Scientific paper Multivariate Regression Modelling of Antifungal Activity of Some Benzoxazole and Oxazolo[4,5-#]pyridine Derivatives Strahinja Z. Kova~evi},* Sanja O. Podunavac Kuzmanovi} and Lidija R. Jevri} University of Novi Sad, Faculty of Technology, Department of Applied and Engineering Chemistry, Bulevar cara Lazara 1, 21000 Novi Sad, Serbia * Corresponding author: E-mail: strahinjakovacevic@hotmail.com; phone: +381642839686; fax: +38121450413 Received: 28-03-2013 Abstract In the present study, principal component analysis (PCA) followed by principal component regression (PCR) and partial least squares (PLS) method was applied in order to identify the most important in silico molecular descriptors and quantify their influence on antifungal activity (expressed as minimal inhibitory concentration) of selected benzoxazole and oxazolo[4,5-b]pyridine derivatives against Candida albicans. PLS regression showed the best statistical performance, according to the lowest value of the standard error (root mean square errors of calibration of 0.02526 and cross-validation of 0.04533), while PCR model was characterized by root mean square errors of calibration of 0.03176 and cross-validation of 0.05661. The most important descriptors in both PLS and PCR model are solubility in water, expressed as AClogS and ABlogS, and lipophilicity, expressed as XlogP2 and ABlogP. Very good predictive ability of the established models, confirmed by corresponding statistical parameters, allows us to estimate antifungal activity of structurally similar compounds. Keywords: QSAR analysis; Principal component regression; Partial least squares; Candida albicans; Heterocyclic compounds. 1. Introduction Candida albicans is one of the most common fungal opportunistic pathogen of humans that can cause local, systemic and superficial mucosal infections (especially gastrointestinal, oral, respiratory and genital infections) in immunocompromised individuals, such as patients suffering from AIDS, leukemia or diabetes. Candidiasis, an infection caused by Candida species, is usually treated with antifungal drugs: amphotericin B, fluconazole, ketocona-zole and nystatin.1 Because of the increasing incidence of both fungal infections and antifungal drug resistance, synthesis and analysis of some novel antifungal compounds are welcome. Benzoxazoles and oxazolo[4,5-b]pyridines, as analogues of benzimidazole, are well known to the chemists, mainly due to their wide spectrum of antimicrobial pro-perties.2-6 It is also determined that these molecules are present in a variety of herbicidal, antihelmintic, antioxi- dant and antitumoral agents.7-10 Antimicrobial activity of studied compounds is expressed as minimal inhibitory concentration (MIC) defined as the lowest concentration of the compound at which no growth of the strain is observed in time and under specified experimental conditions. Prediction of antimicrobial activity of compounds, based on their structural characteristics, is a very important and fundamental issue of pharmaceutical chemistry. Quantitative structure-activity relationships (QSAR) analysis allows us to estimate the biological activity of novel molecules prior to their synthesis, according to statistically significant mathematical models based on a large number of already synthesized molecules.1112 QSAR studies are widely applied in quantitative description of relationships between the chemical structure of a drug molecule and its biological activity, aiming at defining optimal values for some physicochemical properties of the molecule and providing the fundamentals for design of new substances as drug candidates, according to current needs.1314 Beside importance of QSAR modelling in drug design, in silico methods are important contributors to drug discovery processes.15 In the present paper, antifungal activity of eighteen benzoxazole derivatives and six oxazolo[4,5-b]pyridine derivatives was estimated according to their in silico molecular descriptors using principal component analysis (PCA), followed by principal component regression (PCR) and partial least squares (PLS) chemometric methodologies. In our previous work11, we have already studied the influence of some molecular descriptors of ben-zoxazoles on their in vitro antifungal activity against Candida albicans using multiple linear regression (MLR), therefore the novelty of the present study is the extended series of the studied compounds and the application of PCR and PLS for the same purpose. 