Examples of how the genehunter method of counting crossovers can go
wrong. This example is referred to in the Allegro paper.

The first example is fundamental and shows why first finding the MLE
of the recombination fraction and then converting that MLE to
centimorgans can lead to very wrong answers. The example consists of a
single sibpair genotyped at two markers where two recombinations occur
between the two markers. To run this example type

allegro allegro1.opt
gh < genehunter1.opt

Allegro's estimate of the crossoverrate between the two markers is
shown in the file xover1.out and should be 50.665cM. Genehunter's
estimate is shown in the genehunter output and should be
999.90cM(=infinity).

The second example is less important. It shows how Genehunter can
totaling of the number of crossovers is done incorrectly by
genehunter. The example consists of two identical half-sibpairs, both
genotyped at two markers, where one recombinations occurs between the
two markers.

allegro allegro2.opt
gh < genehunter2.opt

The files fxover2.dat and xover2.dat show that Allegro estimates the
number of crossovers between the two markers, in the two informative
meiosis, to be 1.01 in both half-sibships. Allegro therefore states
that the overall number of crossovers between the two markers is 2.03
in 4 informative meiosis. Genehunter on the other hand estimates 1.18
crossovers in 4 informative meiosis, and in the totalling estimates
2.36 crossovers in 4(!) observable meiosis.
