Unsupervised search of low-lying conformers with spectroscopic accuracy: A two-step algorithm rooted into the island model evolutionary algorithm

Pure rotational and rovibrational spectroscopy of cyclopropylamine in the far-infrared region: -NH2 torsion

The pure rotational and rovibrational spectra of the ν27 -NH2 torsion of cyclopropylamine (CPA) in the far-infrared region were measured with a high-resolution Fourier transform infrared coupled to a synchrotron. The complex spectra reflect the presence of both trans and gauche conformers. Analysis of the pure rotational spectra (34-64 cm-1) yielded accurate rotational and centrifugal distortion constants of the ground and first two torsional excited states of trans-CPA. The fundamental, hot bands and weak overtones were identified and assigned in the 200-550 cm-1 range. Global analysis of over 19 000 transitions provides accurate energy levels of the torsional polyads up to vT = 3. The torsional levels and their rotational constants were in agreement with the theoretical results from quasiadiabatic channel reaction path Hamiltonian (RPH) calculations, emphasizing the need for molecular-specific theoretical treatments for large amplitude motions. Tunneling components of the torsional fundamental of gauche-CPA were assigned based on the RPH results and symmetry considerations, differing from previous experimental and theoretical work. This comprehensive spectroscopic characterization of CPA is crucial for its potential detection in the interstellar medium as a precursor to complex prebiotic molecules, providing essential data for future astronomical searches and advancing our understanding of nitrogen-containing organic molecules in space.

Toward Accurate Quantum Chemical Methods for Molecules of Increasing Dimension: The New Family of Pisa Composite Schemes

The new versions of the Pisa composite scheme introduced in the present paper are based on the careful selection of different quantum chemical models for energies, geometries, and vibrational frequencies, with the aim of maximizing the accuracy of the overall description while retaining a reasonable cost for all the steps. In particular, the computation of accurate electronic energies has been further improved introducing more reliable complete basis set extrapolations and estimation of core-valence correlation, together with improved basis sets for third-row atoms. Furthermore, the reduced-cost frozen natural orbital (FNO) model has been introduced and validated for large molecules. Accurate molecular structures can be obtained avoiding complete basis set extrapolation and evaluating core-valence correlation at the MP2 level. Unfortunately, analytical gradients are not available for the FNO version of the model. Therefore, for large molecules, an accurate reduced-cost alternative is offered by evaluation of valence contributions with a double-hybrid functional in conjunction with the same MP2 contribution for core-valence correlation or by means of a one-parameter approximation. The same double-hybrid functional and basis set are employed to evaluate zero-point energies and partition functions. After the validation of the new models for small systems, a panel of molecular bricks of life has been used to analyze their performances for problems of current fundamental or technological interest. The fully black-box implementation of the computational workflow paves the way toward the accurate yet not prohibitively expensive study of medium- to large-sized molecules also by experimentally oriented researchers.

Accurate Structures and Rotational Constants of Steroid Hormones at DFT Cost: Androsterone, Testosterone, Estrone, β-Estradiol, and Estriol

A comprehensive analysis of the structural, conformational, and spectroscopic properties in the gas phase has been performed for five prototypical steroid hormones, namely, androsterone, testosterone, estrone, β-estradiol, and estriol. The revDSD-PBEP86 double-hybrid functional in conjunction with the D3BJ empirical dispersion and a suitable triple-ζ basis set provides accurate conformational energies and equilibrium molecular structures, with the latter being further improved by proper account of core-valence correlation. Average deviations within 0.1% between computed and experimental ground state rotational constants are reached when adding to those equilibrium values vibrational corrections obtained at the cost of standard harmonic frequencies thanks to the use of a new computational tool. Together with the intrinsic interest of the studied hormones, the accuracy of the results obtained at DFT cost for molecules containing about 50 atoms paves the way toward the accurate investigations of other flexible bricks of life.

Quantum chemistry meets high-resolution spectroscopy for characterizing the molecular bricks of life in the gas-phase

Computation of accurate geometrical structures and spectroscopic properties of large flexible molecules in the gas-phase is tackled at an affordable cost using a general exploration/exploitation strategy. The most distinctive feature of the approach is the careful selection of different quantum chemical models for energies, geometries and vibrational frequencies with the aim of maximizing the accuracy of the overall description while retaining a reasonable cost for all the steps. In particular, a composite wave-function method is used for energies, whereas a double-hybrid functional (with the addition of core-valence correlation) is employed for geometries and harmonic frequencies and a cheaper hybrid functional for anharmonic contributions. A thorough benchmark based on a wide range of prototypical molecular bricks of life shows that the proposed strategy is close to the accuracy of state-of-the-art composite wave-function methods, and is applicable to much larger systems. A freely available web-utility post-processes the geometries optimized by standard electronic structure codes paving the way toward the accurate yet not prohibitively expensive study of medium- to large-sized molecules by experimentally-oriented researchers.