2. Materials and Methods The QSAR analysis was performed in the following several steps: molecular structure optimization by computer software, structural descriptors computation, structural descriptors selection, structure-activity model generation using PCR and PLS methods, and statistical validation. 2. 1. Studied Compounds The structures of benzoxazoles and oxazolo[4,5-b] pyridines investigated in this paper are presented in Table 1. The results of their in vitro antifungal activity against Candida albicans (MTCC 183) are presented in literature.16 The logarithm of molar MIC (log(1/cMIC)) was used for further calculations (Supporting information, Table S1). 2. 2. Molecular Modelling and Molecular Descriptors In silico modeling of examined molecules was performed by using following software: CS ChemBioDraw Ultra 12.0 for drawing 2D structures of molecules, and CS ChemBio3D Ultra 12.0 for 3D molecular modelling running on AMD Sempron Processor 3000+.17 The constructed 3D models were subjected to energy minimization using molecular mechanics force field method (MM2). The cutoff for structure optimization was set at a gradient of 0.1 kcal/Amol. The Austin Model 1 (AM1) was used for full geometry optimization of all structures until the root mean square (RMS) gradient reached a value smaller than 0.0001 kcal/Amol using Molecular Orbital Package (MOPAC) program.18 Table 1. The IUPAC names and chemical structures of the compounds studied. Comp. IUPAC name X Ri R2 1 2-phenyl-1,3-benzoxazole CH H H 2 2-(4-tert-butylphenyl)-1,3-benzoxazole CH C(CH3)3 H 3 4-(1,3-benzoxazol-2-yl)aniline CH NH2 H 4 4-(1,3-benzoxazol-2-yl)-N-methylaniline CH NHCH3 H 5* 5-chloro-2-(4-ethylphenyl)-1,3-benzoxazole CH C2H5 Cl 6 N-[4-(5-chloro-1,3-benzoxazol-2-yl)phenyl]acetamide CH nhcoch3 Cl 7 4-(5-chloro-1,3-benzoxazol-2-yl)-N-methylaniline CH NHCH3 3 Cl 8 5-chloro-2-(4-chlorophenyl)-1,3-benzoxazole CH Cl 3 Cl 9* 5-chloro-2-(4-nitrophenyl)-1,3-benzoxazole CH NO2 Cl 10 2-(4-ethylphenyl)-1,3-benzoxazol-5-amine CH C2H5 NH 11* 2-(4-fluorophenyl)-1,3-benzoxazol-5-amine CH F NH 12 5-methyl-2-(4-methylphenyl)-1,3-benzoxazole CH CH3 CH 13 2-(4-ethylphenyl)-5-methyl-1,3-benzoxazole CH C2H5 CH 14 2-(4-methoxyphenyl)-5-methyl-1,3-benzoxazole CH och3 CH 15 2-(4-fluorophenyl)-5-methyl-1,3-benzoxazole CH F3 CH 16 N- [4-(5-methyl-1,3-benzoxazol-2-yl)phenyl] acetamide CH nhcoch3 CH 17* N-methyl-4-(5-methyl-1,3-benzoxazol-2-yl)aniline CH NHCH3 3 CH 18 N,N-dimethyl-4-(5-methyl-1,3-benzoxazol-2-yl)aniline CH N(CH3)32 CH 19 2-(4-methylphenyl)- [ 1,3]oxazolo [4,5-ft]pyridine N CH3 H 20 2-(4-ethylphenyl)-[1,3]oxazolo[4,5-ft]pyridine N C2H5 H 21 2-(4-methoxyphenyl)- [1,3] oxazolo[4,5-ft] pyridine N och3 H 22 2-(4-ethoxyphenyl)- [ 1,3] oxazolo[4,5-ft]pyridine N OC2H5 H 23* 4-{[ 1,3]oxazolo[4,5-ft]pyridin-2-yl}aniline N NH2 H 24* 2-(4-nitrophenyl)-[1,3]oxazolo[4,5-ft]pyridine N NO2 H 2 2 3 3 3 3 3 3 3 *External test set The values of molecular descriptors for each compound in the data set were calculated using the software CS ChemBio3D Ultra 12.0 and ALOGPS 2.1.1719 These data are presented in Table S2 (Supporting information). Determined descriptors of the examined molecules were physicochemical descriptors (relative molecular mass -Mr, boiling point - BP [K], melting point - MP [K], critical temperature - CT [K], critical pressure - CP [bar], critical volume - CV [cm3/mol], partition coefficients for n-octanol/water bi-phase system - AlogPs, ACDlogP, ABlogP, milogP, AlogP, MlogP, XlogP2 and XlogP3, solubility in water - ABlogS and AClogS) and molecular bulkiness descriptors (molar refractivity - MR [cm3/mol], van der Waals surface area - vdWSA [A2]). Other types of molecular descriptors of benzoxazoles and oxazolo[4,5-b]pyridines (electrostatic and topological descriptors) were applied as predictors of their antifungal activity in our other study that includes application of artificial neural networks as chemometric method (not published). 