I have to model IRS-I protein and I use quark for my modelling, but because the protein is too large I only use protein partially around 37 AA based from article that this amount has effect in inhibiting autophosphorylation in IRK. Is it reliable or not? The IRS-I protein have 1242 AA.
did you mean IRS-1 Insulin receptor substrate 1. if this is it, ithink it's very ambicious to make a model with quark.
but if you make a blast with P35568 uniprot code, you can find two structure 1QQG and 1IRS_A from pdb, with more than a hundred residus,it could be a better start for your job.
hope this help
You should use full length protein. Before started you should check the suitability for modelling via BLAST against pdb databse.
If you are going to use Quark, the best place to start is the protein domain containing the binding site. You are unlikely to get any reasonable answer modelling only the binding site itself.
However, I think this binding domain for IRS-1 has already been solved in the apo form
Instead of taking only the binding site for modeling, consider the surrounding region which may influence the binding site area. If you have a suitable template with good identity, you better consider the part which is homologous to your template which can be more appropriate.
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Bioinformatics
- Bioinformatic Tools