Dima Klenchin (klenchin at macc.wisc.edu) wrote:
: In article <4hmra1$c71$1 at mhadf.production.compuserve.com>,
: Eugene A Kapp <100277.2367 at CompuServe.COM> wrote:
: ->I have found that the predicted pI calculations from two WWW
: ->sites differ by approximately 0.4 for a given protein sequence.
: ->Expasy (http://expasy.hcuge.ch/ch2d/pi_tool.html) and the genome
: ->sequencing site at (http://genome1.bio.bnl.gov/bbq.html) are
: ->therefore using slightly different pKR values.
: ->
: ->Could someone suggest the correct pKR values or add some light to
: ->the situation.
: ->
: ->Many thanks.
: Very simple. Ignore both. Predicted pI values are meaningless because
: they treat all AA as if the protein were random coil. It could be OK for
: small proteins and sometimes is OK with some proteins. Real pI of the folded
: protein could be entirely different. So, IMHO, 0.4 unit is not a big deal
: of difference.
It all depends on what you are using the pI value for. If you are
trying to predict some sort of in vivo behaviour, then Dr. Klenchin's
advice is well taken. On the other hand, if you are trying to predict
the behaviour of the protein in an isoelectric focusing gel, the
computer-predicted values are generally on-target.
Keith Robison
Harvard University
Department of Molecular & Cellular Biology
Department of Genetics
robison at mito.harvard.edu
