Two key genomic regions harbour QTLs for salinity tolerance in ICCV 2?×?JG 11 derived chickpea (Cicer arietinum L.) recombinant inbred lines
The detailed results obtained from the unbalanced analysis of variance (ANOVA) for
the phenotyping data, such as mean performance of parental lines, range of trait values
(i.e., maximum and minimum mean values for each trait) across RILs, broad sense heritability
values (H2), F probability values and least significant difference (LSD) of traits across two years
and treatments, are provided in Tables 1 and 2.
Table 1. ANOVA results for the parameters evaluated under control and salinity treatments in
2010
Table 2. ANOVA results for the parameters evaluated under control and salinity treatments in
2011
Variance analysis
In both years and treatments the RILs but not the parents showed significant variation
for DF (days to first flower) and DM (days to maturity). Parents showed variation
for DM in the salinity treatment in both the years. In 2010 with the control treatment,
no significant variation was observed between the two parents for all the yield and
yield-related traits whereas in the salinity treatment they differed significantly
except for the stem?+?leaf dry weight and the harvest index (HI) (Table 1). In 2011, both the control and salinity treatments did not differentiate the parents
for any traits except for filled pod number and empty pod number in the control treatment
(Table 2).
The combined unbalanced ANOVA on two years data, for both of the treatments revealed
that the traits DF, DM and 100-seed weight were significantly influenced by both genotype
and environment, but largely affected by the genetic potential rather than the environment
(larger F statistic value for the genotype than for the genotype?×?year component of the variance).
All the other traits were influenced significantly by the genotype, but not by the
environment component (Additional file 3: Table S3).
Heritability
Heritability estimates were categorized into low (5-10%), medium (10-30%), high (30-60%)
and very high (60%) according to a previous report 17]. In 2010 in the control treatment, the broad-sense heritability (H2) of DF, DM, HI and 100-seed weight was high, whereas all other yield and yield-related
traits had medium heritability (Table 2). In the salinity treatment, the heritability of DF, DM, 100-seed weight, stem?+?leaf
weight was high, whereas heritability of ADM (above ground dry matter), yield, pod
number, seed number and HI had medium heritability values. In 2011, in the control
treatment, the traits DF, DM and 100-seed weight had high heritability values, whereas
all other traits had medium heritability values (Table 2). In salinity treatment, the traits ADM and yield had medium heritability, whereas
all other traits had high to very high heritability values (Table 2). In summary, the phenological traits had high, whereas the yield and yield-related
traits had moderate-to-high, heritability values in the salinity treatment.
Relationships of yield and yield-related traits variables
The seed yield in the salinity treatment correlated significantly to control treatment
in both the years (R2?=?0.23; R2?=?0.21). Similarly, means of all other traits in the salinity treatment significantly
correlated with the control mean of the corresponding trait in both the years (Additional
file 4: Table S4). To understand the importance of the QTLs identified, the mean value of
traits for which QTLs were found was correlated with the mean yield in both the treatments
and across years (Additional file 4: Table S4). Except for DM in the control treatment in 2010 and DF under salinity
in 2011, all the other traits for which QTLs were identified showed significant correlations
with yield. In the salinity treatment, the ADM, pod number, and seed number explained
up to 76%, 75%, and 76% of the variation in yield, respectively. In the control treatment,
the stem?+?leaf weight, filled pod number and seed number explained up to 51%, 56%
and 49% variations in yield. Although the HI and the 100-seed weight were significantly
correlated to seed yield they explained less than 12% of the yield variation in both
treatments [Table 3].
Table 3. Relationship between the traits for which QTLs were identified and yield
As all the traits showed significant correlations between the control and salinity
treatments, indicating that the value of traits in the salinity treatment were influenced
by the potential value in the control treatment, the traits were expressed as relative
values, calculated as the ratio of values in salinity treatment to the mean value
of the trait in the control treatment for each RIL. In 2010 and 2011, the relative
ADM (R2?=?0.86, R2?=?0.76), relative stem?+?leaf weight (R2?=?0.52, R2?=?0.27), relative pod number (R2?=?0.85, R2?=?0.64 and relative seed number (R2?=?0.89, R2?=?0.89) showed significant correlations with relative yield. This indicated that
these traits were important in determining higher yield under salinity in chickpea.
By contrast the relative values of phenological traits, 100-seed weight and HI were
not significantly related to the relative seed yield (Additional file 5: Table S5).
Genetic linkage map and marker correspondence
The intra-specific genetic map developed based on ICCV 2?×?JG 11 spanned 329.6 cM
with 56 markers mapped in 7 out of 8 linkage groups. No markers were mapped on CaLG02.
