Seedling trait performance of RILs and their parents
The difference in seedling traits between the two parental cultivars (‘9311’ and ‘NIP’) was largely affected by seedling density (Additional file 1: Table S2 and Table 1). The two parents differed significantly at P ≤ 0.05 for all the six traits under three seeding densities. However, the degree of difference decreased when SLL and RDW were measured as the increasing of seeding density, while the degree of difference between the parents for SH, FLSL, FLL and SDW increased with increasing density. These results indicated the effect of seeding density on seedling traits was not consistent, thus suggesting significant differences of genetic mechanisms for seedling traits between inter-subspecies. In the RIL population, all the seedling traits showed continuous variation and obvious transgressive segregation, following approximate normal distributions. With the increase of seeding density, the average values of RIL population for SH, FLSL, FLL and SLL gradually increased, while the RDW and SDW gradually decreased (Figure 2). The broad-sense heritabilities ranged from 71.37% to 90.91%, indicating the complexity of the genotypic response to seeding density. The heritability of FLL and SLL were moderate while it was low for RDW and SDW. There was no significant difference in heritability between different densities. The G×E interactions were highly significant (P ≤ 0.01) among the three seeding densities, suggesting the effect of seeding density on seedling traits should not be ignored (Table 1). The correlations among different seedling traits showing similar trends between different seeding densities (Additional file 1: Table S3). FLL and SLL showed the highest correlations in both experiments. RDW was negatively correlated with other traits, but it was positively correlated with SDW. Taken together, these results indicated that seedling quality of the parents and RIL population on plastics nursery tray generally declined with the increase of seeding density, and the influences of seeding density on indica and japonica varied on traits. However, little was known about the genetic loci controlling seedling traits that were affected by seeding density, and the distributions of loci that differ in response to seeding density between indica and japonica rice genomes.
Table 1. Phenotypic variation of the RIL population of 9311/Nipponbare cross under three seeding densities.
Traita
|
Seeding densityb
|
Ex1
|
Ex2
|
Ex3
|
He
|
G×Ef
|
Range
|
Mean±SDc
|
NIP
|
9311d
|
Range
|
Mean±SDc
|
NIP
|
9311d
|
Range
|
Mean±SDc
|
NIP
|
9311d
|
SH
|
LD
|
9.97—27.37
|
18.20±3.53
|
20.16
|
18.2 **
|
11.28—24.58
|
16.59±2.34
|
18.23
|
16.63 **
|
12.11—25.91
|
17.53±2.47
|
19.52
|
16.33 **
|
79.63
|
**
|
MD
|
9.07—33.73
|
19.40±3.91
|
21.17
|
18.53 **
|
11.94—24.73
