Robust nomenclature and software for enhanced reproducibility in molecular modeling of small molecules [article]

Hesam Dashti, Jonathan R. Wedell, Gabriel Cornilescu, Charles D. Schwieters, William M. Westler, John L. Markley, Hamid R. Eghbalnia
2018 bioRxiv   pre-print
AbstractComputational molecular dynamics, energy minimization, and modeling of molecular interactions are widely used in studies involving natural products, metabolites, and drugs. Manually directed computational steps commonly utilize an evolving collection of experimental and computational data, to which new data sources are added or modified as needed. Several software packages capable of incorporating sources of data are available, but the process remains error prone owing to the
more » ... s of preparing and maintaining a consistent set of input files and the proper post-processing of derived data. We have devised a methodology and implemented it using an extensible software pipeline called RUNER (for Robust and Unique Nomenclature for Enhanced Reproducibility) that creates a robust and standardized computational process. The pipeline combines a web service and a graphical user interface (GUI) to enable seamless modifications and verified maintenance of atom force field parameters. The GUI provides an implementation for the widely used molecular modeling software package Xplor-NIH. We describe the RUNER software and demonstrate the rationale for the pipeline through examples of structural studies of small molecules and natural products. The software, pipeline, force field parameters, and file verification data for more than 4,100 compounds (including FDA-approved drugs and natural products) are freely accessible from [http://runer.nmrfam.wisc.edu].Author SummaryWe describe an automated and verifiable computational pipeline for calculating the force field parameters of small molecules. The pipeline integrates several software tools and guarantees reproducibility of the parameters by utilizing a standard nomenclature across multiple computational steps and by maintaining file verification identifiers. We demonstrate the application of this pipeline to (a) processing of more than 4,100 compounds in high-throughput mode, and (b) structural studies of natural products. The graphical user interface (GUI) associated with the pipeline facilitates the manually tedious steps of force field parameters adjustments and supports visualization of the process.
doi:10.1101/429530 fatcat:huiegvgy5rdt5nwso3sgbokmti