Exploring torsional conformer space with physical prior mean function-driven meta-Gaussian processes

We present a novel approach for systematically exploring the conformational space of small molecules with multiple internal torsions. Identifying unique conformers through a systematic conformational search is important for obtaining accurate thermodynamic functions (e.g., free energy), encompassing contributions from the ensemble of all local minima. Traditional geometry optimizers focus on one structure at a time, lacking transferability from the local potential-energy surface (PES) around a specific minimum to optimize other conformers. In this work, we introduce a physics-driven meta-Gaussian processes (meta-GPs) method that not only enables efficient exploration of target PES for locating local minima but, critically, incorporates physical surrogates that can be applied universally across the optimization of all conformers of the same molecule. Meta-GPs construct surrogate PESs based on the optimization history of prior conformers, dynamically selecting the most suitable prior mean function (representing prior knowledge in Bayesian learning) as a function of the optimization progress. We systematically benchmarked the performance of multiple GP variants for brute-force conformational search of amino acids. Our findings highlight the superior performance of meta-GPs in terms of efficiency, comprehensiveness of conformer discovery, and the distribution of conformers compared to conventional non-surrogate optimizers and other non-meta-GPs. Furthermore, we demonstrate that by concurrently optimizing, training GPs on the fly, and learning PESs, meta-GPs exhibit the capacity to generate high-quality PESs in the torsional space without extensive training data. This represents a promising avenue for physics-based transfer learning via meta-GPs with adaptive priors in exploring torsional conformer space.

Molecular Hessian matrices from a machine learning random forest regression algorithm

In this article, we present a machine learning model to obtain fast and accurate estimates of the molecular Hessian matrix. In this model, based on a random forest, the second derivatives of the energy with respect to redundant internal coordinates are learned individually. The internal coordinates together with their specific representation guarantee rotational and translational invariance. The model is trained on a subset of the QM7 dataset but is shown to be applicable to larger molecules picked from the QM9 dataset. From the predicted Hessian, it is also possible to obtain reasonable estimates of the vibrational frequencies, normal modes, and zero point energies of the molecules.

Accurate thermochemistry at affordable cost by means of an improved version of the JunChS-F12 model chemistry

The junChS-F12 composite method has been improved by means of the latest implementation of the CCSD(F12*)(T+) ansatz and validated for the thermochemistry of molecules containing atoms of the first three rows of the periodic table. A thorough benchmark showed that this model, in conjunction with cost-effective revDSD-PBEP86-D3(BJ) reference geometries, offers an optimal compromise between accuracy and computational cost. If improved geometries are sought, the most effective option is to add MP2-F12 core-valence correlation corrections to CCSD(T)-F12b/jun-cc-pVTZ geometries without the need of performing any extrapolation to the complete basis set limit. In the same vein, CCSD(T)-F12b/jun-cc-pVTZ harmonic frequencies are remarkably accurate without any additional contribution. Pilot applications to noncovalent intermolecular interactions, conformational landscapes, and tautomeric equilibria confirm the effectiveness and reliability of the model.

DFT Meets Wave-Function Methods for Accurate Structures and Rotational Constants of Histidine, Tryptophan, and Proline

A new computational strategy has been applied to the conformational and spectroscopic properties in the gas phase of amino acids with very distinctive features, ranging from different tautomeric forms (histidine) to ring puckering (proline), and heteroaromatic structures with non-equivalent rings (tryptophan). The integration of modern double-hybrid functionals and wave-function composite methods has allowed us to obtain accurate results for a large panel of conformers with reasonable computer times. The remarkable agreement between computations and microwave experiments allows an unbiased interpretation of the latter in terms of stereoelectronic effects.

Accurate structures and spectroscopic parameters of α,α-dialkylated α-amino acids in the gas-phase: a joint venture of DFT and wave-function composite methods

Accurate computations of structural, conformational and spectroscopic properties in the gas phase have been performed for two α,α-dialkylated α-amino acids, namely aminoisobutyric acid and cyclopropylglycine. Thanks to the integration of modern double hybrid functionals and wave-function methods, several low-energy structures of the title molecules could be analyzed employing standard computer resources. The computed features of all the most stable conformers of the target amino acids closely match the corresponding spectroscopic parameters issued from microwave spectroscopic studies in the gas-phase. Together with their intrinsic interest, the accuracy of the results obtained with reasonable computer times paves the way for accurate investigations of other flexible bricks of life.