|
17.57±2.26
|
19.07
|
17.01 **
|
11.76—24.58
|
18.38±2.54
|
21.65
|
18.23 **
|
77.76
|
HD
|
10.48—31.94
|
20.12±3.85
|
22.81
|
19.35 **
|
13.33—24.96
|
18.20±2.33
|
19.68
|
17.70 **
|
12.58—25.64
|
18.99±2.56
|
22.59
|
18.59 **
|
80.81
|
FLSL
|
LD
|
1.86—6.89
|
3.71±0.80
|
3.28
|
3.77 **
|
2.20—5.58
|
3.26±0.55
|
2.72
|
3.30 **
|
2.00—4.99
|
3.44±0.55
|
2.89
|
3.54 **
|
77.88
|
**
|
MD
|
1.71—6.47
|
3.95±0.79
|
3.58
|
4.19 **
|
2.27—5.93
|
3.54±0.63
|
3.13
|
3.60 **
|
2.15—5.36
|
3.67±0.58
|
3.47
|
4.07 **
|
82.92
|
HD
|
2.11—6.63
|
4.17±0.83
|
3.74
|
4.77 **
|
2.42—6.41
|
3.83±0.68
|
3.14
|
3.83 **
|
2.13—5.60
|
3.93±0.65
|
3.84
|
4.39 **
|
80.78
|
FLL
|
LD
|
0.70—5.51
|
2.52±0.90
|
1.42
|
3.23 **
|
0.69—4.91
|
2.62±0.87
|
1.56
|
3.15 **
|
0.68—4.60
|
2.64±0.81
|
1.84
|
2.63 **
|
88.94
|
**
|
MD
|
0.76—5.05
|
2.71±0.87
|
1.58
|
3.84 **
|
0.63—4.86
|
2.78±0.93
|
1.57
|
3.33 **
|
0.93—4.79
|
2.79±0.84
|
1.86
|
3.08 **
|
90.91
|
HD
|
1.00—6.81
|
2.96±1.05
|
1.70
|
3.83 **
|
0.72—5.17
|
2.86±0.94
|
1.67
|
3.34 **
|
0.96—5.26
|
2.93±0.90
|
1.89
|
3.67 **
|
90.15
|
SLL
|
LD
|
2.85—15.39
|
8.49±2.36
|
5.03
|
11.23 **
|
4.23—14.25
|
7.61±1.56
|
6.14
|
10.64 **
|
4.08—14.10
|
8.99±1.85
|
8.13
|
11.00 **
|
84.36
|
**
|
MD
|
3.48—16.93
|
9.33±2.56
|
6.70
|
11.55 **
|
4.02—13.37
|
8.34±1.71
|
6.45
|
10.84 **
|
4.11—15.43
|
9.42±2.06
|
8.56
|
11.17 **
|
84.47
|
HD
|
3.30—17.03
|
9.77±2.56
|
7.00
|
11.81 **
|
3.07—16.96
|
8.89±1.87
|
7.18
|
11.39 **
|
4.31—15.41
|
9.83±2.14
|
8.61
|
11.43 **
|
82.43
|
RDW
|
LD
|
5.2—23.2
|
10.6±3.1
|
9.7
|
13.8 **
|
4.8—18.4
|
9.2±2.3
|
8.7
|
12.3 **
|
3.6—18.4
|
7.8±2.2
|
7.6
|
10.3 **
|
81.95
|
**
|
MD
|
3.0—18.7
|
7.1±2.6
|
5.7
|
8.8 **
|
2.9—13.5
|
6.7±2.1
|
6.1
|
8.1 **
|
3.2—11.8
|
5.5±1.4
|
5.0
|
7.1 **
|
71.37
|
HD
|
2.0—14.8
|
5.5±2.1
|
5.3
|
6.5 **
|
2.6—11.3
|
5.3±1.5
|
5.9
|
7.0 **
|
2.4—8.4
|
4.6±1.1
|
4.6
|
6.0 **
|
78.94
|
SDW
|
LD
|
12.7—41.2
|
26.9±5.6
|
24.9
|
28.7 **
|
13.1—38.4
|
22.4±4.5
|
19.5
|
22.7 **
|
15.1—38.2
|
23.1±4.1
|
22.4
|
24.7 **
|
77.72
|
**
|
MD
|
10.1—37.4
|
20.6±4.4
|
20.7
|
24.4 **
|
10.2—35.3
|
18.8±3.7
|
16.3
|
20.3 **
|
11.65—31.8
|
19.4±3.0
|
19.2
|
23.5 **
|
76.63
|
HD
|
6.9—32.9
|
18.1±4.4
|
17.3
|
21.65 **
|
9.5—26.9
|
16.0±3.1
|
14.8
|
18.9 **
|
10.4—26.1
|
17.6±2.8
|
16.2
|
19.8 **
|
71.46
|
a Trait. SH, seedling height (cm); FLSL, first leaf sheath length (cm); FLL, first leaf length (cm); SLL, second leaf length (cm); RDW, root dry weight (mg); SDW, shoot dry weight (mg). b Seeding density, LD, MD, and HD represented low density, medium density, and high density, respectively. c SD, standard deviation. d T-test between two parents, * and **, significant at P ≤ 0.05 and P ≤ 0.01. e Broad-sense heritability (%). f G × E, interaction of genotype and seeding density, * and **, significant at P ≤ 0.05 and P ≤ 0.01.
