Search

The study of protein folding is attractive to scientists from various natural sciences, including physicists/biophysicists [1]. One of them is M. Karplus, Nobel laureate 2013, who modelled protein folding and unfolding of 33 proteins using models selected from 24 generated sequence-based features/descriptors [2]. Protein folding rates correlate with protein length (the total number of amino acids in a protein), the content of secondary structures in the protein, the absolute and relative mean distance of contact order between amino acids, etc. [3]. We have defined and introduced several new sequence-based attributes/features suitable for modelling the folding rates of proteins, such as the features related to the secondary structure of proteins (segments of alpha or beta secondary structure) and the average distance of amino acids. Models were also developed with different combinations of the above features. Improved models have been developed to predict the folding rates of proteins and all features are not dependent on the 3D structure of the protein. References: [1] D.N. Ivankov, A.V. Finkelstein, Biomolecules 10, 250 (2020) [2] A. R. Dinner, S. S. So, M. Karplus, Advan. Chem. Phys. 120, 1 (2002) [3] H. Zhou, Y. Zhou, Biophys. J. 82, 458 (2002)

There is enormous public interest in the EU, but also the interest of the pharmaceutical industry in the possibility of finding significant factors that would make it possible to distinguish between approved drugs and drugs that have been withdrawn after several years of use. The EU responded to this interest by funding a large AI4EU project, which launched calls for proposals to develop various artificial intelligence (AI)-based computing solutions.[1] Among them was the task to develop a predictive AI solution based on the deep neural network method to distinguish drugs from withdrawn drugs based on their structure.[2] In this presentation, the process of defining the problem and creating a dataset will be elaborated, as well as related issues such as the different status of the same drug in different countries or regions, or the problem of cleaning datasets, as many drugs are not pure compounds but complexes. We have calculated comprehensive sets of molecular structure descriptors for the final sets of approved and withdrawn drugs. The analysis of the importance/usefulness of each structure-based molecular descriptor (including those related to ADMETox properties of drugs) in discriminating between approved and withdrawn drugs will be presented. As expected, the classification accuracy of the multivariate models based on descriptors is not high. Nevertheless, it is comparable to (or sometimes better than) more complex models and significantly higher than the accuracy achieved by random guessing (random model). REFERENCES [1] The AI4EU Call for Solutions Seeking Companies to Solve AI Challenges, https://internationaldataspaces.org/the-ai4eu-call-for-solutions-seeking-companies-to-solve-aichallenges/ (assessed on April 15, 2022) [2] The AI-on-demand platform, https://www.ai4europe.eu/ and https://www.ai4media.eu/projectoverview/ (assessed on April 15, 2022)

Despite decades of intensive search for compounds that modulate the activity of particular protein targets, a large proportion of the human kinome remains as yet undrugged. Effective approaches are therefore required to map the massive space of unexplored compound-kinase interactions for novel and potent activities. Here, we carry out a crowdsourced benchmarking of predictive algorithms for kinase inhibitor potencies across multiple kinase families tested on unpublished bioactivity data. We find the top-performing predictions are based on various models, including kernel learning, gradient boosting and deep learning, and their ensemble leads to a predictive accuracy exceeding that of single-dose kinase activity assays. We design experiments based on the model predictions and identify unexpected activities even for under-studied kinases, thereby accelerating experimental mapping efforts. The open-source prediction algorithms together with the bioactivities between 95 compounds and 295 kinases provide a resource for benchmarking prediction algorithms and for extending the druggable kinome. The IDG-DREAM Challenge carried out crowdsourced benchmarking of predictive algorithms for kinase inhibitor activities on unpublished data. This study provides a resource to compare emerging algorithms and prioritize new kinase activities to accelerate drug discovery and repurposing efforts.

Antioxidants play an important role in preserving human health in the fight against cell damage caused by free radicals [1]. Different types of honey and from different geographical origins show different polyphenol content. The aim of this study was to determine the total polyphenol content (TPC) in 96 honey samples from different countries of Central and Eastern Europe (Croatia, Serbia, Bosnia and Herzegovina, Hungary, Slovenia and Italy) by a spectrophotometric method with Folin Ciocalteu reagent [2]. If we analyze the measurements of the same types of honey (e.g. acacia or chestnut) from a number of literature sources, large differences in the values of total phenols are observed, which cannot be attributed only to seasonal or geographical variations. Our previous studies gave the TPC values from 1164.99 mg/kg to 212.9 mg/kg for the dominant uni- and multi-floral honey. In an attempt to decipher the reasons for these excessive differences, we analyzed the impact of sample ageing and seasonal and geographic variations. However, some differences are too large, and the cause seems to be the inconsistency of the calculation of the measurement units. Namely, some authors state that they modified method [1, 2], but do not state in detail which part they modified. Our experience to date shows that the authors took different amounts of honey for analysis and some used a gallic acid for calibration curve in the range of 0–200 μg/mL and some of 0–400 μg/mL. For these reasons, we would suggest a more strict standardisation of the total polyphenol content measurements.

The mineral content of different types of honey reflects the botanical origin of samples. The amount of major and trace minerals present in honey could be related to their arrangement into the soil. Many types of minerals in floral plants first enter into the roots and reached to nectar and finally get into the honey. The increased concentration of trace elements in honey sampled near industrial areas was observed in most cases. Therefore, the quantification of trace toxic mineral elements in honey samples becomes important due to their effects on human health, control of honey quality/safety, as well as because of environmental biomonitoring. We measured mineral contents of 99 floral honey samples taken from Croatia and surrounding countries (Hungary, Serbia and Italy). In all samples, K was identified as the predominant mineral, followed by Ca, Na, Mg, Fe, Mn and Zn. We didn't detect toxic metals (Co, Cr, Pb, Cd and Hg) in our samples.

