Hello everybody! Welcome back to my blog! I just finished my second week working with Vladimir Yarov-Yarovoy to learn the inside work that goes into drug discovery. This week we started off by reviewing and discussing the results of our GALigandDock program. Rosetta is inherently random, so me and the other intern had completely different results. Rosetta scores all of the results and the more negative the score, the more optimal it is. Brandon Harris, Vladmir’s student teacher, helped us to graph our data. Attached below is a picture of mine. The most optimal position for the ligand will have the larked dot and be the furthest down on the y -axis. After completing this project we were given the opportunity to explore whatever part of Rosetta we wanted.
Me and Amough, the other intern, settled on learning how a protein is computationally designed in the first place. To do this, we looked on the PDB, Protein Data Base, and picked out a protein to study. We chose the 7jzu protein. Then, we repeated similar steps from our GALigandDock program and cleaned the protein using Chimera. This made the protein ready for Rosetta, as it only had the data we needed now. Then, we ran the relax protocol we had previously used. This time though, out goal was to run the program for two days. So, we had to do the math and figure out how long it takes to run the program once, and figure out how many times we would need to run it so that it could cover 48 hours. After we had relaxed the protein, we looked through each different version to find the ideal one. I used the 16th protein relaxed using Rosetta as it had the most negative score. Since Rosetta is so randomized it was super interesting to see the differences between me and Amough’s results even though we ran the exact sake program.
After we picked out the best protein from our program it was time to run it through a program called MotifDesign. This would make it so the protein could be visualized. We repeated similar steps like finding out how many times to run the code and picking out the optimal result. Then we send Brandon the code of our best protein and he submitted it to a site that would make it into a tangible image. It took two days, but attached below you can see the protein I constructed. This was our last week working with Vladimir and Brandon and I am so grateful for the opportunity. Next week we start working with Igor Vorobyov and looking at other methods of Computational Drug Discovery. Thanks for reading! See you next week!
Wow! You are learning so much. Proud of you! 🥳