Working on next post
Hello,
Next installment is coming up. I am thinking of writing about energy minimization and structure validation. You might have already notice the trend of this blog, it is not so much about structure determination but about structure analysis.
So, while I was doing some homology modelling for a friend at le Insitut Pasteur that I started thinking about some this. I usually do my modelling using Modeller. I have come to like it and understand it. It is also very powerful and flexible so it is a standard tool for me (yes, I'll review it later). It allows you to evaluate your models as soon as the session is done (maybe thousands of results) and pick your choice. It includes its own tools for evaluating the models giving you a total score as well as a per residue score.
"But, " I asked myself, "what if the is a region that looks (is scored) bad in the model but not in the template? What difference in score is significant?"
With that question in mind I looked up through my old Structure Validation bookmarks and found to sites that I liked: ProQ and ProSA. The first one gives you back two easy to read scores while the second returns you a Z-score, two graph for global quality and a 3D viewer for local scores. Both can be run on their on server or installed locally (Windows and Linux only :-S).
Anyway, Davis Picnic Day got in the way and I have been busy being a dad and a postdoc. I have the data for the next post but I really need to write it so that it is useful. In the meantime, here is something to look at:
Next installment is coming up. I am thinking of writing about energy minimization and structure validation. You might have already notice the trend of this blog, it is not so much about structure determination but about structure analysis.
So, while I was doing some homology modelling for a friend at le Insitut Pasteur that I started thinking about some this. I usually do my modelling using Modeller. I have come to like it and understand it. It is also very powerful and flexible so it is a standard tool for me (yes, I'll review it later). It allows you to evaluate your models as soon as the session is done (maybe thousands of results) and pick your choice. It includes its own tools for evaluating the models giving you a total score as well as a per residue score.
"But, " I asked myself, "what if the is a region that looks (is scored) bad in the model but not in the template? What difference in score is significant?"
With that question in mind I looked up through my old Structure Validation bookmarks and found to sites that I liked: ProQ and ProSA. The first one gives you back two easy to read scores while the second returns you a Z-score, two graph for global quality and a 3D viewer for local scores. Both can be run on their on server or installed locally (Windows and Linux only :-S).
Anyway, Davis Picnic Day got in the way and I have been busy being a dad and a postdoc. I have the data for the next post but I really need to write it so that it is useful. In the meantime, here is something to look at:
 |
| Spring |
Comments