A deep learning approach for the blind logP prediction in SAMPL6 challenge

Water octanol partition coefficient serves as a measure for the lipophilicity of a molecule and is important in the field of drug discovery. A novel method for computational prediction of logarithm of partition coefficient (logP) has been developed using molecular fingerprints and a deep neural network. The machine learning model was trained on a dataset of 12,000 molecules and tested on 2000 molecules. In this article, we present our results for the blind prediction of logP for the SAMPL6 challenge. While the best submission achieved a RMSE of 0.41 logP units, our submission had a RMSE of 0.61 logP units. Overall, we ranked in the top quarter out of the 92 submissions that were made. Our results show that the deep learning model can be used as a fast, accurate and robust method for high throughput prediction of logP of small molecules.     

Quantum chemical predictions of water–octanol partition coefficients applied to the SAMPL6 log P blind challenge

Theoretical approaches for predicting physicochemical properties are valuable tools for accelerating the drug discovery process. In this work, quantum chemical methods are used to predict water–octanol partition coefficients as a part of the SAMPL6 blind challenge. The SMD continuum solvent model was employed with MP2 and eight DFT functionals in conjunction with correlation consistent basis sets to determine the water–octanol transfer free energy. Several tactics towards improving the predictions of the partition coefficient were examined, including increasing the quality of basis sets, considering tautomerization, and accounting for inhomogeneities in the water and n-octanol phases. Evaluation of these various schemes highlights the impact of modeling approaches across different methods. With the inclusion of tautomers and adjustments to the permittivity constants, the best predictions were obtained with smaller basis sets and the O3LYP functional, which yielded an RMSE of 0.79 logP units. The results presented correspond to the SAMPL6 logP submission IDs: DYXBT, O7DJK, and AHMTF.     

Prediction of octanol-water partition coefficients for the SAMPL6-$$log P$$logP molecules using molecular dynamics simulations with OPLS-AA,AMBER and CHARMM force fields

Journal of Computer-Aided Molecular Design - All-atom molecular dynamics simulations with stratified alchemical free energy calculations were used to predict the octanol-water partition coefficient...     

Prediction of cyclohexane-water distribution coefficients for the SAMPL5 data set using molecular dynamics simulations with the OPLS-AA force field

All-atom molecular dynamics simulations were used to predict water-cyclohexane distribution coefficients \(D_{cw}\) of a range of small molecules as part of the SAMPL5 blind prediction challenge. Molecules were parameterized with the transferable all-atom OPLS-AA force field, which required the derivation of new parameters for sulfamides and heterocycles and validation of cyclohexane parameters as a solvent. The distribution coefficient was calculated from the solvation free energies of the compound in water and cyclohexane. Absolute solvation free energies were computed by an established protocol using windowed alchemical free energy perturbation with thermodynamic integration. This protocol resulted in an overall root mean square error in \(\log D_{cw}\) of almost 4 log units and an overall signed error of ?3 compared to experimental data. There was no substantial overall difference in accuracy between simulating in NVT and NPT ensembles. The signed error suggests a systematic error but the experimental \(D_{cw}\) data on their own are insufficient to uncover the source of this error. Preliminary work suggests that the major source of error lies in the hydration free energy calculations.     

Prediction of hydration free energies for the SAMPL4 diverse set of compounds using molecular dynamics simulations with the OPLS-AA force field

All-atom molecular dynamics computer simulations were used to blindly predict the hydration free energies of a range of small molecules as part of the SAMPL4 challenge. Compounds were parametrized on the basis of the OPLS-AA force field using three different protocols for deriving partial charges: (1) using existing OPLS-AA atom types and charges with minor adjustments of partial charges on equivalent connecting atoms and derivation of new parameters for a number of distinct chemical groups (N-alkyl imidazole, nitrate) that were not present in the published force field; (2) calculation of quantum mechanical charges via geometry optimization, followed by electrostatic potential (ESP) fitting, using Jaguar at the LMP2/cc-pVTZ(-F) level; and (3) via geometry optimization and CHelpG charges (Gaussian09 at the HF/6-31G* level), followed by two-stage RESP fitting. The absolute hydration free energy was computed by an established protocol including alchemical free energy perturbation with thermodynamic integration. The use of standard OPLS-AA charges (protocol 1) with a number of newly parametrized charges and the use of histidine derived parameters for imidazole yielded an overall root mean square deviation of the prediction from the experimental data of 1.75 kcal/mol. The precision of our results appears to be mainly limited by relatively poor reproducibility of the Lennard-Jones contribution towards the solvation free energy, for which we observed large variability that could be traced to a strong dependence on the initial system conditions.     

