Physicochemical Descriptors Research Articles

Structure-dependent effects of sweet and sweet taste affecting compounds on their sensorial properties

A reduction in sugar consumption is desirable from a health point of view. However, the sensory profiles of alternative sweet tasting compounds differ from sucrose regarding their temporal profile and undesired side tastes, reducing consumers’ acceptance. The present study describes a sensory characterization of a variety of sweet and sweet taste affecting compounds followed by a comparison of similarity to sucrose and a multivariate regression analysis to investigate structural determinants and possible interactions for the temporal profile of the sweetness and side-tastes. The results of the present study suggest a pivotal role for the number of ketones, aromatic rings, double bonds and the M LogP in the temporal profile of sweet and sweet taste affecting compounds. Furthermore, interactions between aggregated physicochemical descriptors demonstrate the complexity of the sensory response, which should be considered in future models to predict a comprehensive sensory profile of sweet and sweet taste affecting compounds.

Quantitative Structure-Activity Relationship Studies Of Series Of Chalcones Derivatives as Inhibitors Of Tumor Necrosis Factor-Alpha

Quantitative structure-activity relationship (QSAR) teqnique was used to predict the biological activity of a series of chalcones compounds as anti-inflammatory. 26 physicochemical descriptors are tested in QSAR equations configuration to predict biological effectiveness of compounds under study. The values of R2 in Eqs (1–3) ranged from 0.794–0.873, the F values ranged from 14.161–26.206 and the S values ranged from 0.262–0.334. The results demonstrated excellent models based on Eq.3, along with high of R2, F and minimum S by employing three parameters r(C3-C5), (LUMO+1) and (LUMO+2). This signifies that these parameters play a significant role in determining anti-inflammatory characteristics.

Cardiovascular Effects of Polychlorinated Biphenyls and Their Major Metabolites.

Background:Xenobiotic metabolism is complex, and accounting for bioactivation and detoxification processes of chemicals remains among the most challenging aspects for decision making with in vitro new approach methods data.Objectives:Considering the physiological relevance of human organotypic culture models and their utility for high-throughput screening, we hypothesized that multidimensional chemical-biological profiling of chemicals and their major metabolites is a sensible alternative for the toxicological characterization of parent molecules vs. metabolites in vitro.Methods:In this study, we tested 25 polychlorinated biphenyls (PCBs) [PCB 3, 11, 52, 126, 136, and 153 and their relevant metabolites (hydroxylated, methoxylated, sulfated, and quinone)] in concentration–response () for effects in human induced pluripotent stem cell (iPSC)-derived cardiomyocytes (CMs) and endothelial cells (ECs) (iPSC-derived and HUVECs). Functional phenotypic end points included effects on beating parameters and intracellular flux in CMs and inhibition of tubulogenesis in ECs. High-content imaging was used to evaluate cytotoxicity, mitochondrial integrity, and oxidative stress.Results:Data integration of a total of 19 physicochemical descriptors and 36 in vitro phenotypes revealed that chlorination status and metabolite class are strong predictors of the in vitro cardiovascular effects of PCBs. Oxidation of PCBs, especially to di-hydroxylated and quinone metabolites, was associated with the most pronounced effects, whereas sulfation and methoxylation of PCBs resulted in diminished bioactivity.Discussion:Risk characterization analysis showed that although in vitro derived effective concentrations exceeded the levels measured in the general population, risks cannot be ruled out due to the potential for population variability in susceptibility and the need to fill data gaps using read-across approaches. This study demonstrated a strategy for how in vitro data can be used to characterize human health risks from PCBs and their metabolites. https://doi.org/10.1289/EHP7030

Cytochrome-P450-Mediated Drug-Drug Interactions of Substrate Drugs: Assessing Clinical Risk Based on Molecular Properties and an Extended Clearance Classification System.

