3
$\begingroup$

I'm trying to get a list of unique soluble structured proteins that don't have a solved structure. That is, they aren't the usual membrane proteins or some derivative of another protein.

Things that I've been able to think of are low producing transcription factors and novel fusion proteins.

Improve this question
$\endgroup$
4
  • 1
    $\begingroup$ You can build it yourself by comparing protein names (of the species of your choice) and PDB records. UniProt flat files have PDB accessions for some, not all proteins. Extracting the entries with no PDB should be a good starting point. Let me know if this interests you and I can write up an answer with the details. $\endgroup$
    – terdon
    Commented Nov 27, 2012 at 21:10
  • $\begingroup$ @terdon, that pretty much sounds like what I'm looking for. Unfortunately, there might be a significant number of proteins that have analogs in other species but essentially the same structure. Probably the safe test would be to look at the unsolved structures in ecoli. $\endgroup$
    – bobthejoe
    Commented Nov 27, 2012 at 21:13
  • $\begingroup$ Pick a protein, any protein. You are likely to pick one that does not have a solved crystal structure. Also, you'll need to find evidence that the protein is not disordered if you want your structure to mean anything. $\endgroup$
    – user560
    Commented Nov 28, 2012 at 2:59
  • $\begingroup$ @leonardo, well that seems to be a challenge. $\endgroup$
    – bobthejoe
    Commented Nov 28, 2012 at 6:19

3 Answers 3

Reset to default
3
$\begingroup$

There are some funded projects and analyses available just for this purpose.

Structural genomics or high throughput structure projects take all available peptide sequences group them into families and make sure sequence families pointing to most likely novel folds were available.

Here are the status and target list for the joint center for structural genomics. This list is filtered by species as well as by project status.

Their analysis is availble for us to browse through. http://www1.jcsg.org/prod/newscripts/psca/help/document.cgi

Sorry this isn't so thorough. Nature also hosts Target Track, which allows several high throughput structure centers to coordinate their efforts. Each might have resources which could do your work for you.

Improve this answer
$\endgroup$
2
$\begingroup$

This is how I would do it:

  1. Download the UniProt/SWISSPROT flat file for bacteria from here.

  2. After decompressing the files, extract the E. coli protein IDs for which there is no PDB annotation in the file (I am giving you a command line that will work on *ix systems (Linux/Unix/OSX etc)):

    zcat uniprot_sprot_bacteria.dat.gz | gawk '{if($1~/ID/){if($2~/_ECOLI/){id=$2; frag=0; eco=1; want=1}else{eco=0}}  if($1~/DE/ && $0~/Flags: Fragment/){frag=1;}if($1~/DR/ && $2~/PDB/){want=0; } if($1~/\/\// && want==1 && eco==1 && frag==0){print id}}' > no_pdbs.txt

    Explaining the details of this command line is clearly off topic here :). Suffice it to say that it will print those UniProt IDs from the flat file whose name ends in _ECOLI and for which there is no PDB annotation in the file. It also ignores protein fragments. If you need help understanding it, let me know and we can chat or something.

The result of this quick and dirty analysis is a list of 2694 E.coli proteins with no PDB annotations in the UniProt flat files.

CAVEATS:

  • These are only the curated SWISSPROT proteins, you may want to also get the UniProt/TrEMBL proteins from here. I would recommend you stick to SWISSPROT though.

  • As others have pointed out, you should filter this list for homology to other known structures.

  • The fact that there is no PDB annotation in the flat file does not necessarily mean that there is no known structure.

So, this is not perfect but it should serve as a starting point. Good luck!

Improve this answer
$\endgroup$
1
  • $\begingroup$ You can also filter for different protein model databases like "ProteinModelPortal" to get rid of proteins that can be modelled based on homology. $\endgroup$
    – Michael Kuhn
    Commented Nov 29, 2012 at 13:29
2
$\begingroup$

Try looking at structure homology databases - the sequences that they have no annotation for are likely the kind of sequences that you are seeking.

SUPERFAMILY has a comprehensive annotation across almost 2500 fully sequenced cellular genomes. this would be a good place to start ...

Improve this answer
$\endgroup$

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .