Halogen Bonding Hot Spots as a Constraint in Virtual Screening: A Case Study of 5-HT7R.

The recently developed and used molecular modeling approach to search for privileged amino acids for halogen bonding (XB hot spots) through XSAR sets has been applied to 5-HT7R. Herein, among all identified 5-HT7R XB hot spots, the S5x42 was employed in a virtual screening protocol as a constraint. Through a designed virtual screening protocol, 63 XSAR sets (156 compounds) were selected from more than 8 million commercially available compounds and examined using in vitro assay toward 5-HT7R. A 68% accuracy was found in predicting halogenated derivatives with higher affinity for 5-HT7R than their unsubstituted analogs. Moreover, it was observed that a halogen bond formed between S5x42 and a chlorine atom at the 3-position of the arylpiperazine fragment caused the most remarkable, 35.4-fold increase in binding affinity for 5-HT7R when compared to the nonhalogenated analog. Interestingly, molecular dynamics simulations showed the formation of a bifurcated halogen bond with S5x42.

5-HT6 receptor neutral antagonists protect astrocytes: A lesson from 2-phenylpyrrole derivatives.

The serotonin type 6 receptor (5-HT6R) displays a strong constitutive activity, suggesting it participates largely in the physiological and pathological processes controlled by the receptor. The active states of 5-HT6R engage particular signal transduction pathways that lead to different biological responses. In this study, we present the development of 5-HT6R neutral antagonists at Gs signaling built upon the 2-phenylpyrrole scaffold. Using molecular dynamics simulations, we outline the relationship between the exposure of the basic center of the molecules and their ability to target the agonist-activated state of the receptor. Our study identifies compound 30 as a potent and selective neutral antagonist at 5-HT6R-operated Gs signaling. Furthermore, we demonstrate the cytoprotective effects of 30 and structurally diverse 5-HT6R neutral antagonists at Gs signaling in C8-D1A cells and human astrocytes exposed to rotenone. This effect is not observed for 5-HT6R agonists or inverse agonists. In light of these findings, we propose compound 30 as a valuable molecular probe to study the biological effects associated with the agonist-activated state of 5-HT6R and provide insight into the glioprotective properties of 5-HT6R neutral antagonists at Gs signaling.

From NMR to AI: Designing a Novel Chemical Representation to Enhance Machine Learning Predictions of Physicochemical Properties.

A novel approach to the utilization of nuclear magnetic resonance (NMR) spectroscopy data in the prediction of logD through machine learning algorithms is shown. In the analysis, a data set of 754 chemical compounds, organized into 30 clusters, was evaluated using advanced machine learning models, such as Support Vector Regression (SVR), Gradient Boosting, and AdaBoost, and comprehensive validation and testing methods were employed, including 10-fold cross-validation, bootstrapping, and leave-one-out. The study revealed the superior performance of the Bucket Integration method for dimensionality reduction, consistently yielding the lowest root mean square error (RMSE) across all data sets and normalization schemes. The SVR prediction models demonstrated remarkable computational efficiency and low cost, with the best RMSE value reaching 0.66. Our best model outperformed existing tools like JChem Suite's logD Predictor (0.91) and CplogD (1.27), and a comparison with traditional molecular representations yielded a comparable RMSE (0.50), emphasizing the robustness of our NMR data integration. The widespread availability of NMR data in pharmaceutical and industrial research presents an untapped resource for predictive modeling, highlighting the need for accessible methodologies like ours that complement the analytical toolbox beyond conventional 2D approaches. Our approach, designed to leverage the rich spatial data from NMR spectroscopy, provides additional insights and enriches drug discovery and computational chemistry with a freely accessible tool.

The Pivotal Distinction between Antagonists' and Agonists' Binding into Dopamine D4 Receptor-MD and FMO/PIEDA Studies.

