Re: [Jmol-users] Improvement for select sequence range

[Bob H. <ha...@st...>]

I have committed this for testing as 11.5.31_dev. It turns out the 
original rewrite of the sequence range code was done by Miguel in 
Sept/2005 in response to a request by Rolf relating to a strange PDB 
file (1t0n) that had duplicate numbering and no chain designations. As 
far as I can see, that particular PDB file has been fixed, and the 
problem is a non-issue. So the way the (proposed) code works now, is this:

Given two sequence+insertion codes A and B...
  For each chain...
    1)  look for A.
    2)  if A is not found, find the code numerically closest to it and 
larger
    3)  if no such code exists, return 0 atoms
    4)  if B is missing ("select 1-")
             select the last group of the chain
        else
             look for B at or after A in the file
    5) if B is not found, find the code numerically closest to it and 
smaller
    6) if no such code exists, return 0 atoms
    7) return all atoms physically between and including codes A and B

How does that sound?

Bob


Bob Hanson wrote:

>OK, upon rereading this thread, especially Rolf's earliest post, and 
>thinking more about what SELECT a-b should mean, and what insertions 
>are, and all that, I do agree that the first solution I had in mind and 
>put into 11.5.30 is not a good idea. What if we just went with Eric's 
>initial idea, but not in the sense of resno >=/<=. I agree with Rolf 
>that that will really mess up scripts, and I agree with Eric that what 
>we have now is not intuitive. But I disagree with Rolf that the behavior 
>should be different depending upon what the file is like. That to me 
>seems like it would be totally opaque behavior to a user.
>
>It seems to me what we have now is just about right, except we should 
>allow for missing end groups. Then no scripts will be affected, because 
>surely no one has a script that selects 0 atoms (11.5.29 behavior) or 
>expects that.
>
>Taking Eric's example (2rd0):
>
>(505,528,529,530)
>
>Do we all agree that
>
>select 520-530 should select 3 groups (as Rasmol)?
>
>Taking Rolf's example with 1ak4:
>
>(27,1027,2027,3027,4027,5027,28)
>
>
>Do we all agree that 
>
>
>select 27-28 should select all of these?
>
>select 27-2027 should select just the first three?
>
>select 5027-28 should select the last two?
>
>
>And insertions (1jgq):
>
>(1031,1032,1032^A,1032^B,1033)
>
>Do we all agree that
>
>select 1031-1033 should select all?
>
>select 1031-1032^A should select just the first 3?
>
>select 1031-1032^C should select... what? perhaps the first 4?
>
>
>If so, I will just make Jmol do that, that's all.
>
>Bob
>
>
>
>
>  
>


-- 
Robert M. Hanson
Professor of Chemistry
St. Olaf College
Northfield, MN
http://www.stolaf.edu/people/hansonr


If nature does not answer first what we want,
it is better to take what answer we get. 

-- Josiah Willard Gibbs, Lecture XXX, Monday, February 5, 1900