Cytochrome-P450 (P450) isoforms are major drug-metabolizing enzymes implicated in the clearance and drug-drug interactions (DDIs) of diverse small-molecule drugs. Here, we evaluated the association between primary physicochemical descriptors of substrate drugs and their clinical DDI risk with P450 index (probe) inhibitors using an exhaustive clinical data set (n = 397, substrate-inhibitor pairs). Additionally, the ability of extended clearance classification system (ECCS), a categorical clearance mechanism model, to predict P450 DDI risk was assessed. The clinical data set indicated that basic and neutral compounds are probable candidates to show a higher magnitude of DDIs on P450 inhibition (i.e., plasma exposure change > twofold). Additionally, trends with lipophilicity were apparent for P450-based DDIs. ECCS class 2 drugs (high-permeability bases/neutrals) have higher probability to show moderate-to-strong DDIs with probe inhibitors of CYP1A2/2C19/2C9/2D6/3A, while ECCS class 1A/1B drugs are prone to interactions with inhibitors of CYP2C8 and CYP2C9. On the other hand, P450-based DDIs are notably small for classes 3A/3B/4. In conclusion, this study emphasizes the relevance of the ECCS framework in clearance characterization to evaluate victim DDI liabilities and aid chemists in mitigating risk during drug design.

Elaboration of Novel TTK1 Inhibitory Leads via QSAR-Guided Selection of Crystallographic Pharmacophores Followed By In Vitro Assay.

Tyrosine threonine kinase (TTK1) is a key regulator of chromosome segregation. Recently, TTK targeting came into focus for the enhancement of possible anticancer therapies. In this regard, we employed our well-known method of QSAR-guided selection of the best crystallographic pharmacophore(s) to discover considerable binding interactions that transfer inhibitors into TTK1 binding site. Sixty-one TTK1 crystallographic complexes were used to extract 315 pharmacophore hypotheses. QSAR modeling was subsequently used to choose a single crystallographic pharmacophore that, when combined with other physicochemical descriptors, elucidates discrepancy in bioactivity of 55 miscellaneous inhibitors. The best QSAR model was robust and predictive (r2(55) = 0.75, r2LOO = 0.72 , r2press against external testing list of 12 compounds = 0.67), Standard error of estimate (training set) (S)= 0.63 , Standard error of estimate (testing set)(Stest) = 0.62. The resulting pharmacophore and QSAR models were used to scan the National Cancer Institute (NCI) database for new TTK1 inhibitors. Five hits confirmed significant TTK1 inhibitory profiles with IC50 values ranging between 11.7 and 76.6 mM.

QSAR-derived affinity fingerprints (part 2): modeling performance for potency prediction

Affinity fingerprints report the activity of small molecules across a set of assays, and thus permit to gather information about the bioactivities of structurally dissimilar compounds, where models based on chemical structure alone are often limited, and model complex biological endpoints, such as human toxicity and in vitro cancer cell line sensitivity. Here, we propose to model in vitro compound activity using computationally predicted bioactivity profiles as compound descriptors. To this aim, we apply and validate a framework for the calculation of QSAR-derived affinity fingerprints (QAFFP) using a set of 1360 QSAR models generated using Ki, Kd, IC50 and EC50 data from ChEMBL database. QAFFP thus represent a method to encode and relate compounds on the basis of their similarity in bioactivity space. To benchmark the predictive power of QAFFP we assembled IC50 data from ChEMBL database for 18 diverse cancer cell lines widely used in preclinical drug discovery, and 25 diverse protein target data sets. This study complements part 1 where the performance of QAFFP in similarity searching, scaffold hopping, and bioactivity classification is evaluated. Despite being inherently noisy, we show that using QAFFP as descriptors leads to errors in prediction on the test set in the ~ 0.65–0.95 pIC50 units range, which are comparable to the estimated uncertainty of bioactivity data in ChEMBL (0.76–1.00 pIC50 units). We find that the predictive power of QAFFP is slightly worse than that of Morgan2 fingerprints and 1D and 2D physicochemical descriptors, with an effect size in the 0.02–0.08 pIC50 units range. Including QSAR models with low predictive power in the generation of QAFFP does not lead to improved predictive power. Given that the QSAR models we used to compute the QAFFP were selected on the basis of data availability alone, we anticipate better modeling results for QAFFP generated using more diverse and biologically meaningful targets. Data sets and Python code are publicly available at https://github.com/isidroc/QAFFP_regression.