Theoretical-Computational Modeling of CD Spectra of Aqueous Monosaccharides by Means of Molecular Dynamics Simulations and Perturbed Matrix Method

The electronic circular dichroism (ECD) spectra of aqueous d-glucose and d-galactose were modeled using a theoretical-computational approach combining molecular dynamics (MD) simulations and perturbed matrix method (PMM) calculations, hereafter termed MD-PMM. The experimental spectra were reproduced with a satisfactory accuracy, confirming the good performances of MD-PMM in modeling different spectral features in complex atomic-molecular systems, as already reported in previous studies. The underlying strategy of the method was to perform a preliminary long timescale MD simulation of the chromophore followed by the extraction of the relevant conformations through essential dynamics analysis. On this (limited) number of relevant conformations, the ECD spectrum was calculated via the PMM approach. This study showed that MD-PMM was able to reproduce the essential features of the ECD spectrum (i.e., the position, the intensity, and the shape of the bands) of d-glucose and d-galactose while avoiding the rather computationally expensive aspects, which were demonstrated to be important for the final outcome, such as (i) the use of a large number of chromophore conformations; (ii) the inclusion of quantum vibronic coupling; and (iii) the inclusion of explicit solvent molecules interacting with the chromophore atoms within the chromophore itself (e.g., via hydrogen bonds).

Accurate Structures and Spectroscopic Parameters of Phenylalanine and Tyrosine in the Gas Phase: A Joint Venture of DFT and Composite Wave-Function Methods

A general strategy for the accurate computation of conformational and spectroscopic properties of flexible molecules in the gas phase is applied to two representative proteinogenic amino acids with aromatic side chains, namely, phenylalanine and tyrosine. The main features of all the most stable conformers predicted by this computational strategy closely match those of the species detected in microwave and infrared experiments. Together with their intrinsic interest, the accuracy of the results obtained with reasonable computer times paves the route for accurate investigations of other flexible bricks of life.

Computational Protocol for the Identification of Candidates for Radioastronomical Detection and Its Application to the C3H3NO Family of Isomers

The C3H3NO family of isomers is relevant in astrochemistry, even though its members are still elusive in the interstellar medium. To identify the best candidate for astronomical detection within this family, we developed a new computational protocol based on the minimum-energy principle. This approach aims to identify the most stable isomer of the family and consists of three steps. The first step is an extensive investigation that characterizes the vast number of compounds having the C3H3NO chemical formula, employing density functional theory for this purpose. The second step is an energy refinement, which is used to select isomers and relies on coupled cluster theory. The last step is a structural improvement with a final energy refinement that provides improved energies and a large set of accurate spectroscopic parameters for all isomers lying within 30 kJ mol−1 above the most stable one. According to this protocol, vinylisocyanate is the most stable isomer, followed by oxazole, which is about 5 kJ mol−1 higher in energy. The other stable species are pyruvonitrile, cyanoacetaldehyde, and cyanovinylalcohol. For all of these species, new computed rotational and vibrational spectroscopic data are reported, which complement those already available in the literature or fill current gaps.

Protonation of serine: conformers, proton affinities and gas-phase basicities at the "gold standard" and beyond

The potential energy surfaces (PESs) of serine and its protonated counterparts are investigated to determine the structures of the minima. A total of 95 neutral serine, 15 N-(amino-) and 46 O-(carbonyl-)protonated serine conformers are found. Their relative energies, geometries and harmonic vibrational frequencies are determined at the MP2/aug-cc-pVDZ level of theory. To obtain highly accurate thermodynamic values, further computations are performed: the ten conformers with the lowest relative energies from each molecule type (neutral, N- and O-protonated) are further optimized using the explicitly correlated CCSD(T)-F12a/cc-pVDZ-F12 method (for neutral serine, harmonic vibrational frequencies were also computed). In addition, auxiliary corrections were determined: basis-set effects up to CCSD(T)-F12b/cc-pVQZ-F12, electron correlation effects up to CCSDT(Q), core correlation and second-order Douglas-Kroll relativistic effects along with zero-point energy contributions. Two important thermodynamic parameters (at 298.15 K), proton affinity (PA)/gas-phase basicity (GB) are calculated considering the two different protonation sites: 218.05 ± 0.2/209.86 ± 0.6 kcal mol^-1 and 205.87 ± 0.2/196.36 ± 0.3 kcal mol^-1 for the amino and carbonyl sites, respectively. The uncertainty of the determined values was approximated based on various sources including auxiliary corrections, basis-set effects, harmonic vibrational frequencies.