The total phenols are determined by the modified Folin-Ciocalteu method in such a way that the Folin-Ciocalteu reagent (a mixture of phosphotungstate and phosphomolybdic acid), in reaction with phenolic components reduces and changes colour to the blue [1]. During the measurement of the total phenolic content, the concentration is determined by using the calibration curve of gallic or ascorbic acid. Depending on the choice of the reference compound for determination of calibration curve and on the theoretical approach used for determination of calibration curve, the results may vary considerably. Through the measurement of the total phenolic content of the same type of honey (e.g. acacia), it appears that the total phenolic contents of the same type of honey vary depending on the geographical dispersion, the influence of climatic factors in the current year, the difference between the time of sampling and the time of measurement, etc. However, in addition to these expected and normal variations, the increased disagreement between the measurements in different laboratories could be also due to differences in the process of determination of the calibration curve. Namely, analyzing results in the most frequently cited papers [1-3], significant differences in total phenolic contents were observed. These discrepancies are primarily related to the minimal and maximal range of measurement of gallic or ascorbic acid concentrations in which the calibration curve is determined, compared with the total phenolic contents of samples. Thus, it is possible to produce a calibration curve passing through the starting point, while some authors provide the calibration curve without including the starting point. The influence of these factors is investigated in the analysis of total phenolic contents of 99 honey samples from the South- Eastern European region.

Quantitative structure-retention relationships (QSRR) although widespread for prediction of retention time in reversed-phase liquid chromatography (RP-LC) suffer from the same limitation. Typically they are built for a specific set of chromatographic conditions (e.g., stationary phase, mobile phase composition, pH, temperature, total gradient time). To overcome this limitation, in this work we aimed to build global QSRR models for prediction of retention time of synthetic peptides across six RP-LC columns with varied experimental conditions. In this work, QSRR models were based on three a priori selected molecular descriptors: sum of gradient retention times of 20 natural amino acids (logSumAA), van der Waals volume (logvdWvol.), and hydrophobicity (clogP) related to the retention mechanism of RP-LC separation of peptides. A multitude of machine learning methods was compared: random forests (RF), adaptive boosting (ADA), and gaussian process regression (GPR). The models were comprehensively optimized through 3-fold cross- validation (CV) and validated through an external validation set. Chemical domain of applicability was also defined, while statistical significance of the models was tested using CV-ANOVA. All the models were also compared to the conventional linear model built using partial least squares (PLS). Results have shown that all the machine learning methods outperformed PLS with %RMSEP ranging from 14.99 % ; for RF, to 26.35 % for ADA. On the other hand, PLS exhibited a %RMSEP of 40.56 %. The novel ensemble and mixture models revealed mechanisms behind black-box global QSRR models and paved the way to resolving the principal limitation of QSRR modelling. The models have shown the highest feature importance for sum of gradient retention times (logSumAA), followed by van der Waals volume (logvdWvol.), and hydrophobicity (clogP). The promising results of this study show the potential of machine learning for improved peptide identification, retention time standardization and integration into state-of-the-art LC-MS/MS proteomics workflows.

Quantitative structure-retention relationships (QSRR) were employed to build global models for prediction of chromatographic retention time of synthetic peptides across six RP-LC-MS/MS columns and varied experimental conditions. The global QSRR models were based on only three a priori selected molecular descriptors: sum of gradient retention times of 20 natural amino acids (logSumAA), van der Waals volume (logvdWvol.), and hydrophobicity (clogP) related to the retention mechanism of RP-LC separation of peptides. Three machine learning regression methods were compared: random forests (RF), partial least squares (PLS), and adaptive boosting (ADA). All the models were comprehensively optimized through 3-fold cross- validation (CV) and validated through an external validation set. The chemical domain of applicability was also defined. Percentage root mean square error of prediction (%RMSEP) was used as an external validation metric. Results have shown that RF exhibited a %RMSEP of 14.99 % ; PLS exhibited a %RMSEP of 40.561 % ; whereas ADA exhibited a %RMSEP of 26.35 %. The ensemble models considerably outperform the conventional PLS-based QSRR model. Novel methods of treebased model explainability were employed to reveal mechanisms behind black-box global ensemble QSRR models. The models revelead the highest feature importance for sum of gradient retention times (logSumAA), followed by van der Waals volume (logvdWvol.), and hydrophobicity (clogP). The promising results of this study show the potential of machine learning for improved peptide identification, retention time standardization and integration into state- of-the-art LC-MS/MS proteomics workflows.

Among natural food antioxidants, polyphenols are ubiquitously distributed in the vegetable kingdom as plant secondary metabolites. Also, various kinds of antioxidant components in honey may play important roles in a combinative or synergistic contribution to its total antioxidant activity (Meda et al., 2005). As natural antioxidants, phenolic compounds in honey contribute, as an important group of chemicals, to its functional properties and therapeutic significance (Yao et al., 2004). The method for determination polyphenols is based on the reaction of DPPH (2, 2-diphenyl-1- picryl-hydrazyl) which is one of the rare nitrogen radicals (free radical) in a stable form and hydrogen donor (e.g., phenolic group). In reaction with antioxidants, it is reduced to hydrazine, whereby the colour varies from purple to yellow. A change in absorbance (DPPH radical reduction) is measured at 517 nm, spectrophotometrically. The result is proportional to the concentration of antioxidants, while the antioxidant capacity itself is expressed through the IC50. IC50 represents the concentration of antioxidants needed to reduce the initial DPPH concentration by 50%. The higher IC50 values represent a lower antioxidant capacity, which is the lack of this way of expression. Using modified methods for determining the anti-oxidative capacity of the honey (IC50, determined by DPPH), different results are obtained which can cause discrepancies in a comparative analysis of IC50 values in different studies. By performing analysis of 99 samples, we propose standardisation of information that must be given in these studies in the literature and suggest a procedure which could be considered as optimal for determination of IC50 of honey samples. REFERENCES: 1. Meda A., Lamien C.E., Romito, M., Millogo J., Nacoulma O.G. (2005) Determination of the total phenolic, flavonoid and proline contents in Burkina Fasan honey, as well as their radical scavenging activity. Food Chemistry, 91: 571–577. 2. Yao et al. (2004) Falvonoids in food and their health benifits, Plant Foods for Human Nutrition, 59: 113-122.