Comparison of statistical methods for predicting penetration capacity of drugs into human breast milk using physicochemical, pharmacokinetic and chromatographic descriptors

ABSTRACT In silico methods are often used for predicting pharmacokinetic properties of drugs due to their simplicity and cost-effectiveness. This study evaluates the penetration of 83 active pharmaceutical ingredients into human breast milk with an experimental milk-to-plasma ratio (M/P) obtained from the literature. Multiple linear regression (MLR), partial least squares (PLS) and random forest (RF) regression methods were compared to uncover the relationship between physicochemical, pharmacokinetic and membrane crossing properties of active pharmaceutical ingredients (APIs) using their rapid reference measurement value (Rf values), thin-layer chromatography (TLC) data from albumin-impregnated plates. Molecular descriptors of APIs proven to be important for their crossing into breast milk, including protein binding, ionisation state and lipophilicity and TLC data, have been included in the development of the prediction models. The best regression results were achieved by MLR (r 2 = 0.83 and r 2 = 0.86, n = 28) and RF (r 2 = 0.85, n = 58). In addition, the discriminant function analysis (DFA) was performed on acidic, basic and neutral drugs separately and showed a prediction accuracy of 93%, with M/P included as the discriminating variable.

Direct Comparison of Total Clearance Prediction: Computational Machine Learning Model versus Bottom-Up Approach Using In Vitro Assay

The in vitro-in vivo extrapolation (IVIVE) approach for predicting total plasma clearance (CLtot) has been widely used to rank order compounds early in discovery. More recently, a computational machine learning approach utilizing physicochemical descriptors and fingerprints calculated from chemical structure information has emerged, enabling virtual predictions even earlier in discovery. Previously, this approach focused more on in vitro intrinsic clearance (CLint) prediction. Herein, we directly compare these two approaches for predicting CLtot in rats. A structurally diverse set of 1114 compounds with known in vivo CLtot, in vitro CLint, and plasma protein binding was used as the basis for this evaluation. The machine learning models were assessed by validation approaches using the time- and cluster-split training and test sets, and five-fold cross validation. Assessed by five-fold validation, the random forest regression (RF) and radial basis function (RBF) models demonstrated better prediction performance in eight attempted machine learning models. The CLtot values predicted by the RF and RBF models were within two-fold of the observed values for 67.7 and 71.9% of cluster-split test set compounds, respectively, while the predictivity was worse in the time-split dataset. The predictivity of both models tended to be improved by incorporating in vitro parameters, unbound fraction in plasma (fu,p), and CLint. CLtot prediction utilizing in vitro CLint and the well-stirred model, correcting for the fraction unbound in blood, was substantially worse compared to machine learning approaches for the same cluster-split test set. The reason that CLtot is underestimated by IVIVE is not fully explained by considering the calculated microsomal unbound fraction (cfu,mic), extended clearance classification system (ECCS), and omitting high clearance compounds in excess of hepatic blood flow. The analysis suggests that in silico machine learning models may have the power to reduce reliance on or replace in vitro and in vivo studies for chemical structure optimization in early drug discovery.

Comprehensive quantum mechanical studies on three bioactive anastrozole based triazole analogues and their SERS active graphene complex