The dopamine D4 receptor (D4R) is a promising therapeutic target in widespread diseases, and the search for novel agonists and antagonists appears to be clinically relevant. The mechanism of binding to the receptor (R) for antagonists and agonists varies. In the present study, we conducted an in-depth computational study, teasing out key similarities and differences in binding modes, complex dynamics, and binding energies for D4R agonists and antagonists. The dynamic network method was applied to investigate the communication paths between the ligand (L) and G-protein binding site (GBS) of human D4R. Finally, the fragment molecular orbitals with pair interaction energy decomposition analysis (FMO/PIEDA) scheme was used to estimate the binding energies of L-R complexes. We found that a strong salt bridge with D3.32 initiates the inhibition of the dopamine D4 receptor. This interaction also occurs in the binding of agonists, but the change in the receptor conformation to the active state starts with interaction with cysteine C3.36. Such a mechanism may arise in the case of agonists unable to form a hydrogen bond with the serine S5.46, considered, so far, to be crucial in the activation of GPCRs. The energy calculations using the FMO/PIEDA method indicate that antagonists show higher residue occupancy of the receptor binding site than agonists, suggesting they could form relatively more stable complexes. Additionally, antagonists were characterized by repulsive interactions with S5.46 distinguishing them from agonists.

Superiority of the Triple-Acting 5-HT6R/5-HT3R Antagonist and MAO-B Reversible Inhibitor PZ-1922 over 5-HT6R Antagonist Intepirdine in Alleviation of Cognitive Deficits in Rats.

The multifactorial origin and neurochemistry of Alzheimer's disease (AD) call for the development of multitarget treatment strategies. We report a first-in-class triple acting compound that targets serotonin type 6 and 3 receptors (5-HT-Rs) and monoamine oxidase type B (MAO-B) as an approach for treating AD. The key structural features required for MAO-B inhibition and 5-HT6R antagonism and interaction with 5-HT3R were determined using molecular dynamic simulations and cryo-electron microscopy, respectively. Bioavailable PZ-1922 reversed scopolamine-induced cognitive deficits in the novel object recognition test. Furthermore, it displayed superior pro-cognitive properties compared to intepirdine (a 5-HT6R antagonist) in the AD model, which involved intracerebroventricular injection of an oligomeric solution of amyloid-ß peptide (oAß) in the T-maze test in rats. PZ-1922, but not intepirdine, restored levels of biomarkers characteristic of the debilitating effects of oAß. These data support the potential of a multitarget approach involving the joint modulation of 5-HT6R/5-HT3R/MAO-B in AD.

Hydrophobicity modulation via the substituents at positions 2 and 4 of 1,3,5-triazine to enhance therapeutic ability against Alzheimer's disease for potent serotonin 5-HT6R agents.

Alzheimer's disease (AD), a neurodegenerative disorder with a complex aetiology, is the most common memory dysfunction particularly affecting the elderly. Various protein targets have been classified to be involved in the AD treatment, including 5-HT6 receptor (5-HT6R). So far, the 5-HT6R ligands obtained by our research group have become a good basis for hydrophobicity modulation to give a chance for more effective action toward AD by additional influence on target enzymes, e.g. cyclin-dependent kinase 5 (CDK5). In the search for 5-HT6R agents with additional inhibitory action on the enzyme, a series of 25 new 1,3,5-triazines (7-31) as modifications of lead, 4-[1-(2,5-dichlorophenoxy)propyl]-6-(4-methylpiperazin-1-yl)-1,3,5-triazine-2-amine (6), was rationally designed. Molecular modelling, synthesis, crystallographic studies, in vitro biological assays and behavioral studies in vivo were performed. The new triazines showed high affinity (Ki < 100 nM) and selectivity for 5-HT6R. The most effective one, 4-[1-(2,5-difluorophenoxy)propyl]-6-(4-methylpiperazin-1-yl)-1,3,5-triazine-2-amine (8), exhibited the strong antagonistic action towards 5-HT6R (Ki = 5 nM, pKb = 8.16), had an impact on the memory processes in the Novel Object Recognition test and displayed anxiolytic-like activity in the Elevated Plus Maze test in rats. Moreover, it had the antiplatelet effect as well as very good permeability (PAMPA model), high metabolic stability (RLMs) and satisfactory safety in vitro. Although the CDK5 inhibitory effects in vitro for the tested compounds (8, 10, 14, 18, 26-31) missed the potency expected from in silico simulations, the novel antagonist (8) with a very satisfying pharmacological and ADMET profile can serve as a new lead structure in further searches for innovative therapy against AD with accompanying symptoms.

