Prune allows one to eliminate markers or traits. It removes the data from
the file containing the cross and reconstructs the molecular map.
It requires a molecular map that could be a random one produced by
Rmap, or a real one in the same format as the output of
Rmap. The sample could be a randomly generated one from
Rcross or a real one in the same format as the output of
Prune also does bootstraps, permutations and simulations of missing or dominant markers.
See QTLcart(1) for more information on the global options
-h for help, -A for automatic, -V for non-Verbose
-W path for a working directory, -R file to specify a resource
file, -e to specify the log file, -s to specify a seed for the
random number generator and -X stem to specify a filename stem.
The options below are specific to this program.
If you use this program without specifying any options, then you will
get into a menu that allows you to set them interactively.
This requires a filename stem for output. Prune will overwrite the file ending in
.crb if it exists, and create a new file if it does not. If not used, then Prune will use
qtlcart.crb. If the map is recreated, then a new map file will be written to
qtlcart.mpb by default or a file ending in mpb with the specified stem.
Sets the interactive level. A zero means that Prune will do what it needs to
without asking (the default for bootstraps, permutations or missing data simulations).
A one means that the user will be put into a repeating loop to manipulate the
data set. It has a value 1 by default, but using the -b option disables it.
The input format of the molecular map should be the same as that of the
output format from the program Rmap. The input format of the
individual data should be the same as the output format of the program
There are a number of automatic actions that can be performed using
the -b option. You will use one of the numbers below with the
option to tell Prune to do that action. A new dataset is then printed
to a file stem.crb, where stem is the filename stem. Note that
if you give a nonzero value to this option, the interactive flag is
Perform a bootstrap resampling
of the data. Sampling of individuals is done with replacement to
create a sample of the same size as the original.
Permute the the traits against genotype arrays. If there are multiple traits
in the data set, then each trait will be shuffled against the genotype arrays.
Simulate missing markers.
The percent of missing marker data should be specified with the -M option, and it
should be an number in the range of 0 to 100 percent.
Simulate dominant markers.
The percent of dominant marker data should be specified with the -M option, and it
should be an number in the range of 0 to 100 percent. The direction of dominance is
Simulate selective genotyping.
The percent of typed individuals should be specified with the -M option, and it
should be an number in the range of 0 to 100 percent. This will print out individuals with
trait values in the tails of the overall distribution. The value specified will be the sum
of these tails: Each tail will have half of the total. This will apply to whichever
trait was last analyzed, or trait 1 if all the traits had been analyzed. It is probably
best to do this with single trait data sets.
Permute the the traits against genotype arrays. A value of 12 does this as well.
If there are multiple traits
in the data set, then entire trait arrays will be shuffled against the genotype arrays.
This contrasts with option 2 above which permutes the traits independently. If
you think the traits are correlated and you want to maintain that correlation, use this
option. Otherwise, use option 2.
Prune the data back to one trait. Use the -t option with a trait number to select the trait.
The output will have one trait: All individuals with missing values for this
trait will also be deleted.
Prune the data to specified traits. Use the -t option with a trait number to select the trait.
If the original data has t traits, then an integer in the range [1,t] will eliminate all
but the specified trait, that is it will do exactly the same thing as option 7 above. If an integer
less than one is used, then only traits whose names begin with a plus sign will remain in the
output. If greater than the number of traits, then any trait whose name begins with a
minus sign will be eliminated. Once the traits are eliminated, all individuals with missing data
for any of the surviving traits will also be deleted.
Remove all categorical trait information. This is for compatibility
with R/QTL, which can not read categorical trait information as of
8 June 2004.
The -b option creates a new sample from the old. The new sample is
created by resampling the original sample with replacement. Phenotypes
and genotypes are kept together. The new sample will have the same sample
size as the old one. It will be written to exout.crb. No new map
will be written.
Here, the -b option tells Prune to selectively genotype. We specify
20.0 percent with the -M option meaning that those individuals with trait
values falling in the lower and upper 10 percent tails are retained, and the
middle 80 percent are removed.
You can eliminate multiple markers in the interactive loop. You should
be aware that the order marker elimination is important. If all
the markers to be eliminated are on separate chromosomes, the order is
unimportant. If two markers from the same chromosome are to be eliminated,
order should be to eliminate the highest numbered marker.
The same concept holds for
traits: eliminate them in the order of highest to lowest.
Do not try to eliminate any markers or traits AND
do a bootstrap, permutation or simulation of missing markers in the same run.
Christopher J. Basten, B. S. Weir and Z.-B. Zeng
Bioinformatics Research Center, North Carolina State University
1523 Partners II Building/840 Main Campus Drive
Raleigh, NC 27695-7566 USA