Bringing Machine-Learning Enhanced Quantum Chemistry and Microwave Spectroscopy to Conformational Landscape Exploration: the Paradigmatic Case of 4-Fluoro-Threonine

A combined experimental and theoretical study has been carried out on 4-fluoro-threonine, the only naturally occurring fluorinated amino acid. Fluorination of the methyl group significantly increases the conformational complexity with respect to the parent amino acid threonine. The conformational landscape has been characterized in great detail, with special attention given to the inter-conversion pathways between different conformers. This led to the identification of 13 stable low-energy minima. The equilibrium population of so many conformers produces a very complicated and congested rotational spectrum that could be assigned through a strategy that combines several levels of quantum chemical calculations with the principles of machine learning. Twelve conformers out of 13 could be experimentally characterized. The results obtained from the analysis of the intra-molecular interactions can be exploited to accurately model fluorine-substitution effects in biomolecules.

Benchmark Structures and Conformational Landscapes of Amino Acids in the Gas Phase: A Joint Venture of Machine Learning, Quantum Chemistry, and Rotational Spectroscopy

The accurate characterization of prototypical bricks of life can strongly benefit from the integration of high resolution spectroscopy and quantum mechanical computations. We have selected a number of representative amino acids (glycine, alanine, serine, cysteine, threonine, aspartic acid and asparagine) to validate a new computational setup rooted in quantum-chemical computations of increasing accuracy guided by machine learning tools. Together with low-lying energy minima, the barriers ruling their interconversion are evaluated in order to unravel possible fast relaxation paths. Vibrational and thermal effects are also included in order to estimate relative free energies at the temperature of interest in the experiment. The spectroscopic parameters of all the most stable conformers predicted by this computational strategy, which do not have low-energy relaxation paths available, closely match those of the species detected in microwave experiments. Together with their intrinsic interest, these accurate results represent ideal benchmarks for more approximate methods.

Toward Accurate yet Effective Computations of Rotational Spectroscopy Parameters for Biomolecule Building Blocks

The interplay of high-resolution rotational spectroscopy and quantum-chemical computations plays an invaluable role in the investigation of biomolecule building blocks in the gas phase. However, quantum-chemical methods suffer from unfavorable scaling with the dimension of the system under consideration. While a complete characterization of flexible systems requires an elaborate multi-step strategy, in this work, we demonstrate that the accuracy obtained by quantum-chemical composite approaches in the prediction of rotational spectroscopy parameters can be approached by a model based on density functional theory. Glycine and serine are employed to demonstrate that, despite its limited cost, such a model is able to predict rotational constants with an accuracy of 0.3% or better, thus paving the way toward the accurate characterization of larger flexible building blocks of biomolecules.

Benchmark Ab Initio Determination of the Conformers, Proton Affinities, and Gas-Phase Basicities of Cysteine

A systematic conformational mapping combined with literature data leads to 85 stable neutral cysteine conformers. The implementation of the same mapping process for the protonated counterparts reveals 21 N-(amino-), 64 O-(carbonyl-), and 37 S-(thiol-)protonated cysteine conformers. Their relative energies and harmonic vibrational frequencies are given at the MP2/aug-cc-pVDZ level of theory. Further benchmark ab initio computations are performed for the 10 lowest-lying neutral and protonated amino acid conformers (for each type) such as CCSD(T)-F12a/cc-pVDZ-F12 geometry optimizations (and frequency computations for cysteine) as well as auxiliary correction computations of the basis set effects up to CCSD(T)-F12b/cc-pVQZ-F12, electron correlation effects up to CCSDT(Q), core correlation effects, second-order Douglass-Kroll relativistic effects, and zero-point energy contributions. Boltzmann-averaged 0 (298.15) K proton affinity and [298.15 K gas-phase basicity] values of cysteine are predicted to be 214.96 (216.39) [208.21], 201.83 (203.55) [194.16], and 193.31 (194.74) [186.40] kcal/mol for N-, O-, and S-protonation, respectively, also considering the previously described auxiliary corrections.

