53
5.1.1 Isolation and Processing of Microbiome mRNA
Metatranscriptomic analysis involves isolation of total RNA from microbial communities of a particular environmental sample. In eukaryotic samples, messenger
RNA can be selected by cDNA mediated synthesis through application of oligo-d
(T) primers that utilise the characteristic poly-A tail of mRNAs. Prokaryotes have
only 1–5% of mRNA of the total RNA (Peano et al. 2013), even without the poly-A
tail, rendering it unusable for the previously mentioned eukaryotic fashioned cDNA
synthesis. Many technological advancements have been put into practice to offer a
solution to this problem (Sultan et al. 2014; Sharma et al. 2010). Probe application
specific to selective rRNA adhered to magnetic beads provides efficient removal of
rRNA. The procedure engages probe annealing to selected sequences (rRNA) and
subsequent magnetic removal of the rRNA (Sultan et al. 2014). Only an enriched
population of other mRNAs that corresponds to transcriptionally active genes is
used in the method. Furthermore, for massive parallel sequence generation, RNAs
are fractionated for corresponding cDNA synthesis, and then cDNA ends are linked
with adapters followed by end repair, finally generating a library for further amplification and sequencing purposes. Sequence reads are then processed for mapping
and the expressed genes are recognized on the basis of the sequence reads covering
these regions (Bashiardes et al. 2016a, b).
5.1.2 Computational Analysis of Metatranscriptomics Data
The sequenced metatranscriptome dataset possess millions of mRNA molecules,
called RNA-seq reads. With increasing number and sample size of metatranscriptomic studies, highly efficient analysis platforms are required to draw meaningful
inferences from these datasets (Gosalbes et al. 2011; Korf and Rehm 2013). Various
extensive analysis suites such as HUMAnN (Abubucker et al. 2012) and MG-RAST
(Glass et al. 2010) were developed and widely applied for efficiently resolving the
problems related to interpretation of the raw data. These techniques are used in
combination of specialized bioinformatics tools such as BOWTIE (Langmead and
Salzberg 2012) and GEM (Marco-Sola et al. 2012) for mapping, Trimmomatic
(Bolger et al. 2014) for quality filtering, and CuffDuff (Ghosh and Chan 2016) for
differential gene expression, so that variations in the gene expression levels can be
inferred from the raw sequenced mRNA reads (Bashiardes et al. 2016a, b). Certain
analytical steps are necessary for this procedure and, as a result, they are consistently present in almost every metatranscriptomic. These steps are filtering of nonmRNA reads along with the host reads, low-quality reads trimming, open- reading
frames (ORFs), mapping of the reads to a particular database, data normalization,
and computing the expression levels of mRNA with alternative summarized statistics (Wang et al. 2009; Bashiardes et al. 2016a, b).
5.1 Metatranscriptomics
5.1.1 Isolation and Processing of Microbiome mRNA
Metatranscriptomic analysis involves isolation of total RNA from microbial communities of a particular environmental sample. In eukaryotic samples, messenger
RNA can be selected by cDNA mediated synthesis through application of oligo-d
(T) primers that utilise the characteristic poly-A tail of mRNAs. Prokaryotes have
only 1–5% of mRNA of the total RNA (Peano et al. 2013), even without the poly-A
tail, rendering it unusable for the previously mentioned eukaryotic fashioned cDNA
synthesis. Many technological advancements have been put into practice to offer a
solution to this problem (Sultan et al. 2014; Sharma et al. 2010). Probe application
specific to selective rRNA adhered to magnetic beads provides efficient removal of
rRNA. The procedure engages probe annealing to selected sequences (rRNA) and
subsequent magnetic removal of the rRNA (Sultan et al. 2014). Only an enriched
population of other mRNAs that corresponds to transcriptionally active genes is
used in the method. Furthermore, for massive parallel sequence generation, RNAs
are fractionated for corresponding cDNA synthesis, and then cDNA ends are linked
with adapters followed by end repair, finally generating a library for further amplification and sequencing purposes. Sequence reads are then processed for mapping
and the expressed genes are recognized on the basis of the sequence reads covering
these regions (Bashiardes et al. 2016a, b).
5.1.2 Computational Analysis of Metatranscriptomics Data
The sequenced metatranscriptome dataset possess millions of mRNA molecules,
called RNA-seq reads. With increasing number and sample size of metatranscriptomic studies, highly efficient analysis platforms are required to draw meaningful
inferences from these datasets (Gosalbes et al. 2011; Korf and Rehm 2013). Various
extensive analysis suites such as HUMAnN (Abubucker et al. 2012) and MG-RAST
(Glass et al. 2010) were developed and widely applied for efficiently resolving the
problems related to interpretation of the raw data. These techniques are used in
combination of specialized bioinformatics tools such as BOWTIE (Langmead and
Salzberg 2012) and GEM (Marco-Sola et al. 2012) for mapping, Trimmomatic
(Bolger et al. 2014) for quality filtering, and CuffDuff (Ghosh and Chan 2016) for
differential gene expression, so that variations in the gene expression levels can be
inferred from the raw sequenced mRNA reads (Bashiardes et al. 2016a, b). Certain
analytical steps are necessary for this procedure and, as a result, they are consistently present in almost every metatranscriptomic. These steps are filtering of nonmRNA reads along with the host reads, low-quality reads trimming, open- reading
frames (ORFs), mapping of the reads to a particular database, data normalization,
and computing the expression levels of mRNA with alternative summarized statistics (Wang et al. 2009; Bashiardes et al. 2016a, b).
5.1 Metatranscriptomics