Synthesis, computational and experimental pharmacological studies for (thio)ether-triazine 5-HT6R ligands with noticeable action on AChE/BChE and chalcogen-dependent intrinsic activity in search for new class of drugs against Alzheimer's disease.

Alzheimer's disease is becoming a growing problem increasing at a tremendous rate. Serotonin 5-HT6 receptors appear to be a particularly attractive target from a therapeutic perspective, due to their involvement not only in cognitive processes, but also in depression and psychosis. In this work, we present the synthesis and broad biological characterization of a new series of 18 compounds with a unique 1,3,5-triazine backbone, as potent 5-HT6 receptor ligands. The main aim of this research is to compare the biological activity of the newly synthesized sulfur derivatives with their oxygen analogues and their N-demethylated O- and S-metabolites obtained for the first time. Most of the new triazines displayed high affinity (Ki < 200 nM) and selectivity towards 5-HT6R, with respect to 5-HT2AR, 5-HT7R, and D2R, in the radioligand binding assays. For selected, active compounds crystallographic studies, functional bioassays, and ADME-Tox profile in vitro were performed. The exciting novelty is that the sulfur derivatives exhibit an agonistic mode of action contrary to all other compounds obtained to date in this chemical class herein and previously reported. Advanced computational studies indicated that this intriguing functional shift might be caused by presence of chalcogen bonds formed only by the sulfur atom. In addition, the N-demethylated derivatives have emerged highly potent antioxidants and, moreover, show a significant improvement in metabolic stability compared to the parent structures. The cholinesterase study present micromolar inhibitory AChE and BChE activity for both 5-HT6 agonist 19 and potent antagonist 5. Finally, the behavioral experiments of compound 19 demonstrated its antidepressant-like properties and slight ability to improve cognitive deficits, without inducing memory impairments by itself. Described pharmacological properties of both compounds (5 and 19) allow to give a design clue for the development of multitarget compounds with 5-HT6 (both agonist and antagonist)/AChE and/or BChE mechanism in the group of 1,3,5-triazine derivatives.

Synthesis, computational simulations and biological evaluation of new dual 5HT1A/5HT7 receptor ligands based on purine-2,6-dione scaffold.

The new dual 5HT1A/5HT7 receptor ligands were designed based on the purine-2,6-dione scaffold with the fluorine atom. Twenty-one new derivatives were synthesized, and their structure-activity relationship was summarized. Compound 11 (7-(2-(3-fluorophenyl)-2-oxoethyl)-8-((4-(4-(2-methoxyphenyl)piperazin-1-yl)butyl)amino)-1,3-dimethyl-3,7-dihydro-1H-purine-2,6-dione) showed the highest affinity to 5HT1AR and 5HT7R, and was the most potent antagonist of 5-HT1AR (Kb = 0.26 ± 0.1 nM) which activity can be to reference compound NAN-190 (Kb = 0.26 ± 0.1 nM). The experimentally established physicochemical parameters of compound 11 showed that compound, as slightly ionized in the blood, could penetrate the blood-brain barrier. A molecular docking study showed that the fluorine substitution introduces additional stabilization effects on binding to 5HT1A/5HT7Rs. In animal assays of depression and anxiety, compound 11 revealed activity in terms of dosage compared to marketed psychotropics such as fluoxetine, citalopram, and sertraline.