The number of markers mapped per linkage group varied from 2 to 11. On an average
one marker/ 5.88Â cM were mapped in the present study. The linkage group wise marker
correspondence was established between the genetic map constructed in the present
study and previously published genetic maps using CMap (Supplementary figure 2 to
10; http://cmap.icrisat.ac.in/cgi-bin/cmap_public/saved_links?selected_link_group=Pushpavalliaction=saved_links_viewerdata_source=CMAP_PUBLIC). There were no common markers between current study and 15],16], but all the three studies had common markers with other published maps that were
summarised in Table 4.
Table 4. Linkage group correspondence in three studies to published maps
QTLs for salinity tolerance
The genotyping and phenotyping data were analysed for identification of major and
minor QTLs to understand the genetic basis of salinity tolerance. In the mapping population
derived from ICCV 2?×?JG 11, a total of 46 QTLs were identified that included 19 QTLs
for phenological traits (7 QTLs for DF; 12 QTLs for DM) and 27 QTLs for yield and
yield-related traits across years and treatments. The QTL analysis for seven (2010)
and nine (2011) yield and yield-related traits detected 23 major QTLs across treatments
for all traits (3 QTLs for ADM; 1 QTL for seed number; 1 QTL for pod number; 3 QTLs
for yield; 2 QTLs for stem?+?leaf weight; 9 QTLs for HI; 4 for 100-seed weight) except
for filled pod number and empty pod number (Additional file 6: Table S6). In the salinity treatment a few minor QTLs were identified for HI on
CaLG04d in 2010, while in the control treatment minor QTLs were identified for yield,
pod number, filled pod number and seed number on CaLG07 in 2011.
In case when one of the flanking markers was common to more than one QTL, that region
was considered as a single genomic region that contained two or more QTLs. By following
this criterion, the 46 QTLs identified were present in 9 genomic regions (Additional
file 11: Figure S1). QTLs that contributed 10% of the phenotypic variation explained (PVE)
were considered as major QTLs. The PVE by QTLs, in this study, ranged from 6 to 67%.
If in a particular treatment, the QTL for a given trait appeared in the same genomic
region in more than one year, the QTL was considered as stable QTL 14]. A total of 14 stable QTLs for five different traits in control treatment were identified
(Additional file 11: Figure S1).
QTLs for phenological traits
In 2010, for DF neither in control nor in the salinity treatment major QTL was identified
but in 2011, six major QTLs (3 QTLs in the control and 3 QTLs in the salinity treatment),
for DF were identified and explained up to 40% of the PVE. In 2010 no major QTL for
DM in the salinity treatment was identified but 4 major QTLs (up to 67% PVE) for DF
were identified in the control treatment. In 2011, in the salinity treatment, four
major QTLs were identified for DM (up to 67% PVE) and in the control treatment; three
QTLs (up to 65% PVE) were identified. Four stable QTLs for DM in control treatment
were detected, two each in CaLG05 (with flanking markers CaM0463-ICCM272) and in CaLG08
(CKAM1903-CKAM0343) (Additional file 6: Table S6). In any case, since there was no relationship between phenological development
and yield either in the control or salinity treatments, these QTLs were not considered
important for the primary purpose of this study.
Yield and biomass
Four yield QTLs (three major and one minor QTL), were identified across two years
and treatments. In 2010, in the salinity treatment one major QTL was identified on
CaLG07 and explained 17% of the PVE. In 2011, one major QTL in the salinity treatment
that explained 12% PVE was also identified on CaLG05, while one major QTL (16% PVE)
and one minor QTL (8% PVE) were identified on each of CaLG05 and CaLG07 in the control
treatment. The two major QTLs identified in the control and salinity treatments in
2011 were located at the same position on CaLG05 with flanking markers, CaM0463 and
ICCM272.
In the salinity treatment, one major QTL for ADM that explained 12% PVE was identified
in 2011. In the control treatment, two major QTLs for ADM that explained up to 27%
PVE were identified across years. All the three QTLs for ADM were found at the same
loci of CaLG05 (CaM0463-ICCM272). Thus two stable QTLs for ADM in control treatment
were identified. In the salinity treatment, no QTL for stem?+?leaf weight was identified,
whereas in the control treatment two major and stable QTLs for stem?+?leaf weight
were identified on CaLG05 (CaM0463-ICCM272) across years (Additional file 6: Table S6).
QTLs for pod number, filled pod number and seed number
In the salinity treatment in 2010, one major QTL for pod number (25% PVE) was found
on CaLG07 (CaM2031-CKAM0165) while in the control treatment in 2011, one minor QTL
(8% PVE) was found on CaLG07 (ICCM0034-CaM0906). In the control treatment, one more
minor QTL for filled pod number (8% PVE) was found on CaLG07. Again on CaLG07, in
the salinity treatment in 2010, one major QTL for seed number with 17% PVE and in
the control treatment in 2011, one minor QTL (9% PVE) was identified for seed number.
These QTLs were of great interest since the correlation analysis above also showed
a close relationship between seed and pod number and yield across treatments.