Correlation coefficient (or its square) has been used as the most common statistical parameter for estimating the quality of models in chemistry. It is usually calculated in each of three main validation procedure: fitting (r2), cross- validation (q2), and prediction on an external test set ( 2). Wide use of q2 in cross- validation technique was initiated by a very often cited paper by Cramer et al. [1], in which comparative molecular field analysis based on partial least squares method was introduced in chemical modeling. The use of q2 was additionally accelerated by its involvement in the regulatory perspectives of the U.S.A. Environmental Protection Agency (EPA) [2] and in the Guidance document of OECD principles on the validation of (quantitative) structure-activity relationship ((Q)SAR) models [3]. In latter applications of q2 several researchers noticed its strange properties like overestimation or underestimation when applied on external (test) data set. Therefore, two additional alternative variants of q2 [4, 5] were proposed, but the problem remains unsolved. All these results are reviewed and their statistical foundations are re-analysed, considering all three validation procedures, i.e. fitting, cross-validation and prediction. Obtained results will be illustrated on literature data sets. It comes out that the best estimate of the model quality can be obtained by simultaneous calculation and comparison of root-mean-square errors of fit, cross-validation and prediction [1] R. D. Cramer et al., Comparative molecular field analysis (CoMFA). 1. Effect of shape on binding of steroids to carrier proteins. J. Am. Chem. Soc. 110 (1988) 5959–5967. [2] M. Zeeman, et al., U.S. EPA regulatory perspectives on the use of QSAR for new and existing chemical evaluations, SAR QSAR Environ. Res. 3 (1995) 179– 201. [3] OECD guidelines concerning QSARs, 2007, pp. 55–65. http://www.oecd.org/officialdocuments/publicdispla ydocumentpdf/?doclanguage=en&cote=env/jm/m ono(2007)2, accessed: May 25, 2018. [4] G. Schüürmann et al., External validation and prediction employing the predictive squared correlation coefficient - test set activity mean vs training set activity mean. J. Chem. Inf. Model. 48 (2008) 2140– 2145. [5] V. Consonni et al., Comments on the definition of the q2 parameter for QSAR validation. J. Chem. Inf. Model. 49 (2009) 1669–1678.

The thermodynamic preference of different free radical scavenging mechanisms of 3-(4-hydroxy- 3-methoxyphenyl)propanoic acid (HMPPA), a colon catabolite of dietary (poly)phenols, was studied using density functional theory. The role of guaiacyl moiety and carboxyl group of HMPPA in double (2H+/2e-) processes of inactivation of free radicals was theoretically investigated in polar and nonpolar media using Gaussian 09 software package. Obtained results indicate that HMPPA possesses potential for inactivating free radicals of different characteristics (HO•, HOO•, CH3O•, CH3OO•, PhO•, Cl3COO• etc.) via double hydrogen atom transfer (dHAT) and double sequential proton loss electron transfer (dSPLET) mechanisms. Concentration of HMPPA in serum may reach very low micromolar values and may contribute to health benefits associated with regular intake of polyphenol- rich diet by direct scavenging of reactive oxygen species.

Nitroaromatic compounds, like nitrobenzene derivatives can be considered as the typical energetic molecules. Some of them are commonly used as explosives like TNB (1, 3, 5- Trinitrobenzene) or TNT (2, 4, 6- trinitrotoluene). For the suitability of an explosive substance in a specific use, better understanding of physico-chemical attributes of its structure which are related to compound’s stability is a very important task. Many structure-thermal stability models have been developed in the last ten years for this class of compounds. The largest are those published by Fayet et al. based (primarily) on the set of semi- empirical molecular descriptors calculated by the MOPAC program, and by Li et al. selected from the large pool of molecular descriptors calculated by the Dragon 5.5 program. To achieve better physico- chemical interpretation of models, in addition to descriptors calcualted and used in previous models we calculated and used additional descriptors like those given in the PubChem database, experimental or calculated 1H and 13C chemical shifts of nitroaromatics and the number of specific connectivity terms in molecular skeleton. Also, an improved selection of molecular descriptors was performed in attempt to develop simpler structure- property models (having a smaller number of descriptors). Performing descriptor selection, whenever it was possible and statistically reasonable, we considered the possibility of inclusion of simpler molecular descriptors into the final models instead of more complex descriptors or those that are computed from 3D structure. As an interesting results we found that the maximal 1H chemical shift of hydrogen's attached to aromatic ring is a significant descriptor, as well as the number of multiple bonds in a compound. Obtained results and models were compared with other models from literature showing that presented approach is a promising way of selecting statistically good structure-thermal stability models of nitro aromatics, which are more informative and easy to interpret.

Honey is a natural sweetener with a complex composition. Honey features vary depending on the botanical source and geographical origin, as well as climatic, processing and storage conditions. Honey is a highly concentrated solution of a complex mixture of sugars ; namely fructose (38%) and glucose (31%) ; honey also contains other constituents such as minerals, proteins, vitamins, organic acid, flavonoids, phenolic acid, enzymes, and other phytochemicals in small quantities. Honey is a highly concentrated phenolic compounds are biologically active secondary metabolites from plants that act at molecular level and are potent natural antioxidants. Honey samples (n=97) were provided directly from beekeepers in 2017. The samples are from Croatia, Bosnia and Herzegovina, Serbia, Hungary and Italy. The physico-chemistry analysis of honey (HMF, electricar conductivity and moisture) was performing using the official methods of analysis (Codex Alimentarius Commission, 2001). The pH of the honey analysed was in the range 3, 38-6, 05, with the electrical conductivity was in the range 0, 13319-1, 738 mS/cm. Moisture in the range 14, 0-18, 8%. Moisture and pH are of great importance of honey texture, stability and shall life. HMF (Hydroxy Methyl Furfural) have been used as indicator of honey freshness in European because of their sensitivity to heat. An tested samples the honey had low levels HMF (O-6, 26 mg/kg honey). The quantification of total phenol in honey was preformed according to the Folin-Ciocalteu method as described by Singleton and Rossi (1965) using gallic acid as reference. An tested samples the honey was in the range 0, 043-0, 305 mg gallic acid per 100 g honey. The DPPH radical-scavenging assey based oh the abillity of antioxidant to block the 2, 2- difenil-1-picril-hidrazil radical.This relationship is confirm in this study ; the analysed honey present high antioxidant power, with a mean value of 67, 45% a range of 57, 2- 77, 7%. Honey samples were prepared and extracted with ethyl acetate according to the procedure by Ferreros et al., . The purity of the isolated flavonoids was tested by HPLC.