A computational approach for modeling electronic circular dichroism of solvated chromophores

The present study consists in a novel computational protocol to model the UV-circular dichroism spectra of solvated species. It makes use of quantum-chemical calculations on a series of conformations of a flexible chromophore or on a series of chromophore/solvent clusters extracted from molecular dynamic simulations. The protocol is described and applied to the aqueous cationic tripeptide GAG⁺ and to the aqueous neutral decapeptide (GVGVP)₂ . The protocol has proven able to: (i) properly consider the conformational motion of solute in the given environment; (ii) give the actual statistical weight of each conformational state; (iii) provide a reliable quantum mechanical method able to reproduce the spectral features. Temperature effects on conformations and spectral properties are properly taken into account. The role of explicit solvent on the conformational analysis and the spectra calculation is discussed. The comparison of the calculated circular dichroism spectra with experimental ones recorded at different temperatures represents a strict validation test of the method.

Unbiased disentanglement of conformational baths with the help of microwave spectroscopy, quantum chemistry and artificial intelligence: the puzzling case of homocysteine

An integrated experimental-computational strategy for the accurate characterization of the conformational landscape of flexible biomolecule building blocks is proposed. This is based on the combination of rotational spectroscopy with quantum-chemical computations guided by artificial intelligence tools. The first step of the strategy is the conformer search and relative stability evaluation performed by means of an evolutionary algorithm. In this step, last generation semiempirical methods are exploited together with hybrid and double-hybrid density functionals. Next, the barriers ruling the interconversion between the low-lying conformers are evaluated in order to unravel possible fast relaxation paths. The relative stabilities and spectroscopic parameters of the ``surviving' conformers are then refined using state-of-the-art composite schemes. The reliability of the computational procedure is further improved by the inclusion of vibrational and thermal effects. The final step of the strategy is the comparison between experiment and theory without any ad hoc adjustment, which allows an unbiased assignment of the spectroscopic features in terms of different conformers and their spectroscopic parameters. The proposed approach has been tested and validated for homocysteine, a highly flexible non-proteinogenic alpha-amino acid. The synergism of the integrated strategy allowed the characterization of five conformers stabilized by bifurcated N-H-O=C hydrogen bonds, together with an additional conformer involving a more conventional HNH-O hydrogen bond. The stability order estimated from the experimental intensities as well as the number and type of conformers observed in the gas phase are in full agreement with the theoretical predictions. Analogously, a good match has been found for the spectroscopic parameters.

Molecular Dynamics Simulations Enforcing Nonperiodic Boundary Conditions: New Developments and Application to the Solvent Shifts of Nitroxide Magnetic Parameters

Multiscale methods combining quantum mechanics and molecular mechanics (QM/MM) have become the most suitable and effective strategies for the investigation of the spectroscopic properties of medium-to-large size chromophores in condensed phases. In this context, we are developing a novel workflow aimed at improving the generality, reliability, and ease of use of the available computational tools. In this paper, we report our latest developments with specific reference to a general protocol based on atomistic simulations, carried out under nonperiodic boundary conditions (NPBC). In particular, we add to our in house MD engine a new efficient treatment of mean field electrostatic contributions to energy and forces, together with the capability of performing the simulations either in the canonical (NVT) or in the isothermal–isobaric (NPT) ensemble. Next, we provide convincing evidence that the NBPC approach enhanced by specific tweaks for rigid body propagation, allows for the simulation of solute–solvent systems with a minimum number of degrees of freedom and large integration time step. After its validation, this new approach is applied to the challenging case of solvatochromic effects on the electron paramagnetic resonance (EPR) spectrum of a prototypical nitroxide radical. To this end, we propose and validate also an automated protocol to extract and weight simulation snapshots, making use of a continuous description of the strength of solute–solvent hydrogen bridges. While further developments are being worked on, the effectiveness of our approach, even in its present form, is demonstrated by the accuracy of the results obtained through an unsupervised approach characterized by a strongly reduced computational cost as compared to that of conventional QM/MM models, without any appreciable deterioration of accuracy.