Three triazole derivatives, 2-[3-(2-cyanopropan-2-yl)-5-(1,2,4-triazol-1-ylmethyl)phenyl]-2-methylpropanenitrile (anastrozole) (TR1), 3-amino-1-benzyl-5-phenyl-1H-1,2,4-triazole (TR2) and 5-amino-1-benzyl-3-phenyl-1H-1,2,4-triazole (TR3) were analyzed for the structural, nonlinear optical, electronic and biological properties. The functional nature of the compounds were analyzed using vibrational spectra and was compared with the scaled, simulated spectra obtained using DFT with diffused orbitals. Relaxed potential energy scan predicts the stable conformers. Frontier molecular orbital was found and used to generate some important data pertaining to the reactivity and stability of the molecule. Time dependent DFT was used to model the excitation and de-excitation dynamics of these molecules and to predict the use of these molecules as effective photo sensitizer in DSSC. This work further discusses in detail, NBO study for intra molecular interactions, MEP for reactivity preferences and hyperpolarizability calculations for predicting the optical properties. Further molecular docking studies were conducted for the compounds with CYP2C9 inhibitor (PDBID: 4NZ2), cytochrome P450 inhibitor (4D75), tyrosine-protein kinase JAK2 (4YTC), estragen synthase (4GL5) to predict their utility as potential. The compounds were found to interact with graphene monolayer results shows that there is enhancement in various physico-chemical descriptors and surface enhanced Raman spectra (SERS).

Performance comparison of nonlinear and linear regression algorithms coupled with different attribute selection methods for quantitative structure - retention relationships modelling in micellar liquid chromatography

In micellar liquid chromatography (MLC), the addition of a surfactant to the mobile phase in excess is accompanied by an alteration of its solubilising capacity and a change in the stationary phase's properties. As an implication, the prediction of the analytes’ retention in MLC mode becomes a challenging task. Mixed Quantitative Structure – Retention Relationships (QSRR) modelling represents a powerful tool for estimating the analytes’ retention.This study compares 48 successfully developed mixed QSRR models with respect to their ability to predict retention of aripiprazole and its five impurities from molecular structures and factors that describe the Brij - acetonitrile system. The development of the models was based on an automatic combining of six attribute (feature) selection methods with eight predictive algorithms and the optimization of hyper-parameters. The feature selection methods included Principal Component Analysis (PCA), Non-negative Matrix Factorization (NMF), ReliefF, Multiple Linear Regression (MLR), Mutual Info and F-Regression. The series of investigated predictive algorithms comprised Linear Regressions (LR), Ridge Regression, Lasso Regression, Artificial Neural Networks (ANN), Support Vector Regression (SVR), Random Forest (RF), Gradient Boosted Trees (GBT) and K-Nearest neighbourhood (k-NN).A sufficient amount of data for building the model (78 cases in total) was provided by conducting 13 experiments for each of the 6 analytes and collecting the target responses afterwards. Different experimental settings were established by varying the values of the concentration of Brij L23, pH of the aqueous phase and acetonitrile content in the mobile phase according to the Box-Behnken design. In addition to the chromatographic parameters, the pool of independent variables was expanded by 27 molecular descriptors from all major groups (physicochemical, quantum chemical, topological and spatial structural descriptors). The best model was chosen by taking into consideration the Root Mean Square Error (RMSE) and cross-validation (CV) correlation coefficient (Q2) values.Interestingly, the comparative analysis indicated that a change in the set of input variables had a minor impact on the performance of the final models. On the other hand, different regression algorithms showed great diversity in the ability to learn patterns conserved in the data. In this regard, testing many regression algorithms is necessary in order to find the most suitable technique for model building. In the specific case, GBT-based models have demonstrated the best ability to predict the retention factor in the MLC mode. Steric factors and dipole-dipole interactions have proven to be relevant to the observed retention behaviour. This study, although being of a smaller scale, is a most promising starting point for comprehensive MLC retention prediction.

Solvent Selection Scheme Using Machine Learning Based on Physicochemical Description of Solvent Molecules: Application to Cyclic Organometallic Reaction

Abstract A solvent selection scheme for optimization of reactions is proposed using machine learning, based on the numerical descriptions of solvent molecules. Twenty-eight key solvents were represented using 17 physicochemical descriptors. Clustering analysis results implied that the descriptor represents the chemical characteristics of the solvent molecules. During the assessment of an organometallic reaction system, the regression analysis indicated that learning even a small number of experimental results can be useful for identifying solvents that will produce high experimental yields. Observation of the regression coefficients, and both clustering and regression analysis, can be effective when selecting a solvent to be used for an experiment.