2. 3. Multivariate Statistical Methods: PCA, PCR and PLS The main objective of PCA is to substitute the representation of the objects, from the initial representation in the form of the n original intercorrelated variables, into the new principal component coordinate space.20 Therefore, PCA can be defined as a statistical technique for reducing the amount of data when there is a correlation present. It is worth stressing that it is not a useful technique if the variables are uncorrelated.21 In PCA, objects or analy-tes are represented in a multidimensional space, where the variables define the axes, and are projected into a few principal components (PCs) which are linear combination of the original variables and describe the maximum variation within the data. Each PC is characterized by scores, which actually are the new coordinates of the projected objects, and by loadings which reflect the direction with respect to the original variables.22 In addition, PCA is a very useful tool in providing data overview and determination of the outliers among the analytes (data lying outside the Hotelling T2 ellipse). The aim of PCR is to reduce the number of predictor variables by using first few PCs rather than the original variables. This statistical method works well when there is a considerable degree of correlation between the predictor va-riables.21 That can cause mathematical problems with MLR, resulting in unreliable predictions. PCR usually implies three main steps: (1) running the PCA on the table of the explanatory variables, (2) running an ordinary least squares regression (linear regression) on the selected components: the factors that are most correlated with the dependent variable will be chosen, and (3) computation of the parameters of the model for the selected explanatory variables.20 Like PCR, PLS regression uses linear combinations of the predictor variables rather than the original variab- les. In PCR the PCs are chosen so that they describe as much of the variation in the predictor variables as possible, irrespective of the strength of the relationships between the predictor and the response variables. However, in PLS variables that show a high correlation with the response variables are given extra weight because they will be more effective at prediction.21 As a consequence, PLS finds components that both show high variation and are highly correlated with dependent variable. Therefore, PCR and PLS enable analysis of strongly collinear data, reducing the high-dimensional data matrix to a much smaller and interpretable set of latent variables (LVs). The optimal complexity and predictivity of the models are usually determined by cross-validation. In the case of both PCR and PLS regressions, statistical and predictive quality of the models are evaluated by cumulative sum of squares of the Ys explained by all extracted components (R2Ycum), root mean square error of calibration (RMSEC), root mean square error of prediction (RMSEP), root mean square error of cross-validation (RMSECV), determination coefficient of calibration (R2cal), determination coefficient of prediction (R2pred) and determination coefficient of cross-validation (R2cv). The complete PCA, PCR and PLS calculation procedures were conducted by using a demo version of PLS Toolbox statistical package for MATLAB version 7.12.0.635 R2011a.23 The data were separated in two subsets: calibration set with 18 molecules, and external test set with 6 randomly selected molecules (Table 1). 3. Results and Discussion 3. 