Petr wrote:
> Miguel, sorry - some sort of inverse effect.
:-)
> I tried to be very explicit and the result is "Lost in translation".
> Simplification (I hope ... :-)) ) is below in ################'s
Explicit is good ... I just didn't understand the terminology.
> Q: What is N-mer?
> ###################################
> Ooops. N-mer = polymer of N-amino acids = protein.
OK
> Sorry. In plain words:
>
> Jmol loads pdb file=protein.
> Protein is (folded) chain of amino acids, indexed from 1 to N.
> These indices 1..N are associated with each
> monomer=amino acid - usage e.g. in residue labeling.
>
> Now user has some external quantification how in
> say N=130-amino acid protein is amino acid #25
> similar to amino acid #56.
> This is encoded in the external, user-defined file
> containing (in this example) a 130x130 matrix M.
> If user quantifies similarity of amino acids 25
> and 56 by 1.0, then in that
> matrix for our example, element M[25,56]=1.0
> (and for the sake of simplicity, all other
> elements but M[56,25] will be =0).
OK. I understand that there is comething like a correlation coefficient.
> Goal is to have a UI showing graphically 130x130
> plane with a contour around x=25,y=56 and x=56,y=25.
OK.
(Or, if you think that this will be a 'contour' with little mountain
ranges then perhaps you want to show this 130x130 space in a 3D graphical
viewer ... like Jmol)
> That contour should be clickable=by
> clicking there, UI passes 25 and 56 to the rendering window.
OK
> The rendering window takes these and selects
> amino acid 25 and 56 and immediately changes
> their rendering, so user has "real time" view of
> what
> his/her 1.0 generated by some bioinformatics
> analysis of the sequence of that protein means structurally.
This part is trivial.
> You can stop here, below is why:
> #######################################################
>
> Holly grail for protein people is to take the
> (known) sequence of amino acids (genome sequecing
> got them nearly all) and predict from them what
> Jmol shows as the protein structure without relying
> on experiments that are
> the sources of PDB files your program uses.
OK
> So - without experiments, in one type of predictive
> methodology, one has to
> take a sequence and find out the relationships
> between all amino acids i=1..N and j=1..N. Then,
> ideally, matrix M will contain the information that
> amino acid 25 and 56 in our example protein have CA-CA distance 6A and
> similar information for all other pairs.
> If you are testing something like this, you have lots
> of M matrices and you
> are trying to figure out how good/bad your i-j
> relations are by testing them against known structures.
> Thus - one "pretends" for a protein that its PDB file
> does not exist. Then one takes the amino acid sequence
> of that "virtually unknown structure
> protein" and does his "trial magic" with that
> sequence. Result is matrix M.
OK
> Then one wants to know, how good the M reflects the
> real structure ("re-discovered" for this testing).
> So one "re-discovers" the PDB file, loads it in the
> Jmol, looks what 1's in
> M mean structurally, get an idea how to improve "magic",
> and the cycle continues.
> Lots of select click, clicking....
OK
> The Jmol application has a plug-in mechanism
> for adding this type of functionality.
This sounds like an ideal candidate for a Jmol application plugin.
Miguel
-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now.
http://productguide.itmanagersjournal.com/
_______________________________________________
Jmol-users mailing list
Jmol-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/jmol-users