TorsiFlex: an automatic generator of torsional conformers. Application to the twenty proteinogenic amino acids

In this work, we introduce TorsiFlex, a user-friendly software written in Python 3 and designed to find all the torsional conformers of flexible acyclic molecules in an automatic fashion. For the mapping of the torsional potential energy surface, the algorithm implemented in TorsiFlex combines two searching strategies: preconditioned and stochastic. The former is a type of systematic search based on chemical knowledge and should be carried out before the stochastic (random) search. The algorithm applies several validation tests to accelerate the exploration of the torsional space. For instance, the optimized structures are stored and this information is used to prevent revisiting these points and their surroundings in future iterations. TorsiFlex operates with a dual-level strategy by which the initial search is carried out at an inexpensive electronic structure level of theory and the located conformers are reoptimized at a higher level. Additionally, the program takes advantage of conformational enantiomerism, when possible. As a case study, and in order to exemplify the effectiveness and capabilities of this program, we have employed TorsiFlex to locate the conformers of the twenty proteinogenic amino acids in their neutral canonical form. TorsiFlex has produced a number of conformers that roughly doubles the amount of the most complete work to date.

Development and Validation of a Parameter-Free Model Chemistry for the Computation of Reliable Reaction Rates

A recently developed model chemistry (jun-Cheap) has been slightly modified and proposed as an effective, reliable, and parameter-free scheme for the computation of accurate reaction rates with special reference to astrochemical and atmospheric processes. Benchmarks with different sets of state-of-the-art energy barriers spanning a wide range of values show that, in the absence of strong multireference contributions, the proposed model outperforms the most well-known model chemistries, reaching a subchemical accuracy without any empirical parameter and with affordable computer times. Some test cases show that geometries, energy barriers, zero point energies, and thermal contributions computed at this level can be used in the framework of the master equation approach based on the ab initio transition-state theory for obtaining accurate reaction rates.

Integration of theory, simulation, artificial intelligence and virtual reality: a four-pillar approach for reconciling accuracy and interpretability in computational spectroscopy

The established pillars of computational spectroscopy are theory and computer based simulations. Recently, artificial intelligence and virtual reality are becoming the third and fourth pillars of an integrated strategy for the investigation of complex phenomena. The main goal of the present contribution is the description of some new perspectives for computational spectroscopy, in the framework of a strategy in which computational methodologies at the state of the art, high-performance computing, artificial intelligence and virtual reality tools are integrated with the aim of improving research throughput and achieving goals otherwise not possible. Some of the key tools (e.g., continuous molecular perception model and virtual multifrequency spectrometer) and theoretical developments (e.g., non-periodic boundaries, joint variational-perturbative models) are shortly sketched and their application illustrated by means of representative case studies taken from recent work by the authors. Some of the results presented are already well beyond the state of the art in the field of computational spectroscopy, thereby also providing a proof of concept for other research fields.

4-Fluoro-Threonine: From Diastereoselective Synthesis to pH-Dependent Conformational Equilibrium in Aqueous Solution

4-Fluoro-threonine, the only fluoro amino acid of natural origin discovered so far, is an interesting target for both synthetic and theoretical investigations. In this work, we lay the foundation for spectroscopic characterization of 4-fluoro-threonine. First, we report a diastereoselective synthetic route, which is suitable to produce synthetic material for experimental characterization. The addition of the commercially available ethyl isocyanoacetate to benzyloxyacetaldehyde led to the corresponding benzyloxy-oxazoline, which was hydrolyzed and transformed into ethyl (4S*,5S*)-5-hydroxymethyl-2-oxo-4-oxazolidinecarboxylate in a few steps. Fluorination with diethylamino sulfur trifluoride (DAST) afforded ethyl (4S*,5S*)-5-fluoromethyl-2-oxo-4-oxazolidinecarboxylate, which was deprotected to give the desired diastereomerically pure 4-fluoro-threonine, in 8-10% overall yield. With the synthetic material in our hands, acid-base titrations have been carried out to determine acid dissociation constants and the isoelectric point, which is the testing ground for the theoretical analysis. We have used machine learning coupled with quantum chemistry at the state-of-the-art to analyze the conformational space of 4-fluoro-threonine, with the aim of gaining insights from the comparison of computational and experimental results. Indeed, we have demonstrated that our approach, which couples a last-generation double-hybrid density functional including empirical dispersion contributions with a model combining explicit first-shell molecules and a polarizable continuum for describing solvent effects, provides results and trends in remarkable agreement with experiments. Finally, the conformational analysis applied to fluoro amino acids represents an interesting study for the effect of fluorine on the stability and population of conformers.

Machine learning meets chemical physics

Over recent years, the use of statistical learning techniques applied to chemical problems has gained substantial momentum. This is particularly apparent in the realm of physical chemistry, where the balance between empiricism and physics-based theory has traditionally been rather in favor of the latter. In this guest Editorial for the special topic issue on "Machine Learning Meets Chemical Physics," a brief rationale is provided, followed by an overview of the topics covered. We conclude by making some general remarks.