Biology Reference
In-Depth Information
units, by confirming current models or redefining them at their 5
0
,3
0
(or
both) ends (
Fig. 10.4
C). The potential of RNA-Seq coupled to RNA-
PET in term of genome annotation is therefore tremendous.
3.2. NGS provides functional data to analyze TH-induced
differential expression
Amphibian metamorphosis involves extensive modifications/remodeling
of the organism as a whole. Since this biological process is strictly con-
trolled by TH, documenting expression variation of TH-responsive genes
is central to understand the molecular mechanisms controlling metamor-
phosis, decrypt the gene network controlled by TH, and its spatial and
temporal dynamics. In the pre-NGS era, the simultaneous measure of
the expression level of thousand of transcripts was carried out by micro-
arrays. Yet, as discussed in a previous section, this high-throughput tech-
nology is not “naive”
per se
, in that the design of the microarrays is based on
a prior knowledge of the transcripts to be detected. Indeed, RNA-Seq can
benefit from the evolution and ongoing curation of a genome annotation,
whereas detecting novel transcripts with microarrays would require a new
design. In fact, important genes involved in amphibian metamorphosis
might not even be represented on the chip (e.g., TR
b
gene on
X. tropicalis
Affymetrix array, see above). Nowadays, in the NGS era, RNA-Seq rep-
resents a more naive, genome-wide alternative for measuring gene expres-
sion, in that the strength of the method is not based on a prior knowledge
of transcripts, or gene annotation (
Fig. 10.3
). For example, in the case of
TR
b
, even if the gene is not well annotated, the RNA-Seq signal spreads
over 210 kb and can be assigned to it
a posteriori
, whereas this is not possible
with microarrays. RNA-Seq also has the benefit of having a much larger
dynamic range than microarrays (
Łabaj et al., 2011
). In fact, the hybridi-
zation step in RNA-Seq is carried out
in silico
, which provides full control
over it, and does not saturate. This is in contrast with microarrays, with
which the hybridization signal can quickly reach the saturation limit of
the photomultiplier. Once RNA-Seq data have been mapped to either
a reference genome or transcriptome, several layers of analysis can be car-
ried out. The absolute level of gene expression can be measured by com-
puting the reads per kilobase of exon model per million of mapped reads
(RPKM) value. Additionally, if several samples have been subject to the
same analysis, one can perform differential analysis in a manner similar
to that of microarray data, but using read count instead of photons. As
shown in
Fig. 10.5
A, the expression of the TH regulated gene THb/
ZIP is strongly induced upon T
3
treatment. Although the THb/ZIP gene
