Difference between revisions of "Talk:Canavan Task 2 - Sequence alignments"

From Bioinformatikpedia
(Review)
(Review)
Line 16: Line 16:
 
I did not have the time to really rethink your approach in detail, but you worked nicely and especially very efficiently.
 
I did not have the time to really rethink your approach in detail, but you worked nicely and especially very efficiently.
 
Best, Alice
 
Best, Alice
  +
  +
A few things in addition to what alice said
  +
  +
* The read markings in the sequence at the beginning are explained only towards the very end
  +
* What does this figure signify? http://i12r-studfilesrv.informatik.tu-muenchen.de/wiki/index.php/File:Seq_overlap.png
  +
* I do not agree with your reasoning that it is unlikely to find another structure for a class of proteins, once one structure has been solved for them. In fact the knowledge of how to solve the structure gained from the first experiment might lead to a higher probability, that more structures e.g. of homologs in another species or mutants, will be solved
  +
*

Revision as of 09:03, 8 May 2012

  • Your PSI-Blast search seems to have hit the reporting limits. Please increase the cutoffs, so you can really compare the results. andrea 10:32, 3 May 2012 (UTC)

Review

  • Some figures are not referenced properly in the text, tables are not referenced or mentioned at all
  • Your cutoff for the blast searches seems very low, and I think you didn't increase the number of results like andrea wrote. My guess is that you consequently can not really compare the number of hits.
    • Agreed, this might also solve you problem of no pdb structures -jonas
  • You could have tried to use the whole big database for PSI-Blast and PDB for HHBlits to make the results more comparable.
  • You could elaborate a little bit more, e.g. on the MSA's.
  • I like your figures, very nice, just get rid of the ??.
  • Making the link to 3D structure is very interesting but I would have enjoyed reading more about it :)


I did not have the time to really rethink your approach in detail, but you worked nicely and especially very efficiently. Best, Alice

A few things in addition to what alice said

  • The read markings in the sequence at the beginning are explained only towards the very end
  • What does this figure signify? http://i12r-studfilesrv.informatik.tu-muenchen.de/wiki/index.php/File:Seq_overlap.png
  • I do not agree with your reasoning that it is unlikely to find another structure for a class of proteins, once one structure has been solved for them. In fact the knowledge of how to solve the structure gained from the first experiment might lead to a higher probability, that more structures e.g. of homologs in another species or mutants, will be solved