For each compound a score is computed ranging from 0 (unfavourable) to 1 (favourable). This approach involves eight descriptors. (2012) reported a new approach, the so-called QED method that does not involve sharp thresholds. ![]() Well-known examples applying sharp thresholds on several PC properties are the Lipinski’s RO5 ( Lipinski et al., 1997) or the Gleeson’s RO4 ( Gleeson, 2008). PC properties can be computed to investigate the notion of chemical quality. The tool can also search for small molecules that could potentially make a covalent bond with a macromolecular target. In addition, and compared to the FAF-Drugs3 version of the server, various other improvements were also implemented and include (see Supplementary Material), for instance, optimization of the standardization–neutralization procedure, refinement of the SMARTS definitions for 137 structural alerts and better detection of the 515 PAINS, the generation of the filtered subsets protonated at physiological pH and finally the possibility for the users to obtain a PDF report for the selected compounds. During this parallelized core procedure, which allowed us to increase by at least four times the speed of the calculations, PC properties and rules are computed and toxicophores and PAINS are identified. The procedure now employs the Python ‘multi-process’ module which controls a pool of worker processes to which jobs can be distributed with timeouts and callbacks and has a parallel map implementation. Efforts to increase the speed of the calculations focused on the second core part of the process. To this end, we fully revised the underlying Python package of FAF-Drugs4 ( ) and now the protocol involves three distinct steps: (i) a preliminary step of data curation before the filtering process that is needed to standardize the molecular structures and remove salts, counterions, inorganics, mixtures and duplicate compounds, (ii) a core procedure that computes the different parameters and (iii) a final part that collects the results, generates CSV tables (computed PC descriptors, structural alerts, PAINS), and writes SDF filtered output files. The updated online package was developed with the aim of improving several aspects our tool including increasing the speed of the computations. Computed data can be visualized online and corresponding output files downloaded 2 The FAF-Drugs4 web-server standardized, salts and duplicates are removed, etc.) prior to the filtering steps. For both services, the molecules are curated (e.g. Users can upload or paste molecular files and run FAF-Drugs4 or FAF-QED. Both services are available via the RPBS’ Mobyle portal, users can upload compounds in SDF/SMILES format or draw a compound with the ChemAxon’s Marvin Sketch applet ( and carry out computations appropriate for their projects. Here, we report the last version of our web server named FAF-Drugs4 and a new service, FAF-QED, that implements the quantitative estimate of drug-likeness (QED) method ( Bickerton et al., 2012) ( Fig. We reported the first version of FAF-Drugs for Free-ADMET-Filtering tool in 2006 ( Miteva et al., 2006) and the web server was significantly modified with the development of FAF-Drugs2 ( Lagorce et al., 2011) and FAF-Drugs3 ( Lagorce et al., 2015). In general, when performing such computations, one aims at enhancing the chances of finding compounds with an acceptable absorption, distribution, metabolism, excretion and toxicity (ADMET) profile ( Leeson, 2016 Leeson and Springthorpe, 2007). ![]() The tools can also flag unwanted reactive chemical groups or molecules that interfere with experimental readouts ( Baell and Holloway, 2010 Cumming et al., 2013 McGovern et al., 2002). Different computations can be performed to help select quality hit compounds, some tools analyse the physicochemical (PC) properties and filter compounds based on PC-derived rules, others search for the presence of toxicophores (potentially toxic chemical groups). Among the different technologies that can assist the process, it is known that in silico tools play an important role ( Hillisch et al., 2015). The search and the development of quality chemical probes acting on novel targets or on disease pathways are both time consuming and costly.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |