abc
package for more details).data(human) loads in four R objects: stat.voight is a
data frame with 3 rows and 3 columns and contains the observed summary
statistics for three human populations, stat.3pops.sim is also a
data frame with 150,000 rows and 3 columns and contains the simulated
summary statistics, models is a vector of character strings of
length 150,000 and contains the model indices, par.italy.sim is a
data frame with 50,000 rows and 4 columns and contains the parameter
values that were used to simulate data under a population bottleneck
model. The corresponding summary statistics can be subsetted from the
stat.3pops.sim object as subset(stat.3pops.sim,
subset=models=="bott").
data(human)The stat.voight data frame contains the following columns:
piThe mean nucleotide diversity over 50 loci in 3 human populations, Hausa, Italian, and Chinese.
TajD.mThe mean of Tajima's D statistic over 50 loci in 3 human populations, Hausa, Italian, and Chinese.
TajD.vThe variance of Tajima's D statistic over 50 loci in 3 human populations, Hausa, Italian, and Chinese.
Each row represents a simulation. Under each model 50,000 simulations were performed. Row names indicate the type of demographic model.
The stat.3pops.sim data frame contains the following columns:
piThe mean of nucleotide diversity over 50 simulated loci under 3 demographic scenarios: constant size population, population bottleneck, and population expansion.
TajD.mThe mean of Tajima's D statistic over 50 simulated loci under 3 demographic scenarios: constant size population, population bottleneck, and population expansion.
TajD.vThe variance of Tajima's D statistic over 50 simulated loci under 3 demographic scenarios: constant size population, population bottleneck, and population expansion.
Each row represents a simulation. Under each model 50,000 simulations were performed. Row names indicate the type of demographic model.
The par.italy.sim data frame contains the following columns:
NeThe effective population size.
aThe intensity of the bottleneck (i.e. the ratio of the population sizes before and during the bottleneck).
durationThe duration of the bottleneck.
startThe start of the bottleneck.
Each row represents a simulation.
models contains the names of the demographic models.
Data is provided to estimate the posterior probabilities of classical
demographic scenarios in three human populations: Hausa, Italian, and
Chinese. These three populations represent the three continents:
Africa, Europe, Asia, respectively. par.italy.sim may then used
to estimate the ancestral population size of the European population
assuming a bottleneck model.
It is generally believed that African human populations are expanding, while human populations from outside of Africa have gone through a population bottleneck. Tajima's D statistic has been classically used to detect changes in historical population size. A negative Tajima's D signifies an excess of low frequency polymorphisms, indicating population size expansion. While a positive Tajima's D indicates low levels of both low and high frequency polymorphisms, thus a sign of a population bottleneck. In constant size populations, Tajima's D is expected to be zero.
With the help of the human data one can reach these expected
conclusions for the three human population samples, in accordance with
the conclusions of Voight et al. (2005) (where the observed statistics
was taken from), but using ABC.
B. F. Voight, A. M. Adams, L. A. Frisse, Y. Qian, R. R. Hudson and A. Di Rienzo (2005) Interrogating multiple aspects of variation in a full resequencing data set to infer human population size changes. PNAS 102, 18508-18513.
Hudson, R. R. (2002) Generating samples under a Wright-Fisher neutral model of genetic variation. Bioinformatics 18 337-338.