1. PCA PCA was carried out in order to determine the presence of outliers among the analytes with 0.95 confidence level for T2 Hotelling limit for outliers and to overview the examined compounds for similarities and dissimilarities. PCA resulted in a three-component model explaining 92.56% of the data variation. The cumulative variance explained by the first two PCs is 87.39% (PC1 comprises 60.45% and PC2 comprises 26.94% of the total data variability). The addition of more PCs did not significantly change the distribution of the molecules on the score plot. Figure 1 shows score values and the mutual projections of the loading vectors for the first two PCs. It is obvious that all the compounds are lying inside the Hotelling T2 ellipse, suggesting that there are no outliers among the analy-tes. Unfortunately, score plot did not reveal any classification of the compounds, except significant distance of compound 1, 6 and 16 from the other compounds going along the PC2 direction. It could be assumed that this separation of mentioned compounds is caused by their structural characteristics: compound 1 is unsubstituted, and compounds 6 and 16 are the only acetamides in the series with significantly different values of certain mole- a) [ig A * 2- I i V23 •22 S 0 § S -2 h to i- .+.-----»21-vn--- •3 *4*20l •19 1*10- I _________| '14 •2 ! »12 13 ♦ 15 •1 X 1_,_,_i_ -10 -S -6 -4 -2 0 2 4 5 S Scores on PC 1 (60.45%) b) Figure 1. Score plot (a) and factor loadings (b) of molecular descriptors for the first two PCs. cular descriptors. Loading plot shows that partition coefficients have the highest positive influence going along the PC1 axis, while water solubility descriptors and CP express the highest negative influence. The highest positive influence on the PC2 score values have MP, BP and CT. was preformed by splitting the entire calibration set into three random subsets and one iteration. After building the calibration model and application of cross-validation, the obtained model was applied on external data set and calculated RMSEP amounts to 0.04143 followed by R2pred of 0.9180. Figure 2a shows the plot of the experimentally measured log(1/cMIC) versus predicted log(1/cMIC) for PCR model. a) b) tojrtUCniKÏ Experwnental Figure 2. The plot of the experimentally measured versus predicted log(1/cMIC) values obtained from (a) the PCR and (b) PLS model. Low scattering of the points around the linear relationship curve, low values of errors and high determination coefficients indicate a very good concurrence between experimental and predicted log(1/cMIC) values, as well as a very good predictive power of the model. This statement is also confirmed by the residual versus predicted log(1/cMIC) values plot, which is presented in Figure 3a. This plot is very informative regarding model fitting to a data set. If the residual values are randomly distributed (low correlation), then it implies that the model fits the da-del shows R2Ycum = 92.99%, RMSEC = 0.03176, RMSECV ta well.25 Therefore, Figure 3a indicates that PCR model = 0.05661, R2cal = 0.9299, R2cv = 0.8105. Cross-validation fits adequately to all the data. 3. 2. PCR Simple methods for selection of a good set of PCA scores for PCR are (1) selection of the first PCA scores that cover a certain percentage of the total variance of X (i.e. 95%) and (2) selection of the PCA scores with maximum correlation to Y.24 In this paper, we applied first approach and selected the first six PCA scores which cover 98.62% of the total variance of X. The obtained PCR mo- a) f TEST SET • CALIBRATION SET m log(1/CmK|Pr«sclM! b) Figure 3. The residual versus predicted log(1/cMIC) values plot for (a) the PCR and (b) PLS model. 3. 