMBER03, and OPLS-AA), the united-atom force field (GROMOS96) predicts a broader distribution of the peptide OCNH dehedral angle. A map constructed by fitting the npi* and pipi* transition energies as quadratic functions of the NMA geometric variables is used to simulate the excitation energy fluctuations. GROMOS96 predicts blue-shifted npi* and pipi* energies and stronger fluctuations compared to the other three force fields, which indicates that different force fields may predict different spectral lineshapes for proteins.
A Dataset of Simple 1-D and 2-D NMR Spectra of Peptides, Including all Encoded Amino Acids, for Introductory Instruction in Protein Biomolecular NMR Spectroscopy

Neal J. Zondlo,Geoffrey M. Sametz,Nicholas V. Costantini,Noah J. Daniecki`

DOI: https://doi.org/10.26434/chemrxiv-2024-zww4n

2024-03-11

Abstract:NMR spectroscopy is the most important technique for understanding the structure of peptides and proteins in solution, providing information at the single-residue and single-atom level. However, written instruction in the interpretation of NMR spectra of peptides and proteins is generally focused on advanced techniques and highly complex spectra, with a lack of simple spectra and guides available for beginning students. In order to address this instructional limitation, we have generated a dataset of 1H NMR spectra of a series of simple peptides that include all canonical amino acids. Peptides examined include Ac-X(S/pS)-NH2, Ac-X(T/pT)-NH2, and Ac-XPPGY-NH2, where X = all encoded amino acids, pS = phosphorylated Ser, and pT = phosphorylated Thr. The characterization of each peptide includes a 1-D spectrum and a TOCSY spectrum, with both the raw and processed available. The spectra can be used for instructional applications including analysis of regions of the spectra (e.g. amide HN, aromatic, Hα, and aliphatic regions); identification of spin systems and residue assignment via TOCSY spectra; analysis of conformational features including amide HN chemical shift dispersion and changes due to hydrogen bonding or post-translational modifications; the 3JαN coupling constant that reports on the φ torsion angle and on order versus disorder at a given residue; conformational preferences at Hα via chemical shift index analysis; understanding of diastereotopic hydrogens; dynamic processes, including hydrogen exchange; and identification of proline cis-trans isomerism. In addition, for a limited number of peptides, NOESY spectra are included to allow sequential resonance assignment and for assignment of trans versus cis proline conformations. Spectra from closely related peptides allow the analysis of the relative effects of single amino acid changes. The paper is written to be directly accessible to students as a tutorial guide. In addition, the data can be used by instructors for problem sets and exams.

Chemistry
Protein NMR assignment by isotope pattern recognition

Uluk Rasulov,Harrison K Wang,Thibault Viennet,Maxim A Droemer,Srđan Matosin,Sebastian Schindler,Zhen-Yu J Sun,Luca Mureddu,Geerten W Vuister,Scott A Robson,Haribabu Arthanari,Ilya Kuprov

DOI: https://doi.org/10.1126/sciadv.ado0403

2024-09-06

Abstract:The current standard method for amino acid signal identification in protein NMR spectra is sequential assignment using triple-resonance experiments. Good software and elaborate heuristics exist, but the process remains laboriously manual. Machine learning does help, but its training databases need millions of samples that cover all relevant physics and every kind of instrumental artifact. In this communication, we offer a solution to this problem. We propose polyadic decompositions to store millions of simulated three-dimensional NMR spectra, on-the-fly generation of artifacts during training, a probabilistic way to incorporate prior and posterior information, and integration with the industry standard CcpNmr software framework. The resulting neural nets take [1H,13C] slices of mixed pyruvate-labeled HNCA spectra (different CA signal shapes for different residue types) and return an amino acid probability table. In combination with primary sequence information, backbones of common proteins (GB1, MBP, and INMT) are rapidly assigned from just the HNCA spectrum.
First-Principles Simulation of Amide and Aromatic Side Chain Ultraviolet Spectroscopy of A Cyclic Dipeptide

Zhenyu Li,Shaul Mukamel

DOI: https://doi.org/10.1021/jp075515s

2007-01-01

Abstract:By combining time-dependent density functional theory (TDDFT) and molecular dynamics (MD) simulations, we calculate the ultraviolet absorption and circular dichroism (CD) of a cyclic dipeptide, cyclo(L-Pro-DTyr), in the 185-300 nm region. The absorption is dominated by the phenol chromophore of tyrosine. The CD spectrum shows both phenol and amide units transitions. A crude coherent two-dimensional ultraviolet spectrum (2DUV) calculated by neglecting the two-excitation states shows a cross-peak between two transitions of the phenol in the tyrosine side chain. Additional cross-peaks between the side chain and the backbone are observed when using a chirality-induced pulse polarization configuration.
Ultraviolet Spectroscopy of Protein Backbone Transitions in Aqueous Solution: Combined QM and MM Simulations.

