Dear Ashish

I would start by performing a BLAST using e.g. Uniprot.org. That will tell you if your protein is similar to anything in the database, and if so, it is likely that the relevant catalytic site for the similar proteins are annotated. Alternatively (or subsequently), you could also attempt doing a templated model with e.g. SwissModel, but the resulting model may not have a satisfactory fit. If a good template exists, you can try to overlay the structures if the model and the template to identify the catalytic site. What is the protein you wish to predict the structure of?

Hi Ashish,

So, I would think the situation is next. You got the sequence and you have no idea what that sequence represents.

1. First, as Simon recommended, find the name of your protein, if your sequence is a protein sequence.

It could be that you got some sequence of letters, which do not represent any protein.

2. Second, after you got positive result from the step 1, run your sequence against "paper blast" (http://papers.genomics.lbl.gov/cgi-bin/litSearch.cgi). It will search for publications related to your sequence, or proteins with similar sequences.

You have to find the publication about function of the protein with the highest sequence similarity. Read the publication carefully and make a list of residues of the active and/or binding sites.

It is possible that you will get hints (publications) with crystal structures. Look for highest sequence homology structures.

3. If your search in "paper blast" did not give any information about crystal structures of similar proteins, I would recommend just in case, repeat search at the Protein Data Bank. In the advance search you can find structures based on the sequence homology (BLAST).

4. If you got a sequence or the structure with higher than 75% homology in the sequence, your protein most likely does the same thing. You can align two sequences and check if positions for active/binding sites match.

If you got sequence homology between 40% and 75%, there is some probability that your protein has similar function. But in this case I would run multiple sequence alignment, just to make sure, that you are checking correct patterns.

If sequence homology is below 40% the predicted function just from sequence would be not reliable. You have to have the structure of your protein to compare with the homologous structure (structural alignment).

5. If the BLAST search gave you result "Unknown function", there is no match you can do.

6. If you did not get any matches in PDB for your sequence. You have a unique protein with the "Unknown function" and the "Unknown structure". Only way would be to solve the crystal structure to get more information about your protein (but that is different story)...

Hope this will help you to identify your sequence.

Good luck!!!

George.

Thank you everyone.

I would like to go through the protocol suggested by Dr. Simon and Dr. George, and I will try Modeller as well.

Name of the protein is RuBisCO Simon Gregersen Echers

Dear Ashish Chauhan

RuBisCO is probably the most abundant protein on earth, so there are tons of resources available (in fact, I am also working a little with RuBisCO from primarily spinach, alfalfa and Sugar beets). You should have no problem finding useful things for it. It has a large and a small chain that, in vivo, forms a homotetramer. Here you see, for instance the UniProt accession for the spinach large chain

https://www.uniprot.org/uniprot/P00875

You can refer to below attached paper full-text, which has all the methodology detials and link to webservers.

Hello

After you find the sequence you should find the crystal structure of your protein in (Protein Data Bank [PDB] ) and after that you should download 3D structure of your protein in PDB format. At end you can download Chimera or BIOVIA Discovery Studio software and import your PDB file ... and see the active site and all amino acids in cavity ...

Good luck.