Biology Reference
In-Depth Information
humans.
1
Scientists therefore face the problem of deriving the function
and/or structure of an extremely complex biological system from a
limited set of experimental data that frequently contain large
uncertainties. The quantitative methods developed to support these
efforts can be broadly divided into two types. Statistical methods are
data-driven and provide an objective approach to deriving hormone
secretion patterns from individual hormone time series. Mathematical
modeling methods are based on physiological evidence and hypotheses
regarding hormone secretion mechanisms. These two methods are not
unrelated. Although some statistical methods are model-independent,
other statistical methods frequently utilize mathematical models as part
of the data analyses, and the results of a mathematical model are often
validated through statistical data analyses.
In this chapter, we focus on the statistical approach. Mathematical
modeling of hormone networks is discussed in the next chapter.
II. EXPERIMENTAL DESIGN, DATA COLLECTION,
AND ERRORS OF MEASUREMENT
The primary objective when analyzing hormone data is to characterize
the intrinsic pulsatile and nonpulsatile nature of the secretory process.
These characteristics include the number, times, and masses of the
pulsatile secretory events that increase the hormone concentration in the
blood and the amount of basal (nonpulsatile) secretion. Before data
collection is initiated, a carefully designed experimental protocol should
be in place. Among other things, the protocol should specify the detailed
mechanism of data collection, including whether repeated
measurements will be used and specifying how often the measurements
will be taken.
Because direct data collection at the endocrine glands where the
hormones are secreted is often impossible, measurements of the
hormone concentration in the blood are used to reconstruct the secretion
patterns. This reconstruction is necessitated by the fact that the observed
hormone concentrations in blood as a function of time are the result
of a combination of hormone secretion into the blood and
pharmacokinetic removal of the hormone from the blood. In addition,
because the exact timing of hormone secretion events changes between
closely matched subjects, and even when the same subject is resampled
days or months later, considering repeated measurements is
appropriate. Finally, an extremely important property of the data is the
precision of the measured hormone concentrations, also known as the
1. Recall that hormones such as GnRH are secreted in tiny amounts by the
hypothalamus, which is part of the brain. It is not feasible to directly measure the
GnRH as it is secreted.
