Here, we detail a protocol for simultaneous mapping of TSSs
and RNA cleavage sites transcriptome-wide in mycobacteria using
an Illumina sequencing platform (see Note 1). Data from this
protocol have been published for both Mycobacterium tuberculosis
and Mycobacterium smegmatis [1, 2]. The protocol has two major
components. The first describes in detail a procedure for construction of Illumina sequencing libraries (Subheadings 3.1–3.7), and
the second describes bioinformatic analysis of the resulting data
(Subheadings 3.8–3.12). This protocol is not designed for comparative gene expression analysis, although it is appropriate for comparing alternative promoter use and RNA cleavage patterns in
different strains or conditions. These data are highly complementary to RNA-seq gene expression datasets that can be generated
using published protocols [3, 4] or commercially available kits.
To capture RNA 5
0 ends, adapters are ligated directly to intact
RNAs. In order to facilitate subsequent discrimination between
primary RNA 5
0 ends (which have 5
0 triphosphates and correspond
to TSSs) and 5
0 ends produced by endonucleolytic cleavage (which
have 5
0 monophosphates or 5
0 hydroxyls), two parallel libraries are
constructed from each RNA sample. The “converted” library captures all RNA 5
0 ends that have 5
0 triphosphates or 5
0 monophosphates, and the “nonconverted” library captures only RNA 5
0 ends
that have 5
0 monophosphates (see Fig. 1). The relative sequence
coverage from each of the two libraries at a given RNA 5
0 end can
then be used to determine if it was a primary 5
0 end (reflecting a
TSS) or an endonucleolytic cleavage product.
The major steps of library construction consist of differential
treatment of RNA 5
0 ends, ligation of adapters, fragmentation of
RNA, cDNA synthesis, PCR to add further adapter sequences, and
size selection. The resulting libraries consist of DNA inserts flanked
by adapters equivalent to those in standard genomic DNA pairedend TruSeq© Illumina libraries and can be sequenced on any
Illumina sequencer (see Fig. 1).
The major steps of data analysis are mapping reads to a reference genome, determining normalized coverage values for 5
0 ends
mapping to each coordinate of the genome, selecting coordinates
with above-background coverage, and using mixture modeling to
distinguish TSSs from cleavage sites. Here, we present scripts and
bioinformatics tools that can be used to perform these analyses.
Note that this bioinformatics pipeline represents a functional path
to achieve the desired results but that an experienced bioinformatician could use these conceptual steps to produce a more efficient
and streamlined pipeline.
514
M. Carla Martini et al.
and RNA cleavage sites transcriptome-wide in mycobacteria using
an Illumina sequencing platform (see Note 1). Data from this
protocol have been published for both Mycobacterium tuberculosis
and Mycobacterium smegmatis [1, 2]. The protocol has two major
components. The first describes in detail a procedure for construction of Illumina sequencing libraries (Subheadings 3.1–3.7), and
the second describes bioinformatic analysis of the resulting data
(Subheadings 3.8–3.12). This protocol is not designed for comparative gene expression analysis, although it is appropriate for comparing alternative promoter use and RNA cleavage patterns in
different strains or conditions. These data are highly complementary to RNA-seq gene expression datasets that can be generated
using published protocols [3, 4] or commercially available kits.
To capture RNA 5
0 ends, adapters are ligated directly to intact
RNAs. In order to facilitate subsequent discrimination between
primary RNA 5
0 ends (which have 5
0 triphosphates and correspond
to TSSs) and 5
0 ends produced by endonucleolytic cleavage (which
have 5
0 monophosphates or 5
0 hydroxyls), two parallel libraries are
constructed from each RNA sample. The “converted” library captures all RNA 5
0 ends that have 5
0 triphosphates or 5
0 monophosphates, and the “nonconverted” library captures only RNA 5
0 ends
that have 5
0 monophosphates (see Fig. 1). The relative sequence
coverage from each of the two libraries at a given RNA 5
0 end can
then be used to determine if it was a primary 5
0 end (reflecting a
TSS) or an endonucleolytic cleavage product.
The major steps of library construction consist of differential
treatment of RNA 5
0 ends, ligation of adapters, fragmentation of
RNA, cDNA synthesis, PCR to add further adapter sequences, and
size selection. The resulting libraries consist of DNA inserts flanked
by adapters equivalent to those in standard genomic DNA pairedend TruSeq© Illumina libraries and can be sequenced on any
Illumina sequencer (see Fig. 1).
The major steps of data analysis are mapping reads to a reference genome, determining normalized coverage values for 5
0 ends
mapping to each coordinate of the genome, selecting coordinates
with above-background coverage, and using mixture modeling to
distinguish TSSs from cleavage sites. Here, we present scripts and
bioinformatics tools that can be used to perform these analyses.
Note that this bioinformatics pipeline represents a functional path
to achieve the desired results but that an experienced bioinformatician could use these conceptual steps to produce a more efficient
and streamlined pipeline.
514
M. Carla Martini et al.