The study of nitrobenzene derivatives like nitroaromatic compounds is a very important task, due to their increased influence on living organisms and environment, in general [1]. These compounds and their derivatives and metabolic products show mutagenic and carcinogenic effects, as well as allergic reactions, endocryne system impairment and skin irritation [2]. In recent study the authors performed a comprehensive quantitative structure-toxicity relationship study on rats using a set of 90 nitroaromatics whose structures were described by three set having ~ 800, 20000 and 1500 different descriptors (based on 1D, 2D and 3D structures) calculated by the PaDEL, HiTQSAR and Dragon program, respectively [2]. Among them Dragon set contains the largest portion of 3D descriptors. Additionally, two smaller sets of 3D descriptors based on semiempirical and DFT structure were also calculated and all descriptors were selected into Multivariate Linear Regression (MLR) models by the method based on genetic algorithm, as implemented in QSARIns [3]. In modeling, experimental values of concentrations which had letal toxic outcome in 50% tested organisms were used as endopints (LD50). In summary, better models were obtained from sets of descriptors which are composed mostly of 1D or 2D descriptors (HiTQSAR, PaDEL) and if data sets contain only smaller sub-sets of 3D descriptors (Dragon). The worst models were those based on semiempirical and DFT 3D descriptors. To test the possibility of improvement of models obtained in [2] we performed an analysis using only Dragon 5.5 set of descriptors and extended previous study by: (1) enlarging data set by 100 novel nitroaromatics obtained from the LD50 database involved into the program TEST 4.2.1 [4] (given on U.S Env. Protection Agency web site), (2) the use of an additional algorithm which performs detailed descriptor selection by considering all combinations of fixed size sub- sets of descriptors [5], and (3) by the use of more robust Partial Least Square (PLS) method for model generation [6]. The novel set of 100 nitroaromatics contains compounds having high similarity with 90 compounds from ref. [2]. This set was used as an additional set for testing external predictivity of developed models. As the main results we have found that the Root Mean Squared Error (RMSE) of prediction by U.S. EPA method TEST 4.2.1 for 10 compounds (out of 90 nitroaromatics from [2]) which are not involved in the TEST 4.2.1 database is ~1.17 log units. This fact justify the exictence of need for improvement of existing models of nitroaromatics. RMSEs of fit or cross-validation of LD50 values for 90 nitroaromatics were between 0.28 - 0.7 on log scale in all considered PLS and MLR models. But, for the novel test set of 100 nitroaromatics RMSEs are ~0.75 - 1.0 (for all models). However, the lowest RMSE of prediction on 100 compounds of complex PLS models having 13 components (that are formed by ~40 initial descriptors) is 0.75, and the best one-descriptor (the number of phosphorous atoms) MLR model gives RMSE of 0.89. Almost all obtained MLR and PLS models have RMSEs in prediction on large external set comparable or even lower than the TEST 4.2.1 program, thus confirming that there is a space for improvement of structure-toxicity models of nitroaromatics. [1] P.-Z. Lang, G.-H Lu, Chem. J. Chin. Univ. 16 (1995) 1083–1087. [2] A. Gooch et al., Environ. Toxicol. Chem. 36 (2017) 2227–2233. [3] P. Gramatica et al., J. Comput. Chem. 34 (2013) 2121- 2132. [4] TEST 4.2.1 program, 2016, https://www.epa.gov/chemical-research/toxicity- estimation-software-tool-test. [5] B. Lučić, N. Trinajstić, J. Chem. Inf. Comput. Sci. 39 (1999) 121-132. [6] K. Roy, P. Ambure, Chemom. Intell. Lab. Syst. 159 (2016) 108–126.

Thermodynamics of 2H+/2e- free radical scavenging mechanisms of dihydrocaffeic acid (3- (3’, 4’-dihydroxyphenyl)propionic acid), an abundant colon catabolite of complex dietary (poly)phenols [1], was studied by DFT method using the Gaussian 09 program package [2]. Geometry optimizations and frequency calculations were carried out using the M06- 2X/6-311++G(d, p) level of theory, in conjunction with the SMD continuum solvation model. The role of catechol moiety (reaction path a) and carboxyl moiety (reaction path b) in free radical scavenging processes was investigated by considering double hydrogen atom transfer (dHAT), double electron transfer- proton transfer (dET-PT) and double sequential proton loss electron transfer (dSPLET) mechanisms [3]. The Gibbs free energies of reactions indicate that dihydrocaffeic acid possesses potential for inactivating free radicals of different nature (HO•, HOO•, CH3O•, CH3OO•, CH2=CH–OO•, PhO• etc.) via dHAT and dSPLET mechanisms. 3’- Methylated, 3’- glucuronidated and 3’-sulfated conjugates of dihydrocaffeic acid may retain their appreciable activity via reaction path b. Because dihydrocaffeic acid is produced in colon in high concentrations (up to 90 mM) [1], and could be much more abundant in the circulation than its precursor molecules, it has potential to contribute to health benefits associated with regular intake of polyphenol- rich diet by direct scavenging of reactive oxygen species. Acknowledgments: This work has been supported by The Foundation of the Croatian Academy of Sciences and Arts, under the project No. 10- 102/244-1-2016. - “Investigations of the antioxidant mechanisms of polyphenols and their metabolites.” [1] B. Halliwell, J. Rafter, A. Jenner, Am. J. Clin. Nutr., 2005, 81, 268S. [2] M. J. Frisch, G. W. Trucks et al., Gaussian 09. Revision A.02, Gaussian, Inc., Wallingford CT, 2009. [3] A. Amić, B. Lučić, V. Stepanić, Z. Marković, S. Marković, J. M. Dimitrić Marković, D. Amić, Food Chem., 2017, 218, 144.

