      Welcome to CROSAFIN!

     A new type of ligand-binding experiment called cross-titration
experiment is being proposed for characterization of tissues
containing multiple receptor subtypes. It is an extension of
traditional inhibition experiment employing two different
inhibitors at once: binding of a constant amount of labeled ligand
is measured in approximately 10x10 array of wells where concentration
of one inhibitor is varying in rows and concentration of another
one - in columns.
     The program CROSAFIN recovers from such data a two-dimensional
distribution of receptor subtypes on the coordinate plane with
affinity constant to one inhibitor on X-axis and to the second
inhibitor - on Y-axis. Routinely, this method ensures a problem-free
dealing with tissues containing 4-5 receptor subtypes, though, of
course, still much depends upon availability of inhibitors with
appropriate affinity profiles.

   This is a really easy-to-use program. You should just enter
the data from cross-titration experiment in 'Edit' spreadsheet-like
environment and press F9. The only way to regulate calculations is
changing 'Spectrum density' in 'Option' menu: it defines density of
the grid of calculated specrum.

   The design of experiment
seems to be absolutely obvious.                ...
Specific demands to experiment
are: 1. It should be performed      i  1/4  ^  0   0   0
in "Cheng-Prusoff" conditions,      n       !
i.e. only negligible fractions      h       !
of ligands should be bound          i  1/2  !  0   0   0
throughout experiment.              b       !
2. The measurement error should     2       !
be truely random, i.e. there            1   !  0   0   0   ...
should be no serious "systematic
distortions" of dataset.                       --------->
   No exquisit strictness in fol-
lowing these demands is necessary.             1  1/2  1/4
Any number of "holes" in this                   inhibitor1
array of data is allowed.
   Perhaps, it is reasonable also to have pronounced plateu at small
concns. of ligands, the value of dilution factor seems reasonable to
choose between 2 and 3.
   Accurate and abundant dataset is, of course, an advantage. Yet to a
bigger extent success in identification of receptor subtypes depends
availability of a pair of inhibitors ensuring maximal distances on
LogKd1-LogKd2 plane between subtypes you want to resolve. That is,
it appears that search for inhibitors with appropriate affinity
profiles ought to be the the major concern in using CROSAFIN.

         MODEL function
  This function allows to generate simulated datasets to get acquainted
with abilities of this program. In addition to parameters of experiment
which are defined in 'Edit' environment you should draw the actual
spectrum and to define the magnitude of error in simulated data.
Error should be defined at two points: at maximal signal (B=Bmax) and
at no signal (B=0). For example, if you wish to generate data with
constant coefficient of variation enter 0 to upper cell and 5 to lower
cell.

         FILES
   There are three types of files used by CROSAFIN. Data are load/saved
in .DAT files, calculated spectrum - in .SPC files. Both these files
are in ASCII format. And you may save any picture drawn on the screen
into .PCX file (i.e. Paintbrush picture format) to export it later to
your favorite text editor. You may operate with these files only when
it is  allowed by the line in the bottom of screen.

         ANALYZING CALCULATED SPECTRUM
   CROSAFIN calculates in 'Spectrum' function the concentrations and
positions of crests for every fractions in calculated spectrum which is
a real 'island'. If you want to subdivide picture into more fractions
cut "channels" in 'Model' function and then return to 'Spectrum'.


   For more theory underlying work of CROSAFIN see documentation to
my previous program AFFINOGEN (J.Immunol.Meth.(1991),139:297;
jim.mscs.mu.edu under directory pub/Yuryev). Note, that in contrast
to that program CROSAFIN works in 'no smoothing' mode.

   Cite, if necessary, notice on CROSAFIN in:
   Life Sciences 56(17):I-III (1995)

   This version is completely ready for practical use, yet, I think,
some theoretical aspects may remain vague for non-mathematical user.
I think, the meaning of peripheral areas (within 1 order of magnitude
from border) of calculated spectrum representing numerous types of non-
specific binding and factors influencing resolving power are difficult
to perceive adequtelely.

   This is a shareware program. If you wish to pay anything contact me at:

   Dmitriy K.Yuryev
   ul.Tuchkovskaya, d.11, kv.48
   121087 Moscow RUSSIA

   e-mail yur77@glas.apc.org

_______________________________
   The program file in crsaf_02.zip is the same as in crsaf_01.zip.
To the present version two critical articles on Scatchard plot analysis
amd on interpretation of inhibition curves are added as documentation:
"Absurd Trivial Errors in Scatchard Plot Analysis" and
"Refining Cheng-Prusoff Equation"