QTL analysis of seedling traits under three seeding densities
QTL analysis based on SNP bin map under three seeding densities in three experiments identified 37 QTLs. Four QTLs for SH, 9 QTLs for FLSL, 6 QTLs for FLL, 5 QTLs for SLL, 5 QTLs for RDW and 8 QTLs for SDW were identified above the LOD thresholds under three seeding densities (Additional file 1: Table S4). We further analyzed in detail the QTLs that detected in at least two experiments (Table 2).
Table 2. Comparative analysis of QTLs for seedling traits under three seeding densities.
Chr
|
Position a
|
QTL
|
Traits/
experimentb
|
LD
|
MD
|
HD
|
Addc
|
LOD
|
Vard
|
Addc
|
LOD
|
Vard
|
Addc
|
LOD
|
Vard
|
1
|
bin72
|
qRDW1.1
|
RDW/Ex1
|
-1.02
|
4.73
|
8.25
|
|
|
|
|
|
|
|
|
qRDW1.1
|
RDW/Ex2
|
-0.61
|
3.79
|
6.82
|
-0.48
|
3.45
|
5.08
|
|
|
|
|
|
qRDW1.1
|
RDW/Ex3
|
-0.63
|
7.42
|
7.64
|
-0.4
|
4.79
|
8.09
|
|
|
|
|
bin191
|
qRDW1.2
|
RDW/Ex1
|
|
|
|
|
|
|
-0.47
|
3.28
|
4.7
|
|
|
qRDW1.2
|
RDW/Ex2
|
|
|
|
|
|
|
-0.35
|
4.45
|
5.72
|
|
bin311
|
qSH1.1
|
SH/Ex1
|
0.68
|
6.42
|
7.12
|
1.09
|
9.45
|
7.34
|
1.10
|
8.69
|
7.35
|
|
|
qSH1.1
|
SH/Ex2
|
0.99
|
6.74
|
16.0
|
1.07
|
13.52
|
20.2
|
1.07
|
13.46
|
19.07
|
|
|
qSH1.1
|
SH/Ex3
|
0.96
|
14.70
|
13.46
|
1.13
|
17.5
|
17.91
|
1.12
|
15.53
|
17.24
|
|
|
qFLSL1.1
|
FLSL/Ex1
|
0.16
|
4.79
|
6.75
|
0.11
|
3.82
|
4.12
|
0.13
|
4.13
|
4.75
|
|
|
qFLSL1.1
|
FLSL/Ex2
|
0.18
|
5.22
|
9.74
|
0.23
|
8.53
|
11.73
|
0.25
|
8.91
|
11.99
|
|
|
qFLSL1.1
|
FLSL/Ex3
|
0.18
|
6.27
|
10.68
|
0.19
|
10.31
|
9.85
|
0.2
|
8.23
|
8.52
|
2
|
bin411
|
qFLL2.1
|
FLL/Ex1
|
|
|
|
-0.31
|
4.87
|
12.97
|
-0.35
|
3.84
|
10.86
|
|
|
qFLL2.1
|
FLL/Ex2
|
|
|
|
-0.30
|
4.11
|
10.6
|
|
|
|
|
|
qFLL2.1
|
FLL/Ex3
|
|
|
|
-0.31
|
5.84
|
13.41
|
-0.35
|
4.67
|
14
|
|
|
qSLL2.1
|
SLL/Ex1
|
|
|
|
-0.55
|
3.28
|
9.55
|
-0.56
|
3.98
|
9.73
|
|
|
qSLL2.1
|
SLL/Ex3
|
-0.50
|
3.36
|
7.47
|
-0.58
|
3.4
|
8.01
|
-0.61
|
3.83
|
8.16
|
|
|
qSDW2.1
|
SDW/Ex2
|
1.13
|
4.58
|
6.29
|
0.90
|
3.32
|
5.83
|
0.88
|
3.42
|
7.77
|
|
|
qSDW2.1
|
SDW/Ex3
|
0.73
|
3.66
|
5.11
|
|
|
|
|
|
|
|