The thermodynamic preference of different radical scavenging mechanisms of 3, 5- dihydroxy- 4-methoxybenzyl alcohol (DHMBA), a natural product identified in oysters, was studied using DFT. Geometry optimizations and frequency calculations were carried out using the M06-2X functional and the 6-311++G(d, p) basis set, in conjunction with the SMD continuum solvation model [1]. The involvement of guaiacyl moiety in multiple radical scavenging processes was investigated considering reaction energetics of triple hydrogen atom transfer (tHAT), triple electron transfer-proton transfer (tET-PT) and triple sequential proton loss electron transfer (tSPLET) mechanisms. All these mechanisms result in production of phenoxyl radical of 6- (hydroxymethyl)-1, 3-benzodioxol-4-ol (sesamol like structure – sesamol: potent plant antioxidant with 1, 3-benzodioxole core). The Gibbs free energy change for reactions of inactivation of radicals of different chemical nature (HO•, HOO•, CH3O•, CH3OO•, CH2=CH–O–O•, PhO•, Cl3COO• etc.) indicate tHAT and tSPLET mechanisms as thermodynamically feasible. Exceptional antioxidant activity of DHMBA can be explained by consideration of triple processes (based on involvement of guaiacyl moiety, i.e., vicinal phenolic –OH and –OCH3 group) in addition to single processes (dealing only with single phenolic –OH group of DHMBA) recently studied by Villuendas-Rey et al. [2]. References: [1] M.J. Frisch, G.W. Trucks, H.B. Schlegel et al., Gaussian 09, Wallingford, CT, 2009. [2] Y. Villuendas-Rey, J.R. Alvarez- Idaboy, A. Galano, J. Chem. Inf. Model. 55 (2015) 2552- 2561. Acknowledgment Financial support from The Foundation of the Croatian Academy of Sciences and Arts (project No. 10- 102/244-1-2016.: Investigations of the antioxidant mechanisms of polyphenols and their metabolites) is gratefully acknowledged.

Health benefits of polyphenol-rich diet are related to intestinal metabolites, rather than to polyphenolic compounds originally present in food [1, 2], which usually possess very low bioavailability and low systemic concentration [3, 4]. Particular intestinal metabolites are produced in high M concentrations and may promote health via several possible in vivo mechanisms [5]. 4-hydroxyphenylpropionic acid (4-HPPA) is among the most abundant products of polyphenols degradation in large intestine [5]. Its concentration in colon may reach value of 200 M [5], which could be sufficient to exert biological activity and effective direct radical inactivation. Radical scavenging mechanisms of 4-HPPA, in water and pentyl ethanoate as a solvent, were studied by DFT method using Gaussian 09 package [6]. Geometry optimizations and frequency calculations were carried out using the M06– 2X/6-311++G(d, p) level of theory, in conjunction with the SMD continuum solvation model. Hydrogen atom transfer (HAT) and sequential proton loss electron transfer (SPLET) mechanisms were found to be thermodynamically probable and competitive processes in both media. The Gibbs free energy change for reaction of inactivation of radicals indicates 4-HPPA as a potent scavenger. 4-HPPA possesses potential for inactivating radicals of different characteristics by direct scavenging via HAT and SPLET mechanisms. Because 4-HPPA is produced in high M concentrations and is usually better absorbed than its precursor molecules, it may contribute to health benefits associated with regular intake of polyphenol- rich diet. References [1] Y.S. Chiou, J.C. Wu, Q. Huang, F. Shahidi, Y.J. Wang, C.T. Ho and M.H. Pan, J. Funct. Foods 7 (2014) 3-25. [2] A. Duda-Chodak, T. Tarko, P. Satora and P. Sroka, Eur. J. Nutr. 54 (2015) 325-341. [3] D. Del Rio, A. Rodriguez-Mateos, J.P.E. Spencer, M. Tognolini, G. Borges and A. Crozier, Antioxid. Redox Signal. 18 (2013) 1818-1892. [4] P.C.H. Hollman, Arch. Biochem. Biophys. 559 (2014) 100-105. [5] B. Halliwell, J. Rafter and A. Jenner, Am. J. Clin. Nutr. 81 (2005) 268S−276S. [6] M.J. Frisch, G.W. Trucks, H.B. Schlegel et al., GAUSSIAN 09, Wallingford, CT, 2009. Acknowledgments We gratefully acknowledge the financial support to this work from The Foundation of the Croatian Academy of Sciences and Arts, under the project No. 10-102/244-1- 2016. – "Investigations of the antioxidant mechanisms of polyphenols and their metabolites"

Colon polyphenolic metabolites can reduce activity of enzymes involved in human carcinogenesis [1], for instance by inhibition of COX-2 [2]. Recent studies have shown the importance of selective inhibition of COX-2 for the anti-inflammatory and anticancer therapy [3], indicating COX-2 as a valid molecular target for cancer prevention and treatment [4]. Theoretical investigations of active site of COX-2 affirmed some natural phenolic antioxidants as potential inhibitors of this enzyme [5]. Radical scavenging mechanisms of selected polyphenolic metabolites were studied in water and pentyl ethanoate as a solvent, by DFT method using Gaussian 09 package [6]. Geometry optimizations and frequency calculations were carried out using the M06- 2X/6-311++G(d, p) level of theory, in conjunction with the SMD continuum solvation model. Inhibitory potency against COX- 2 by colon polyphenolic metabolites and their mono- and di-anionic forms, were theoretically studied. Free energy of binding and inhibition constant for these ligands at the most favourable binding positions were estimated. Hydrogen atom transfer and sequential proton loss electron transfer mechanisms were found to be thermodynamically probable and competitive processes in both media. The Gibbs free energy change for reaction of inactivation of radicals indicate selected metabolites as potent scavengers. Docking analysis with structural forms of selected metabolites indicates dianionic ligands as potent inhibitors of COX- 2. Obtained results indicate that, because polyphenolic metabolites are produced in high mM concentrations and are usually better absorbed than their precursor molecules, they may contribute to health benefits associated with regular intake of polyphenol-rich diet. Acknowledgments: We gratefully acknowledge the financial support to this work from The Foundation of the Croatian Academy of Sciences and Arts, under the project No. 10-102/244-1- 2016.- “Investigations of the antioxidant mechanisms of polyphenols and their metabolites.” [1] C. Miene, A. Weise, M. Glei, Nutr. Cancer, 2011, 63, 653. [2] P.C. Karlsson, U. Huss, A. Jenner, B. Halliwell, L. Bohlin, J.J. Rafter, J. Nutr., 2005, 135, 2343. [3] G. Dannhardt, W. Kiefer, Eur. J. Med. Chem., 2001, 36, 109. [4] R.A. Gupta, R.N. DuBois, Nat. Rev. Cancer, 2001, 1, 11. [5] M. Amaravani, N.K. Prasad, V. Ramakrishna, Springerplus, 2012, 1, 58. [6] M.J. Frisch, G.W. Trucks, H.B. Schlegel et al., GAUSSIAN 09, Wallingford, CT, 2009.

