could not separate for lack of distinctive
morphological traits and/or time.
Quantitively was where NGS metabarcoding
did not performed as well. More
specifically, the amount of DNA reads
(how much DNA was retrieved) from the
different plant species did not match
the pollen grains counted using standard
microscopy on the same sample.
The work of Bell et al., 2019 (https://doi.
org/10.1111/mec.14840) showed that
in a nice experiment. Reasons for this
inconsistency are diverse, including: copy
number, preservation, DNA isolation and
amplification biases.
The need for better
database?
Currently for plants, the largest databases
for matching sequencing data (or DNA barcoding
sequence reference libraries) are:
• NCBI – US National Center for Biotechnology
Information or GenBank®
(www.ncbi.nlm.nih.gov/genbank/)
• EMBL – The European Bioinformatics
Institute 7
• DDBJ – The DNA Data Bank of Japan
Center 8
In particular, in the NCBI database, over
30 million vascular plant nucleotide
sequences can be found from the entire
world. However, it is common knowledge
that the NCBI database is regarded
as problematic because of the lack of a
rigorous quality check of data, which are
commonly entered and published with
misidentifications, intraspecific variation
and sequencing errors.
Good databases with improved quality are:
• BOLD – The Barcode of Life Data System
9
• ITS2 – The internal transcribed spacer 2
database 10
However, these databases lack the global
widespread abundance of sequences and
data of the NCBI.
Using customized
bioinformatic pipelines
One way to help solve this issue is by
creating an adhoc library specifically to
match a list of species found in a certain
area. We could for example create
a library from all common honey plants
in Germany using only rigorously checked
data. Unfortunately, this approach does
not work for fraud analysis, where a
more extensive world database like the
NCBI works better.
A more practical alternative, is to
validate the results post-analysis using
informatic tools and process (a so-called
bioinformatic pipeline). For example, cutoff
quality values are used to remove all
reads that are not passing a desirable
threshold defined by experts.
Despite initial difficulties, rather
than being discouraged, scientists and
experts are seizing this new opportunity.
With the message that the use of NGS
metabarcoding can complement standard
methods.
Where does my honey
come from? Finding out
mislabeling
In the past decades, the market for
honey constantly increased and honeybees
were exported from Europe, where
they originated, to almost every corner
around the world. Increasingly more
varieties of honey are produced for the
joy of honey consumers, who can now
choose from a large selection of local
and exotic honey sources. Different
sources produce honey with characteristics
that are more or less appreciated
or more or less common/rare. The price
in honey varies greatly, from few to hundreds
of euro per kilo. Food fraud has
become a common issue. Honey, from
less desirable origins, is sold using false
labelling at a higher price or to circumvent
import tax as it is the case for
imports from China to the USA. Such
fraud can be discovered by identification
of taxa or cultivars that grow in a specific
region (pollen markers). For this purpose,
the identification must be carried
out down to species level or the lowest
taxonomic level possible, something not
possible in most of the cases by conventional
melissopalynology. In addition,
many of these marker pollen grains are
rare or unknown to palynologists, so
that they are often overlooked. Recently
a collaborative project lead by a group
in Quality Services International GmbH,
applied DNA metabarcoding and floristic
literature reviews to correctly identified
geographical origin of honey 11.
The entomological fingerprint
of honeydew honey
Commonly honeydew honeys in Europe
are produced by honeybees feeding on
excretions of insects like Metcalfa pruinosa
and Cinara pectinatae feeding on
Food analytics and services 23
oak, spruce and fir trees. Standard pollen
analysis cannot successfully be used to
identify these honeydew sources, as the
marker is not pollen but the honeydew
produced by the plant sucking insects.
Other markers are used, such as the presence
of increase amount of other biological
particles (e.g. fungi spores, plant
remains, etc.). However, a clear distinction
between different kinds of honeydews
cannot be made. In 2018, a group
of scientists from the University of Bologna
had the idea to investigate if honeydew
honey contains residual DNA from
the insects, which honeybees fed on.
The results were very promising, showing
that NGS metabarcoding can be used to
successfully answer specific questions on
origin of honey 12.
In short, how can NGS
Metabarcoding help us
answer our questions?
NGS metabarcoding and bio-informatic
technologies attached are becoming
more and more accessible for standard
laboratories working on food analysis
and especially on confirmation of authenticity
of foods. While not being able to
completely replicate the results obtained
from the standard microscopic analysis,
they can complement them providing:
• A completer and more extensive list of
plants contained in the honey (in technical
terms, a lower taxonomic resolution)
• Answer to specific questions related
to what is the biological material contained
in the honey (example of the
honeydew and aphid eDNA)
References
1 Pfister, R., 1895. Versuch einer
Mikroskopie des honing. Forschber.
Lebensmitt. U. ihre Bez. Z. Hygiene,
Forens. Chem Pharmakogn 2, 29–35.
2 Fritzsche, C.J., 1837. Ueber den Pollen.
Kais. Academie der Wissenschaften,
St. Petersburg.
3 Manten, A.A., 1967. Lennart
Von Post and the foundation of
modern palynology. Rev. Palaeobot.
Palynol. 1, 11–22.
https://doi.org/10.1016/0034-
6667(67)90105-4
4 Von Der Ohe, W., Persano Oddo,
L., Piana, M.L., Morlot, M., Martin,
P., 2004. Harmonized methods of
melissopalynology. Apidologie 35,
/www.bilacon.com
/www.bilacon.com
/doi
/)
/0034-