The MultiMap program currently uses a few algorithms that require
CEPH-style pedigrees.  CEPH (Centre d'Etude Polymorphisme Humain)
pedigrees are standard human reference pedigrees commonly used for
genetic mapping in humans.  They consist of 4 grandparents, 2 parents,
and any number of children (though some pedigrees do not have
grandparents).  

The algorithms that require CEPH-style pedigrees are not crucial to
the mapping process, but are used to rank the markers prior to mapping
such that more informative markers are added to the map before less
informative markers.  I am considering ways to change this
requirement, but for now, if you wish to map in non-CEPH pedigrees
(such as experimental crosses) please use the following set of
additional instructions.


          Instructions for Mapping in non-CEPH-style pedigrees

One of the initial algorithms of MultiMap is written exclusively for
CEPH-style pedigrees (3 generations with up to 4 grandparents, 2
parents, many offspring.)  These are the algorithms that compute
heterozygosity and joint-PIC values.  Either the joint-PIC values or
the heterozygosities (depending on the value of the parameter
*use-joint-pic* which is set in the .input file) are used to rank the
remaining markers to determine the order by which MultiMap will
attempt to place them on the framework map.  

If you know the heterozygosities of your markers, you can rank them
yourself.  Otherwise, they can be added to the map randomly, which
would eliminate MultiMap's capability to add the most informative
marker to the map first.  However, this is not a crucial step in map
construction.

Its not clear to me whether MultiMap will necessarily bomb when
computing the heterozygosities in non-CEPH pedigrees.  You are welcome
to try running your non-CEPH peds exacly as if they were CEPH peds,
with one exception.  BE SURE TO SET *use-joint-pic* to NIL in the
chrZ.input file.  You will probably see this message toward the
beginning of the run - you can ignore it - its not correct "NOTE WELL:
The following markers have NO GENOTYPE DATA!!".  MultiMap cannot
properly handle the non-CEPH data so it incorrectly concludes there is
no data!

If the program does not bomb, then you can be confident in the
results.  However, using this approach, the program will attempt to
add the markers to the map in the reverse order that they appear in
the datafile.  If you wish to have more heterozygous markers added
first and you have know the marker heterozygosities, follow the
instructions below for setting up the chrZ.ordh file with your
preferred marker ranking.  Or, should MultiMap bomb when trying the
regular approach, follow the directions below.

SPECIAL INSTRUCTIONS FOR NON-CEPH PEDIGREES

You will need 3 files - chrZ.ordh, chrZ.ordj and chrZ.hash.  Also, in
the .input file, you must set *use-joint-pic* to NIL.

A.  make a chrZ.ordj file

You must provide a dummy .ordj file and set *use-joint-pic* to nil in
the .input file (forcing the order as given in the .ordh file to be
used).  Simply create a file called chrZ.ordj (substituting your
chromosome number for Z) and put in it this: (()) That's left paren,
left paren, right paren, right paren.  Also known as an empty list
inside a list.  THat's it.

B.  make a chrZ.hash file.

If a correct version of this file does not exist MultiMap will try to
make its own .ordh and .ordj files, and thus may bomb.  This hash
file should have 2 things in it, a list containing the checksum value
of the chrZ.dat file, and an empty list.  Try using the UNIX function
sum (i.e. 'sum chrZ.dat') to get the checksum value.  (Let me know if
you can't do this).  (Or, if you already have a .hash file from before
it may be in there as the first item.)  The next list in the hash file
will eventually contain CRI-MAP results (in ugly format), so if you
are starting from scratch this should be empty.  If you have an old
.hash file, as long as it corresponds to the same .dat file, its
probably fine.

example chrX.hash file:

(12345)()

if the checksum value is 12345.

C. - make/get a chrZ.ordh file.  Although MultiMap cannot compute
accurate heterozygosities for your data, it tries anyway and ends up
computing inaccurate values (these will be in chrZ.freq1).  Thus, it
does rank your markers according to these innacurate heterozygosities
and will produce a valid chrZ.ordh file (though not actually ranked by
informativeness).  If you have no way of ranking the markers youself,
you can let MultiMap create and use this file as just explained.  In
this case, do not worry about making this file, just make the .ordj
and .hash files.

If you wish to provide a ranking for MultiMap, simply create a
chrZ.ordh file which has in it a list of your marker numbers in your
preferred ranking order.  The marker numbers can be found in the .loc
file.  Each marker must be represented in this list, if there are 20
markers in your data file, there should be 20 markers in this list.
For example,if you have 10 markers and you know wish to have them
ranked in this way: 9 4 5 2 0 1 3 7 8 6, then you would create a
chrZ.ordh file with this in it:

(9 4 5 2 0 1 3 7 8 6)


Once you have done these steps, you should be able to run MultiMap as
you normally would.