I'm working on a protein.I've found no suitable template for homology modelling for that protein. So, I used I-tasser for modelling and after validation, I got around 78-80% validation in Ramachandran plot. Will it be ok for publication?
I need some low impact factor journal name for bioinformatics publication with scopus & pubmed index in.
Thanks in advance
It depends on the nature of your publication. If the model helped you to solve some interesting scientific problem and the outcome of this study might interest a broad scientific community, then it would be published along with the other stuff.
If you mean that you would like to publish a description of how did you generate a model of a protein using some standard tools, then it is not a publishable matter whether your model is "validated" or not.
did you validate the protein with prosa, errat, & prove ? see what is the result their and the most important thing did u minimize the protein? because it affects the score. i am not sure about the percentage because it is usually mentioned under the graph that a good/average protein structure has around 90% score .
Dear Niloy,
Validation is must, since you have constructed a homology model. Moreover, you must perform molecular dynamics simulations on the validated model prior to predictions.
Dear Niloy,
I will start from the beginning, did you try different prediction websites so here is list of 3D structure prediction portals
http://molbiol-tools.ca/Protein_tertiary_structure.htm
if you still have the same results you can check WHAT IF http://swift.cmbi.ru.nl/servers/html/index.html
to fix some errors of the prediction and raise the validation...
Best Regards,
If Ramachandran is your only way of "validation" than it's probably way to weak for publication. Adding other computational validation protocols like prosa or errat is a good idea, but it doesn't really validate you i-tasser structure - it only says, that it's more or less likely that the obtained structure is correct.
In the current state of art no homology model means usually no way of obtaining a reasonable model of the protein, that's the harsh truth. For such proteins the only true validation of the structure can be made by experiment (xray/nmr structure or at least low-quality electron density) or at least by showing agreement with some known experimental data (e.g. ligand binding - recreating the pose and/or the Ki/Kd/EC50 values). If there are no experimental data at all, there is no way to perform any kind of validation, so there is no sense in doing any computational prediction.
It depends on the nature of your publication. If the model helped you to solve some interesting scientific problem and the outcome of this study might interest a broad scientific community, then it would be published along with the other stuff.
If you mean that you would like to publish a description of how did you generate a model of a protein using some standard tools, then it is not a publishable matter whether your model is "validated" or not.
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Bioinformatics