A theoretical evaluation on free radical scavenging activity of 3-styrylchromone derivatives: the DFT study.

Chromone (4H-chromen-4-one, 4H-1-benzopyran-4-one) and related compounds are important pharmacophores and privileged structures in medicinal chemistry because of their important biological activities such as anti-tumor, anti-HIV, and antioxidant. In the study, the density functional theory (DFT) calculations were performed for radical scavenging activity evaluation of a series of 3-styrylchromone derivatives. The reaction enthalpies related to the steps in the radical scavenging action mechanisms and several physicochemical descriptors such as global hardness, softness, and electronegativity were computed in gas phase and in water. The solvation effect of water on the antioxidant activity was taken into account by using the conductor-like polarizable continuum model. The calculated results were discussed by considering all physicochemical properties of molecules: thermodynamic, orbital, and structural. The results obtained were consistent with the experimental results.

QSAR modeling, molecular docking and ADMET/pharmacokinetic studies: a chemometrics approach to search for novel inhibitors of norepinephrine transporter as potent antipsychotic drugs

Chemometrics study that relates biological activity to physicochemical descriptors of a molecule and the prediction of absorption, distribution, metabolism, excretion and toxicity (ADMET) properties in advance are important steps in drugs discovery. In this study, a chemometrics approach was employed on some molecules (inhibitors) of norepinephrine transporter to assess their inhibitory potencies, interactions with the receptor and predict their ADMET/pharmacokinetic properties for identification of novel antipsychotic drugs. The molecules were optimized by using density functional theory at the basis set of B3LYP/6-31G*. The genetic function algorithm technique was used to generate a statistically significant model with a good correlation coefficient R2Train = 0.952 Cross-validated coefficient Q2cv = 0.870, and adjusted squared correlation coefficient R2adj = 0.898. The molecular docking simulation using a neurotransmitter transporter receptor (PDB Code 2A65) revealed that three inhibitors (molecule No 38, 44 and 12) exhibited the highest binding affinity of − 10.3, − 9.9 and − 9.3 kcal/mol, respectively, were observed to inhibit the target by forming strong hydrogen bonds with hydrophobic interactions. The physicochemical and ADMET/pharmacokinetic properties result showed that these three molecules are orally bioavailable, high gastrointestinal absorption, good permeability and non-inhibitors of CYP3A4 and CYP2D6 except for molecule No 38. Also, Molecules No 38 and 44 proved to be non-substrate of P-glycoprotein and nontoxicity to a human ether-a-go-go-related gene with predicted hERG toxicity endpoints (pIC50 < 6) and low ADMET_Risk (< 7.0). The results of this study would provide physicochemical and pharmacokinetics properties needed to identify potent antipsychotic drugs and other relevant information in drug discovery.

Comprehensive investigation of selectivity landscape of glycogen synthase kinase-3 inhibitors

Interaction signatures of drug candidates are characteristic to off-target (neutral) and antitarget (negative) effects, inferring reduced efficiency, side-effects and high attrition rate. Today’s retroactive scaled-down virtual screening (VS) experiments relying on benchmarking datasets are extensively involved to assess ligand enrichment in the real-world problem. In recent years, unbiased benchmarking sets turned into a tremendous need to assist virtual screening methodologies for emerging drug targets. To date, the benchmarking datasets are quite limited, whereas glycogen synthase kinase-3 (GSK-3) is not included into directories of benchmarking datasets such as DUD-e, MUV, etc. Herein we introduced our in-house algorithm to build an unbiased benchmarking dataset, including highly selective, moderately selective and nonselective inhibitors for a significant therapeutic target - GSK-3, suitable for both ligand-based and structure-based VS approaches. These datasets are unbiased in terms of physico-chemical properties and topological descriptors, as resulted from mean(ROC–AUC) leave-one-out cross-validation (LOO CV). and additional 2 D similarity search. Moreover, we investigated the gradual selectivity dataset by application of multiple 2 D similarity coefficients and distances, 3 D similarity and docking. Besides the resulted links between the enrichment of selective GSK-3 inhibitors and their chemical structures, a database of compounds and their 3 D similarity signatures including cut-off thresholds for enhanced selectivity was generated. 2 D similarity space analysis revealed that selectivity problem cannot be evaluated appropriately with 2 D similarity searching alone. The current analysis provided useful, comprehensive insights, which may facilitate the knowledge-based identification of novel selective GSK-3 inhibitors. Communicated by Ramaswamy H. Sarma

