Ozren Jović

We used the extreme gradient boosting (XGB) algorithm to predict the experimental solubility of chemical compounds in water and organic solvents and to select significant molecular descriptors. The accuracy of prediction of our forward stepwise top-importance XGB (FSTI-XGB) on curated solubility data sets in terms of RMSE was found to be 0.59–0.76 Log(S) for two water data sets, while for organic solvent data sets it was 0.69–0.79 Log(S) for the Methanol data set, 0.65–0.79 for the Ethanol data set, and 0.62–0.70 Log(S) for the Acetone data set. That was the first step. In the second step, we used uncurated and curated AquaSolDB data sets for applicability domain (AD) tests of Drugbank, PubChem, and COCONUT databases and determined that more than 95% of studied ca. 500,000 compounds were within the AD. In the third step, we applied conformal prediction to obtain narrow prediction intervals and we successfully validated them using test sets’ true solubility values. With prediction intervals obtained in the last fourth step, we were able to estimate individual error margins and the accuracy class of the solubility prediction for molecules within the AD of three public databases. All that was possible without the knowledge of experimental database solubilities. We find these four steps novel because usually, solubility-related works only study the first step or the first two steps.

Due to sedentary lifestyle and harsh environmental conditions, gorgonian coral extracts are recognized as a rich source of novel compounds with various biological activities, of interest to the pharmaceutical and cosmetic industries. The presented study aimed to perform chemical screening of organic extracts and semi-purified fractions obtained from the common Adriatic gorgonian, sea fan, Eunicella cavolini (Koch, 1887) and explore its abilities to exert different biological effects in vitro. Qualitative chemical evaluation revealed the presence of several classes of secondary metabolites extended with mass spectrometry analysis and tentative dereplication by using Global Natural Product Social Molecular Networking online platform (GNPS). Furthermore, fractions F4 and F3 showed the highest phenolic (3.28 ± 0.04 mg GAE/g sample) and carotene (23.11 ± 2.48 mg β-CA/g sample) content, respectively. The fraction F3 inhibited 50% of DPPH (2,2-diphenyl-1-picryl-hydrazyl-hydrate) and ABTS (2,2′-azino-bis (3-ethylbenzthiazolin-6-yl) sulfonic acid) radicals at the concentrations of 767.09 ± 11.57 and 157.16 ± 10.83 µg/mL, respectively. The highest anti-inflammatory potential was exhibited by F2 (IC₅₀ = 198.70 ± 28.77 µg/mL) regarding the inhibition of albumin denaturation and F1 (IC₅₀ = 254.49 ± 49.17 µg/mL) in terms of soybean lipoxygenase inhibition. In addition, the most pronounced antiproliferative effects were observed for all samples (IC₅₀ ranging from 0.82 ± 0.14–231.18 ± 46.13 µg/mL) against several carcinoma cell lines, but also towards non-transformed human fibroblasts pointing to a generally cytotoxic effect. In addition, the antibacterial activity was tested by broth microdilution assay against three human pathogenic bacteria: Escherichia coli, Pseudomonas aeruginosa, and Staphylococcus aureus. The latter was the most affected by fractions F2 and F3. Finally, further purification, isolation and characterization of pure compounds from the most active fractions are under investigation.

In this work we introduce a novel filtering and molecular modeling pipeline based on a fingerprint and descriptor similarity procedure, coupled with molecular docking and molecular dynamics (MD), to select potential novel quoinone outside inhibitors (QoI) of cytochrome bc1 with the aim of determining the same or different chromophores to usual. The study was carried out using the yeast cytochrome bc1 complex with its docked ligand (stigmatellin), using all the fungicides from FRAC code C3 mode of action, 8617 Drugbank compounds and 401,624 COCONUT compounds. The introduced drug repurposing pipeline consists of compound similarity with C3 fungicides and molecular docking (MD) simulations with final QM/MM binding energy determination, while aiming for potential novel chromophores and perserving at least an amide (R1HN(C=O)R2) or ester functional group of almost all up to date C3 fungicides. 3D descriptors used for a similarity test were based on the 280 most stable Padel descriptors. Hit compounds that passed fingerprint and 3D descriptor similarity condition and had either an amide or an ester group were submitted to docking where they further had to satisfy both Chemscore fitness and specific conformation constraints. This rigorous selection resulted in a very limited number of candidates that were forwarded to MD simulations and QM/MM binding affinity estimations by the ORCA DFT program. In this final step, stringent criteria based on (a) sufficiently high frequency of H-bonds; (b) high interaction energy between protein and ligand through the whole MD trajectory; and (c) high enough QM/MM binding energy scores were applied to further filter candidate inhibitors. This elaborate search pipeline led finaly to four Drugbank synthetic lead compounds (DrugBank) and seven natural (COCONUT database) lead compounds—tentative new inhibitors of cytochrome bc1. These eleven lead compounds were additionally validated through a comparison of MM/PBSA free binding energy for new leads against those obtatined for 19 QoIs.