Prediction of hydration free energies for the SAMPL4 data set with the AMOEBA polarizable force field

Hydration free energy calculations are often used to validate molecular simulation methodologies and molecular mechanics force fields. We use the free-energy perturbation method together with the AMOEBA polarizable force field and the Poltype parametrization protocol to predict the hydration free energies of 52 molecules as part of the SAMPL4 blind challenge. For comparison, similar calculations are performed using the non-polarizable General Amber force field. Against our expectations, the latter force field gives the better results compared to experiment. One possible explanation is the sensitivity of the AMOEBA results to the conformation used for parametrization.     

Mercury porosimetry in mesoporous glasses: a comparison of experiments with results from a molecular model

We present results from experiments and molecular modeling of mercury porosimetry into mesoporous Vycor and controlled pore glass (CPG) solid materials. The experimental intrusion/extrusion curves show a transition from a type H2 hysteresis for the Vycor glass to a type H1 hysteresis for the CPG. Mercury entrapment is observed in both materials, but we find that the amount of entrapped mercury depends on the chosen experimental relaxation time. No additional entrapment is found in a second intrusion/extrusion cycle, but hysteresis is still observed. This indicates that hysteresis and entrapment are of different origin. The experimental observations are qualitatively reproduced in theoretical calculations based on lattice models, which provide significant insights of the molecular mechanisms occurring during mercury porosimetry experiments in these porous glasses.     

Prediction of cyclohexane-water distribution coefficient for SAMPL5 drug-like compounds with the QMPFF3 and ARROW polarizable force fields

We present the performance of blind predictions of water—cyclohexane distribution coefficients for 53 drug-like compounds in the SAMPL5 challenge by three methods currently in use within our group. Two of them utilize QMPFF3 and ARROW, polarizable force-fields of varying complexity, and the third uses the General Amber Force-Field (GAFF). The polarizable FF’s are implemented in an in-house MD package, Arbalest. We find that when we had time to parametrize the functional groups with care (batch 0), the polarizable force-fields outperformed the non-polarizable one. Conversely, on the full set of 53 compounds, GAFF performed better than both QMPFF3 and ARROW. We also describe the torsion-restrain method we used to improve sampling of molecular conformational space and thus the overall accuracy of prediction. The SAMPL5 challenge highlighted several drawbacks of our force-fields, such as our significant systematic over-estimation of hydrophobic interactions, specifically for alkanes and aromatic rings.     

A comparison of different contractions for molecular calculations with gaussian-type functions

The effect of various possible contractions of a gaussian basis set is investigated for atomic and molecular calculations. The gaussian basis set used consists of 11s-type functions and 7p-type functions. Atomic calculations for the atoms Li to F are reported with fourteen different contractions of the s orbitals. The effect of the same contractions has also been investigated for molecular calculations of LiH, BH, CH₂, NH ₂ ^– , H₂O, and FH, together with the effect of the contraction for the p orbitals and for the s orbitals of the hydrogen atom. It is shown that the contraction in itself does not affect seriously the quality of a molecular calculation, but that a wrong choice of the contraction can produce a poor result.

---

Zusammenfassung Der Einfluß der verschiedensten Kontraktionen einer Basis von Gauß-Funktionen bei Berechnung atomarer oder molékularer Systeme wird untersucht. Diese Basis besteht zunächst aus 11 s-Funktionen und 7 p-Funktionen. 14 verschiedene Kontraktionen werden für die Atome Li bis F sowie für die Moleküle LiH, BH, CH₂, NH ₂ ^– , H₂O und FH getestet. Wie zu erwarten wird die Genauigkeit nicht wesentlich beeinflußt, solange man nur die richtigen Kontraktionen wählt.

---

Résumé On étudie l'effet des différentes contractions possibles d'une base de fonctions gaussiennes dans le cas d'atomes et de molécules. La base de fonctions gaussiennes comprend 11 fonctions du type s et 7 fonctions du type p. On donne les résultats de quatorze différentes contractions des orbitales s pour les atomes du Li à F ainsi que pour les molécules LiH, BH, CH₂, NH _- ² , H₂O et FH. L'effet de la contraction des orbitales p ainsi que des orbitales s des atomes d'hydrogène est également discuté pour les molécules CH₂ et H₂O. On montre que la contraction par elle-même n'affecte pas sensiblement les résultats obtenus a condition d'effectuer un choix judicieux parmi les contractions possibles.