Development of a Gaussian Process - feature selection model to characterise (poly)dimethylsiloxane (Silastic® ) membrane permeation.

The current study aims to determine the effect of physicochemical descriptor selection on models of polydimethylsiloxane permeation. A total of 2942 descriptors were calculated for a data set of 77 chemicals. Data were processed to remove redundancy, single values, imbalanced and highly correlated data, yielding 1363 relevant descriptors. For four independent test sets, feature selection methods were applied and modelled via a variety of Machine Learning methods. Two sets of molecular descriptors which can provide improved predictions, compared to existing models, have been identified. Best permeation predictions were found with Gaussian Process methods. The molecular descriptors describe lipophilicity, partial charge and hydrogen bonding as key determinants of PDMS permeation. This study highlights important considerations in the development of relevant models and in the construction and use of the data sets used in such studies, particularly that highly correlated descriptors should be removed from data sets. Predictive models are improved by the methodology adopted in this study, notably the systematic evaluation of descriptors, rather than simply using any and all available descriptors, often based empirically on in vitro experiments. Such findings also have clear relevance to a number of other fields.

Detailed quantum mechanical studies on bioactive benzodiazepine derivatives and their adsorption over graphene sheets.

Estazolam (Z1) and related derivatives, adinazolam (Z2), alprazolam (Z3), 4-hydroxyalprazolam (Z4) and triazolam (Z5) have been studied by using various computational tools to analyze their geometry and spectral characteristics. The compounds were found to interact with graphene monolayer results shows that there is enhancement in various physico-chemical descriptors and surface enhanced Raman spectra (SERS). The various reactive descriptors obtained from the FMO analysis predict the reactive nature of the compound. The various lone pair/sigma to pi conjugation was analyzed using NBO formalism, which provides valuable information about intra molecular electron transfer which is vital in predicting the inherent stability of the molecule. Simulated electronic spectra using TD-DFT and CAM-B3LYP functional are discussed in detail with respect to electronic transitions and light harvesting efficiency. Suitability of candidates as a photo sensitizer in dye sensitized solar cells was studied and 4-Hydroxyalprazolam is identified as a suitable candidate. Nucleophilic and electrophilic regions of the molecules are identified using MESP, which adds to the reactivity information. It can be seen that the highest interaction energy has been obtained in the case of the Z5-graphene system, while the lowest interaction energy has been obtained in the case of the Z1-graphene system. Docking indicates that the ligands adsorbed over graphene also form stable complexes with the receptors as indicated by the high binding affinity energy values.

Updating the portfolio of physicochemical descriptors related to permeability in the beyond the rule of 5 chemical space

Beyond rule of 5 (bRo5) molecules are attracting significant interest in modern drug discovery mostly because many novel targets require large and more flexible structures.The main aim of this paper is the identification of ad hoc bRo5 physicochemical descriptors of ionization, lipophilicity, polarity and chameleonicity and their measurement. We used different methods to collect ionization (pKa measures and log k’80 PLRP-S trends), lipophilicity (in octanol/water, in apolar systems and in biomimetic environments), polarity (Δlog Poct−tol, EPSA and Δlog KWIAM) and chameleonicity (ChameLogD) descriptors for 26 bRo5 drugs. A second aim was to check the relationship between physicochemical descriptors and permeability for a subset of compounds for which solid permeability values are reported in the literature.Results showed that the physicochemical profile in the bRo5 chemical space is often experimentally accessible, albeit more tools are required to overcome limitations of individual methods. For the investigated compounds, permeability is governed by Δlog Poct−tol and preliminary data support that chameleonicity could also have an impact.

