Molecular Methods - Research Methods in Human Skeletal Biology - page 452

Descriptive statistics are used to summarize a set of observations, with the purpose of
communicating as much as possible and as simply as possible about a sample. No models
or hypotheses are necessarily being tested. For example, take the hypothetical statement,
“95% of survey takers in California claim to like chocolate.” The use of a percentage imme-
diately leads one to understand that an overwhelming proportion of individuals in the hypo-
thetical sample are chocolate fans.
Let's say that we want to know whether or not such a response is unusual. To begin with,
we want to know what the likelihood is that that response result could have been observed
from any random set of observations. We might also want to know if that response result is
unusual with respect to worldwide data on chocolate preferences, for example. Addressing
both these and similar questions requires the construction of models d including models of
randomness d against which to compare the data. Many of the statistical methods currently
in use have already developed robust comparative models, and can be used to generate
indices of statistical significanced that is, a value (known as a probability or p-value ) that
expresses how unusual or unexpected the research result is, given the statistical model.
Researchers often incorporate statistics into simulation-based methodologies. Simulation
modeling is an approach that predicts probability distributions based on given hypotheses.
This approach allows researchers to take into account the complexity of multiple interacting
variables. In general, modeling methods force researchers to formalize their assumptions,
allow researchers to explore the interactions of the parameters structuring their models,
and help convert models of process into patterns that can be compared with real world,
observed data.
To help researchers in the task of developing interpretive statements of genetics-based
data, software packages are now available, many of which are freely downloadable. Depend-
ing on the software, you will be able to generate descriptive statistics for your data, and eval-
uate your data against various models, including population genetic models. For example,
a popular and widely used software package for use with genetic data is Arlequin , developed
by Laurent Excoffier and colleagues (the latest version, Arlequin v. 3.5.1.2, is described in
Excoffier and Lischer, 2010 ). Arlequin is designed to provide the average user with a large
set of basic methods and statistical tests to derive population genetic information on a collec-
tion of population samples. For a listing of more of the widely utilized software packages, see
the review article by Excoffier and Heckel (2006) .
DEGRADED DNA
How does DNA Decay?
All organisms, whether living or dead, sustain damage to their DNA. In life, an organism
has active biomolecules that repair DNA as needed. That repair necessarily involves biomol-
ecules that both excise and stitch together nucleotides, as damaged nucleotides need to be
removed and then replaced. In death, these repair molecules stop working but the excising
molecules continue to function, chopping up DNA into fragments. This process is called
autolysis , and is capable of completely degrading an organism's DNA unless hindered by
extreme desiccation or freezing.