80
M.S. Clark et al.
between alleles and loci across individuals. This leads to a form of homoplasy (similarity due to convergent evolution), a phenomenon that can be encountered for any
genetic marker (e.g. Ellegren 2004). For instance, Wares and Blakeslee (2007) found
this issue was the major drawback to their analysis of the invasive status of the
periwinkle Littorina littorea along northern American coasts. As AFLP markers
are dominant markers, they do not really comply with detection methods of outlier
loci developed so far. Such methods (reviewed in Guinand et al. 2004) rely on the
estimation of allelic frequencies assuming Hardy-Weinberg proportions. Dominant
markers do not allow verification of this assumption, but, as noted by Bonin et al.
(2006), locus-specific deviations from the Hardy-Weinberg equilibrium may arise
and therefore bias the results of outlier detection in an unpredictable way. Finally, a
major drawback of AFLP markers is that they are anonymous markers in non-coding
regions. As they are not identified per se and not easily linked to genes it is difficult to use them to fill in the gaps between phenotype and genotype and therefore
between observed diversity and fitness. AFLPs are best used to improve knowledge
in some ecological areas such as better stock or species delineation and the topics mentioned in Table 1. However they do not particularly allow links be made
with regard to individual fitness, as investigating this issue with any type of genetic
marker is controversial (Balloux et al. 2004, Pemberton 2004, Slate et al. 2004,
DeWoody and DeWoody 2005). The AFLP technique is now being combined with
new technologies based on pyrosequencing of AFLP fragments to develop SNPs
and microsatellites.
3.1.5 DNA Studies: High Through-Put Sequencing
Sanger sequencing has become a routine laboratory technique, but recent major
advances with the next-generation of sequencing technologies such as massively
parallel sequencing signature (MPSS; Brenner et al. 2000) and pyrosequencing
(reviewed in Margulies et al. 2005, Langaee and Ronaghi 2005) (also called 454
sequencing) are revolutionising this technique, allowing simultaneous processing
of millions of short sequence reads. Although challenging from a bioinformatic
perspective, such technologies offer several opportunities for addressing ecology
and evolution issues, among them the possibility to carry out biodiversity analysis
(Venter et al. 2004). Moreover, such methods are less prone to errors due to mishandling, recovery of missing or rare transcripts or clones that are unstable when
cloned into bacteria. Technical improvements make such methods increasingly reliable (e.g. Hamady et al. 2008), and expression of most transcripts, including rare
variants, can be accurately and precisely quantified (e.g. Stolovitzky et al. 2005).
This methodology will become increasingly used as the length of the sequence read
increases (450 bp reads are now possible with 454 pyrosequencing), increasing the
chance of identifying genes in non-model species (Hudson 2008). The initial technology was restricted to tens of bases which is really only of use in sequenced model
organisms (Hudson 2008).
M.S. Clark et al.
between alleles and loci across individuals. This leads to a form of homoplasy (similarity due to convergent evolution), a phenomenon that can be encountered for any
genetic marker (e.g. Ellegren 2004). For instance, Wares and Blakeslee (2007) found
this issue was the major drawback to their analysis of the invasive status of the
periwinkle Littorina littorea along northern American coasts. As AFLP markers
are dominant markers, they do not really comply with detection methods of outlier
loci developed so far. Such methods (reviewed in Guinand et al. 2004) rely on the
estimation of allelic frequencies assuming Hardy-Weinberg proportions. Dominant
markers do not allow verification of this assumption, but, as noted by Bonin et al.
(2006), locus-specific deviations from the Hardy-Weinberg equilibrium may arise
and therefore bias the results of outlier detection in an unpredictable way. Finally, a
major drawback of AFLP markers is that they are anonymous markers in non-coding
regions. As they are not identified per se and not easily linked to genes it is difficult to use them to fill in the gaps between phenotype and genotype and therefore
between observed diversity and fitness. AFLPs are best used to improve knowledge
in some ecological areas such as better stock or species delineation and the topics mentioned in Table 1. However they do not particularly allow links be made
with regard to individual fitness, as investigating this issue with any type of genetic
marker is controversial (Balloux et al. 2004, Pemberton 2004, Slate et al. 2004,
DeWoody and DeWoody 2005). The AFLP technique is now being combined with
new technologies based on pyrosequencing of AFLP fragments to develop SNPs
and microsatellites.
3.1.5 DNA Studies: High Through-Put Sequencing
Sanger sequencing has become a routine laboratory technique, but recent major
advances with the next-generation of sequencing technologies such as massively
parallel sequencing signature (MPSS; Brenner et al. 2000) and pyrosequencing
(reviewed in Margulies et al. 2005, Langaee and Ronaghi 2005) (also called 454
sequencing) are revolutionising this technique, allowing simultaneous processing
of millions of short sequence reads. Although challenging from a bioinformatic
perspective, such technologies offer several opportunities for addressing ecology
and evolution issues, among them the possibility to carry out biodiversity analysis
(Venter et al. 2004). Moreover, such methods are less prone to errors due to mishandling, recovery of missing or rare transcripts or clones that are unstable when
cloned into bacteria. Technical improvements make such methods increasingly reliable (e.g. Hamady et al. 2008), and expression of most transcripts, including rare
variants, can be accurately and precisely quantified (e.g. Stolovitzky et al. 2005).
This methodology will become increasingly used as the length of the sequence read
increases (450 bp reads are now possible with 454 pyrosequencing), increasing the
chance of identifying genes in non-model species (Hudson 2008). The initial technology was restricted to tens of bases which is really only of use in sequenced model
organisms (Hudson 2008).