Inclusion of molecular descriptors in predictive models improves pesticide soil-air partitioning estimates

The soil-air exchange of pesticides is one potential fate and exposure pathways, and this process is generally thought to be governed by soil properties and environmental conditions. The experimental determination of soil-air partitioning coefficient (Ksa) is laborious and costly and typically, Ksa’s are predicted from a semiempirical or a simple linear regression approach with soil and environmental variables. Here we developed a model that combined linear regression of soil, environmental and molecular parameters with the quantitative structural-property relationship (QSPR) to predict Ksa for pesticides. The values of theoretical descriptors of pesticides were calculated and the best descriptors selected using the Boruta Algorithm. Seventy-six experimental logKsa values for 17 pesticides were used in model development. Multiple linear regression (MLR) with a soil (organic carbon fraction), physicochemical (octanol-air partitioning coefficient), environmental (temperature and humidity) and molecular descriptor (Gmin, a 2D E-state molecular parameter), called as MLR-QSPR combined model exhibited better predictability (adj. r2 = 0.95) of logKsa compared to MLR (adj. r2 = 0.87) or QSPR (adj. r2 = 0.82) itself. MLR-QSPR also showed the best performance in five-fold cross-validation (adj. r2 = 0.94) and test set verification (adj. r2 = 0.96). The developed model was validated and characterized by the applicability domain. Results showed that the proposed MLR-QSPR approach is highly predictive and statistically robust with >95% of predictions within ±0.5 log unit of the measured Ksa. Therefore, this approach can be used in estimating the soil-air partitioning of pesticides to better predict it’s fate and transport in environments.

Polymorphism of monotropic forms: relationships between thermochemical and structural characteristics.

In this work, a database containing thermochemical and structural information about 208 monotropic polymorphic forms has been created and analyzed. Most of the identified compounds (77 cases) have been found to have two polymorphs, 14 compounds have three forms and there are only three examples of systems with four polymorphs. The analysis of density distribution within the database has revealed that only 62 out of 114 metastable polymorphs (referred to as group I) obey the `density rule' proposed by Burger and Ramberger [(1979), Mikrochim. Acta, 72, 259-271], while the remaining 45% of the monotropic systems (group II) violate the rule. A number of physicochemical, structural and molecular descriptors have been used to find and highlight the differences between group I and group II of the polymorphs. Group II is characterized (on average) by higher values of descriptors, which are responsible for conformational flexibility of molecules. An algorithm has been proposed for carrying out bivariate statistical analysis. It implies partitioning the database into structurally related clusters based on Tanimoto similarity coefficients and subsequent analysis of each cluster in terms of the number of hydrogen bonds per molecule.

Prediction of peptide binding to MHC using machine learning with sequence and structure-based feature sets

Selecting peptides that bind strongly to the major histocompatibility complex (MHC) for inclusion in a vaccine has therapeutic potential for infections and tumors. Machine learning models trained on sequence data exist for peptide:MHC (p:MHC) binding predictions. Here, we train support vector machine classifier (SVMC) models on physicochemical sequence-based and structure-based descriptor sets to predict peptide binding to a well-studied model mouse MHC I allele, H-2Db. Recursive feature elimination and two-way forward feature selection were also performed. Although low on sensitivity compared to the current state-of-the-art algorithms, models based on physicochemical descriptor sets achieve specificity and precision comparable to the most popular sequence-based algorithms. The best-performing model is a hybrid descriptor set containing both sequence-based and structure-based descriptors. Interestingly, close to half of the physicochemical sequence-based descriptors remaining in the hybrid model were properties of the anchor positions, residues 5 and 9 in the peptide sequence. In contrast, residues flanking position 5 make little to no residue-specific contribution to the binding affinity prediction. The results suggest that machine-learned models incorporating both sequence-based descriptors and structural data may provide information on specific physicochemical properties determining binding affinities.