“…The advantage of selecting this specific dataset is that the data have been obtained in the same experimental conditions. To verify how representative our dataset is, we collected data for 131 compounds available in the literature for a total of 277 Caco-2 cell permeability values (different values have been obtained for a number of these drugs in different experimental conditions) ( Artursson, 1990 ; Artursson and Karlsson, 1991 ; Artursson and Magnusson, 1990 ; Augustijns et al , 1996 ; Aungst et al , 2000 ; Chong et al , 1997 ; Collett et al , 1996 ; Gres et al , 1998 ; Haeberlin et al , 1993 ; Hilgendorf et al , 2000 ; Hou et al , 2004 ; Hovgaard et al , 1995 ; Lentz et al , 2000 ; Liang et al , 2000 ; Rubas et al , 1993 ; Ruiz-Garcia et al , 2002 ; Saha and Kou, 2002 ; Schipper et al , 2001 ; Wu et al , 2000 ; Yee 1997 ; Zhu et al , 2002 ) and compared both their Caco2 experimental values ( Supplementary Figure S1 ) and their structural features.…”