Question: Novel Strategy For Improving Sequence Alignment
5
gravatar for Pals
8.1 years ago by
Pals1.3k
Finland
Pals1.3k wrote:

I have been involved in modeling proteins that have initial %identity of 22, 26 and 23. I tried to imply a couple of strategies.

  1. I filtered the homologous sequences that are only related to my model sequence, made MSA and combined with the MSA that was made with homologous sequences of the used template.

  2. I used the mixture of non-redundant protein sequences of both template and target (50/50) and then made structure based sequence alignment using 3D-COFFEE.

  3. I made MSA of model sequence using NR-sequences derived from BLAST and another MSA of template derived from secondary structure prediction method (JPRED). I combined those two MSAs to make final MSA.

However, I am not satisfied with any of the above strategies. The thing is that the ligand-substrate binding site is almost conserved with any of the methods but the overall structure always worse.

I am using discovery studio for making sequence alignment that uses modified CLUSTAL W algorithm.Model building is a never ending process and when the sequence identity is below 30%, there is always a problem. However, I still want to improve the alignment so it would be a great help if anyone can suggest me the fourth approach that would benefit my problem.

Thanking you in advance!!

alignment sequence • 1.6k views
ADD COMMENTlink modified 8.1 years ago by Alastair Kerr5.2k • written 8.1 years ago by Pals1.3k
1
gravatar for Alastair Kerr
8.1 years ago by
Alastair Kerr5.2k
The University of Edinburgh, UK
Alastair Kerr5.2k wrote:

Looks like you have a fairly advanced understanding of what is going on.

A few suggestions:

  1. Have a look at ProbCons which is my favourite alignment program for proteins identities in the 'twilight zone' (at least when there are no structures for 3D coffee).
  2. Look at performing local alignments of just the problems regions: jalview might be able to help with this.
  3. Try using PSI-BLAST against a relevant protein dataset. It can be good at getting the key conserved residues if the protein family is large enough
ADD COMMENTlink written 8.1 years ago by Alastair Kerr5.2k

Thank you very much for you suggestions. I will of course try with first two suggestions. The third one, I had tried already.

Thanks once again for your helpful suggestions!!

ADD REPLYlink written 8.1 years ago by Pals1.3k
Please log in to add an answer.

Help
Access

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 2.3.0
Traffic: 1219 users visited in the last hour