Basic HTML Version
Triticeae Genomics and Genetics 2012, Vol.3, No.2, 9
-
24
http://tgg.sophiapublisher.com
18
affected by the cumulative effect of QTLs for PHST
and its associated traits GC and dormancy. Therefore,
a breeder will have to select appropriate MQTLs for
MAS to improve PHST. For this purpose, a breeder
may select one or more MQTLs (for PHST with or
without one or more other associated trait), which
resulted from a large number of original QTLs and has
a narrow CI. Keeping this in view, all 8 MQTLs
(excluding MQTL3 for seed dormancy) are interesting
and markers associated with these MQTL may be used
for marker-aided introgression of independent QTL
and MQTLs in any adaptive genetic background for
development of PHS tolerant wheat cultivars.
3 Material and Methods
Following four steps were involved in meta-QTL
analysis. First, bibliographic review was done to
collect data related to the mapped QTLs; second, a
consensus map of individual chromosomes based on
all the markers available in the framework maps used
in earlier studies was constructed and the reported
QTLs were projected on to these consensus maps;
third, an overview was computed to identify the
genomic regions carrying all the QTLs in an
individual experiment and fourth meta-QTL analysis
was conducted to identify true and reliable QTLs
based on QTLs reported earlier. Methods involved in
these four individual steps are described.
3.1 Bibliographic review and selection of QTL
studies for meta-QTL analysis
During bibliographic review, information on genetic
maps and details about QTL (including QTL positions,
CIs,
R
2
and LOD values) were collected from 24
independent studies involving QTL analysis for PHST,
dormancy and the associated trait grain colour. We
know that CIs and
R
2
values for all the QTLs are
needed for meta-QTL analysis, but in some studies,
CIs for individual QTLs are not available. For this
purpose, and to have uniformity in data used for
meta-QTL analysis, a 5% confidence interval (CI) was
worked out for each reported QTL using the following
formula (Darvasi and Soller, 1997): C.I.=530/(N×
R
2
).
Where 530 is a constant value obtained from
simulations, N is the size of population and
R
2
is the
proportion of phenotypic variance explained (PVE) by
the QTL.
3.2 Construction of consensus maps and QTL
projection
In earlier studies involving QTL analysis for PHST/
dormancy in bread wheat, a number of framework
genetic maps (each with 4 to 25 markers) was utilized,
one each for an individual mapping population that
was used for QTL interval mapping in a particular
study. However, the mapped markers on these different
maps differ, and not many markers are common to all
maps. Also raw data for marker genotyping is seldom
available through literature or databases. Therefore,
two relatively dense reference maps published earlier
and the available framework maps were used for
developing a consensus map, which carried maximum
number of markers from each of the framework map.
Both the reference maps used in the present study
were developed in 2004, one belonging to Somers et
al. (2004) and the other being the wheat composite
map (2004), available at GrainGene 2.0. Together, the
two maps had a much larger number of markers than
an individual reference map. Therefore these two
maps were first integrated to provide a pre-consensus
map, which was then used for developing a consensus
map, where it was possible to project all the QTLs
reported in 15 earlier studies (for details see Table 1)
to facilitate meta-QTL analysis.
Computation for developing a consensus map was
performed using BioMercator 2.1 (www.genoplante.
org) applying a weighted least square method (Arcade
et al., 2004). This allowed arrangement of markers in
a linear order, and positioning of these markers on the
consensus map. BioMercator 2.1 facilitates projection
of the mapped markers included in the framework
maps and also the corresponding reported QTLs on to
the pre-consensus map prepared as above, so that the
loci which were present in a framework map and
absent in the pre-consensus map were added, keeping
in view the linear order of loci that were common
between a framework map and the pre-consensus map.
In this process, common loci that occurred in an
inverted order with respect to each other are discarded.
For each interval flanked by common markers, a
specific distance ratio was computed (using interval
lengths in the framework map and pre-consensus map)
Triticeae Genomics and Genetics Provisional Publishing