Regression Analysis (RA) The application of the biochromatographic data in QSAR analysis of some thiazole derivatives with H-antihistamine

activity was described in the previous papers ( !"# Additionally$ the lipophilicity data of solutes were applied as independent variables in the regression analysis# %n the basis of described results it was found that log & is a crucial indicator of the H'-antihistamine effect of thiazole derivatives# An increase in the log & value favors higher biological activity of the tested compounds# (umerous significant multivariate relationships of the antihistamine effect involved log & values ( )!"# The present RA started with intercorrelation study of the used independent variables# %nly the uncorrelated physicochemical data can be used in the multivariate relationships (see Table !"# (e*t$ the systematic analysis was performed# As result$ over +, statistically significant relationships were determined# The obtained relationships of H-antihistamine effect and molecular descriptors values e*plained --).!/ of the variance# The best univariate and multivariate relationships are shown in Table 0# 1t is evident in QSAR investigation that the best correlations obtained by the RA method for thiazole derivatives 1-19 with H'-antihistamine activity underline important role of hydrophobic and steric parameters for this 2ind of activity# The same parameters (A 34 54 log &4 log 64 7R4 a" with strong factor loadings built the most significant factor$ which was obtained in factor analysis of the investigated compounds 1-19 ('"# 1t is also evident that some of the electronic parameters are very important# These parameters were lin2ed to higher and lower compounds activity via 68A (a$ A9$ HeH%7%$ QAr3" and &:A (a$ log &$ Hh$ H%7%$ Q" methods# All calculated significant relationships can be applied to predict the pharmacological activity of new drug candidates# The best of these relationships can be e*pressed by the e;uation 20 (Table 0"$ which e*plains .!/ of the total variance< pA= !#'0(> ,#,!" ? ,#-'(> ,#,0" log & ? ,#'@@(>,#,0" m ) ,#'--(> ,#,!" Q ArArh$ 1t was concluded that the high lipophilicity and dipole moment combined with less negative charge on aromatic nitrogen atom are the properties of active H-antihistamine thiazole derivatives# 1t is clearly seen that e;uation 20 may have predictive value for the design of new thiazole derivatives as the H -antihistamine drugs (Table and 8ig# @"# The correlation of calculated pA'" values of the tested compounds predicted by the use of the above mentioned e;uation versus their pA (H" values obtained from biological tests was significant (R'= ,#.!"# However$ the range of pA data of the e*amined compounds obtained from biological tests clustered around two sets (compounds 1-11 have pA values between A#,, and !#,A4 compounds 12-19 have pA values between !#+and 0#@+"# 8or the two-point data distribution the possibility of coincidence in the model presented in the figure cannot be eliminated# CONCLUSION The dimensionality of physicochemical parameters was reduced by the &:A and 68A methods$ And the subset of variables more effective for classification the thiazole derivatives according to their degree of anti-H activity were determined# The &:A method can be useful as an efficient tool for initial selection of the parameters which significantly enhance anti-H' activity# The analysis determined the direction of the lead compounds modification# The results of 68A method showed that $ A9$ B'$Hh$ eH%7% and Q parameters are 2ey properties for e*plaining the HAr-antihistamine activity of thiazole derivatives 1-19 but log & is also important for design of new thiazoles e*hibiting antihistamine activity# The determined discrimination function for groups A and 9 can be an efficient tool in further investigations# Cood univariate and multivariate relationships obtained by the use of RA method can be used for predicting the ;uantitative effect of H ' antihistamine activity of different thiazole derivatives# These relationships involved the parameters determined via 68A and &:A methods#

Analisis Regresi ( RA ) Penerapan data biochromatographic dalam analisis QSAR beberapa turunan tiazol dengan aktivitas H1 - antihistamin digambarkan pada paper sebelumn a ( !-" ) # Selain itu $ data lipophilicit zat terlarut ang diterapkan digunakan sebagai variabel independen dalam analisis regresi # Atas dasar hasil ang ditemukan bah%a log P merupakan indikator penting dari e&ek H1 - antihistamin dari derivati& thiazole # Peningkatan nilai log P akan men okong aktivitas biologis ang lebih tinggi dari sen a%a ang diu'i # (an ak hubungan multivariat ang signi&ikan dari e&ek antihistamin melibatkan nilai log P ( !-" ) # Pemaparan RA dimulai

