In article <v0211020fabc8fbfbe88c@[132.236.236.189]> gma2 at CORNELL.EDU (Gregory M. Acland) writes:
>By now, J. J. Campanella (jjc4 at po.CWRU.Edu) has been appropriately
>castigated for his assaults on the various linkage programs, and the rigor
>with they have been written and maintained, especially as they are all
>public domain.
I jolly well hope he has been - programs to do linkage analysis are hard
enough to write without people wandering in slagging them off for being
badly-written. I certainly wouldn't want to attempt writing one of these
programs, and the people who have produced the LINKAGE package have made an
immense contribution to genetics research for which they deserve praise and
gratitude rather than gratuitous insults.
>However, the point does remain that the programs are not user friendly for
>naive or casual users, and the demand/need for access to such programs by
>such users is going to increase.
>What is the hope of someone setting up a client server system for linkage
>analysis on the Net?. That way a user could have a Mac (or other warm and
>fuzzy) client interface locally for converting pedigrees into a format
>suitable for submission to a remote server running (eg:) the Linkage
>package. The pedigree analysis at the client end could be restricted to
People have been working on producing this kind of system and some packages to
do this have been published. Indeed I produced something in that direction
myself, called DOLINK. I'm not particularly proud of it, but I find it
helpful. Other people have other systems. John Attwood and Steve Bryant wrote
LINKSYS, but I gather they feel it's rather outdated now, especially in terms
of limitations on the number of loci it can handle. Steve is currently working
on a mega-system which will intelligently set up and perform analyses without
the user having to get his or her hands dirty at all. A year or two a system
was written up in Am J Hum Genet which was available as portable C source - I
don't recall its name. And there's LCP itself, which comes with the Pascal
(though not C) versions of LINKAGE. I agree though that there's an unmet
demand here. Although this kind of package may be very useful, it doesn't come
across as a "scientific advance" which may make it difficult to get funding or
persuade people with the necessary expertise to focus on the problem.
Dave Curtis (dcurtis at hgmp.mrc.ac.uk)
http://www.iop.bpmf.ac.uk/home/depts/psychmed/general/dcurtis/dcurtis.htm
