FAF-Drugs2 client version requires as input the compound collection in SD or SMILES file format and two parameter files, one containing among others physicochemical thresholds and another one listing the chemical substructures that have to be investigated. SDF filtered files or a tabulated file reporting all the computed values/descriptors.
The main improvements on the server version as compared with the FAF-Drugs2 stand-alone tool include: (i) a step before the filtering process to prepare and clean the electronic input molecular data file, with removal of empty structures, salts, counterions, inorganics, mixtures, duplicates.
Further, we apply a standardization procedure on eight common chemical groups using SMARTS search and the Chem Axon Standardizer utility (see , 2008).
(ii) Novel, optimized and user-defined filtering rules, with a major improvement of the log P calculation through the use of the XLOGP3 program enhanced with experimental log P values extracted from the PHYSPROP database (Syracuse Research Database).
1c), the compound projection (magenta) onto the first plan of the principal component analysis of the orally bioavailable (blue) Drug Bank Small molecules (Fig.
1d) and a radar plot depicting some molecular properties important for oral absorption (Fig. Various screenshots from Internet browser windows that screenplay the FAF-Drugs2 results.
(vi) Finally, an intuitive visualization of the results through any web browser is now implemented.
More explanations about our filtering engine protocol are detailed in the , 2009).
The first two, in particular, are time consuming and costly, and as such, the preparation of the compound collections is critical.
Analyses of past failures have led to a much better understanding of crucial properties that distinguish any chemical from an interesting drug-candidate or a relevant chemical probe.
Thus, the concept of screening high-quality compound collections in terms of improved ADMET (absorption, distribution, metabolism, excretion, toxicity) properties and containing a reduced number of ‘nuisance compounds’ is gaining momentum.
1a), a pie chart summarizing the results for the substructures search (Fig.
1b), a table detailing the results for each compound: the filtering options, the compound's 2D depiction via the Chem Axon molconvert tookit (Fig.
This online toolkit has been designed through a user-centered approach with emphasis on user-friendliness.