2 Metagenome Analysis
45
2.3.5 Construction and Analysis of Metagenomic Libraries
2.3.5.1 Small Insert Metagenomic Libraries
Small insert metagenomic libraries (generally 1.5–3 kbp insert size) in plasmid or
bacteriophage λ derived vectors are a repository of the metagenome, which can be
screened by sequence-based and, more often, function-based assays. Early studies
of such small insert libraries (Cottrell et al. 1999, Henne et al. 1999) focused on
the analysis of single metabolic genes of uncultured microorganisms. Small insert
libraries in reporter gene constructs proved to be valuable to identify novel catabolic
operons in a substrate-induced gene expression screening (Uchiyama et al. 2005).
In a variety of studies small insert libraries have been constructed as a prerequisite
for Sanger-based metagenome sequencing. High throughput sequencing of small
insert libraries has proved to be a powerful method for the assessment of genome
drafts (e.g. Tyson et al. 2004, Venter et al. 2004, Woyke et al. 2006), and for the
study of the phylogenetic and metabolic diversity in complex habitats (Venter et al.
2004, Rusch et al. 2007). The construction of small insert libraries is straightforward. Nevertheless, there are several major drawbacks when small insert libraries
are used in sequence-based metagenomic studies. First of all, a phylogenetic marker,
which allows the reliable assignment of a fragment to a certain phylogenetic group,
is generally missing on small inserts. If insufficient sequence data is available or
if microbial diversity is high within the sample, as is the case for soil samples
(Handelsman et al. 2002, Tringe et al. 2005), the reconstruction of draft genomes
or even the assignment of genetic capabilities to certain species is almost impossible. Moreover, the in silico assembly of larger fragments involves a risk of creating
chimeras, especially when the overlaps of the reassembled fragments are too short
or repetitive elements are present.
2.3.5.2 Large Insert Metagenomic Libraries
Metagenomic libraries with large DNA inserts of > 30 kbp (large insert libraries)
have the advantage that genome fragments can often be assigned to a specific
species. This assignment can be made based on the presence of a phylogenetic
marker gene, such as the 16S rRNA gene (Schleper et al. 1997, Beja et al. 2002), on
the same DNA fragment. Moreover, an in silico assignment of large fragments to a
reference sequence carrying a phylogenetic marker can be attempted, e.g., by comparing nucleotide frequencies (Krüger et al. 2003, Teeling et al. 2004). Large insert
libraries can give access to complete operons (Krüger et al. 2003) or genomic islands
(Schübbe et al. 2003, Mußmann et al. 2005), along with a lower risk of chimera formation and a reasonable sequencing effort. Therefore, large insert libraries were
favoured in early metagenomic studies. Improved single-cell and sequencing technologies along with a decrease of sequencing costs per base pair have changed the
focus within the past few years. It needs to be seen whether large insert repositories
of metagenomes will be dispensable in future metagenome analyses.
45
2.3.5 Construction and Analysis of Metagenomic Libraries
2.3.5.1 Small Insert Metagenomic Libraries
Small insert metagenomic libraries (generally 1.5–3 kbp insert size) in plasmid or
bacteriophage λ derived vectors are a repository of the metagenome, which can be
screened by sequence-based and, more often, function-based assays. Early studies
of such small insert libraries (Cottrell et al. 1999, Henne et al. 1999) focused on
the analysis of single metabolic genes of uncultured microorganisms. Small insert
libraries in reporter gene constructs proved to be valuable to identify novel catabolic
operons in a substrate-induced gene expression screening (Uchiyama et al. 2005).
In a variety of studies small insert libraries have been constructed as a prerequisite
for Sanger-based metagenome sequencing. High throughput sequencing of small
insert libraries has proved to be a powerful method for the assessment of genome
drafts (e.g. Tyson et al. 2004, Venter et al. 2004, Woyke et al. 2006), and for the
study of the phylogenetic and metabolic diversity in complex habitats (Venter et al.
2004, Rusch et al. 2007). The construction of small insert libraries is straightforward. Nevertheless, there are several major drawbacks when small insert libraries
are used in sequence-based metagenomic studies. First of all, a phylogenetic marker,
which allows the reliable assignment of a fragment to a certain phylogenetic group,
is generally missing on small inserts. If insufficient sequence data is available or
if microbial diversity is high within the sample, as is the case for soil samples
(Handelsman et al. 2002, Tringe et al. 2005), the reconstruction of draft genomes
or even the assignment of genetic capabilities to certain species is almost impossible. Moreover, the in silico assembly of larger fragments involves a risk of creating
chimeras, especially when the overlaps of the reassembled fragments are too short
or repetitive elements are present.
2.3.5.2 Large Insert Metagenomic Libraries
Metagenomic libraries with large DNA inserts of > 30 kbp (large insert libraries)
have the advantage that genome fragments can often be assigned to a specific
species. This assignment can be made based on the presence of a phylogenetic
marker gene, such as the 16S rRNA gene (Schleper et al. 1997, Beja et al. 2002), on
the same DNA fragment. Moreover, an in silico assignment of large fragments to a
reference sequence carrying a phylogenetic marker can be attempted, e.g., by comparing nucleotide frequencies (Krüger et al. 2003, Teeling et al. 2004). Large insert
libraries can give access to complete operons (Krüger et al. 2003) or genomic islands
(Schübbe et al. 2003, Mußmann et al. 2005), along with a lower risk of chimera formation and a reasonable sequencing effort. Therefore, large insert libraries were
favoured in early metagenomic studies. Improved single-cell and sequencing technologies along with a decrease of sequencing costs per base pair have changed the
focus within the past few years. It needs to be seen whether large insert repositories
of metagenomes will be dispensable in future metagenome analyses.