bin560
|
qFLL2.2
|
FLL/Ex2
|
-0.29
|
3.13
|
11.48
|
|
|
|
-0.29
|
3.62
|
8.69
|
|
|
qFLL2.2
|
FLL/Ex3
|
-0.28
|
6.54
|
11.82
|
|
|
|
-0.30
|
3.25
|
10.54
|
3
|
bin650
|
RDW3.1
|
RDW/Ex1
|
-0.46
|
3.44
|
5.07
|
|
|
|
|
|
|
|
|
RDW3.1
|
RDW/Ex2
|
-0.67
|
4.16
|
8.3
|
-0.57
|
3.87
|
7.03
|
|
|
|
|
|
RDW3.1
|
RDW/Ex3
|
-0.46
|
4.11
|
4.3
|
-0.25
|
3.75
|
3.53
|
|
|
|
|
bin758
|
qSLL3.1
|
SLL/Ex2
|
|
|
|
|
|
|
-0.59
|
4.51
|
8.16
|
|
|
qSLL3.1
|
SLL/Ex3
|
|
|
|
-0.77
|
3.3
|
11.8
|
-0.85
|
3.4
|
12.28
|
4
|
bin1135
|
qSH4.1
|
SH/Ex1
|
-0.92
|
3.66
|
6.73
|
|
|
|
-1.07
|
4.51
|
7.81
|
|
|
qSH4.1
|
SH/Ex3
|
-0.52
|
4.14
|
4.41
|
-0.56
|
4.61
|
4.83
|
-0.52
|
3.52
|
4.13
|
|
|
qSDW4.1
|
SDW/Ex1
|
-2.07
|
6.08
|
9.49
|
-1.36
|
3.96
|
8.81
|
-1.27
|
4.07
|
8.3
|
|
|
qSDW4.1
|
SDW/Ex2
|
-0.97
|
5.36
|
4.61
|
-0.73
|
4.39
|
3.79
|
-0.58
|
5.07
|
3.49
|
|
|
qSDW4.1
|
SDW/Ex3
|
-1.21
|
6.06
|
8.79
|
-0.96
|
6.39
|
10.14
|
-0.77
|
3.81
|
7.6
|
5
|
bin1218
|
qSDW5.1
|
SDW/Ex2
|
0.91
|
4.28
|
4.08
|
|
|
|
|
|
|
|
|
qSDW5.1
|
SDW/Ex3
|
0.74
|
4.89
|
5.2
|
|
|
|
|
|
|
|
bin1281
|
qFLSL5.1
|
FLSL/Ex1
|
|
|
|
|
|
|
-0.20
|
5.93
|
5.82
|
|
|
qFLSL5.1
|
FLSL/Ex2
|
|
|
|
-0.10
|
3.56
|
4.44
|
-0.09
|
4.00
|
4.54
|
|
|
qFLSL5.1
|
FLSL/Ex3
|
|
|
|
|
|
|
-0.11
|
3.92
|
4.84
|
|
bin1325
|
qSDW5.2
|
SDW/Ex1
|
|
|
|
-1.55
|
3.92
|
11.63
|
|
|
|
|
|
qSDW5.2
|
SDW/Ex3
|
-0.99
|
3.81
|
5.91
|
-0.88
|
5.75
|
8.67
|
|
|
|
6
|
bin1565
|
qFLSL6.1
|
FLSL/Ex1
|
-0.25
|
3.25
|
9.49
|
|
|
|
|
|
|
|
|
qFLSL6.1
|
FLSL/Ex3
|
|
|
|
-0.13
|
4.44
|
4.46
|
|
|
|
|
|
qFLL6.1
|
FLL/ Ex1
|
-0.38
|
9.18
|
17.98
|
-0.38
|
9.58
|
18.67
|
-0.41
|
7.57
|
15.08
|
|
|
qFLL6.1
|
FLL/Ex2
|
-0.36
|
8.22
|
17.12
|
-0.42
|
7.48
|
20.12
|
-0.40
|
6.47
|
17.2
|
|
|
qFLL6.1
|
FLL/Ex3
|
-0.36
|
7.05
|
19.83
|
-0.41
|
9.18
|
23.74
|
-0.42
|
10.16
|
20.32
|
|
|
qSLL6.1
|
SLL/Ex1
|
-0.75
|
10.01
|
19.18
|
-0.74
|
6.05
|
15.88
|
-0.72
|
7.72
|
15.08
|
|
|
qSLL6.1
|
SLL/Ex2
|
-0.52
|
6.96
|
11.13
|
-0.59
|
7.57
|
11.76
|
-0.45
|
7.13
|
5.84
|
|
|
qSLL6.1
|
SLL/Ex3
|
-0.71
|
12.2
|
14.64
|
-0.80
|
6.57
|
15.13
