MIA-QSAR: a simple 2D image-based approach for quantitative structure-activity relationship analysis

An accessible and quite simple QSAR method, based on 2D image analysis, is reported. A case study is carried out in order to compare this model with a previously reported sophisticated methodology. A well known set of (S)-N-[(1-ethyl-2-pyrrolidinyl)methyl]-6-methoxybenzamides, compounds with affinity to the dopamine D2 receptor subtype, was divided in 40 calibration compounds and 18 test compounds and the descriptors were generated from pixels of 2D structures of each compound, which can be drawn with aid of any appropriate program. Bilinear (conventional) PLS was utilized as the regression method and leave-one-out cross-validation was performed using the NIPALS algorithm. The good predicted Q2 value obtained for the series of test compounds (0.58), together with the similar prediction quality obtained to other data sets (nAChR ligands, HIV protease inhibitors, COX-2 inhibitors and anxiolytic agents), suggests that the model is robust and seems to be as applicable as more complex methods.