producing the model is called the training set.
Models built using the training data can then be
independently validated using the hold-out set.An
alternative method of independent model validation
is to use permutation testing.19 More robust methods
include confirming the observations with a
complementary technique and replicating the
experiment in a different sample set.20
There are many publications, across all the
biological sciences, pointing out the potential folly
of using profiling techniques such as metabolomics,
proteomics, transcriptomics and genomics in order
to discover clinically significant biomarkers.21–25
These areas of experimental design, sample
preparation, analytical techniques and data analysis
are covered in greater detail in a number of review
articles.13,21,26–30
Omic research in obstetrics
and gynaecology
While omic studies in the area of obstetrics and
gynaecology are relatively small in number, with
advancing technology, it is a rapidly expanding field
of research. It is not possible in this article to give an
in-depth discussion of all omic research in
obstetrics and gynaecology.As may be anticipated,
most of the omic research in gynaecology is centred
on oncology and cancer screening and much of the
work in obstetrics on identifying biomarkers for
complications of pregnancy such as pre-eclampsia
and preterm birth.
A wide range of biological samples has been used
for omic research, including plasma/serum, urine,
amniotic fluid, cultured trophoblasts and
cervicovaginal and follicular fluid.
Pre-eclampsia
Studies have found gene expression to differ
between pregnancies with pre-eclampsia and
uncomplicated pregnancies in peripheral blood,31
first-trimester placentas32 and placentas at delivery.33
Much of the proteomic research in pregnancy has
been in the area of pre-eclampsia. Studies have
shown differences in women with pre-eclampsia
compared with women with uncomplicated
pregnancies, including differing serum levels of
clusterin34 and ficolin35 but these were time-ofdisease
samples. One study36 showed differences in
five proteins at 26 weeks of gestation but these
proteins could not be identified. Recently the
plasma proteome at 20 weeks in women who
subsequently developed pre-eclampsia (n = 39) was
compared with that in normal healthy controls
(n = 57): 39 proteins were identified and two protein
clusters identified as fibrinogen gamma-chain and
alpha-1-antichymotrypsin accurately classified
women at risk of developing pre-eclampsia.37