Alignment Guide

 Education Page


    Finding a Structural Template for a Protein Sequence

The steps below are a general procedure for finding a structural template for a protein sequence of unknown structure. The links to tools will all open in the same (new) browser window.

Step 1: Determine if your sequence aligns to a curated CD

Is your sequence in Entrez Protein?

  • YES
    • Next to the links menu, click Conserved Domains if it exists; then click Show Details
  • NO
Did you find a curated CD?
  • YES
    • Click on the colored bar representing the curated CD.
    • STOP. You're done for this part of the protein!
  • NO
    • Continue with step 2.
Step 2: Find the most sequence-similar structure to your protein

Is your sequence in Entrez Protein?
  • YES
    • In the links menu, click Related Structure if it exists
  • NO (or there is no Related Structure link)
    • Run BLASTp against the pdb database

Choose the structure with the smallest E-value as the most sequence-similar structure. If multiple structures have the same E-value, choose the structure with the largest bit score.

Step 3: Use VAST to build a structure-based template

  • Retrieve the most sequence-similar structure in Entrez Structure
  • Find VAST neighbors for the structure by clicking on the appropriate chain bar (Find neighbors for the whole chain if you found no curated CDs, or if you did find curated CDs, click on 3D domains that occur in portions of the sequence that did not align to a curated CD.)
  • Load a set of VAST neighbors into Cn3D (preferably at least 3 structures)
Step 4: Align the query to the template
  • Import the query sequence into Cn3D
  • Align using the algorithms in the following order (stop whenever an algorithm produces an alignment with no block errors shaded in red):
    • BLAST/PSSM
    • Block Align in local alignment mode (uncheck Global Alignment)
    • Threader

Revised April 10, 2006