Tuning the Biological Activity of PI3Kδ Inhibitor by the Introduction of a Fluorine Atom Using the Computational Workflow.

As a member of the class I PI3K family, phosphoinositide 3-kinase δ (PI3Kδ) is an important signaling biomolecule that controls immune cell differentiation, proliferation, migration, and survival. It also represents a potential and promising therapeutic approach for the management of numerous inflammatory and autoimmune diseases. We designed and assessed the biological activity of new fluorinated analogues of CPL302415, taking into account the therapeutic potential of our selective PI3K inhibitor and fluorine introduction as one of the most frequently used modifications of a lead compound to further improve its biological activity. In this paper, we compare and evaluate the accuracy of our previously described and validated in silico workflow with that of the standard (rigid) molecular docking approach. The findings demonstrated that a properly fitted catalytic (binding) pocket for our chemical cores at the induced-fit docking (IFD) and molecular dynamics (MD) stages, along with QM-derived atomic charges, can be used for activity prediction to better distinguish between active and inactive molecules. Moreover, the standard approach seems to be insufficient to score the halogenated derivatives due to the fixed atomic charges, which do not consider the response and indictive effects caused by fluorine. The proposed computational workflow provides a computational tool for the rational design of novel halogenated drugs.

Impact of the Substitution Pattern at the Basic Center and Geometry of the Amine Fragment on 5-HT6 and D3R Affinity in the 1H-Pyrrolo[3,2-c]quinoline Series.

Salt bridge (SB, double-charge-assisted hydrogen bonds) formation is one of the strongest molecular non-covalent interactions in biological systems, including ligand-receptor complexes. In the case of G-protein-coupled receptors, such an interaction is formed by the conserved aspartic acid (D3.32) residue and the basic moiety of the aminergic ligand. This study aims to determine the influence of the substitution pattern at the basic nitrogen atom and the geometry of the amine moiety at position 4 of 1H-pyrrolo[3,2-c]quinoline on the quality of the salt bridge formed in the 5-HT6 receptor and D3 receptor. To reach this goal, we synthetized and biologically evaluated a new series of 1H-pyrrolo[3,2-c]quinoline derivatives modified with various amines. The selected compounds displayed a significantly higher 5-HT6R affinity and more potent 5-HT6R antagonist properties when compared with the previously identified compound PZ-1643, a dual-acting 5-HT6R/D3R antagonist; nevertheless, the proposed modifications did not improve the activity at D3R. As demonstrated by the in silico experiments, including molecular dynamics simulations, the applied structural modifications were highly beneficial for the formation and quality of the SB formation at the 5-HT6R binding site; however, they are unfavorable for such interactions at D3R.

Isomeric Activity Cliffs-A Case Study for Fluorine Substitution of Aminergic G Protein-Coupled Receptor Ligands.

Currently, G protein-coupled receptors (GPCRs) constitute a significant group of membrane-bound receptors representing more than 30% of therapeutic targets. Fluorine is commonly used in designing highly active biological compounds, as evidenced by the steadily increasing number of drugs by the Food and Drug Administration (FDA). Herein, we identified and analyzed 898 target-based F-containing isomeric analog sets for SAR analysis in the ChEMBL database-FiSAR sets active against 33 different aminergic GPCRs comprising a total of 2163 fluorinated (1201 unique) compounds. We found 30 FiSAR sets contain activity cliffs (ACs), defined as pairs of structurally similar compounds showing significant differences in affinity (≥50-fold change), where the change of fluorine position may lead up to a 1300-fold change in potency. The analysis of matched molecular pair (MMP) networks indicated that the fluorination of aromatic rings showed no clear trend toward a positive or negative effect on affinity. Additionally, we propose an in silico workflow (including induced-fit docking, molecular dynamics, quantum polarized ligand docking, and binding free energy calculations based on the Generalized-Born Surface-Area (GBSA) model) to score the fluorine positions in the molecule.

