There has been some confusion w.r.t. the new Ramachandran plot
core area definition (GJ Kleywegt & TA Jones, Structure 4,
1395-1400 (1996)).
First, someone asked if we can show the contour lines from the
Structure paper instead of making the core areas grey. This can
be done (instructions are on the Web page) if you really want to,
but you have to realise that the smooth contours hide the fact that
the underlying grid is still made up of 10 degree by 10 degree
squares.
A related problem is the "left shift" of the contours versus the
core areas observed by Joachim Meyer. Again this is an artefact
of the contouring. For example, the values for the square
phi=-60 to -50, psi=100 to 110 are stored at the point corresponding
to -60,100. Also, phi and psi have period 360 degrees, so the
values at +180 are identical to those at -180. This makes that
the contouring algorithm will produce some contours in the top
right hand corner (near phi=180 and psi=150-180). And, even though
none of the points with phi=170 belongs to the core area, it may
*look* as if there ought to be grey "core" squares.
It would have been easiest if we had simply left out the contour
plot ;-)
As for the residues rejected by the B-factor cut-off: this cut-off
was applied on-the-fly by a little jiffy program, so the results
are not stored, although they should be simple to reproduce. I
have put an ftp-able list of *all* residues that were included in
the analysis on one of our servers. This list, and a list of all
PDB entries/chain IDs that were used, can be accessed through
the Ramachandran page at:
http://alpha2.bmc.uu.se/~gerard/rama/rama.html
--gerard
