In our modeling group we develop several new open source software tools that are made available through GitHub:
LEADD (Lamarckian Evolutionary Algorithm for de novo Drug Design) is a tool for molecular design and optimization. Molecules are represented as meta-graphs of molecular fragments. Fragments are extracted by fragmenting an input virtual library, and broken bonds are converted to labelled attachment points (a.k.a. connectors). Molecules are reassembled by genetic operators that combine fragments through said connectors. Knowledge-based connectivity rules, extracted from the same input library, are enforced throughout the process. A population of molecules is evolved stochastically through use of genetic operators, with the goal of optimizing a user-provided scoring function. A Lamarckian evolutionary mechanism adjusts the reproductive behaviour of the molecules based on the outcome of the previous generation.
FEPprep is a Python-based alignment tool for free energy perturbation (FEP) preparation. The tool takes an input and reference molecules, and alters the conformation of the input molecule so that it aligns this molecule onto the reference structure.
A spectrophore is a one-dimensional molecular descriptor consisting of 48 real numbers. The technology and its applications have been described in Journal of Cheminformatics (2018) 10, 9.
QED stands for quantitative estimation of drug-likeness and the concept has been introduced by Richard Bickerton and coworkers Bickerton, G.R.; Paolini, G.V.; Besnard, J.; Muresan, S.; Hopkins, A.L. (2012) ‘Quantifying the chemical beauty of drugs’, Nature Chemistry, 4, 90-98. The code allows the user to calculate QED values and relies on RDKit as chemoinformatics toolkit.