Biology Reference
In-Depth Information
identified. The identification of the actual muta-
tion, and therefore gene, responsible for a QTL
requires a level of marker density well beyond
that typically employed by even association map-
ping procedures, but is still nevertheless well
within reach of the data produced by, for exam-
ple, the Illumina system. Combined with appro-
priate data mining and candidate gene identifi-
cation, it even becomes possible to efficiently
process quite large QTL intervals.
As a case in point, it has been mentioned that
resequencing data from five
O. sativa
genotypes
have been obtained in an attempt to identify the
causal mutations underlying QTLs from vari-
ous populations derived from these genotypes.
One significant QTL identified from a number
of mapping populations is located on the long
arm of chromosome 3 (Figure 3.1b). This has
been identified in populations derived from the
crosses Capsule (tolerant)
as potential candidate genes for the QTL (our
unpublished data).
Clearly, it is not practical to examine all of
these candidates via traditional cloning tech-
niques. Cloning of the most likely candidates
(based on annotated function) revealed those
that showed little or no polymorphism between
IR29 and/or Azucena and the tolerant donors.
Thus, a very long list of secondary candidates
remained. Using expression data obtained from
Affymetrix arrays (Walia et al. 2005, 2007; Cot-
saftis et al. 2011), it was possible to narrow this
list to just 30 to 40 plausible candidates, but this
number is still large to validate via traditional
approaches. However, combined with polymor-
phism data derived from whole-genome rese-
quencing, only three showed patterns of poly-
morphism fully consistent with the pattern of
presumed donor/recipient combinations, located
at 27, 29, and 35 Mb (Figure 3.1b). These posi-
tions clearly coincide well with the noted QTL
peaks, although this does not help in determin-
ing whether one or two QTLs are present. But
this number of candidate genes is quite feasi-
ble for traditional cloning and transgenic valida-
tion approaches. Thus, next-generation sequenc-
ing data make it possible to search through even
quite large and complex QTL regions in an effi-
cient manner and successfully identify good can-
didate genes for further validation.
As an adjunct to the identification of candi-
date genes, a very large number of SNP polymor-
phisms were identified within the QTL region.
Once these SNPs and positions are known, it
becomes possible to filter these and select a set
of high-quality, robust SNPs for further fine-
mapping of the QTL. As an example, SNP data
can be post-processed in MS Access to (A)
retrieve SNPs within genic (or coding) regions,
(B) filter for SNPs that are present in donor lines
but NOT present in any recipient line, and (C)
retrieve SNPs with few or no polymorphisms in
surrounding areas (e.g., a 500-bp window cen-
tered on the SNP site) and that are therefore likely
to produce good markers. Filtering for SNPs that
are present in most or all donor lines but not
×
BRRI dhan29 (sen-
sitive), FL478
×
Azucena, FL478
×
IR29, and
×
Pokkali
IR29. In all cases, the tolerant par-
ent is likely to be the donor of the tolerant
allele, so that FL478, Pokkali, and Capsule are
likely donors, and IR29, Azucena, and BRRI
dhan29 are recipients. The QTL was also iden-
tified in a population derived from a Nippon-
bare
Kasalath cross, in which Nipponbare was
the donor (Takehisa et al. 2004). Traits mapped
to this location are quite diverse, but typically
include parameters related to biomass accumula-
tion and injury scores, together with some effects
on Na
+
and/or K
+
contents. Comparison of QTL
positions mapped onto the Nipponbare reference
genome shows that the QTL interval spans from
approximately 26 Mb down to the end of the
chromosome, with a peak between 29 and 31 Mb
depending on the population (it is also possible
that two QTLs exist in coupling, with ranges of
25 - 30 Mb and 29 Mb to the end, and peaks of
27 Mb and 33 Mb, respectively). Thus, the inclu-
sive QTL interval spans around 10 Mb of the
reference genome. Within this region, 1729 non-
redundant gene models are found within either
(or both) of the MSU v. 6.1 and RAP 4 anno-
tations, of which 424 (
×
∼
25%) can be classified
Search WWH ::
Custom Search