Thermodynamics of 2H+/2e- radical scavenging mechanisms of dihydroferulic acid, a naturally occurring plant compound and colon catabolite of (poly)phenols, was studied by DFT method using SMD/M06–2X/6–311++G(d, p) level of theory with the Gaussian 09 program package [1]. Until recently, radical inactivation by (poly)phenolic compounds has been investigated by considering involvement of only phenolic –OH group(s) via 1H+/1e- mechanisms. Here, we studied the role of guaiacyl moiety (reaction path a [2]) and carboxyl group (reaction path b [3]) (Figure 1) in radical scavenging by dihydroferulic acid via double hydrogen atom transfer (dHAT), double electron transfer- proton transfer (dETPT) and double sequential proton loss electron transfer (dSPLET) mechanisms in polar and non-polar solvents [4]. Obtained results indicate that dihydroferulic acid possesses potential for inactivating radicals of different characteristics (HO•, HOO•, CH3O•, CH3OO•, PhO•, Cl3COO• etc.) what can be ascribed to 2H+/2e- mechanisms such as dHAT and dSPLET, rather than to a single 1H+/1e- processes. We found that reaction path a is less energy demanding than reaction path b. References [1] M.J. Frisch, G.W. Trucks et al., Gaussian 09. Revision A.02. Gaussian, Inc., Wallingford CT, 2009. [2] D. Kozlowski, P. Trouillas, C. Calliste, P. Marsal, R. Lazzaroni and J.-L. Duroux, J. Phys. Chem. A 111 (2007) 1138–1145. [3] H. Iwasaki, L.A. Cohen and B. Witkop, J. Am. Chem. Soc. 85 (1963) 3701–3702. [4] A. Amić, B. Lučić, V. Stepanić, Z. Marković, S. Marković, J.M. Dimitrić Marković and D. Amić, Food Chem. 218 (2017) 144–151. Acknowledgment This work has been supported by The Foundation of the Croatian Academy of Sciences and Arts, under the project No. 10-102/244-1-2016. – “Investigations of the antioxidant mechanisms of polyphenols and their metabolites.”

Phloretic acid (3-(4’-hydroxyphenyl)propanoic acid) is one of the most abundant colon metabolites of various classes of polyphenols (e.g., polymeric proanthocyanidins, tea catechins, and ellagitannins). Its concentration in fecal water may reach value of 210mol/L [1], which could be high enough to exert at least in situ antioxidant activity [2]. Thermodynamics of 2H+/2e free radical scavenging mechanisms of phloretic acid was studied by DFT method using Gaussian 09 package [3]. Geometry optimizations and frequency calculations were carried out using the M06-2X/6-311++G(d, p) level of theory, in conjunction with the SMD continuum solvation model. For the first time direct involvement of carboxylic group in free radical scavenging mechanisms was investigated considering double hydrogen atom transfer (HAT), double single electron transferproton transfer (SET-PT) and double sequential proton loss electron transfer (SPLET) processes producing dienone lactone [4]. Obtained results indicate that phloretic acid possesses potential for inactivating free radicals of different characteristics (HO•, HOO•, CH3O•, CH3OO•, CH2=CH–O–O•, PhO•, Cl3COO• etc.) by direct scavenging via double HAT and double SPLET mechanisms. In this way, because phloretic acid is produced in highM concentrations and it is usually better absorbed than its precursor molecules, it may contribute to health benefits associated with regular intake of polyphenol-rich diet. References: [1] B. Halliwell, J. Rafter, A. Jenner, Am. J. Clin. Nutr. 81 (2005) 268S−276S. [2] M. Galleano, S.V. Verstraeten, P.I. Oteiza, C.G. Fraga, Arch. Biochem. Biophys. 501 (2010) 23−30. [3] M.J. Frisch, G.W. Trucks, H.B. Schlegel et al., Gaussian 09, Wallingford, CT, 2009. [4] H. Iwasaki, L.A. Cohen, B. Witkop, J. Am. Chem. Soc. 85 (1963) 3701-3702.

It has been suggested that the rise in plasma antioxidant potency observed after the consumption of fruits and vegetables is not caused by the flavonoids (polyphenols), but is the consequence of increased uric acid levels. Polyphenol concentrations in human plasma are much lower ˂ 5 micromolar) than that of uric acid (~ 300 micromolar)[1]. Uric acid is the product of purine metabolism with both beneficial and harmful effects on human health [2]. This endogenous antioxidant may protect against oxidative stress by free radical scavenging [3]. Reaction enthalpies related to hydrogen atom transfer (HAT), single electron transfer- proton transfer (SET-PT), and sequential proton loss electron transfer (SPLET) mechanisms were studied using MOPAC2012 COSMO PM7 and Gaussian 09 DFT/M05-2X/6-311++G(d, p) SMD methods in water as a solvent [4]. Due to the lowest values of proton affinity (PA) and electron transfer enthalpy (ETE) both methods indicate double SPLET mechanism as thermodynamically preferred process. As the scavenging mechanisms are highly influenced by the properties of the scavenged free radical species, we used DFT method for calculating the free energy (∆G) of reactions of uric acid with ·OH, ·OOH, ·OCH3, ·O-O-CH=CH2, O2·- and ·OOCCl3 free radicals [5]. Exergonicity of double SPLET mechanism reactions in aqueous solution indicates that uric acid is efficient for scavenging of·OH and ·OOCCl3 free radicals.

Most Cited Papers in Croatica Chemica Acta

This report presents the best possible QSPR models for predicting the solubility of aliphatic alcohols in water that can be obtained with non-orthogonalized valence-connectivity basis. The corresponding models with orthogonalized basis, ordered in the usual manner, are also given. However, both models are statistically equivalent. Therefore, we proposed a novel approach to the QSPR modelling, which is based on the consideration of all possible orthogonalization orderings of the valence-connectivity basis and dominant descriptor analysis. This novel procedure produced models that are better than the empirical models of Amidon et al. and the Kier-Hall models when considering the same source of experimental data.

A novel multivariate linear regression, based on the ordered orthogonalized connectivity basis and dominant descriptor analysis, is applied to several physicochemical properties of alpha-amino acids. Our results compare favourably with those based on the non-orthogonalized connectivity basis by Pogliani.