QTLs for harvest index and 100-seed weight
The QTL analysis identified nine QTLs for HI across years and treatments. In 2010,
in the salinity treatment a minor QTL (6% PVE) for HI was identified on CaLG04d while
in the control treatment two major QTLs for HI were identified, one each on CaLG05
(46% PVE) and CaLG08 (10% PVE). In 2011, in the salinity and control treatment, three
major QTLs per treatment for HI explaining PVE of 30-49% and 32 to 56%, one each on
CaLG05, CaLG04d and CaLG08 were identified. Four stable QTLs for HI under control
treatment were identified. Four major QTLs for 100-seed weight, one each per treatment
and per year, were identified on CaLG05. Three of the four QTLs for 100-seed weight
were identified at the same locus of CaLG05 (CaM0463-ICCM272) and explained PVE up
to 40%. Two stable QTLs for 100-seed weight under control treatment were identified.
The fourth QTL was also identified on CaLG05, but at a different position which explained
17% of the PVE. Again, although these QTLs were significant, they had limited importance
for the primary scope of this study since there was only limited or no significant
relationship between 100-seed weight or HI and yield in any of the treatments, especially
under salinity (Additional file 5: Table S5).
Genomic regions harbouring QTLs for salinity tolerance identified
The genomic region of CaLG05 flanked by markers CaM0463 and ICCM272 contained 17 major
QTLs for seven different traits (DF, DM, ADM, stem?+?leaf weight, 100-seed weight,
HI and yield) across treatments (Figure 1). Furthermore, one major QTL for DF, DM, ADM, HI, 100-seed weight and yield in the
salinity treatment was found in this region. Another genomic region, on CaLG07, harboured
seven QTLs, out of which 5 QTLs were identified in the salinity treatment for five
different traits (DF, DM, seed number, pod number and yield), but none of these QTLs
were stable (Figure 2). A genomic region on CaLG08 harboured eight QTLs (6 in the control treatment and
2 in the salinity treatment) for three traits, DF, DM and HI. Out of these three genomic
regions, the genomic regions on CaLG05 and CaLG07 were of greatest interest as they
hold QTLs for traits that were significantly related to yield under salinity (Additional
file 11: Figure S1).
Figure 1. QTLs for seven different traits were identified across years and treatments on CaLG05.
A. Genomic region on CaLG05 that harboured the 17 QTLs for traits that conferred salinity
tolerance in ICCV 2?×?JG 11 population were identified using QTL cartographer. B. CaLG05 in ICCV 2?×?JG 11 population corresponded to LG 5 in Thudi et al. 2011 and
LG7 in Vadez et al. 2012.
Figure 2. QTLs for five different traits were identified across years and treatments on CaLG07.
A. Genomic region on CaLG07 that harboured the 9 QTLs for traits that conferred salinity
tolerance in ICCV 2?×?JG 11 population were identified using QTL cartographer. B. CaLG07 in ICCV 2?×?JG 11 population corresponded to LG 7 in Thudi et al. 2011 and
LG5 in Vadez et al. 2012.
Mining candidate genes in salinity stress responsive genomic regions
The BES-SSRs (CaM0463 and CaM0123) on CaLG05 were mapped on Ca5, chickpea reference
genome, over a 11.7Â Mb (33.1Â Mb and 44.8Â Mb) distance between the markers. Similarly
the BES-SSRs CaM2031 and CaM1942 markers on CaLG07 were mapped on Ca7 over a 12.5Â Mb
(36.3Â Mb and 48.9Â Mb) distance between the markers on the chickpea reference genome.
A total of 1129 and 440 genes were identified on CaLG05 and CaLG07 respectively (Additional
file 7: Table S7). All the identified 1569 genes could be assigned to three functional categories:
(i) molecular function, (ii) cellular component and (iii) biological processes.
Though the total number of genes found on CaLG05 and CaLG07 were 1569, the sum of
genes assigned to different functional categories (2710) was higher. This is because
a given gene may fall in more than one category (Additional file 8: Table S8). In the molecular function category, the highest number of genes fell
into binding (575) followed by catalytic activity (501) whereas in cellular component
category, the highest number of genes fell into cell part (765) followed by membrane
(335). Similarly, in the biological processes category, a maximum number of genes
fell into metabolic process (747) followed by cellular process (727) and biological
regulation (336) (Additional file 7: Table S7).
Based on gene ontology (GO) annotation, from 1569 genes, 48 putative candidate genes
were found to have reported to have a reponse in several plant species to salinity
stress (31 on CaLG05 and 17 on CaLG07). These 48 genes were located in a distance
of 11.1Â Mb (33.6Â Mb to 44.7Â Mb) and 8.2Â Mb (starting at 37.9Â Mb and ending at 46.1Â Mb)
on CaLG05 and CaLG07 respectively.