-------------------------------------------------------------------------------------------------------------------------------
-------------------------------------------------------------------------------------------------------------------------------

This direcoty contains data published in the QSAR article:

Kiralj R., Ferreira M. M. C., "A priori molecular descriptors in QSAR: a case of HIV-1 protease 
inhibitors. I. The Chemometric approach". J. Mol. Graph. Mod., 21, 435-448 (2003).

-------------------------------------------------------------------------------------------------------------------------------

The data are suitable for validation and testing of various chemometric methodologies.


The data are in Tables 2 and 3 in the article:

-1 dependent variable Y (biological activity, pIC50=-log(IC50), where IC50 is inhibitor's molar concentration at 50% viral
 inhibition);

-14 independent variables (molecular descriptors of different nature, made by the a priori approach):
 X1, X2, X3, X4, X5, X6, X7, X8, X9, X10, X11, X12, X13 and X14;

-48 samples (peptidic HIV-1 integrase inhibitors).



The meaning and definition of all the variables can be found in Table 1 in the paper.

-------------------------------------------------------------------------------------------------------------------------------

The data are reorganized into 3 datasets (in .txt format), with matrices of dimensions 48x15.

The columns in each .txt file are:
-the first column: sample's position in the original dataset, being equal to the name of sample, as published in the article;
-columns 2-15: independent variables (molecular descriptors) X1-X14, respectively.
-the last column: the dependent variable (biological activity) Y. 


HIV1-QSAR-DATASET-A
-------------------
-Used in the article:
 Kiralj R., Ferreira M. M. C., "A priori molecular descriptors in QSAR: a case of HIV-1 protease 
 inhibitors. I. The Chemometric approach". J. Mol. Graph. Mod., 21, 435-448 (2003).
-Three models were built:
   a priori I: all 48 samples (1-48) are training samples, there are no external validation samples;
   a priori II: first 32 (1-32) are training samples, the rest (33-48) are external validation samples.


HIV1-QSAR-DATASET-B
-------------------
-Used in the article:
 Teófilo R. F., Martins J. P. A., Ferreira M. M. C. "Sorting variables by using informative vectors as 
 a strategy for feature selection in multivariate regression". J. Chemometr., 23, 32-48 (2009).
-One model was built:
   the QSAR model: first 32 (1-32) are training samples, the rest (33-48) are external validation samples;
                   the training samples are scrambled prior to any analysis.


HIV1-QSAR-DATASET-C
-------------------
-Used in the article:
 Hernández N., Kiralj R., Ferreira M. M. C., Talavera I., "Critical comparative analysis, validation and
 interpretation of SVM and PLS regression models in a QSAR study on HIV-1 protease inhibitors". 
 Chemom. Intell. Lab. Syst., 98, 65-77 (2009).
-Two models were built:
   ordinary PLS model: first 32 are training samples, the rest are external validation samples;
                       the data split is different than in the previous two datasets; the training 
                       samples are scrambled prior to any analysis.
   OPS-PLS model: the same data split and scrambling as for the ordinary PLS model; only five descriptors
                  were used (X3, X6, X9, X10 and X13).
   
-------------------------------------------------------------------------------------------------------------------------------
-------------------------------------------------------------------------------------------------------------------------------