An automated iterative method is developed for predicting secondary conformation in membrane proteins. The initial set of parameters are a- helix preferences and associated conformational preference functions extracted from the data set of known soluble protein structures. The secondary structure segments are assigned to each of 14 tested membrane proteins by using the prediction method, which evaluates and compares preference functions in the tested protein. A new set of parameters are then calculated which is based on the predicted protein structure from the previous iterative cycle. The method takes advantage of the similarities in local sequence patterns found in the tested proteins. Residues in membrane proteins are predicted with 84% accuracy and with the correlation coefficient for the a- helix structure equal to 0.68, which is a considerably better performance than that of neural network programs or Garnier-Robson’s algorithm.

U disertaciji je razvijen novi računalni postupak procjene ekotoksikološkog rizika vodenih staništa uzrokovan bioaktivnim kemijskim onečišćivalima sastavljen od tri cjeline: utvrđivanja stanja, određivanja rizika od postojećim metodama, te razvoja i primjene novih QSAR modela za procjenu rizika. Utvrđeno je da su vodena staništa rijeke Save opterećena kemijskim onečišćivalima od rastućeg značaja za okoliš. S obzirom na nedostatke postojećih QSAR metoda za procjenu rizika razvijenih na malom broju spojeva, novi robusniji QSAR modeli razvijeni su na nekoliko tisuća kemikalija iz baza ToxCast i Tox21 s brojnim i osjetljivijim toksičnim učincima izmjerenim na staničnim linijama i embrijima zebrice. Novi modeli temeljeni su na molekularnim deskriptorima i strukturnim otiscima, a dobiveni su logističkom regresijom, neuronskim mrežama i slučajnim šumama koje su dale ponajbolje modele. Dobiveni modeli pokazuju dobru generalizaciju na vanjskim skupovima onečišćivala, a njihova kvaliteta provjerena je novouvedenim metrikama. Procjene ekotoksikološkog rizika od spojeva utvrđenih u rijeci Savi temeljena na novim modelima toksičnosti pokazuju dobro slaganje s postojećim metodama, uz značajno povećanje kemijskog strukturnog prostora kojeg pokrivaju. A novel computational approach for assessing the ecotoxicological risk for water habitats caused by bioactive chemical pollutants is developed in this work. It consists of three parts: the determination of current ectoxicological status, the risk assessment based on existing methods, and the development and application of new QSAR models for risk assessment. Water habitats of the Sava River were found to be burdened with chemical pollutants of growing importance to the environment. Given the shortcomings of existing QSAR risk assessment methods developed on a small number of compounds, new, more robust QSAR models were developed herein on several thousand chemical compounds from the ToxCast and Tox21 databases which were assessed on more sensitive toxic endpoints measured on zebrafish embryos and cell lines. The newly developed models are based on molecular descriptors and structural fingerprints and are obtained by logistic regression, neural networks and random forests which gave the best models. The trained models show good generalization on external sets of chemical compounds. Their quality is validated by newly introduced metrics. Ecotoxicological risk assessment of compounds identified in the Sava River, is based on the new toxicity models which agree well with the existing methods, supported by a significant increase of structural space covered.

As an important part of the innate immune system of all organisms, antimicrobial peptides, are considered as a possible solution for fighting bacteria resistant to standard antibiotics.Crucial characteristic of peptide antibiotic, as drug candidate is its high selectivity, parameterized as the ratio of concentration causing 50% haemolysis (HC50) against erythrocytes and the minimal inhibitory concentration (MIC) against the reference bacterium Escherichia coli. Using the “Designer” algorithm adepantins were designed - highly selective artificial glycine and lysine rich peptides in predominant α helical conformation, having less than 50% homology of primary sequence to any known sequence of antimicrobial peptides. The algoritham used our database of anuran antimicrobial peptides with known selectivity index. Structure and activity of adepantins were experimentally and computationally tested in their monomeric, dimeric and fluorescently labelled form. Experimental investigations performed on different bacteria strains showed high selectivity of adepantins for Gram-negative bacteria ( MIC = 0.5 - 4 μM ), especially E. coli. In membrane permeabilization measurements, different membrane models were used and adepantins showed rapid permeabilization of both membranes of E. coli. These tests provided insight in their mode of action. All monomeric adepantins have exceptionally low haemolytic activity, while dimers expressed certain toxicity against host cells. It is proven that adepantins bind efficiently to the cell surface of the host cell membranes without subsequent membrane damage. Gathered results confirmed that adepantins are indeed highly selective artificial peptide antibiotics., Kao važan dio urođenog imunološkog sustava svih živih bića, antimikrobni peptidi, smatraju se možebitnim riješenjem u brobi protiv bakterija otpornih na standardne antibiotike. Važna osobina dobrog kandidata za budući lijek njegova je selektivnost, koja se određuje kao omjer koncentracije pri kojoj se opaža 50%-tna hemolitična aktivnost prema eritrocitima i minimalne inhibitorne koncentracije pri kojoj se opaža 100%-tna inhibicija rasta bakterije Escherichije coli. Algoritmom „Designer“ dizajnirani su adepantini–veoma selektivni umjetni peptid i u konformaciji α uzvojnice, bogati glicinom i lizinom, a sličnost njihovih sekvenci s poznatim antimikrobnim peptidima manja je od 50%. Algoritam je koristio našu bazu podataka antimikrobnih peptida iz anura s poznatim indeksom selektivnosti. Eksperimentalnim i računalnim metodama istražene su strukture i aktivnosti monomera, dimera i fluorescentno označenih oblika adepantina. Eksperimentalna istraživanja provedena na različitim sojevima bakterija pokazala su visoku selektivnost adepantina prema Gram-negativnim bakterijama ( MIC = 0.5 - 4 μM ), naročito prema E. coli. U mjerenju propusnosti membrana rabljeni su različiti modeli membrana, a pokazalo se da adepantini vrlo brzo povećavaju propusnost obje membrane E. coli. Time je dobiven uvid u mehanizme djelovanja adepantina. Svi monomeri adepantina imaju izrazito malu hemolitičnost prema eritrocitima, dok su dimeri pokazali određenu toksičnost prema ljudskim stanicama. Dokazano je da se adepantini vežu na staničnu površinu ljudskih stanica bez popratnog oštećenja membrane. Prikupljeni rezultati potvrdili su adepantine kao jako selektivne umjetne peptidne antibiotike.