3. PLS The PLS model calibration was carried out on the entire calibration data set. The obtained model was evaluated using random subsets (three splits and one iteration) cross-validation method. The number of the LVs was chosen on the basis of the minimum RMSECV value, which was obtained for five LVs model. Five LVs capture 92.89% of the variance in the descriptor variables space, indicating that the information contained in the descriptors are effectively used in the calibration model. The obtained PLS model is also characterized by R2cal = 0.9557 and R2cv = 0.8606. The explained calibration variance R2Ycum for the dependent variable log(1/cMIC) was 94.16% and RMSEC was 0.02526 units of log(1/cMIC). The error of validation, expressed as RMSECV, was 0.04533 units of log(1/cMIC). After calibration, established model was tested by external test set. The obtained RMSEP was 0.04068 and R2pred was 0.9328. Comparison of the experimentally obtained log(1/cMIC) values and predicted log(1/cMIC) values is showed in Figure 2b. It shows low scattering of the points around the linear relationship cur- ve indicating very good concurrence between predicted and measured data. Residual versus predicted log(1/cMIC) values plot for PLS model is presented in Figure 3b. Randomly distributed residual values imply excellent fitting of the data. Depicted results indicate a very good predictive power of the established PLS model, which is better than PCR model. The reason why the PLS model has better statistical measures and predictive power than PCR model lies in the fact that PLS technique uses both criteria: the direction of the highest variance in the data set and the best correlation with the dependent variable scores in its search for the latent variables in the X-variable and Y-variable domain. However, both methods PCR and PLS are very important in chemometrics because of ability to model a large number of highly intercorrelated variables. The appearance of outliers in PCR and PLS models could be explained with significant differences in molecular structure as well as with prediction error of the model. 3. 4. Contribution of the Molecular Descriptors to the MIC of Studied Compounds The assessment of the contribution of variables (molecular descriptors) on Y (log(1/cMIC)) in both the PCR and PLS models, was done based on variable importance in the projection (VIP) scores. The descriptors with a VIP score higher than 1 were considered as relevant for explaining Y, and those significantly lower than 1 (arbitrarily, the value lower than 0.5 was taken) had little or almost no influence. The descriptors characterized by VIP > 1.5 were considered as the most significant. The variables versus VIP scores for Y plot is presented in Figure 4. Figure 4. Plot of the variables versus VIP scores for log(1/cMIC) The plots of the regression coefficients of descriptors in PCR and PLS models are showed in Figure 5, where the descriptors with VIP > 1.5 are denoted by asterisks. a) : b) tunately, PCA did not reveal any significant classification of studied compounds according to calculated molecular descriptors. However, PCR and PLS methods showed that antifungal activity of benzoxazoles and oxazolo[4,5-b]pyridines is mostly influenced by their lipophilicity, expressed as ABlogP and XlogP2, as well as solubility in water, expressed as AClogS and ABlogS. In both PCR and PLS models, mentioned descriptors are the most important, based on VIP factors and regression coefficients. Taking into account the statistical parameters that represent the quality and predictivity of the model, it can be concluded that PLS regression resulted as better than PCR. The present study allows us to estimate antifungal activity for similar compounds and to understand their biological behaviour. 5. Acknowledgement These results are the part of the projects No. 172012, No.172014 and TR 31055 supported by the Ministry of Education, Science and Technological Development of the Republic of Serbia and the project No. 114451-2373/2011, financially supported by the Provincial Secretariat for Science and Technological Development of Vojvodina. 6. References Figure 5. Plot of the coefficients of molecular descriptors in (a) the PCR and (b) PLS model. Hence, the most significant descriptors influencing the log(1/cMIC) values of studied benzoxazoles and oxazo-lo[4,5-b]pyridines are: solubility in water (AClogS, ABlogS) and partition coefficients (ABlogP, XlogP2). It is evident that the same descriptors are included in both PCR and PLS models. These descriptors are characterized by the highest regression coefficients and VIP values. Solubility in water and lipophilicity of molecules are very important factors of biological activity,26 in this case antifungal activity. Therefore, on the basis of mentioned descriptors, antifungal activity of studied compounds could be successfully predicted by using appropriate mathematical model. 