© Benaki Phytopathological Institute
Abstracts - 16th Hellenic Phytopathological Congress
37
I
NVITED
L
ECTURE
Deep sequencing and the identification without prior
knowledge of phytoviruses: applications to plant disease etiology and to
metagenomics
T. C
ANDRESSE
1
, A. M
ARAIS
1
, C. F
AURE
1
, L. S
VANELLA
-D
UMAS
1
, B. B
ERGEY
1
, Y. L
AIZET
1
, M. C
AMBRA
2
and A. O
LMOS
2
1
UMR 1332 BFP, INRA and Université de Bordeaux, BP81, 33883 Villenave d’Ornon
cedex, France.
2
Instituto Valenciano de Investigaciones Agrarias (IVIA), 46113
Moncada, Valencia, Spain
Similar to viruses infecting other organisms,
plant viruses are collectively characterized
by an amazing diversity, encoding their ge-
nome on various kinds of nucleic acids and
sharing as a group no single common gene
or determinant. The direct consequence of
this situation is that although the detection
of a known virus is now generally straight-
forward, the detection, without prior knowl-
edge, of all viruses present in a plant sample
remains a significant challenge. The devel-
opment of novel sequencing technologies
(NGS: next generation sequencing) allows
unprecedented sequence data generation
at a fraction of the cost of previous technol-
ogies and has drastically altered this situa-
tion. Indeed, it is now possible to generate
vast amounts of sequence information from
a plant sample and to then use bioinformat-
ics tools to sift through this data in order to
identify any virus that might be present. Var-
ious templates have been sequenced in such
approaches, including messenger RNAs (mR-
NAs), small interfering RNAs (siRNAs), dou-
ble stranded RNAs (dsRNAs) or nucleic acids
extracted from semi-purified viral particle
preparations. In the case of DNA plant virus-
es, which to date all have circular genomes,
the sequencing of RCA (Rolling Circle Ampli-
fication) products has been used.
The first area in which these approaches are
likely to have a major impact concerns the
field of etiology and the conceptually linked
certification and quarantine fields. In each
case, the key problem is to be able to quickly
identify with great sensitivity any viral agent
that may be present in a sample. We have re-
cently used siRNA and dsRNA sequencing in
an effort to identify viruses in
Prunus
stone
fruit material, illustrating the power and the
potential interest of these approaches. In
our hands, the best results so far were ob-
tained with dsRNA which can be analyzed in
a multiplexed format in order to reduce in-
dexing costs. Although a detailed sensitivity
comparison with biological indexing has yet
to be performed, the cost of this new tech-
nique already compares favourably.
The second area where these strategies are
arousing wide interest is the global anal-
ysis of viral populations associated with
plants in an environment (metagenomics).
Such questions were formerly not accessi-
ble to experimentation but are now feasible
thanks to the democratization of NGS. We
are developing this approach in two con-
trasted environments, the simplified eco-
system of the Kerguelen islands, the second
most isolated archipelago on earth, and in a
less constrained but anthropized temperate
horticultural setting. The first results indi-
cate a very high proportion in dsRNA viruses
in the Kerguelen islands and, conversely, an
enrichment in pathogenic ssRNA viruses in
the agricultural context. They also provide
a first glimpse at virus biodiversity in these
two highly contrasting situations.
Outside of these first results demonstrating
the interest and the potential of these nov-
el approaches, it is clear that new technical
and conceptual developments are to be ex-
pected in the coming years. On the techno-
logical side, NGS is progressing, allowing
for ever increasing volumes of sequence at
lower cost. Challenges are therefore to be
found in the processing of samples (autom-
