Molecular Modeling
- Isaiah Liu
- Nov 3, 2020
- 2 min read
What is molecular modeling? It is a bunch of computer based techniques that are for manipulating and representing the structures and reactions of molecules (or compounds) and properties associated with them that are dependent on three dimensional structures. Molecular modeling is used for chemistry and biochemistry as well as many professions that fall into that category. One major profession that benefits greatly from molecular modeling is pharmaceuticals or companies that make drugs. In the past, companies have used traditional high throughput screening which uses robots to quickly test compound interactions and process data. But, companies are now using what’s called a high throughput virtual screen which involves simulating different tests on a computer rather than using a robot to process data. The advantages of using a high throughput virtual screen (HTVS) over a traditional one is mainly cost and time because it takes less time and money to use a computer to simulate rather than a robot to test actual compounds.
Before going into the different techniques used for HTVS, you must understand the drug design flow. Below is a chart showing the workflow for designing a drug:
A hit is simply a compound that is found to have the desired chemical activity during the HTVS. Eventually, the hits are evaluated and optimized based on promising properties. Now, these companies are not starting from scratch and completely designing the drug themselves. What they do is use HTVS or the traditional version to search through a compound library which can hold up to millions or billions of compounds. The reason companies search through these libraries instead of making compounds from scratch is that these already known compounds may have properties or sections that can be used. Instead of trying to design a whole new compound, they are using data to base their designs off of.
Now, in the first step, target identification, companies will use a consensus screen. That means they use the results of many tests, such as shape screening, docking, QSAR, fingerprint searching, and pharmacophore screening, to filter the compounds of a large compound library down to anywhere between 3%-10% of the compounds. All those tests use the data about the known receptors (chemical structures that receive and send signals) and known binders (a material that sticks to other material or other structures) to run their calculations. What they all do essentially is look for structure similarity and test how well each compound in the library “binds” to the known receptor. Each one has their own set of calculations and can vary but they are all looking for which compounds work well with the receptor.
Then, in the next step, compound screening, they will use more extensive tests to narrow down the compounds they already have. And then will use something like an SAR analysis (where a chemical group responsible for starting the target biological effect in the structure of a molecule) to optimize different properties. Once those chemical groups are identified, each compound will be optimized by changing around the chemical groups so that they will have better test results.
In conclusion, drug discovery uses molecular modeling to filter a compound library and then optimize those compounds. It saves time and money because of how advanced computing has become.
Sources:
Introduction to Molecular Modeling in Drug Discovery: a Schrodinger Online Course
Comments