The sib_map program generates two-point and multipoint maximum
likelihood estimates of the map distances between markers, based on
marker data from nuclear families. It will also determine support
intervals. Allele frequencies will be used to reconstruct missing
parents. All sibling marker data is used to determine likelihoods,
without ever being broken down into sib pairs.
The sib_map program uses the same likelihood engine used in
sib_phase, so it also has a strict limit of 16 siblings per
family.
The multipoint maximization algorithm determines the complete set of distances that gives the global maximum likelihood for the marker data across all markers. The two-point algorithm considers each pair of adjacent markers separately, ignoring information from more distant markers.
The support intervals calculated for multipoint distance estimates assume that all other distances are fixed at their maximum likelihood positions while one target interval is varied. In a future version, we may implement a method which will optimize the other distances as the target interval is varied.
If error_freq is non-zero, then sib_map will attempt to
identify likely typing errors using provisional distance maps using
the multipoint algorithm. The map is recalculated iteratively until
no new errors are discovered.
By default, sib_map calculates a single sex-averaged map. If
sex_split is enabled, then separate maps are calculated based on
the observed paternal and maternal recombination rates.
In a second mode of operation, sib_map calculates three-point
distances for one marker against all other pairs of adjacent markers
along a map. This data can be used to verify map orders, or to
position new markers on an already determined map. This mode always
uses sex-averaged distances.
sib_map [-v]
[-c cmd]
[-f file]
[marker_data ...]
[> map_data]
One or more marker data files can be listed on the command line. If no files are specified, marker data will be read from standard input. The map(s) will be written to standard output.
The following additional parameters can be specified using TCL commands via
either the -c or -f mechanisms:
Specifies the LOD score cutoff for support intervals. The default is 1.0 log units, which means that support intervals will cover the range of distances that give LOD scores within 1.0 of the LOD score of the most likely distance. The default is 0.6 LOD units.
Specifies the convergence criterion for the iterative distance refinement routines. The default is 0.00001 Morgans.
A boolean value: specifies if sib_map should perform the map
shuffling function, to verify map order and/or place new markers on an
existing map. The default is false.
A boolean value: if set, then separate paternal and maternal recombination maps will be calculated. The default is false.
A boolean value: if set, then allele frequencies are reported.
Here is a sample parameter file for sib_map:
set nloc 4
set loc { "M1" "M2" "M3" "M4" }
set blank "0"
set mapping Kosambi
The normal output of sib_map consists of the two-point and
multipoint distances between each pair of adjacent markers, and the
corresponding support intervals. If two markers are determined to be
unlinked, their distance will be reported as ``[inf]'' (for
``infinity''). In this case, the support interval will give a lower
bound on the most likely distance between the pair. A LOD score is
reported for each interval, giving the likelihood for the most likely
distance, compared to the likelihood of the two markers being
unlinked.
From simulations, we estimate that for support levels of 0.2, 0.6, and 0.8 LOD units, the true distance should be within the support interval about 70%, 90%, and 95% of the time, respectively. These are not strict confidence intervals, however, so these probabilities should be used only as rough guidelines.
If verbose (-v) output is selected, then tables of LOD scores
versus distance for each marker pair will also be generated.
If do_shuffle is true, then the output is, for each marker, a
table giving three-point distance estimates for that marker with every
other pair of adjacent markers along the map. The total distance
spanning the three markers, and the corresponding LOD score, is
generated for all possible orders of markers (XAB, AXB, ABX). Thus, a
comparison of the LOD scores indicates where the test marker is likely
to be in relation to the pair.
Following the distances and LOD scores, sib_map will print one of
several symbols based on a comparison of the LOD scores. If the test
marker appears to be to the left of the specified pair, then ``<'' or
``<<'' will be printed: the number of arrows indicates that the LOD
score difference exceeds that number times the value of support.
Similarly, ``>'' or ``>>'' will be printed if the marker is to the
right of the pair. ``+'' or ``++'' will be printed if the marker is
most likely to be between the specified pair. If the map order is
correct and well supported by the data, the symbols for each marker
should show a pattern of ``>'' rows, then one ``+'' row, then ``<''
rows, as the marker is shuffled through its true position. To use
the output to position new loci on a map, in the parameter file for
sib_map, list the new loci first in the map, followed by all the
already-mapped loci in their proper order.