4. Conclusions The present paper focuses on identifying the most significant molecular descriptors that have the highest influence on antifungal activity of some benzoxazole and oxazolo[4,5-b]pyridine derivatives against Candida albicans. For this purpose, PCA followed by PCR and PLS regression chemometric methodologies was applied. Unfor- 1. O. Ursu, A. Costescu, M. V. Diudea, B. Parv, Carpatian J. Math. 2004, 20, 267-274. 2. O. Temiz-Arpaci, Turk. J. Med. Sci. 2001, 31, 493-497. 3. Z. A. Kaplancikli, G. Turan-Zitouni, G. Revial, K. Guven, Arch. Pharm. Res. 2004, 27, 1081-1085. 4. O. Temiz, I. Oren, E. Sener, I. Yalgin, N. Ugarturk, Farmaco. 1998, 53, 337-341. 5. M. Arisoy, O. Temiz-Arpaci, I. Yildiz, F. Kaynak-Onurdag, E. Aki, I. Yalgin, U. Abbasoglu, SAR QSAR Environ Res. 2008, 19, 589-612. 6. T. Ertan, I. Yildiz, B. Tekiner-Gulbas, K. Bolelli, O. Temiz-Arpaci, S. Ozkan, F. Kaynak, I. Yalgin, E. Aki, Eur. J. Med. Chem. 2009, 44, 501-510. 7. L. Ouyang, Y. Huang, Y. Zhao, G. He, Y. Xie, J. Liu, J. He, B. Liu, Y. Wei, Bioorg Med Chem Lett. 2012, 22, 3044-3049. 8. E. I. Khizhan, A. I. Khizhan, A. N. Nikolaevskii, O. P. Kniga, T. N. Ivleva, G. A. Tikhonova, Russ J Gen Chem, 2011, 81, 122-127. 9. I. Yildiz-Oren, B. Tekiner-Gulbas, O. Temiz-Arpaci, I. Yal-gin, E. Aki-Sener, Asian J. Chem. 2004, 16, 1359-1366. 10. R. Gudipati, R. N. Anreddy, S. Manda, J Enzyme Inhib Med Chem. 2011, 26, 813-818. 11. S. O. Podunavac-Kuzmanovic, L. R. Jevric, S. Z. Kovacevic, N. D. Kalajdzija, APTEFF 2012, 43, 273-282. 12. S. O. Podunavac-Kuzmanovic, D. D. Cvetkovic, L. R. Jevric, N. U. Uzelac, Hem. Ind. 2013, 67, 27-33. 13. S. W. Dietrich, in: M. E. Wolff (ed.): Burger's Medicinal Chemistry and Drug Discovery: Principles and Practice, Vol. 1, 5th ed. John Wiley, New York, USA, 1995, pp. 415-496. 14. C. Hansch, A. Leo, Exploring QSAR: Fundamentals and Applications in Chemistry and Biology, American Chemical Society, Washington DC, USA, 1995. 15. S. O. Podunavac-Kuzmanovic, D. D. Cvetkovic, L. R. Jevric, N. J. Uzelac, Acta Chim. Slov. 2013, 60, 26-33. 16. O. Ursu, A. Costescu, M. V. Diudea, B. Parv, Croat. Chem. Acta 2006, 79, 483-488. 17. ChemBioOffice 2010, PerkinElmer Informatics, http://www.cambridgesoft.com/ 18. M0PAC2012, James J. P. Stewart, Stewart Computational Chemistry, Colorado Springs, CO, USA, http://0pen-MOPAC.net 19. VCCLAB, Virtual Computational Chemistry Laboratory, http://www.vcclab.org/ 20. K. H. Esbensen, Multivariate Data Analysis - In Practice: An Introduction to Multivariate Data Analysis and Experimental Design, 5th Edition, CAMO Software AS, USA, 2009. 21. J. N. Miller, J. C. Miller, Statistics and Chemometrics for Analytical Chemistry, 6th Edition, Pearson Education Limited, Harlow, UK, 2010. 22. R. G. Brereton, Chemometrics, Data Analysis for the Laboratory and Chemical Plant, Wiley, Chichester, England, 2003. 23. The MathWorks Inc, Natick, MA, USA, http://www.math-works.com 24. K. Varmuza, P. Filzmaster, Introduction to Multivariate Statistical Analysis in Chemometrics, CRC Press, Taylor & Francis Group, Boca Raton, 2008. 25. J. Trifkovic, F. Andric, P. Ristivojevic, D. Andric, Z. Lj. Tesic, D. M. Milojkovic-Opsenica, J. Sep. Sci. 2010, 33, 2619-2628. 