Biology Reference
In-Depth Information
that survive to reproduce in years
t
2, and so on (
Figure 8.1
A).
These proportions determine the average age to adulthood (i.e. genera-
tion length,
T
).
65
Thus, even though
Ascaris
adults may live only a year in
their host,
11
generation length is likely several years longer due to the fact
that eggs can persist 6
1,
t
þ
þ
9 years in the environment.
12
Interestingly,
Ascaris
life history closely approximates that of semelparous, age-structured
species such as annual plants with seed banks and Pacific salmon. A
detailed theoretical treatment of estimating
N
e
in the latter groups of
organisms is given by Waples.
65
In short, generational
N
e
is
e
T
˜
t
,
, where
T
is generation time in years and
˜
t
is the harmonic mean of the
N
t
's
within a generation.
65
An important point to recognize is that a sample of
adult worms of a given breeding year will contain individuals of mixed
ages, i.e. there are overlapping generations (
Figure 8.1
A). With the LD-
N
e
and SA-
N
e
methods, the estimated
N
e
reflects that of the breeders that
produced the sampled adult worms (i.e. the parents of the sampled
worms) and not the sampled worms themselves (i.e. not
N
t
). While
cautioning that testing is needed, Waples and Do
56
conjectured that
a mixed-aged sample that includes a number of consecutive age classes
approximately equal to generation length should produce an estimate
roughly corresponding to generational
N
e
. Thus, throughout the chapter, I
will assume that the sample of adult worms from each household
provides an estimate of generational
N
e
of each subpopulation
(
Figure 8.1
A). I will return to the estimation of
N
t
(
Figure 8.1
B) in my
concluding remarks.
I used two single-sample, contemporary estimators, LD-
N
e
and
SA-
N
e
,
55
e
57
as implemented in the programs
LDNE
66
and
COLONY
v2.0.2.1
67
,
respectively. Both of these methods provide estimates that are related to
the inbreeding
N
e
.
56,57
The LD-
N
e
method can be sensitive to rare alleles,
thus I followed the recommendations of Waples and Do
56
for using alleles
with frequencies above a cutoff given the sample size (see
Table 8.1
). The
random mating system option was used. In
COLONY
, I selected the male
and female polygamy options without inbreeding. These latter options in
the two programs seem reasonable given the current state of knowledge
about
Ascaris
mating systems.
9,10
Length of run and likelihood precision
(full-likelihood) were set to medium in
COLONY
. I used the update allele
frequency option and the complexity prior, which should result in
a higher
N
e
estimate (compared to not using it) as this prior discourages
complex pedigree inference.
Table 8.1
provides the estimates of
N
e
per household-by-year where 10
or more worms were genotyped (
n
z
26). There are several important
patterns and questions that emerge from these data. First, sample size
matters in obtaining estimates that are not infinite or do not have an upper
confidence interval of infinity. Infinity estimates (negative values in the
LD-
N
e
method or the 2.15
¼
10
9
values in the SA-
N
e
method) result when
