I want to stitch two different proteins. My next step is to generate my data using HHblits server. By putting both the proteins in one sequence, HHblits will only result in one part and will ignore the other. Does anyone know about other protein stitching methods? I have to mention that my genes are not in the same operon too.
I am uncertain what you are trying to achieve. Do the two sequences correspond to known 3D structures. Will you try to express the stitched proteins ?
Well, I intend to gather data from hhblits/HMMER/ or simply you may say BLAST. I want to use that data to generate native contacts, for that i am using DCA (Direct coupling analysis), or plmDCA. Now the real problem is that these programs are really powerful but they can handle one sequence. I am using very large data set, i have almost 56000 sequences for protein A, and almost 50,000 of B. They both are interacting in biochemical experiments (GST pull down). I belong to structure lab and wish to structure this complex again sad news are that both are membrane proteins and facing lots of issue in this case). I want to use computational power to detect some interacting residues and to mutate them an blah blah.. To run DCA i need to combine sequence A and B as DCA can read only one sequence, now its almost impossible for me to combine such a huge data manually. Some asked me you can stitch them and run a blast i would say so you can have your data. If i keep both the proteins side by side under one header, then BLAST will pick one of them not both. So i am here, to search any possible option or any literature. Currently, if this method is of a high value to the person who want to share with me, our group is open in any kind of collaboration too because we are interested in answering a research question more.
Regards,
One thing I want to add that my both proteins A and B are not in the same operon too, which is a real problem i think. As i read few papers in Cell, there the author used the advantage of two proteins located in the same operon.
You did not say if there is a homology PDB model for A and B separately.
Have you tried HHPRED on the sequence of A+B. If you find PDB hits for A and B portions you can then put these through Modeller which will build a 3D model. However if A and B are separate peptides then this will probably not work. If A and B are both emembrane proteins then if you can get a homology fold for both seperately then the problem reduces to a 2d positioning of both in the membrane and reduces the likely contact points to mutate.
Your splicing problem in DCA is probably because you are putting both sequences into the online form with their FASTA headers. Remove the second header so that one sequence runs into the second and DCA will then see the the full A+B sequence
- Badges
- Science topic
- Similar topics
- Biological Science
- Molecular Biology
- Biomolecules
- Proteins