Algoritam za izbor najboljih multivarijatnih modela (prema koeficijentu korelacije) razvijen je za potrebe primjena u istraživanju lijekova. Realiziran je u programskom jeziku Visual Basic i povezan je s bazom podataka. U radu je istražen odnos između statističkih parametara izračunatih na podacima u postupku prilagodbe na skupu za učenje, križne provjere ispuštanjem određenog postotka podataka skupa za učenje i u postupku predviđanja na vanjskom skupu podataka. Postupak križne provjere u kojem se u svakom koraku izbacuju veći podskupovi podataka pokazuje bolje slaganje s rezultatima dobivenim u predviđanju na vanjskom skupu nego najčešće korišteni postupak križne provjere u kojem se u svakom koraku izbacuje po jedna molekula. Nadalje, slaganje između statističkih parametara izračunatih u postupku prilagodbe na skupu za učenje i odgovarajućih parametara izračunatih u postupcima križnih provjera i na vanjskome skupu, znatno je bolje nego slaganja objavljivana u literaturi. To potvrđuje dobre strane primijenjenog postupka početne eliminacije deskriptora, smanjenja korelacije između deskriptora, postupka modeliranja i primijenjenoga algoritma za izbor modela. Za razliku od uobičajenih postupaka križne provjere koji se rabe u literaturi istražile su se učestalosti pojavljivanja pojedinih deskriptora u najboljim modelima. Na temelju tih analiza načinio se redoslijed važnosti pojedinih deskriptora u najboljim modelima, što je važna informacija pri interpretaciji i uporabi najboljih modela. Dobiveni rezultati i razvijena aplikacija MR_QSAR vrijedan su znanstveni i stručni doprinos području razvoja i primjene računalnih algoritama za modeliranja svojstava i aktivnosti molekula koja se provode u znanstvenim krugovima, u istraživanjima novih lijekova u farmaceutskoj industriji, te u zaštiti okoliša u postupcima procjene toksičnosti molekula. An algorithm for the selection of best multivariate regression models (according to correlation coefficient) was developed for the use in drug research. It is realized in the programming language Visual Basic and connected to a database. This work explores the relationship between the statistical parameters calculated in the process of fitting on the training data set, cross-validation by dropping a certain percentage of data study sets (k-fold cross-validation), and in the process of prediction on an external data set. Cross-validation procedure, in which a larger sub-set of data is omitted in each step, shows better agreement with results obtained in prediction on external set than the leave-one-out cross-validation procedure. Additionally, correlations between statistical parameters calculated in the fitting procedure on the training set and corresponding parameters calculated in the cross-validation procedures and in validations on external data sets are much better than it was published in literature. Such an analysis confirmed very good attributes of applied procedures for initial elimination of descriptors, reduction of inter-correlations between descriptors, and very good attributes of modeling method and algorithm used for model selection. Unlike a regular cross-validation procedures used in the literature, frequencies of occurences of individual descriptors in the best models were explored. Based on this analysis an order of importance of individual descriptors in best models was made. This information is important in the interpretation and use of best models. Obtained results and developed application MR_QSAR are valuable scientific and technical contribution to the development and application of computer algorithms for modeling molecular properties and activities carried out in scientific circles, in studies of new drugs in the pharmaceutical industry, and in environmental protection in procedures for assessment the toxicity of the molecules.

Cilj većine magnetoecefalografskih (MEG) studija je odrediti aktivna kortikalna područja i redoslijed njihove aktivacije. Pri tome se javlja problem razlučivanja aktivnih izvora, posebno ako su oni nalaze na maloj udaljenosti, imaju sličnu orijentaciju i istovremeno su aktivni (npr. Supek i Aine, 1993, 1997). Neuromagnetska istraživanja koja su koristila jednostavne vidne stimuluse malih dimenzija identificirali su 3-5 strujnih izvora tijekom rane aktivnosti do 170 ms poslije prezentiranja stimulusa (Aine i dr., 1996, Supek i dr., 1999). Mi smo u našim mjerenjima koristili lica kao kompleksne vidne stimuluse i to većih dimenzija. Prethodne studije s licima su uspjele identificirati najviše tri strujna dipola u prvih 200 ms nakon podražaja (Sams i dr., 1997 ; Linkenkaer-Hansen i dr., 1998 ; Swithenby i dr., 1998 ; Watanabe i dr., 1999 ; Halgren i dr., 2000). Budući da su empirijski podaci kompleksni a rješenje inverznog bioelektromagnetskog problema nije jedinstveno, numeričke simulacije MEG mjerenja omogućuju kontrolirano proučavanje utjecaja raznih parametara (npr. broja dipola, širine vremenskog intervala i sl.) na točnost rješenja. Prostorno-vremenska analiza naših empirijskih podataka iz MEG mjerenja s licima sugerira aktivnost više od tri strujna izvora u prvih 200 ms. Na osnovu preliminarnih rezultata analize empirijskih podataka jednog ispitanika zadane su lokacije i dinamika izvora za simulirane podatke. Na tim smo simuliranim mjerenjima proučavali utjecaj duljine vremenskog intervala na identifikaciju aktivnih izvora. Za dugačke vremenske intervale teško je odrediti broj izvora aktivnih tijekom cijelog intervala, a ponekad nije moguće ni razlučiti veliki broj izvora. Dijeljenjem intervala na manje odsječke dobiva se manji broj aktivnih izvora, potreban je manji broj početnih točaka i utrošeno računalno vrijeme je kraće. Pomoću simulacija smo ispitivali je li moguće variranjem duljine vremenskog intervala optimizirati točnost izračunatih parametara i utrošeno računalno vrijeme. Empirijski podaci jednog ispitanika upućuju na aktivnost čak 7 područja od 80 – 200 ms nakon početka prikazivanja nasmijanog lica. Simulacije potvrđuju mogućnost identifikacije tako velikog broja dipola uz povoljne okolnosti da su najbliži izvori dovoljno daleko (u našem slučaju minimalna udaljenost je 1.6 cm), da nisu istovremeno aktivni i da šum nije velik što je u skladu sa prijašnjim numeričkim simulacijama manjeg (Supek i Aine, 1997) ali i većeg broja izvora (Ranken i dr., 2004). U tom slučaju, i uz pretpostavku koja je u našim simulacijama bila zadovoljena, a to je da je adekvatan red modela jednak stvarnom broju aktivnih izvora u svakom od odsječaka, točnost rješenja ne ovisi o izboru vremenskog intervala.