322
reads allowed to build a rank-abundance curve (Fig. 19.3 )
showing that only 7 OTUs had an abundance greater than
1 % belonging to Haptophyceae, Chlorophyta and Fungi. We
defi ned as abundant the OTUs characterized by a proportion
of DNA reads greater than 0.1 %. The other OTUs belonging
to the rare biosphere were divided in two classes: the OTUs
that are always rare in all samples rare and the others (cycling
fraction) being sometimes rare (≤0.1 %) but becoming other
times not rare (>0.1 %). The rare and the cycling fraction rep-
resented 77.2 % and 20.4 % respectively of the 6446 OTUs.
When the data were available, the proportion of rare taxa was
lower for marine bacteria (44.7 %) (Campbell et al. 2011 )
and Archaea (40.1 %) (Hugoni et al. 2013 ). The contribution
of the rare fraction of picoeukaryotes sequences in the lacus-
trine ecosystems is congruent with the results obtained in
marine system, which showed that the rare taxa ranged
between 66.2 % and 76.6 % (Logares et al. 2014 ).
19.3.1 Is There New Taxa in the Rare
Biosphere?
Many investigations using NGS have uncovered a large
number of unclassifi ed sequences with low abundances
(Bartram et al. 2011 ). However very few studies explore the
link between rarity of an OTU and its phylogenetic distance
to the nearest neighbor reference (sequences from cultured
or uncultured taxa available in public database, ie Silva,
NCBI, ...) whereas this diversity could represent new lin-
eages with new physiological properties. The always rare
Archaea in the coastal surface waters were characterised by
a group of OTU with low identity with the public database
sequences (Hugoni et al. 2013 ). For picoeukaryotes, in the
always rare category, 13.9 % of OTUs have an identity < 92 %
whereas this proportion is only 6.4 % for the abundant OTUs.
In addition, the lowest similarities found by BLAST search
were observed for MGs including only always rare OTUs.
For the phylogenetic indices (MNNDs and patristic dis-
tances), the greatest values corresponded to the MGs includ-
ing only always rare OTUs. Therefore, these results
underlined that these MGs presented the most divergent indi-
ces from the nearest neighbour and could represent newly
detected lineages of which the SSU rDNA phylogeny is
indicative that they represent previously undescribed taxo-
nomic groups at the level of Genus. Interestingly, these
clades were active and distributed throughout all phyloge-
netic groups, and thus belonged to various functional groups
(parasites, autotrophs, saprotrophs, etc.). Therefore, they are
present in groups particularly studied by the cultural
approach (Fungi, phytoplankton), which highlights that an
important fraction of the eukaryotes’ biodiversity likely
remains unknown. Finally, the presence of the unknown taxa
among the rare biosphere and the existence of a biogeogra-
phy among the rarest picoeukaryotes (Lepère et al. 2013 )
suggest that their richness is certainly higher than expected.19.3.2 An Active BiosphereThere are numerous debates about the fact that the rare bio-
sphere is active or is composed by inactive taxa (dormant
cells, dead cells, pseudogenes etc...). From DNA analysis it
is not possible to confi rm that the rare and highly divergent
18S rRNA gene sequences amplifi ed represent active micro-
organisms. By studying the presence or not of rRNA in each
OTU, it is possible to defi ne active and inactive OTUs. The
abundant, the cycling fraction and always rare OTUs repre-
sented 40.2 %, 32.3 % and 27.5 % of the active biosphere
depicted by rRNA respectively, whereas the most abundant
OTUs (>1 %) represented only 13.3 % of the total activity.
The temporal changes of the active OTUs on lake Pavin
showed that composition of the always rare OTUs differedFig. 19.4 Dynamics of the active
monophylogenetic groups (MGs)
including only always rares OTUs
(AllwR) considered as highly
frequent (observed at least in 70 %
of the samples) in Pavin. Each MG
is expressed as normalized rRNA
reads according to the following
formula: (x–μ)/⌠ where x is the
number of rRNA reads at a given
date, μ and ⌠ are the mean and the
standard deviation computed for all
the dates for the MG considered. A
MG is defi ned by its taxonomy and
an number in the brackets
C. Lepère et al.