The synthesis of novel thioderivative chalcones and their influence on NF-κB, STAT3 and NRF2 signaling pathways in colorectal cancer cells.

This study aimed to synthesize new thioderivative chalcones and analyze their impact on the NF-κB, STAT3, EGFR and Nrf2 signaling pathways in colorectal cancer cells. Among the studied compounds, derivatives 4 and 5 decreased the activation of NF-κB and the expression of the target gene COX-2. In the case of STAT3, we observed the inhibition of activation of this signaling pathway after influencing derivative 4. Increased activation of the Nrf2 signaling pathway was demonstrated for derivatives 5 and 7 in DLD-1 and HCT116 cells. The results of this study indicated that new chalcone derivatives, especially compounds 4, 5, and-to some degree-7, possess potential applications in the prevention of colorectal cancer.

Fast and Noninvasive Hair Test for Preliminary Diagnosis of Mood Disorders.

The main objective of this study was to develop a test for the fast and noninvasive prediagnosis of mood disorders based on the noninvasive analysis of hair samples. The database included 75 control subjects (who were not diagnosed with depression) and 40 patients diagnosed with mood disorders such as depression or bipolar disorder. Both women and men, aged 18-65 years, participated in the research. After taking the hair samples, they were washed (methanol-water-methanol by shaking in a centrifuge for two min) and air-dried in a fume hood. Each hair collection was analyzed using Fourier transform infrared spectroscopy attenuated total reflection (ATR-FTIR) spectroscopy. Subsequently, the results obtained were analyzed based on chemometric methods: hierarchical cluster analysis (HCA) and principal component analysis (PCA). As a results of the research conducted, potential differences were noticed. There was a visible change in the spectra intensity at around 2800-3100 cm-1 and smaller differences around 1460 cm-1; the bands can be assigned to protein vibrations. However, these are preliminary studies that provide a good basis for the development of a test for the initial diagnosis of mood disorders.

Mining anion-aromatic interactions in the Protein Data Bank.

Mutual positioning and non-covalent interactions in anion-aromatic motifs are crucial for functional performance of biological systems. In this context, regular, comprehensive Protein Data Bank (PDB) screening that involves various scientific points of view and individual critical analysis is of utmost importance. Analysis of anions in spheres with radii of 5 Å around all 5- and 6-membered aromatic rings allowed us to distinguish 555 259 unique anion-aromatic motifs, including 92 660 structures out of the 171 588 structural files in the PDB. The use of a scarcely exploited (x, h) coordinate system led to (i) identification of three separate areas of motif accumulation: A - over the ring, B - over the ring-substituent bonds, and C - roughly in the plane of the aromatic ring, and (ii) unprecedented simultaneous comparative description of various anion-aromatic motifs located in these areas. Of the various residues considered, i.e. aminoacids, nucleotides, and ligands, the latter two exhibited a considerable tendency to locate in region Avia archetypal anion-π contacts. The applied model not only enabled statistical quantitative analysis of space around the ring, but also enabled discussion of local intermolecular arrangements, as well as detailed sequence and secondary structure analysis, e.g. anion-π interactions in the GNRA tetraloop in RNA and protein helical structures. As a purely practical issue of this work, the new code source for the PDB research was produced, tested and made freely available at https://github.com/chemiczny/PDB_supramolecular_search.

Overcoming undesirable hERG affinity by incorporating fluorine atoms: A case of MAO-B inhibitors derived from 1 H-pyrrolo-[3,2-c]quinolines.

