targets and was considered to obtain molecules capable of acting on multiple Mtb
targets including CmaA1. The third part of the dataset, i.e. a set of 11,089 highly
active anti-HIV molecules (<1 lM activity on HIV cell lines/targets) was taken to
screen molecules that can inhibit both Mtb-CmaA1 and HIV simultaneously. After
subjecting these three subsets of molecules parallelly through the four screening
filters, 12 compounds were obtained as potential anti-CmaA1 hits. As analysed
from the Glide XP docking results, all of the identified hits made strong interactions
with the important CmaA1 active site residues. Figure 4 shows virtual screening
workflow with various levels of filters.
Virtual screening is usually a highly ordered approach combining diverse
computational screening methods, where at each consecutive step, the filter criteria
become more and more stringent, thus retaining the most promising compounds for
experiments. As the steps proceed, the approaches used go on being more thorough
and computationally expensive. So, being simple and fast by nature, pharmacophore models are usually implemented at the beginning of a hierarchical protocol
to eliminate the compounds which do not even fulfil bare simple spatial and
chemical requirements of the query, before subjecting the compound libraries to
more complicated and computationally demanding docking calculations.
Fig. 4 Virtual screening workflow with structure and ligand-based pharmacophore models
44
C. Choudhury and G. Narahari Sastry
targets including CmaA1. The third part of the dataset, i.e. a set of 11,089 highly
active anti-HIV molecules (<1 lM activity on HIV cell lines/targets) was taken to
screen molecules that can inhibit both Mtb-CmaA1 and HIV simultaneously. After
subjecting these three subsets of molecules parallelly through the four screening
filters, 12 compounds were obtained as potential anti-CmaA1 hits. As analysed
from the Glide XP docking results, all of the identified hits made strong interactions
with the important CmaA1 active site residues. Figure 4 shows virtual screening
workflow with various levels of filters.
Virtual screening is usually a highly ordered approach combining diverse
computational screening methods, where at each consecutive step, the filter criteria
become more and more stringent, thus retaining the most promising compounds for
experiments. As the steps proceed, the approaches used go on being more thorough
and computationally expensive. So, being simple and fast by nature, pharmacophore models are usually implemented at the beginning of a hierarchical protocol
to eliminate the compounds which do not even fulfil bare simple spatial and
chemical requirements of the query, before subjecting the compound libraries to
more complicated and computationally demanding docking calculations.
Fig. 4 Virtual screening workflow with structure and ligand-based pharmacophore models
44
C. Choudhury and G. Narahari Sastry