Retention behaviour of twenty-one chalcones synthesized in our laboratory was tested in three thin-layer chromatography (RP-TLC) systems (acetonitrile–water, ethanol–water and acetone–water) and chromatography parameters \( {R}_M^0 \), S and C₀ were calculated. The most suitable RP-TLC system (acetonitrile–water) and chromatography parameter (C₀) for lipophilicity prediction of tested compounds were selected on the basis of the highest correlations with calculated logP values. In selected system, compound 12 had the highest, whereas 47 had the lowest C₀ value. QSRR analysis was performed and three models representing relationships between C₀ and selected molecular descriptors were created—MLR(C₀), PLS(C₀) and SVM(C₀). Interpretation of molecular descriptors which form statistically the most reliable SVM(C₀) model identified the most important structural and physico-chemical properties that influence retention behaviour of tested compounds. In addition, descriptors with the highest influence on \( {R}_M^0 \) as well as on C₀ calculated in the remaining two RP-TLC systems were identified and interpreted.

中文翻译：

---

RP-TLC和计算方法评估查耳酮的亲脂性

我们在三个薄层色谱（RP-TLC）系统（乙腈-水，乙醇-水和丙酮-水）和色谱参数\（{R} _M ^ 0 \ ），计算出S和C ₀。根据与计算出的log P值的最高相关性，选择最合适的RP-TLC系统（乙腈-水）和色谱参数（C ₀）预测被测化合物的亲脂性。在选定的系统中，化合物12最高，而47的C ₀最低值。进行了QSRR分析，并创建了三个表示C ₀与所选分子描述符之间关系的模型-MLR（C ₀），PLS（C ₀）和SVM（C ₀）。解释统计上最可靠的SVM（C ₀）模型的分子描述符，可以确定影响测试化合物保留行为的最重要的结构和理化性质。另外，识别并解释了对\（{R} _M ^ 0 \）以及对在其余两个RP-TLC系统中计算出的C _0的影响最大的描述符。