     

Synthesis and comparison of the structure–property relationships of symmetric and asymmetric water‐soluble poly(p‐phenylene)s

Sodium salts of water‐soluble polymers poly{[2,5‐bis(3‐sulfonatopropoxy)‐1,4‐phenylene]‐alt‐[2,5‐bis(hexyloxy)‐1,4‐phenylene]} ( P1 ), poly{[2,5‐bis(3‐sulfonatopropoxy)‐1,4‐phenylene]‐alt‐[2,5‐bis(dodecyloxy)‐1,4‐phenylene]} ( P2 ), poly{[2,5‐bis(3‐sulfonatopropoxy)‐1,4‐phenylene]‐alt‐[2,5‐bis(dibenzyloxy)‐1,4‐phenylene]} ( P3 ), poly[2‐hexyloxy‐5‐(3‐sulfonatopropoxy)‐1,4‐phenylene] ( P4 ), and poly[2‐dodecyloxy‐5‐(3‐sulfonatopropoxy)‐1,4‐phenylene] ( P5 )] were synthesized with Suzuki coupling reactions and fully characterized. The first group of polymers ( P1 – P3 ) with symmetric structures gave lower absorption maxima [maximum absorption wavelength (λ_max) = 296–305 nm] and emission maxima [maximum emission wavelength (λ_em) = 361–398 nm] than asymmetric polymers P4 (λ_max = 329 nm, λ_em = 399 nm) and P5 (λ_max = 335 nm, λ_em = 401 nm). The aggregation properties of polymers P1 – P5 in different solvent mixtures were investigated, and their influence on the optical properties was examined in detail. Dynamic light scattering studies of the aggregation behavior of polymer P1 in solvents indicated the presence of aggregated species of various sizes ranging from 80 to 800 nm. The presence of alkoxy groups and 3‐sulfonatopropoxy groups on adjacent phenylene rings along the polymer backbone of the first set hindered the optimization of nonpolar interactions. The alkyl chain crystallization on one side of the polymer chain and the polar interactions on the other side allowed the polymers ( P4 and P5 ) to form a lamellar structure in the polymer lattice. Significant quenching of the polymer fluorescence upon the addition of positively charged viologen derivatives or cytochrome‐C was also observed. The quenching effect on the polymer fluorescence confirmed that the newly synthesized polymers could be used in the fabrication of biological and chemical sensors. © 2006 Wiley Periodicals, Inc. J Polym Sci Part A: Polym Chem 44: 3763–3777, 2006     

Structure–property relationships of poly(vinyl alcohol). II. The influence of molecular regularity on the crystallization–dissolution temperature relationships of poly(vinyl alcohol)

Dissolution temperatures T_s have been determined for poly(vinyl alcohol) (PVA) samples of varying tacticity as a function of crystallization temperatures T_c. From the values of T_s and T_c, one can obtain values of (T_m)_∞, the dissolution temperature of crystals of infinite stepheight. (T_m)_∞ is a characteristic property of a given sample. This method of characterization is very sensitive and reliable for detecting differences in molecular regularity among PVA samples. The variation of (T_m)_∞ with stereoregularity is attributed in part to differences in hydrogen-bonding characteristics. Determinations of the crystallinities of solution-crystallized PVA have shown that stereoregularity in PVA does not result in higher crystallizability.     

Molecular dynamics of 18-crown-6 complexes with alkali–metal cations and urea: Prediction of their conformations and comparison with data from the cambridge structural database

Complexes of 18-crown-6 with alkali–metal cations (Na⁺, K⁺, and Rb⁺), urea, and the uncomplexed crown ether were studied in vacuo with the molecular dynamics method. Conformational data from these calculations (simulation times in the range from 6–15 ns) was compared with information from the Cambridge Structural Database. Despite the differences in condition between the simulations and the solid state, a number of interesting similarities are observed. © 1993 John Wiley & Sons, Inc.     

Energy-conformation studies of HCN,HNC and CN−: A comparison of results from EH-SCC and SCF molecular orbital calculations

We have made an Extended Hückel Self Consistent Charge (EH-SCC) molecular orbital calculation for hydrogen cyanide, hydrogen isocyanide and cyanide ion. The main purpose of this calculation was to compare the EH-SCC and the more accurate SCF MO calculations for HCN in order to evaluate the method we used here for future use. Specifically, we have calculated and compared the following properties of HCN: total energy, binding energy, variation of ground state energy with geometric conformation, ionization potential and dipole moment. In addition, we have extended previous calculations of HCN by also considering its energy variation with bond angle for two excited state configurations and deducing some of the characteristics of its electronic spectra. Finally we have also made an MO calculation of the isocyanide isomer HNC and CN^– ion to compare with and add to the known characterization of the H, C, N, system.

---

Zusammenfassung Rechnungen nach der erweiterten Hückeltheorie werden für HCN, HNC und CN^– durchgeführt und mit ab initio Resultaten verglichen. Im einzelnen wurden Gesamtenergie, Bindungsenergie in Abhängigkeit von der geometrischen Struktur, Ionisierungspotential und Dipolmoment von HCN berechnet und außerdem die Energie für zwei doppelt angeregte Konfigurationen in Abhängigkeit vom Bindungswinkel bestimmt. Darüber hinaus sind MO-Rechnungen für HNC und CN^– gemacht worden.