Documentation for (get-likelihoods) and (get-many-likes)

8-13-97: Documentation is now available for two functions that can be used to show the results of
likelihood analysis of one or many markers against a map. These functions have been in MultiMap since 1995. The type of output provided by these functions is similar to that now written to the output file chrXX.unmapped (see additional documentation).

Functions to try to place a marker on a map, and simply report the
results.  If the marker can be placed on the map with the given odds
threshold, this is noted in the results.  If not, the possible map
positions within odds <= to the odds threshold specified in the
chrXX.input file, and the respective odds for each such placement,
are given in the results.

To use these functions, you will need to provide a map, and either a
marker or list of markers (which will be analyzed against the map).
This map can either be any map you specify (the map must consist of a
list of marker numbers) or might be the final map reached at the end
of a mapping construction (this map is always stored in the variable
*mapped*).  The marker(s) to be analyzed must also be given as marker
numbers.  Some examples are given below.

A)  When you only want to do this for a few markers, use this function:

(get-likelihoods map marker)

The function paremeters are a map and a marker number.  The marker
will be run against the map, and within the user-specified
log10 odds-threshold, all possible map positions will be returned.  In
Example 1 below, we see that marker ATRC1 was mapped versus the map
stored in *mapped*.  The additional function (get-marker-number) was
also used to convert the marker name into its corresponding marker
number.  Several equivalent ways of running the same analysis are
listed later.  For each interval to which ATRC1 can be localized, the
interval is specified by marker names, and the likelihood for each
position is given.  Following this, the odds of the position versus
the best map position are given.  These results are ranked from most
likely interval to least likely, therefore, the odds for the first
interval are always 1.0.  

EXAMPLE 1:

> (get-likelihoods *mapped* (get-marker-number 'ATRC1))

 Marker "ATRC1":
        D13S120 -   D13S1299       -1053.068    1.000
       D13S1299 -   D13S1246       -1053.072    1.009
       D13S1246 -    D13S289       -1053.431    2.307
        D13S289 -    D13S260       -1053.926    7.211


In this example, ATRC1 maps to 4 map positions.  The most likely is
between D13S120 and D13S1299.  The next most likely position is
between D13S1299 and D13S1246.  The odds in support of the best
position versus the second-best position are only 1.009:1.  The last
position listed, between D13S289 and D13S260, is 7.2 times less likely
than the most likely position.

This example could also have been run like this:

> (setf a (get-marker-number 'ATRC1))       
	-- this assigns the marker number to a variable, "a".

> (get-likelihoods *mapped* a)

Example 2.

> (setf c '(10 15 20 25 35))
	-- this assigns the marker numbers for the map 10-15-20-25-35 to 
           the variable "c"

> (get-likelihoods c 5)

 Marker "D13S289":
         TEL -    D13S220         -59.871    1.000

In this case, marker D13S289, which has marker number 5, maps to only
1 position, between the telomere and D13S220, with odds greater than
the user-specified log10 odds-threshold.  If this threshold is
specified in the .input file as 3.0, then any other map position is
more than 1000 times less likely than the one presented here.  Since
there is only 1 map position for this marker, the relative odds listed
above are 1.0.


B) When you want to run this analysis for a several markers versus the
same map, use this function: 

(get-many-likes map markers).

Its use and results are identical to the previous description, except
the markers to be analyzed are provided to the function as a list of
marker numbers instead of a single marker number.

For example, at the end of a map construction, you might want to know
the map positions and relative odds for all markers that did not map
to a single location.  If you have not quit Multimap, then the final
map is held in the variable *mapped*, and the remaining "unmapped"
markers are held in the variable *unmapped*.

Example 3.

In this example, at the end of map construction, there were 3 markers
that did not place to single map positions, but mapped to larger
intervals.

> (get-many-likes *mapped* *unmapped*)

 Marker "D13S192": 
    D13S1254 -   D13S1250       -1121.089    1.000
    D13S1304 -   D13S1254       -1122.594    31.947

 Marker "D13S170": 
    D13S1255 -    D13S317       -1090.479    1.000
     D13S317 -    D13S231       -1090.479    1.000

 Marker "D13S1274": 
     D13S168 -    D13S201       -1090.438    1.000
     D13S201 -   D13S1319       -1090.438    1.000

Note that for 2 of these markers, the likelihood for each of 2
possible positions is the same, meaning that the marker is equally
likely in either position.  They are probably quite closely linked to the
marker in the middle of the interval, such that recombination events
necessary for ordering are very unlikely.

As mentioned above, the analysis described in B) is now automatically written to the output file chrXX.unmapped at the end of a mapping run. See additional documentation for more details.