Jun Jiang,Darius Abramavicius,Benjamin M. Bulheller,Jonathan D. Hirst,Shaul Mukamel

DOI: https://doi.org/10.1021/jp101980a

IF: 3.466

2010-01-01

The Journal of Physical Chemistry B

Abstract:A generalized approach combining quantum mechanics (QM) and molecular mechanics (MM) calculations is developed to simulate the n --> pi* and pi --> pi* backbone transitions of proteins in aqueous solution. These transitions, which occur in the ultraviolet (UV) at 180-220 nm, provide a sensitive probe for secondary structures. The excitation Hamiltonian is constructed using high-level electronic structure calculations of N-methylacetamide (NMA). Its electrostatic fluctuations are modeled using a new algorithm, EHEF, which combines a molecular dynamics (MD) trajectory obtained with a MM forcefield and electronic structures of sampled MD snapshots calculated by QM. The lineshapes and excitation splittings induced by the electrostatic environment in the experimental UV linear absorption (LA) and circular dichroism (CD) spectra of several proteins in aqueous solution are reproduced by our calculations. The distinct CD features of alpha-helix and beta-sheet protein structures are observed in the simulations and can be assigned to different backbone geometries. The fine structure of the UV spectra is accurately characterized and enables us to identify signatures of secondary structures.
Automated Fragmentation QM/MM Calculation of Amide Proton Chemical Shifts in Proteins with Explicit Solvent Model.

Tong Zhu,John Z. H. Zhang,Xiao He

DOI: https://doi.org/10.1021/ct300999w

2013-01-01

Journal of Chemical Theory and Computation

Abstract:We have performed a density functional theory (DFT) calculation of the amide proton NMR chemical shift in proteins using a recently developed automated fragmentation quantum mechanics/molecular mechanics (AF-QM/MM) approach. Systematic investigation was carried out to examine the influence of explicit solvent molecules, cooperative hydrogen bonding effects, density functionals, size of the basis sets, and the local geometry of proteins on calculated chemical shifts. Our result demonstrates that the predicted amide proton ((1)HN) NMR chemical shift in explicit solvent shows remarkable improvement over that calculated with the implicit solvation model. The cooperative hydrogen bonding effect is also shown to improve the accuracy of (1)HN chemical shifts. Furthermore, we found that the OPBE exchange-correlation functional is the best density functional for the prediction of protein (1)HN chemical shifts among a selective set of DFT methods (namely, B3LYP, B3PW91, M062X, M06L, mPW1PW91, OB98, OPBE), and the locally dense basis set of 6-311++G**/4-31G* is shown to be sufficient for (1)HN chemical shift calculation. By taking ensemble averaging into account, (1)HN chemical shifts calculated by the AF-QM/MM approach can be used to validate the performance of various force fields. Our study underscores that the electronic polarization of protein is of critical importance to stabilizing hydrogen bonding, and the AF-QM/MM method is able to describe the local chemical environment in proteins more accurately than most widely used empirical models.
Raman spectra of amino acids and peptides from machine learning polarizabilities

Ethan Berger,Juha Niemelä,Outi Lampela,André H. Juffer,Hannu-Pekka Komsa

2024-04-29

Abstract:Raman spectroscopy is an important tool in the study of vibrational properties and composition of molecules, peptides and even proteins. Raman spectra can be simulated based on the change of the electronic polarizability with vibrations, which can nowadays be efficiently obtained via machine learning models trained on first-principles data. However, the transferability of the models trained on small molecules to larger structures is unclear and direct training on large structures in prohibitively expensive. In this work, we first train two machine learning models to predict polarizabilities of all 20 amino acids. Both models are carefully benchmarked and compared to DFT calculations, with neural network method found to offer better transferability. By combining machine learning models with classical force field molecular dynamics, Raman spectra of all amino acids are also obtained and investigated, showing good agreement with experiments. The models are further extended to small peptides. We find that adding structures containing peptide bonds to the training set greatly improves predictions even for peptides not included in training sets.

Computational Physics,Biological Physics

Machine Learning Simulation Protocol for Protein Amide II Spectra

Artificial Intelligence-based Amide-II Infrared Spectroscopy Simulation for Monitoring Protein Hydrogen Bonding Dynamics.

Machine Learning Spectroscopy Using a 2-Stage, Generalized Constituent Contribution Protocol

Machine Learning Spectroscopy Based on Group Contribution and Molecule Contribution Methods

A Machine Learning Protocol for Predicting Protein Infrared Spectra.

A neural network protocol for electronic excitations of <i>N</i> -methylacetamide

A neural network protocol for electronic excitations of N-methylacetamide.

Machine Learning Molecular Dynamics for the Simulation of Infrared Spectra

Simulation Study of Chiral Two-Dimensional Ultraviolet Spectroscopy of the Protein Backbone

A Machine-Learning Protocol for Ultraviolet Protein-Backbone Absorption Spectroscopy under Environmental Fluctuations

Two-Dimensional Electronic Correlation Spectroscopy of the N Pi* and Pi Pi* Protein Backbone Transitions: A Simulation Study

Multipolar Force Fields for Amide-I Spectroscopy from Conformational Dynamics of the Alanine-Trimer

HSQC Spectra Simulation and Matching for Molecular Identification

Accurate Machine Learning Prediction of Protein Circular Dichroism Spectra with Embedded Density Descriptors

Geometry and Excitation Energy Fluctuations of NMA in Aqueous Solution with CHARMM, AMBER, OPLS, and GROMOS Force Fields: Implications for Protein Ultraviolet Spectra Simulation.

A Dataset of Simple 1-D and 2-D NMR Spectra of Peptides, Including all Encoded Amino Acids, for Introductory Instruction in Protein Biomolecular NMR Spectroscopy

Protein NMR assignment by isotope pattern recognition

First-Principles Simulation of Amide and Aromatic Side Chain Ultraviolet Spectroscopy of A Cyclic Dipeptide

Ultraviolet Spectroscopy of Protein Backbone Transitions in Aqueous Solution: Combined QM and MM Simulations.

Automated Fragmentation QM/MM Calculation of Amide Proton Chemical Shifts in Proteins with Explicit Solvent Model.

Raman spectra of amino acids and peptides from machine learning polarizabilities