
Things to do for R/qtl
----------------------
 
o summary.cross() should check that markers are in increasing order.

o Allow use of covariates in scanone() function (ie, CIM).

o Make third row (marker positions) in csv file optional.

o Implement multiple interval mapping (MIM).

o Fix problem in create.map for when map doesn't start at 0.

o Write drop.ind(), to drop some set of individuals from a cross.
  [make general: subset.cross rather than remove.markers, drop.ind,
   pull.chr, etc.]

o pull.chr() should be more flexible in the chr argument.

o phenotypes should be a data.frame.

o Add appropriate functions to analyze RI lines and advanced
  intercrosses. 
  
o Non-parametric interval mapping (Kruglyak and Lander 1995)

o Starting values in vbscan().

o Fix up details on model in help files for vbscan() and vbscan.perm().

o Re-write vbscan.perm() in C code (like scanone.perm()).

o Revise the details section of the help file for read.cross.mm().

o Individual numbers in plot.geno() function.

o plot.geno() to plot only individuals with apparent genotyping
  errors, with a mix of chromosomes.

o plot.pxg(): dotplot of phenotypes split by genotype at a particular
  marker.  Also do interaction plots for a pair of markers.

o Thoroughly test argmax.geno, calc.genoprob, sim.geno, est.map,
  est.rf, calc.errorlod.  Compare these to mapmaker.  Compare scanone
  to mapmaker/qtl.

o Forward selection or MCMC with the BIC-delta criterion.

o Forward selection with permutation tests.

o Add a scantwo() function, scanning for pairs of loci simultaneously,
  calculating LOD scores for the main effects, interactions, and
  total. 

o Implement Sen and Churchill's pseudomarker (imputation) algorithms.

----------------------------------------------------------------------
This file is intended to contain a list of many of the additions that
are planned for the R/qtl package.  

If you any additions to suggest, please send an email to Karl Broman,
kbroman@jhsph.edu 
