I have developed some software, orginally for epitope mapping
using ELISA data, which can provide a means of segregating
similar "entities" into clusters.

It seems to work in any situation where a number of subjects
(e.g. antibodies) are tested against a panel of "objects" (e.g.
antigens), and can segregate both subjects and objects: the
resolution of subjects depends on the number/scope/specificity of
objects; the resolution of objects depends on the
number/scope/specificity of subjects.  It will provide a map of
the segregated entities.  

At present it is PC based and works on DBase files set up with
the first field as the subject identity and the remaining fields
as the object results, and each record is a different subject.  I
can cope with comma-delimited text files also.  

If you are interested in running some data through this, or in
helping develop the software to a more user-friendly state,

--
Robin Barclay
Edinburgh, Scotland