Frequently asked questions
(The typos in the Guo and Thompson paper were first brought to our
attention by Dr. Chris Triggs of University of Auckland, New Zealand.)
1. Does the Guo and Thompson (1992) paper contain any errors/typos?
Yes, it does contain some typos.
2. Does the program, hwe.c or cmc.c contain any errors/bugs?
Fortunately, no (not we are aware of as of June, 2000).
3. What are those typos in the paper?
There are several of them:
a) In the definition of f_i, the number of A_i
alleles in the sample (three lines above expression (1) on page 363)
the limit of summation should be m, not k.
b) In expression (1), p 363, the denominator should be
\prod_{j\geq i} f_{ij}! not just the \prod_{j > i}.
The > is correct in the exponent of 2. Another way to write that
subexpression is
2^{n - \sum_{i=1}^{m} f_{ii}}
because the subexpression 2^n can be taken out when evaluating Pr({\bf
f}). This formulation also makes it easier to find the probability
ratios in Table 1.
c) There is a typo in the definition of \delta in
line -2 on page 364. As currently defined
\delta = \delta_{i_{1}j_{1}} \delta_{i_{2}j_{2}}
which by our
calculations always takes the value 0 if \Delta = 1. Thus \gamma
always takes the value 1/2. This has the consequence that the
probability ratios for D_1 and R_1 switches in Table 1 can be
incorrect.
For example take i_1 = 1, i_2 = 4, j_1 = 2, and j_2 = 4. Direct
calculation of the probability ratio for a D_1 switch shows that
\gamma = 2 (not 1/2).
The correct definition should be:
\delta = \delta_{i_1j_1} + \delta_{i_2j_2} -
\delta_{i_1j_1}\delta_{i_2j_2}
that is
\delta = 1 if either i_1 = j_1 or i_2 = j_2
= 0 otherwise
The gamma should be
\gamma = 2^{\Delta}\delta + 2^{-\Delta}(1-\delta)
which should be in D_0, R_0, D_2, and R_2 rows, too, in addition to
D_1 and R_1 rows (same place as in gamma and 4).