This repository contains the setup to perform lead optimization with bioisosteres, other modules for a local energy minimization and derivative alignment are available as well.
There are two options for Linux, including a docker container or anaconda environment (recommended).
For the docker, pull the latest LOSDT image from dockerhub and run the container.
In case you want to use an anaconda environment, run:
conda env create --file environment.yml
conda activate LOSDT
streamlit run Home.py
Install docker.desktop, search for the LOSDT image, pull it and run it with: Optional settings > Host port > 8501
After, you can open the local URL link and use the application.
An OpenMM based energy minimization of complexes with a restrained environment, allowing relatively quick optimization of the ligand-protein-solvent complex. An option is provided to protonate the ligand using pKaLearn at pH 7.0. The following forcefields are used:
- Protein: ff19SB
- DNA: OL21
- Lipids: lipids21
- Waters: OPC3
- Small molecules: Sage 2.3.0 with AshGC neural network charge model
An open-source tool for lead design and optimization, built on the principle of safety by design. This tool serves as both an ADMET optimizer and an idea generator for navigating beyond patent space. It integrates multiple features, including ADMET-AI and ShEPhERD-score, leveraging a comprehensive bioisosteric reaction library for functionality.
Options for input are 2D compound or 3D ligand in bioactive confirmation inside protein complex. In case of the latter, the bioisosteric derivatives are automatically aligned to the input confirmation using the MCSAlign module. An option for protonation with pKaLearn at pH 7.0 and energy minimization with the CCM module are given.
An RDKit based alignment tool using an iterative approach of finding Maximum Common Substructures between the 3D template and SMILES-derivatives, copy-pasting the common substructure and iteratively repeating on the leftover template.
Swanson, K.; Walther, P.; Leitz, J.; Mukherjee, S.; Wu, J. C.; Shivnaraine, R. V.; Zou, J. ADMET-AI: A Machine Learning ADMET Platform for Evaluation of Large-Scale Chemical Libraries. Bioinformatics 2024. https://doi.org/10.1093/bioinformatics/btae416
Adams, K.; Abeywardane, K.; Fromer, J.; Coley, C. W. ShEPhERD: diffusing shape, electrostatics, and pharmacophores for bioisosteric drug design. Arxiv.org. https://arxiv.org/html/2411.04130v1
Genzling, J., Luo, Z., Weiser, B. et al. Development of a pKa predictor (pKaLearn) by leveraging teaching experience to improve machine learning. Commun Chem (2026). ; https://www.nature.com/articles/s42004-026-01983-y
A high-performance toolkit for molecular simulation. Use it as an application, a library, or a flexible programming environment. We include extensive language bindings for Python, C, C++, and even Fortran.
The RDKit is a collection of cheminformatics and machine-learning software written in C++ and Python.
A swiss army knife for manipulating and editing PDB files.
Python-based webserver builder.
Streamlit components that adds the ability to draw chemical compounds.
Mol* is a modern web-based open-source toolkit for visualisation and analysis of large-scale molecular data.