26. R. Mannhold, Molecular Drug Properties, Measurement and Prediction, Vol. 37, Wiley-VCH, Darmstadt, Germany, 2008. Povzetek QSAR metoda je bila uporabljena v seriji benzoksazolskih in oksazolo[4,5-b]piridinskih derivatov različnih struktur s ciljem ugotavljanja njihovih inhibitorskih aktivnosti na glivo Candida albicans. Analiza glavnih komponent (Principal Component Analysis - PCA), regresija glavnih komponent (Principal Component Regression - PCR) in metod delnih najmanjših kvadratov (Partial Least Squares - PLS) so bili uporabljeni za identifikacijo najustreznejše deskriptorjev in za modeliranje odnosa med različnimi fizikalno-kemičnimi molekularnimi deskriptorji in antifungalne aktivnosti preizkušenih derivatov. Ugotovljeno je da imajo na inhibitorsko aktivnost derivatov benzoksazola in oksazolo[4,5-b]piridina na C. albicans največji vpliv deskriptorji lipofilnosti (ABlogP in XlogP2) in topnost v vodi (AClogS in ABlogS). PLS regresija je pokazala statistično najboljšo učinkovitost (RMSEC = 0.02526 in RMSECV = 0.04533) od PCR (RMSEC = 0.03176 in RMSECV = 0.05661). SUPPORTING INFORMATION Multivariate Regression Modelling of Antifungal Activity of Some Benzoxazole and Oxazolo[4,5-#]pyridine Derivatives Strahinja Z. Kova~evi},* Sanja O. Podunavac Kuzmanovi} and Lidija R. Jevri} University of Novi Sad, Faculty of Technology, Department of Applied and Engineering Chemistry, Bulevar cara Lazara 1, 21000 Novi Sad, Serbia * Corresponding author: E-mail: strahinjakovacevic@hotmail.com; phone: +381642839686; fax: +38121450413 Received: 28-03-2013 Table S1. In vitro antifungal activity (log(1/cMIC)) of studied compounds against Candida albicans (MTCC 183)16 Comp. IUPAC name log(1/cMIC) 1 2-phenyl-1,3-benzoxazole 3.892 2 2-(4-tert-butylphenyl)-1,3-benzoxazole 4.001 3 4-(1,3-benzoxazol-2-yl)aniline 3.924 4 4-(1,3-benzoxazol-2-yl)-N-methylaniline 3.952 5* 5-chloro-2-(4-ethylphenyl)-1,3-benzoxazole 4.013 6 N-[4-(5-chloro-1,3-benzoxazol-2-yl)phenyl]acetamide 4.059 7 4-(5-chloro-1,3-benzoxazol-2-yl)-N-methylaniline 4.015 8 5-chloro-2-(4-chlorophenyl)-1,3-benzoxazole 4.024 9* 5-chloro-2-(4-nitrophenyl)-1,3-benzoxazole 4.040 10 2-(4-ethylphenyl)-1,3-benzoxazol-5-amine 3.979 11* 2-(4-fluorophenyl)-1,3-benzoxazol-5-amine 3.960 12 5-methyl-2-(4-methylphenyl)-1,3-benzoxazole 4.005 13 2-(4-ethylphenyl)-5-methyl-1,3-benzoxazole 3.950 14 2-(4-methoxyphenyl)-5-methyl-1,3-benzoxazole 3.977 15 2-(4-fluorophenyl)-5-methyl-1,3-benzoxazole 3.980 16 N-[4-(5-methyl-1,3-benzoxazol-2-yl)phenyl]acetamide 3.958 17* N-methyl-4-(5-methyl-1,3-benzoxazol-2-yl)aniline 4.027 18 N,N-dimethyl-4-(5-methyl-1,3-benzoxazol-2-yl)aniline 3.979 19 2-(4-methylphenyl)-[1,3]oxazolo[4,5-b]pyridine 4.225 20 2-(4-ethylphenyl)-[1,3]oxazolo[4,5-b]pyridine 4.253 21 2-(4-methoxyphenyl)- [ 1,3]oxazolo [4,5-b]pyridine 4.257 22 2-(4-ethoxyphenyl)-[1,3]oxazolo[4,5-b]pyridine 4.283 23* 4-{[1,3]oxazolo[4,5-b]pyridin-2-yl}aniline 4.227 24* 2-(4-nitrophenyl)-[1,3]oxazolo[4,5-b]pyridine 4.285 H3 U A -a -o -a o .c H o - o H O - « < in £ T3 > ce M _0 EE < cc M (5 o - M (5 * - M (5 * - M O - M O - M O - M _0 EE < - M (5 S o < - M O - s s a • ogo o az o o o o o o o o VO o o o O o o o o o o IN CO o ■a IN t- CO i—i in VO VO p~ o\ IN o CO IN CO CO in o o\ CD P-; 00 IN 00 00 CO CM 00 00 CM 00 VO 00 I/O 00 P^ IN p^ CO I/O ,—H p^ cK CO CO cK CM p^ CO cK I/O CM p^ CD CO in t- VO VO t- VO VO VO VO p~ VO p~ p~ p- VO VO VO VO VO VO o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o 1/0 1/0 I/O I/O I/O I/O 1/0 I/O I/O 1/0 1/0 1/0 I/O 1/0 1/0 I/O I/O I/O I/O I/O I/O I/O I/O ^ o cm cm CD cK \D CM 00 CO cK CO p^ 00 p^ CD \D 00 ■'t CO CD vo t- o\ in IN in o in in o o p~ IN a\ in in o VO IN 00 00 o in t- in VO t- VO VO VO VO P~ VO VO p~ p~ VO VO VO VO in VO o o o o o o o o r- o o o O o o o o o o o o o o ^ CO i—i o\ 1—1 o o\ in in o CO o\ o\ VO in IN o\ IN 1—1 in o o ay vo p^ 1/^ CM p^ c^ c^ c^ T-H a) 1/^ V a) CM a) p^ cK cm p^ ,—H cK CD \D 00 V1D 00 CD cK cK I/O V1D CM I/O —H p^ CO IN CO CO IN CO CO CO CO CO CO IN IN IN CO IN IN IN CO CO CO CO o o o o O o o o o o o O o o O o o o o o o o 1—1 CO in in i—i CO 00 in CO o\ in VO P~ 00 in p~ VO IN 00 00 o vo co c^ 00 c^ c^ c^ c^ c^ V 00 CM c^ CO c^ c^ ^ CD 00 ^ cK VO V1D 00 CD CM CM ■'t p^ cK \D CD CM ■'t p^ I/O ^ vo t- o\ t- t- 00 00 o\ o 00 VO VO VO ^ c^ P~ p- o 00 00 00 00 00 00 00 00 00 o\ 00 00 00 00 00 o\ 00 00 00 00 00 00 o\ 00 o o o o o o o o o o o o o o o o o o o o o o o o o IN 00 in o 1—1 o\ 00 in o\ 00 in 1—1 o VO IN VO CO o\ VO in 00 t- t- ■