Prioritizing Virtual Screening with Interpretable Interaction Fingerprints.

Machine learning-based drug discovery success depends on molecular representation. Yet traditional molecular fingerprints omit both the protein and pointers back to structural information that would enable better model interpretability. Therefore, we propose LUNA, a Python 3 toolkit that calculates and encodes protein-ligand interactions into new hashed fingerprints inspired by Extended Connectivity FingerPrint (ECFP): EIFP (Extended Interaction FingerPrint), FIFP (Functional Interaction FingerPrint), and Hybrid Interaction FingerPrint (HIFP). LUNA also provides visual strategies to make the fingerprints interpretable. We performed three major experiments exploring the fingerprints' use. First, we trained machine learning models to reproduce DOCK3.7 scores using 1 million docked Dopamine D4 complexes. We found that EIFP-4,096 performed (R2 = 0.61) superior to related molecular and interaction fingerprints. Second, we used LUNA to support interpretable machine learning models. Finally, we demonstrate that interaction fingerprints can accurately identify similarities across molecular complexes that other fingerprints overlook. Hence, we envision LUNA and its interface fingerprints as promising methods for machine learning-based virtual screening campaigns. LUNA is freely available at https://github.com/keiserlab/LUNA.

Modeling the α- and ß-resorcinol phase boundary via combination of density functional theory and density functional tight-binding.

The ability to predict not only what organic crystal structures might occur but also the thermodynamic conditions under which they are the most stable would be extremely useful for discovering and designing new organic materials. The present study takes a step in that direction by predicting the temperature- and pressure-dependent phase boundary between the α and ß polymorphs of resorcinol using density functional theory (DFT) and the quasi-harmonic approximation. To circumvent the major computational bottleneck associated with computing a well-converged phonon density of states via the supercell approach, a recently developed approximation is employed, which combines a supercell phonon density of states from dispersion-corrected third-order density functional tight binding [DFTB3-D3(BJ)] with frequency corrections derived from a smaller B86bPBE-XDM functional DFT phonon calculation on the crystallographic unit cell. This mixed DFT/DFTB quasi-harmonic approach predicts the lattice constants and unit cell volumes to within 1%-2% at lower pressures. It predicts the thermodynamic phase boundary in almost perfect agreement with the experiment, although this excellent agreement does reflect fortuitous cancellation of errors between the enthalpy and entropy of transition.

Overcoming the difficulties of predicting conformational polymorph energetics in molecular crystals via correlated wavefunction methods.

Molecular crystal structure prediction is increasingly being applied to study the solid form landscapes of larger, more flexible pharmaceutical molecules. Despite many successes in crystal structure prediction, van der Waals-inclusive density functional theory (DFT) methods exhibit serious failures predicting the polymorph stabilities for a number of systems exhibiting conformational polymorphism, where changes in intramolecular conformation lead to different intermolecular crystal packings. Here, the stabilities of the conformational polymorphs of o-acetamidobenzamide, ROY, and oxalyl dihydrazide are examined in detail. DFT functionals that have previously been very successful in crystal structure prediction perform poorly in all three systems, due primarily to the poor intramolecular conformational energies, but also due to the intermolecular description in oxalyl dihydrazide. In all three cases, a fragment-based dispersion-corrected second-order Møller-Plesset perturbation theory (MP2D) treatment of the crystals overcomes these difficulties and predicts conformational polymorph stabilities in good agreement with experiment. These results highlight the need for methods which go beyond current-generation DFT functionals to make crystal polymorph stability predictions truly reliable.

Improving Predicted Nuclear Magnetic Resonance Chemical Shifts Using the Quasi-Harmonic Approximation.

Ab initio nuclear magnetic resonance chemical shift prediction plays an important role in the determination or validation of crystal structures. The ability to predict chemical shifts more accurately can translate to increased confidence in the resulting chemical shift or structural assignments. Standard electronic structure predictions for molecular crystal structures neglect thermal expansion, which can lead to an appreciable underestimation of the molar volumes. This study examines this volume error and its impact on 68 13C- and 28 15N-predicted chemical shifts taken from 20 molecular crystals. It assesses the ability to recover more realistic room-temperature crystal structures using the quasi-harmonic approximation and how refining the structures impacts the chemical shifts. Several pharmaceutical molecular crystals are also examined in more detail. On the whole, accounting for quasi-harmonic expansion changes the 13C and 15N chemical shifts by 0.5 and 1.0 ppm on average. This, in turn, reduces the root-mean-square errors relative to experiment by 0.3 ppm for 13C and 0.7 ppm for 15N. Although the statistical impacts are modest, changes in individual chemical shifts can reach multiple ppm. Accounting for thermal expansion in molecular crystal chemical shift prediction may not be needed routinely, but the systematic trend toward improved accuracy with the experiment could be useful in cases where discrimination between structural candidates is challenging, as in the pharmaceutical theophylline.

Identifying pragmatic quasi-harmonic electronic structure approaches for modeling molecular crystal thermal expansion.

