R/qtl version 1.09-43

[Karl Broman]

A new version of R/qtl (1.09-43) has been released at its web site
(http://www.rqtl.org).

See http://www.rqtl.org/STATUS.txt for a list of changes. The major
changes are listed below.

As always, I'd appreciate hearing any suggestions or comments or
complaints.

Karl Broman
kbr...@biostat.wisc.edu

------

Major changes:

Added a function stepwiseqtl() for performing forward/backward
selection to identify a multiple QTL model, with model choice made
via a penalized LOD score, with separate penalties on main effects
and interactions.

The documents "Brief tour of R/qtl" and "New functions for exploring
multiple-QTL models" were revised to discuss the stepwiseqtl
function.

calc.penalties() uses permutation results for a 2-dimensional, 2-QTL
scan to derive penalties for the penalized LOD scores used by
stepwiseqtl().

plotModel() is a new function for creating a simple graphical
representation of a QTL model.

Added an additional cross type, "dh", for doubled haploids. This is
treated like a backcross, though genotypes will be indicated as
homozygotes. Changed a whole bunch of functions very slightly to
accommodate this.

The argument pheno.col in many functions can now be a vector of
numeric phenotypes. This could be useful for studying the results
with various transformations of a phenotype. The vector has to be
numeric, has to have the length equal to the number of individuals
in the cross, and has to contain either non-integers or values
outside the range 1,2,3,...n.phe. (The revised functions are
scanone, scantwo, addqtl, addint, addpair, cim, effectplot, fitqtl,
plot.pxg, plot.pheno, refineqtl, scanoneboot, scanqtl, stepwiseqtl.)

lodint and bayesint were revised to accept qtl objects output by
refineqtl (with keeplodprofile=TRUE). An additional argument,
qtl.index, was added to indicate for which QTL (within such qtl
objects) the approximate confidence intervals should be derived.
For scanone output, the functions were modified so that, if the
results concern just a single chromosome, the chromosome argument is
not needed.

In refineqtl, the default for the argument keeplodprofile is now
TRUE. The LOD profiles contained in the output of refineqtl are now
of class "scanone".