Hi there 👋

In the previous post I wrote about why protonation state determination is important and the various method that can be used to determine protonation state. In this post I would like to do a simple case study on how to use empirical-based pK or protonation state prediction using PROPKA and how to use this tool to help preparing the input structure and topology for GROMACS. As a disclaimer, this post does not contain the whole step on how to use pK prediction tool result to guide the preparation of protein structure prior to MD simulation, rather this is only a guide that could be used as pointer and give insight on how straightforward it is to integrate pK prediction tool like PROPKA in preparing “correctly” protonated protein structure for MD simulation. ...

The structure, function, and dynamics of proteins are dictated by the molecular interaction between its residues and the interaction with its environment. Some of the most important interactions are hydrogen bond, electrostatic, and non polar interations. And it is important to note that for some residues these interactions were affected by their protonation states. For example the deprotonation of aspartate, glutamate, and cysteine residue will cause them unable to act as hydrogen donor. Also, the protonation or deprotonation of a residue can cause formal charge change, which could modulate the short range and long range electrostatic interactions. Thus, in order to properly model and simulate a protein it is important to calculate the protonation state of the titratable residues in the protein. ...

What is a Domain-Specific Language? It is more common than you think. Few of us – computational chemists – have heard the term Domain-Specific Language (DSL for short), even though this is somewhat ubiquitous in our workflow, especially when dealing with molecule visualisation. Don’t you believe me? Check out the command you type on VMD, PyMOL, YASARA, or Chimera. YASARA command: LoadPDB 1crn, download=yes AutoPosOriObj 1, x = 1, y = 10, z = 0, alpha = 0, beta = 0, gamma = 0 ColorRes cys, green ShowRes cys BallStickRes cys AutoRotateObj 1, Y = 1 wait 180 AutoRotateObj 1, Y = 0 ZoomRes Cys 16, step = 50 wait 100 ZoomRes all VMD Tcl/Tk command: ...

Prerequisite Basic command prompt Basic Python (variable type, collection, loop, decision, and functions) Python virtual environment (conda/mamba) Jupyter notebook/lab Create new environment So first as always, start with creating new environment (preferably mamba). If you have never used mamba before and have some experience with conda then great, you can simply change every conda command with mamba and vice versa. Create new environment for this series using this command: mamba create -n learn-open-babel python=3.9 openbabel pandas jupyterlab rich Brief explanation for the command above: ...

There are many ways to improve structure-based virtual screening, from something as simple as screening the ligand based on the physico-chemical properties (MW, number of H-donor, number of aromatic rings, pKa, etc.), using consensus docking, or even post-analysis using FEP-based method. One of the simplest and yet fruitful approach to enhance the virtual screening workflow is using interaction fingerprinting. It has been known for long time that certain interaction from certain residue holds more significance than others. And this is especially true when it comes to receptor, where some residue holds the key for agonism while some others for antagonism (or reverse agonism). This key idea, in turn can be used to analyze the interaction pattern (hence, interaction fingerprinting) to figure out if the ligand is a potential agonist, antagonist, or just a dud. ...

So yesterday I spent at least 3 hours to deal with Python package development in Conda environment. So I wrote this post as a self-reminder, since the steps are somewhat long. And also for those of you who is facing the same problem as I do. Why do I need to use pip when I can use Conda to install Open Babel If you’re an experienced Python package developer you should have known the following trick: ...