The limited number of medicinal products available to treat of fungal infections makes control of fungal pathogens problematic, especially since the number of fungal resistance incidents increases. Given the high costs and slow development of new antifungal treatment options, repurposing of already known compounds is one of the proposed strategies. The objective of this study was to perform in vitro experimental tests of already identified lead compounds in our previous in silico drug repurposing study, which had been conducted on the known Drugbank database using a seven-step procedure which includes machine learning and molecular docking. This study identifies siramesine as a novel antifungal agent. This novel indication was confirmed through in vitro testing using several yeast species and one mold. The results showed susceptibility of Candida species to siramesine with MIC at concentration 12.5 µg/mL, whereas other candidates had no antifungal activity. Siramesine was also effective against in vitro biofilm formation and already formed biofilm was reduced following 24 h treatment with a MBEC range of 50–62.5 µg/mL. Siramesine is involved in modulation of ergosterol biosynthesis in vitro, which indicates it is a potential target for its antifungal activity. This implicates the possibility of siramesine repurposing, especially since there are already published data about nontoxicity. Following our in vitro results, we provide additional in depth in silico analysis of siramesine and compounds structurally similar to siramesine, providing an extended lead set for further preclinical and clinical investigation, which is needed to clearly define molecular targets and to elucidate its in vivo effectiveness as well.

Novel machine learning and molecular modelling filtering procedures for drug repurposing have been carried out for the recognition of the novel fungicide targets of Cyp51 and Erg2. Classification and regression approaches on molecular descriptors have been performed using stepwise multilinear regression (FS-MLR), uninformative-variable elimination partial-least square regression, and a non-linear method called Forward Stepwise Limited Correlation Random Forest (FS-LM-RF). Altogether, 112 prediction models from two different approaches have been built for the descriptor recognition of fungicide hit compounds. Aiming at the fungal targets of sterol biosynthesis in membranes, antifungal hit compounds have been selected for docking experiments from the Drugbank database using the Autodock4 molecular docking program. The results were verified by Gold Protein-Ligand Docking Software. The best-docked conformation, for each high-scored ligand considered, was submitted to quantum mechanics/molecular mechanics (QM/MM) gradient optimization with final single point calculations taking into account both the basis set superposition error and thermal corrections (with frequency calculations). Finally, seven Drugbank lead compounds were selected based on their high QM/MM scores for the Cyp51 target, and three were selected for the Erg2 target. These lead compounds could be recommended for further in vitro studies.

Rosemary (Rosmarinus officinalis), garden sage (Salvia officinalis), summer savory (Satureja hortensis), laurel (Laurus nobilis), and other aromatic plants were put in olive oil and exposed to ultrasounds for different duration. Filtrated oils were dissolved in cyclohexane, and UV-Vis measurements were carried out. Absorbance values corresponding to chlorophylls, carotenoids, flavonoids (370 nm), and polyphenols (around 300 nm) were measured. In addition, for some samples, total phenols were determined using Folin-Denis reagent and compared with the similar maceration procedure in water solvent (instead of olive oil). Maceration without ultrasound in olive oil for each plant was also compared with ultrasound-assisted extraction. The results show that significant amount of aromatic content can be extracted in olive oil by applying ultrasounds for only few minutes, especially for Salvia officinalis powder. The use of UV-Vis measurements is simple but enough to examine the extent of phenol content extraction through such maceration procedure.

Adulteration of hempseed (H) oil, a well‐known health beneficial nutrient, is studied in this work by mixing it with cheap and widely used oils such as rapeseed (R) oil and sesame (Se) and sunflower (Su) oil. Many samples of different geographic origins were taken into account. Binary mixture sets of hempseed oil with these 3 oils (HR, HSe, and HSu) were considered. FTIR spectra of pure oils and their mixtures were recorded, and quantitative analyses were performed using partial least squares regression (PLS) and first‐break forward interval PLS methods (FB‐FiPLS). The obtained results show that each particular oil can be very successfully quantified (R²(val) > 0.995, RMSECV 0.9%–2.9%, RMSEP 1.0%–3.2%). This means that FTIR coupled with multivariate methods can rapidly and effectively determine the level of adulteration in the adulterated hempseed oil for these studied and frequently used adulterant oils. Also, the relevant variables selected by FB‐FiPLS could be used for verification of hempseed oil adulteration.