The incorporation of the fluorine motif is a strategy widely applied in drug design for modulating the activity, physicochemical parameters, and metabolic stability of chemical compounds. In this study, we attempted to reduce the affinity for ether-à-go-go-related gene (hERG) channel by introducing fluorine atoms in a group of 1H-pyrrolo[3,2-c]quinolines that are capable of inhibiting monoamine oxidase type B (MAO-B). A series of structural modifications guided by in vitro evaluation of MAO-B inhibition and antitargeting for hERG channels were performed, which led to the identification of 1-(3-chlorobenzyl)-4-(4,4-difluoropiperidin-1-yl)-1H-pyrrolo[3,2-c]quinoline (26). Compound 26 acted as a reversible MAO-B inhibitor exhibiting selectivity over 45 targets, enzymes, transporters, and ion channels, and showed potent glioprotective properties in cultured astrocytes. In addition, the compound demonstrated good metabolic stability in rat liver microsomes assay, a favorable safety profile, and brain permeability. It also displayed procognitive effects in the novel object recognition test in rats and antidepressant-like activity in forced swim test in mice. The findings of the study suggest that reversible MAO-B inhibitors can have potential therapeutic applications in Alzheimer's disease.

An exit beyond the pharmacophore model for 5-HT6R agents - a new strategy to gain dual 5-HT6/5-HT2A action for triazine derivatives with procognitive potential.

This research allowed us to find the first highly potent 5-HT6/5-HT2A receptor (5-HT6/5-HT2AR) dual antagonists in a group of 1,3,5-triazine compounds as a result of an exit beyond the hydrophobic feature of the pharmacophore model for 5-HT6R antagonists. Design and synthesis of the series (2-16) of new O- and S-containing ether derivatives of 1,3,5-triazines with the double-ring aromatic region have been performed. The new compounds were examined within the comprehensive pharmacological screening, including: radioligand binding assays, functional and ADMET studies in vitro as well as behavioral tests in rats. Crystallographic aspects and computer-aided structure-activity relationship were analyzed, as well. The comprehensive approach led to selection of compound 12 (4-(4-methylpiperazin-1-yl)-6-(2-(naphthalen-2-ylthio)propan-2-yl)-1,3,5-triazin-2-amine) with the most significant dual 5-HT6/5-HT2AR antagonistic action (5-HT6R Ki = 11 nM, 5-HT2AR Ki = 39 nM). Moreover, the compound 12 has satisfactory ADMETox properties in vitro, i.e.: the high permeability through biological membranes, high metabolic stability, neither mutagenic nor hepatotoxic effects, and moderate ability to inhibit CYP3A4. Above all, 12 showed ability to reverse the pharmacologically-induced (MK-801) memory impairment at low doses (1-3 mg/kg) in Novel Object Recognition (NOR) test in rats. Our results indicate a promising potency of dual 5-HT6/5-HT2AR antagonism in the search for novel strategy to fight Alzheimer's disease, which remains an unmet clinical need.

Hydrogen Bonds with Fluorine in Ligand-Protein Complexes-the PDB Analysis and Energy Calculations.

Fluorine is a common substituent in medicinal chemistry and is found in up to 50% of the most profitable drugs. In this study, a statistical analysis of the nature, geometry, and frequency of hydrogen bonds (HBs) formed between the aromatic and aliphatic C-F groups of small molecules and biological targets found in the Protein Data Bank (PDB) repository was presented. Interaction energies were calculated for those complexes using three different approaches. The obtained results indicated that the interaction energy of F-containing HBs is determined by the donor-acceptor distance and not by the angles. Moreover, no significant relationship between the energies of HBs with fluorine and the donor type was found, implying that fluorine is a weak HB acceptor for all types of HB donors. However, the statistical analysis of the PDB repository revealed that the most populated geometric parameters of HBs did not match the calculated energetic optima. In a nutshell, HBs containing fluorine are forced to form due to the stronger ligand-receptor neighboring interactions, which make fluorine the "donor's last resort".

Data-Driven Analysis of Fluorination of Ligands of Aminergic G Protein Coupled Receptors.

