SNPs are found significantly more frequently in group of genotypes (accessions) with a certain phenotype (trait) than in the
general population, the mutations are said to be “associated”
with the trait. The GWAS analysis, represented in a Manhattan
plot with significance (Àlog10 (P value) on the y-axis, and
genomic position shown as chromosomes in the x-axis. This
approach has several major advantages over conventional QTL
mapping. First, a much larger and more representative gene
pool can be surveyed. Second, it consumes less time and
expense than mapping studies and enables the mapping of
many traits in one set of genotypes. Third, a much finer
mapping resolution can be achieved, resulting in small confidence intervals of the detected loci compared to classical
mapping, where the identified loci need to be fine-mapped.
Finally, it has the potential not only to identify and map QTLs
but also to identify the causal polymorphism within a gene that
is responsible for the difference in two alternative phenotypes
[39]. A major issue with association studies is, however, a high
rate of false positives, with the main source of these being the
linkage between causal and non-causal sites and the confounding effects of population structure [40, 41]. A high rate of false
negatives, wherein loci with previous experimental validation
for specific traits are not found in GWAS, may also appear; this
is due to epistasis and lack of natural variation [34, 35]. A
quantitative genetics approach combined with metabolomics,
whereupon metabolite levels are regarded as traits, can help
unravel the genetic architecture of metabolic networks, identifying enzymes and regulatory genes that take part in specific
metabolic pathways. Network analysis based on multiple omics
data, as part of a top-down, complexity-reduction approach,
can lead to original hypothesis generation about metabolic
pathway regulation. Based on first principles, the integration
of these two orthogonal approaches can accelerate the discovery of novel genes involved in plant metabolism with increased
statistical confidence.
12. Cross-validation using RIL mapping: Immortal mapping
populations consisting of homozygous individual have also
been much used to map loci for complex metabolic traits
(cite references). Recombinant inbred lines (RILs) can be
obtained relatively easily and are produced by successively selfing the progeny of individual F2 plants (single seed descent
method), from which the F8 generation and onward are practically homozygous lines that will produce further progeny that
is essentially identical to the previous generation. Such a population can also be produced by induced chromosomal doubling
of haploids, such as for doubled haploids (DHs; [28]). However RILs are advantageous over DHs given they are normally
406
Si Wu et al.
general population, the mutations are said to be “associated”
with the trait. The GWAS analysis, represented in a Manhattan
plot with significance (Àlog10 (P value) on the y-axis, and
genomic position shown as chromosomes in the x-axis. This
approach has several major advantages over conventional QTL
mapping. First, a much larger and more representative gene
pool can be surveyed. Second, it consumes less time and
expense than mapping studies and enables the mapping of
many traits in one set of genotypes. Third, a much finer
mapping resolution can be achieved, resulting in small confidence intervals of the detected loci compared to classical
mapping, where the identified loci need to be fine-mapped.
Finally, it has the potential not only to identify and map QTLs
but also to identify the causal polymorphism within a gene that
is responsible for the difference in two alternative phenotypes
[39]. A major issue with association studies is, however, a high
rate of false positives, with the main source of these being the
linkage between causal and non-causal sites and the confounding effects of population structure [40, 41]. A high rate of false
negatives, wherein loci with previous experimental validation
for specific traits are not found in GWAS, may also appear; this
is due to epistasis and lack of natural variation [34, 35]. A
quantitative genetics approach combined with metabolomics,
whereupon metabolite levels are regarded as traits, can help
unravel the genetic architecture of metabolic networks, identifying enzymes and regulatory genes that take part in specific
metabolic pathways. Network analysis based on multiple omics
data, as part of a top-down, complexity-reduction approach,
can lead to original hypothesis generation about metabolic
pathway regulation. Based on first principles, the integration
of these two orthogonal approaches can accelerate the discovery of novel genes involved in plant metabolism with increased
statistical confidence.
12. Cross-validation using RIL mapping: Immortal mapping
populations consisting of homozygous individual have also
been much used to map loci for complex metabolic traits
(cite references). Recombinant inbred lines (RILs) can be
obtained relatively easily and are produced by successively selfing the progeny of individual F2 plants (single seed descent
method), from which the F8 generation and onward are practically homozygous lines that will produce further progeny that
is essentially identical to the previous generation. Such a population can also be produced by induced chromosomal doubling
of haploids, such as for doubled haploids (DHs; [28]). However RILs are advantageous over DHs given they are normally
406
Si Wu et al.