From 4D-QSAR paradigm to virtual high-throughput screening and molecular similarity.

摘要

Different 3D-QSAR methods have been greatly discussed in rational drug design. 4D-QSAR is one de novo method to study the QSAR properties and also overcome the traditional QSAR limit. Molecular similarity is another field of molecular properties measurement.; Different sets of HSA analogs binding to Site II are analyzed using 4D-QSAR, 4D-molecular similarity measures with both AMS and RMS. 4D-MS, either with absolute similarity measures or relative similarity measures can provide a good tool of exploring molecular chemical information space before getting into a QSAR model construction.; Different 4D-QSAR analysis are applied to the resulting 4D-QSAR models. Usually, an optimal 4D-QSAR model contains very specific information provided by the GCOD occupancy data. In the HSA non-specific binding series, the analysis demonstrates handling of (1) very narrow range of activity available of a training set, and (2) non-specific multiple binding were discussed and presented as well. Also the application of VHTS screening are done through the 4D-QSAR method by using GCOD occupancy data for compound activity prediction. Consensus modeling of VHTS for non-specific multiple binding type of 4D-QSAR models is also applied for partitioning VHTS library compounds. High dimensional data visualization of 4D-molecular similarity is also discussed for better understanding the distribution (profile) of molecule pairs.; Overall, 4D-QSAR and 4D-MS, two methods together can demonstrate a set of powerful tool and a good paradigm for QSAR analysis.