Supplementary figure 1' a Phylogeny of the wild species in series Sativae of Oryza A genome with inf

1
a
b
Supplementary figure 1. (a) Phylogeny of the wild
species in series Sativae of Oryza (A genome)
with inferred character evolution. Numbers
associated with branches are bootstrap support
taken from athe phylogeny inferred based on a
combined analysis (Fig. 3 of Zhu and Ge 2005) and
bthe phylogeny inferred from the combined
sequences of three nuclear genes, Adh1, Adh2, and
RgRc2 (Sang Lab unpublished). Geographic
distribution, life history, and photoperiod
sensitivity of the species are indicated. Based
on character variation in the entire genus (the
rest of the genus not shown Oka and Chang 1960
Oka 1988 Vaughan 1989, 1994 Sang, pers. obs.),
it is more parsimonious to conclude that the
common ancestor of the series Sativae as well as
that of O. nivara and O. rufipogon were perennial
and photoperiod sensitive. The annual habit and
photoperiod insensitivity of O. nivara are
derived states. (b) Schematic drawing of
evolution of O. nivara and the indica cultivar.
Blue background associated with the O.
rufipogon-like ancestor and O. rufipogon
indicates persistently wet habitat, brown
background associate with O. nivara indicates
seasonally dry habitat. Oryza nivara may have
served as the primary genome donor for the indica
cultivar while introgression from O. rufipogon is
like to have occurred during rice domestication
(Li et al. 2007b Sang and Ge 2007a, b).
2
a
b
PN
PR
F1A
F1B
F2A
F2B
Locus 1
Locus 2
Supplementary figure 2. Sckematic illustration of
cross design and genotyping. (a) Cross design.
An individual of O. nivara (PN) was crossed with
an individual of O. rufipogon (PR). Each
successful hand-pollination of a recipient flower
(spikelet) produced a single seed of a hybrid,
here an F1 seed. From four F1s screened for
genetic variation, two individuals, F1A and F1B,
with the highest divergence of PR alleles were
selected for the development of F2 mapping
populations. Each of the two F1 plants was
allowed for self-pollination and production of
several hundred F2 seeds. 192 individuals of
each of the two F2 populations, F2A and F2B, was
used as a mapping population. (b) Genotyping
with SSR markers. PN was homozygous at all SSR
loci examined, whereas PR was heterozygous at 59
of the loci, for examples, at loci 1 and 2, where
alleles of O. nivara and O. rufipogon are shown
in red and blue, respectively. At locus 1, the
same O. rufipogon allele was passed to both F1
individuals at locus 2, different O. rufipogon
alleles were passed to different F1s. As a
consequence, the two F2 populations segregated
different O. rufipogon alleles at locus 2. In
the separate QTL analysis of each of the two F2
populations using program QTL Cartographer,
homozygous O. nivara and O. rufipogon genotypes
were score as 2 and 0, respectively, and the
heterozygous genotype was scored as 1. In the
combined analysis of both populations, the
genotypes were scored in the same way by not
distinguishing heterozygous O. rufipogon alleles
(see text for explanation). Note For QTL
analyses involving outcrossing plant species with
relatively high levels of heterozygosity using
co-dominant markers such as SSR, two cross
designs are commonly adopted. One is to
self-pollinate the outcrossing parent (if it is
self-compatible) for at least five generations so
that it becomes almost homozygous across the
genome. Analysis of F2 populations derived from
such parents avoids complications caused by the
allelic variation. The drawback of this design,
however, is the inability to evaluate the
influence of allelic variation on QTL detection
for natural variations. Additionally, it may
take many years to complete the process of
self-pollination especially for plants with
relatively long life cycles (It would take us
four years to complete the self-pollination of
the O. rufipogon parent if we were to do so).
The second design is to conduct cross pollination
to generate F1s, and then generate mapping
populations by either backcrossing the F1s to the
parents or crossing between the F1s. This
approach requires that each cross yields a large
number of seeds for developing the mapping
population. However, this is inapplicable to the
rice species where one successful
hand-pollination produces one seed. Given these
reasons as well as the fact that one of the
mapping parents, O. nivara, is self-pollinated,
we came up with another cross design as
illustrated above. The comparison between the
separate analyses of two F2 populations allowed
us to identify QTL of relative large effect with
higher confidence and to evaluate the impact of
allelic variation on QTL detection. The combined
analysis further provided an opportunity to
evaluate the power and accuracy of QTL detection
with a larger mapping population.
3
Supplementary figure 3. Variation in panicle
shape and exsertion between parents (A, O.
rufipogon E, O. nivara) and selected F2
individuals (B, C, D). Panicle shape was
measured by the angles between primary branches
and rachis (C). Panicle exsertion was measured
by the distance between the flag leaf node (FLN)
and the first primary branch node (PBN). O.
rufipogon has open and exserted panicles, whereas
O. nivara has compact and partially inserted
panicles. The traits segregated in the F2
population, with individuals having open but
partially enveloped panicles (B) and individuals
having compact but exserted panicles (D).
4
55-bp insertion
Zinc finger
Exon 2
Exon 1
100 bp
Supplementary figure 4. Sckematic drawing of rice
Hd1 gene. The positions of the zinc finger
domain and the 55-bp insertion found in the O.
nivara mapping parent are indicated.
5
Supplementary figure 5. Distribution of magnitude
of effect of QTL identified in two F2 mapping
populations. Black bar, population A Gray bar,
population B.