Currently, G protein-coupled receptors are the targets with the highest number of drugs in many therapeutic areas. Fluorination has become a common strategy in designing highly active biological compounds, as evidenced by the steadily increasing number of newly approved fluorine-containing drugs. Herein, we identified in the ChEMBL database and analysed 1554 target-based FSAR sets (non-fluorinated compounds and their fluorinated analogues) comprising 966 unique non-fluorinated and 2457 unique fluorinated compounds active against 33 different aminergic GPCRs. Although a relatively small number of activity cliffs (defined as a pair of structurally similar compounds showing significant differences of activity -ΔpPot > 1.7) was found in FSAR sets, it is clear that appropriately introduced fluorine can increase ligand potency more than 50-fold. The analysis of matched molecular pairs (MMPs) networks indicated that the fluorination of the aromatic ring showed no clear trend towards a positive or negative effect on affinity; however, a favourable site for a positive potency effect of fluorination was the ortho position. Fluorination of aliphatic fragments more often led to a decrease in biological activity. The results may constitute the rules of thumb for fluorination of aminergic receptor ligands and provide insights into the role of fluorine substitutions in medicinal chemistry.

Structure-Based Design and Optimization of FPPQ, a Dual-Acting 5-HT3 and 5-HT6 Receptor Antagonist with Antipsychotic and Procognitive Properties.

In line with recent clinical trials demonstrating that ondansetron, a 5-HT3 receptor (5-HT3R) antagonist, ameliorates cognitive deficits of schizophrenia and the known procognitive effects of 5-HT6 receptor (5-HT6R) antagonists, we applied the hybridization strategy to design dual-acting 5-HT3/5-HT6R antagonists. We identified the first-in-class compound FPPQ, which behaves as a 5-HT3R antagonist and a neutral antagonist 5-HT6R of the Gs pathway. FPPQ shows selectivity over 87 targets and decent brain penetration. Likewise, FPPQ inhibits phencyclidine (PCP)-induced hyperactivity and displays procognitive properties in the novel object recognition task. In contrast to FPPQ, neither 5-HT6R inverse agonist SB399885 nor neutral 5-HT6R antagonist CPPQ reversed (PCP)-induced hyperactivity. Thus, combination of 5-HT3R antagonism and 5-HT6R antagonism, exemplified by FPPQ, contributes to alleviating the positive-like symptoms. Present findings reveal critical structural features useful in a rational polypharmacological approach to target 5-HT3/5-HT6 receptors and encourage further studies on dual-acting 5-HT3/5-HT6R antagonists for the treatment of psychiatric disorders.

Pharmacoprint: A Combination of a Pharmacophore Fingerprint and Artificial Intelligence as a Tool for Computer-Aided Drug Design.

Structural fingerprints and pharmacophore modeling are methodologies that have been used for at least 2 decades in various fields of cheminformatics, from similarity searching to machine learning (ML). Advances in in silico techniques consequently led to combining both these methodologies into a new approach known as the pharmacophore fingerprint. Herein, we propose a high-resolution, pharmacophore fingerprint called Pharmacoprint that encodes the presence, types, and relationships between pharmacophore features of a molecule. Pharmacoprint was evaluated in classification experiments by using ML algorithms (logistic regression, support vector machines, linear support vector machines, and neural networks) and outperformed other popular molecular fingerprints (i.e., ECFP4, Estate, MACCS, PubChem, Substructure, Klekota-Roth, CDK, Extended, and GraphOnly) and the ChemAxon pharmacophoric features fingerprint. Pharmacoprint consisted of 39 973 bits; several methods were applied for dimensionality reduction, and the best algorithm not only reduced the length of the bit string but also improved the efficiency of the ML tests. Further optimization allowed us to define the best parameter settings for using Pharmacoprint in discrimination tests and for maximizing statistical parameters. Finally, Pharmacoprint generated for three-dimensional (3D) structures with defined hydrogens as input data was applied to neural networks with a supervised autoencoder for selecting the most important bits and allowed us to maximize the Matthews correlation coefficient up to 0.962. The results show the potential of Pharmacoprint as a new, perspective tool for computer-aided drug design.