dengan studi intercorrelation variabel independen ang digunakan # Han a data &isikokimia tidak korelasi ang dapat digunakan dalam hubungan multivariat ( lihat )abel " ) # Selan'utn a$ analisis sistematis akan ditampilkan# *an hasiln a$ lebih dari +, hubungan ang signi&ikan secara statistik ditentukan# Hubungan ang diperoleh dari e&ek H1 - antihistamin dan nilai-nilai deskriptor molekul men'elaskan ---." / dari varians # Hubungan univariat dan multivariat terbaik ditun'ukkan pada )abel 0 # Hal ini terbukti dalam pen elidikan QSAR bah%a korelasi terbaik diperoleh dengan metode RA untuk thiazole derivati& 1-1. dengan aktivitas H1 - antihistamin menggaris ba%ahi peran penting parameter hidro&obik dan sterik untuk kegiatan semacam ini # Parameter ang sama ( A1 $ 2% $ log P $ log * $ 3R $ a) dengan &aktor loadings ang kuat dibangun &aktor ang paling signi&ikan $ ang diperoleh dalam analisis merupakan &aktor sen a%a diselidiki komponen 1-1. ( 1 ) # Hal ini 'uga 'elas bah%a beberapa parameter elektronik sangat penting # Parameter ini terkait dengan sen a%a aktivitas ang lebih tinggi dan lebih rendah melalui *4A (a $ A( $ HeH535 $ QAr1 ) dan P6A (a $ log P $ Hh $ H535 $ Q ) metode # Semua hubungan ang signi&ikan dihitung dapat diterapkan untuk memprediksi aktivitas &armakologi dari calon obat baru# 7ang terbaik dari hubungan ini dapat din atakan dengan persamaan !, ()abel 0)$ ang men'elaskan ."/ dari total varian8
pA= !#'0(> ,#,!" ? ,#-'(> ,#,0" log & ? ,#'@@(>,#,0" m ) ,#'--(> ,#,!" Q ArArh$

*isimpulkan bah%a lipophilicit tinggi dan momen dipol dikombinasikan dengan muatan kurang negati& pada atom nitrogen aromatik merupakan si&at akti& derivati& thiazole H1-antihistamin# Hal ini 'elas terlihat bah%a persamaan !, mungkin memiliki nilai predikti& untuk desain baru derivati& thiazole sebagai obat H-(antihistamin )abel - dan 9ambar# :)# ;orelasi dihitung PA1) nilai-nilai sen a%a ang diu'i diperkirakan dengan menggunakan persamaan ang disebutkan di atas dibandingkan pA! mereka (H) nilai ang diperoleh dari tes biologis signi&ikan (R1 < ,$.")# =amun$ berbagai data sen a%a pA! ang diperiksa diperoleh dari tes biologis berkerumun di sekitar dua set (1-11 sen a%a memiliki nilai pA! antara >$,, dan "$,>? sen a%a 1!-1. memiliki nilai pA antara "#+- dan 0#:+)# @ntuk distribusi data dua titik kemungkinan kebetulan dalam model disa'ikan dalam gambar tidak bisa dihilangkan# ;ASB3P@CA= *imensi dari parameter &isika berkurang oleh P6A dan metode *4A $ *an subset dari variabel ang lebih e&ekti& untuk klasi&ikasi turunan tiazol menurut tingkat aktivitas anti - H ditentukan # 3etode P6A dapat berguna sebagai alat ang e&isien untuk seleksi a%al parameter ang secara signi&ikan meningkatkan aktivitas anti - H1 # Analisis menentukan arah sen a%a timbal modi&ikasi # Hasil metode *4A menun'ukkan bah%a parameter $ A( $ A1 $ Hh $ eH535 dan Q adalah properti kunci untuk men'elaskan aktivitas H1 - antihistamin dari thiazole turunan 1-1. tapi log P 'uga penting untuk desain baru thiazoles ang menun'ukkan aktivitas antihistamin # 4ungsi diskriminasi ditentukan untuk kelompok A dan ( dapat men'adi alat ang e&isien dalam pen elidikan lebih lan'ut # Hubungan univariat dan multivariat ang baik diperoleh dengan menggunakan metode RA dapat digunakan untuk memprediksi e&ek kuantitati& dari aktivitas H1-antihistamin derivati& thiazole ang berbeda # Hubungan ini melibatkan parameter determinasi melalui metode *4A dan P6A #