Quasi-harmonic approaches provide an economical route to modeling the temperature dependence of molecular crystal structures and properties. Several studies have demonstrated good performance of these models, at least for rigid molecules, when using fragment-based approaches with correlated wavefunction techniques. Many others have found success employing dispersion-corrected density functional theory (DFT). Here, a hierarchy of models in which the energies, geometries, and phonons are computed either with correlated methods or DFT are examined to identify which combinations produce useful predictions for properties such as the molar volume, enthalpy, and entropy as a function of temperature. The results demonstrate that refining DFT geometries and phonons with single-point energies based on dispersion-corrected second-order Møller-Plesset perturbation theory can provide clear improvements in the molar volumes and enthalpies compared to those obtained from DFT alone. Predicted entropies, which are governed by vibrational contributions, benefit less clearly from the hybrid schemes. Using these hybrid techniques, the room-temperature thermochemistry of acetaminophen (paracetamol) is predicted to address the discrepancy between two experimental sublimation enthalpy measurements.

Theoretical predictions suggest carbon dioxide phases III and VII are identical.

Solid carbon dioxide exhibits a rich phase diagram at high pressures. Metastable phase III is formed by compressing dry ice above ∼10-12 GPa. Phase VII occurs at similar pressures but higher temperatures, and its stability region is disconnected from III on the phase diagram. Comparison of large-basis-set quasi-harmonic second-order Møller-Plesset perturbation theory calculations and experiment suggests that the long-accepted structure of phase III is problematic. The experimental phase III and VII structures both relax to the same phase VII structure. Furthermore, Raman spectra predicted for phase VII are in good agreement with those observed experimentally for both phase III and VII, while those for the purported phase III structure agree poorly with experimental observations. Crystal structure prediction is employed to search for other potential structures which might account for phase III, but none are found. Together, these results suggest that phases III and VII are likely identical.

Superlinear composition-dependent photocurrent in CVD-grown monolayer MoS2(1-x)Se2x alloy devices.

Transition metal dichalcogenides (TMDs) have emerged as a new class of two-dimensional materials that are promising for electronics and photonics. To date, optoelectronic measurements in these materials have shown the conventional behavior expected from photoconductors such as a linear or sublinear dependence of the photocurrent on light intensity. Here, we report the observation of a new regime of operation where the photocurrent depends superlinearly on light intensity. We use spatially resolved photocurrent measurements on devices consisting of CVD-grown monolayers of TMD alloys spanning MoS2 to MoSe2 to show the photoconductive nature of the photoresponse, with the photocurrent dominated by recombination and field-induced carrier separation in the channel. Time-dependent photoconductivity measurements show the presence of persistent photoconductivity for the S-rich alloys, while photocurrent measurements at fixed wavelength for devices of different alloy compositions show a systematic decrease of the responsivity with increasing Se content associated with increased linearity of the current-voltage characteristics. A model based on the presence of different types of recombination centers is presented to explain the origin of the superlinear dependence on light intensity, which emerges when the nonequilibrium occupancy of initially empty fast recombination centers becomes comparable to that of slow recombination centers.

Controlled argon beam-induced desulfurization of monolayer molybdenum disulfide.

Sputtering of MoS2 films of single-layer thickness by low-energy argon ions selectively reduces the sulfur content of the material without significant depletion of molybdenum. X-ray photoelectron spectroscopy shows little modification of the Mo 3d states during this process, suggesting the absence of significant reorganization or damage to the overall structure of the MoS2 film. Accompanying ab initio molecular dynamics simulations find clusters of sulfur vacancies in the top plane of single-layer MoS2 to be structurally stable. Measurements of the photoluminescence at temperatures between 175 and 300 K show quenching of almost 80% for an ~10% decrease in sulfur content.

What foods are US supermarkets promoting? A content analysis of supermarket sales circulars.

This study compared the types of foods advertised in supermarket newspaper circulars across geographic region (US Census regions: northeast [n=9], midwest [n=15], south [n=14], and west [n=13]), obesity-rate region (i.e., states with CDC adult obesity rates of <25% [n=14], 25 to <30% [n=24], and ≥ 30% [n=13]), and with MyPlate recommendations. All food advertisements on the first page of each circular were measured (±0.12-in.) to determine the proportion of space occupied and categorized according to food group. Overall, ≥ 50% of the front page of supermarket sales circulars was devoted to protein foods and grains; fruits, vegetables, and dairy, combined, were allocated only about 25% of the front page. The southern geographic region and the highest obesity-rate region both devoted significantly more advertising space to sweets, particularly sugar-sweetened beverages. The lowest obesity-rate region and western geographic region allocated the most space to fruits. Vegetables were allocated the least space in the western geographic region. Grains were the only food group represented in ads in proportions approximately equal to amounts depicted in the MyPlate icon. Protein foods exceeded and fruits, dairy, and vegetables fell below comparable MyPlate proportional areas. Findings suggest supermarket ads do not consistently emphasize foods that support healthy weight and MyPlate recommendations. More research is needed to determine how supermarket newspaper circulars can be used to promote healthy dietary patterns.