Hi all,
I'm quite new to homology modeling so I would appreciate some help.
I'm yielding complete different search results when I submit an identical sequence with the same initial parameters to PSI-BLAST via embl and via ncbi.
The ncbi results seem "ok" (with respects to bit-score), whilst those returned by the ncbi host of PSI-BLAST won't always get past the second iteration. I am using the same E(), PSSM E-value cut off, database (pdb), matrix score, word size, etc.
Further more to this issue. I have managed to identify some homologies using the embl implementation of blast. How can I then generate a 3D pdb structure? I have tried using Maestro, which I was very successful with when I had an identity of 85% of almost the entire sequence from a past project. However, this current project, I believe the maestro homology toolkit uses values of the PSI-BLAST API similar to ncbi, and hence I don't get similar results to the embl implementation.
Thanks
Yes indeed! It looks like EMBL-EBI is using a very old version: 2.2.29+.
Whereas NCBI is using the most recent: 2.6.0+
Also, NCBI has a much smaller database (only 88,936) whereas the EMBL-EBI version has a more appropriate number (350,681).
So, neither is ideal.
Brilliant, thank you, Christian. I noticed there was a version difference late last night. I also assumed that both services would be pointing to the same database, and given that I requested Protein Structural Database (denoted: pdb), I thought they would both link in with the Protein Data Bank. I am guessing this is not the case given those numbers you mentioned are different from one another.
Thanks
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Bioinformatics
- Bioinformatic Tools