78
M.S. Clark et al.
to the situation for microsatellites, the development of SNPs from a zero baseline in non-model organisms can be particularly time-consuming and expensive
(Kim and Misra 2007). However, these markers have also benefited from the everexpanding public genome databases (Kim and Misra 2007, Phillips 2007, Hayes
et al. 2007). For example, SNPs were successfully obtained in the Pacific salmon, a
non-model species, by using sequences from two sister taxa, the rainbow trout and
the Atlantic salmon (Smith et al. 2005, Campbell and Narum 2008, respectively; see
also Ryynanen and Primmer 2006 for a review). Also 318 segregating SNPs were
recently isolated from 17,056 EST sequences in cod (Moen et al. 2008) and the
first results based on cDNA library screening at a population level have also been
obtained in marine mollusc species (Tanguy et al. 2008, Faure et al. 2007, 2008).
Identification of SNPs is not solely reliant upon screening EST databases and they
can be generated from scratch. In-depth methodologies for the generation of SNP
markers will not be presented here (reviewed in Kim and Misra (2007) and Hudson
2008). However, one particularly useful method is that of GMPD technology (glycosylase mediated polymorphism detection; O’Leary et al. 2006, Vaughan 2000). This
approach seems to be particularly useful for targeting specific genes with SNPs, and
as such, the results can contrast with patterns of genetic diversity using other marker
types.
As with all marker types, there are disadvantages as well as advantages: Luikart
et al. (2003) highlighted the possible drawbacks associated with the use of SNPs
in population genomics as they are prone to severe ascertainment bias (bias in
estimating population parameters) due to the criteria used for their selection (i.e.
polymorphism) and the way they are usually tested (e.g. few individuals). Indeed,
in Chinook salmon (Oncorhynchus tshawytscha) some ascertainment bias has been
reported for SNPs compared to anonymous microsatellites and allozymes (Smith
et al. 2007). The main point is to be aware of the advantages and limitations of
using a specific technique and construct the experiment and analyses accordingly
(see Section 3.2.1 for further discussion).
3.1.4 DNA Studies: Amplified Fragment Length Polymorphisms
(AFLPs)
AFLP data sets are now relatively easy to produce and reliable. This technique
involves the use of restriction enzymes to cut genomic DNA. Adaptors are then
ligated onto the sticky ends and a sub-set of these amplified using primers designed
to the adaptor and part of the restriction site. The resulting DNA fragments are
separated by size (length of sequence) using polyacrylamide gel electrophoresis or
capillary sequencers. Only two allelic states are counted (presence or absence), and a
band of a specific length represents the presence of such an allele at an AFLP locus.
The AFLP technique is a particularly well-established molecular technique in
plants (Meudt and Clarke 2007). It is relatively easy to produce >100 AFLP markers,
compared to the microsatellite studies in most non-model organisms where often
M.S. Clark et al.
to the situation for microsatellites, the development of SNPs from a zero baseline in non-model organisms can be particularly time-consuming and expensive
(Kim and Misra 2007). However, these markers have also benefited from the everexpanding public genome databases (Kim and Misra 2007, Phillips 2007, Hayes
et al. 2007). For example, SNPs were successfully obtained in the Pacific salmon, a
non-model species, by using sequences from two sister taxa, the rainbow trout and
the Atlantic salmon (Smith et al. 2005, Campbell and Narum 2008, respectively; see
also Ryynanen and Primmer 2006 for a review). Also 318 segregating SNPs were
recently isolated from 17,056 EST sequences in cod (Moen et al. 2008) and the
first results based on cDNA library screening at a population level have also been
obtained in marine mollusc species (Tanguy et al. 2008, Faure et al. 2007, 2008).
Identification of SNPs is not solely reliant upon screening EST databases and they
can be generated from scratch. In-depth methodologies for the generation of SNP
markers will not be presented here (reviewed in Kim and Misra (2007) and Hudson
2008). However, one particularly useful method is that of GMPD technology (glycosylase mediated polymorphism detection; O’Leary et al. 2006, Vaughan 2000). This
approach seems to be particularly useful for targeting specific genes with SNPs, and
as such, the results can contrast with patterns of genetic diversity using other marker
types.
As with all marker types, there are disadvantages as well as advantages: Luikart
et al. (2003) highlighted the possible drawbacks associated with the use of SNPs
in population genomics as they are prone to severe ascertainment bias (bias in
estimating population parameters) due to the criteria used for their selection (i.e.
polymorphism) and the way they are usually tested (e.g. few individuals). Indeed,
in Chinook salmon (Oncorhynchus tshawytscha) some ascertainment bias has been
reported for SNPs compared to anonymous microsatellites and allozymes (Smith
et al. 2007). The main point is to be aware of the advantages and limitations of
using a specific technique and construct the experiment and analyses accordingly
(see Section 3.2.1 for further discussion).
3.1.4 DNA Studies: Amplified Fragment Length Polymorphisms
(AFLPs)
AFLP data sets are now relatively easy to produce and reliable. This technique
involves the use of restriction enzymes to cut genomic DNA. Adaptors are then
ligated onto the sticky ends and a sub-set of these amplified using primers designed
to the adaptor and part of the restriction site. The resulting DNA fragments are
separated by size (length of sequence) using polyacrylamide gel electrophoresis or
capillary sequencers. Only two allelic states are counted (presence or absence), and a
band of a specific length represents the presence of such an allele at an AFLP locus.
The AFLP technique is a particularly well-established molecular technique in
plants (Meudt and Clarke 2007). It is relatively easy to produce >100 AFLP markers,
compared to the microsatellite studies in most non-model organisms where often