|
-0.82
|
7.17
|
14.01
|
9
|
bin2139
|
qFLL9.1
|
FLL/Ex2
|
-0.28
|
5.06
|
10.73
|
-0.36
|
6.65
|
14.92
|
-0.34
|
3.99
|
11.87
|
|
|
qFLL9.1
|
FLL/Ex3
|
-0.32
|
4.33
|
15.33
|
|
|
|
-0.35
|
5.67
|
13.85
|
|
bin2176
|
qFLL9.1
|
FLL/Ex1
|
-0.34
|
7.41
|
14.45
|
-0.31
|
6.66
|
13.03
|
-0.33
|
4.91
|
10.07
|
|
|
qFLL9.1
|
FLL/Ex3
|
-0.29
|
4.62
|
13.14
|
-0.32
|
5.77
|
14.81
|
-0.32
|
4.56
|
11.87
|
11
|
bin2518
|
qSH11.1
|
SH/Ex1
|
-1.13
|
5.74
|
10.13
|
-1.13
|
4.92
|
8.39
|
-0.95
|
4.37
|
6.58
|
|
|
qSH11.1
|
SH/Ex3
|
-0.51
|
4.3
|
4.19
|
-0.51
|
3.62
|
3.99
|
-0.54
|
4.36
|
4.51
|
|
|
qSDW11.1
|
SDW/Ex1
|
-2.06
|
3.79
|
9.38
|
-1.53
|
5.05
|
11.24
|
-1.30
|
4.4
|
8.78
|
|
|
qSDW11.1
|
SDW/Ex3
|
-1.05
|
5.18
|
6.72
|
-0.90
|
4.82
|
9.04
|
-0.75
|
3.38
|
7.36
|
12
|
bin2760
|
qFLSL12.1
|
FLSL/Ex1
|
|
|
|
0.12
|
4.12
|
4.63
|
|
|
|
|
|
qFLSL12.1
|
FLSL/Ex2
|
0.10
|
4.02
|
4.28
|
|
|
|
|
|
|
|
|
qFLSL12.1
|
FLSL/Ex3
|
|
|
|
0.12
|
6.46
|
4.78
|
|
|
|
a The position of the LOD peak of each QTL. b Trait/experiment. SH, seedling height; FLSL, first leaf sheath length; FLL, first leaf length; SLL, second leaf length; SDW, shoot dry weight; RDW, root dry weight. Ex1, Ex2 and Ex3 indicated the three experiments, respectively. c Addictive effect. Positive values indicate that the allele from Nip increase trait values. d Variance (%) explained by the QTL.
For seedling height, three QTLs distributed on chromosomes 1 (qSH1.1), chromosomes 4 (qSH4.1) and chromosomes 11 (qSH11.1) were both identified in at least two seeding densities, respectively. At qSH1.1, the allele from ‘NIP’ increased the seedling height, while the allele from ‘9311’ had positive effect at qSH4.1 and qSH11.1. Moreover, the phenotypic variances explained by qSH1.1 slightly increased with the increase of seedling density. For FLSL, four QTLs were detected at chromosomes 1, 5, 6 and 12 (qFLSL1.1, qFLSL5.1, qFLSL6.1 and qFLSL12.1), respectively. qFLSL1.1 was detected in both three seeding densities, while qFLSL5.1 was specially detected at MD and HD. On the other hand, qFLSL6.1 and qFLSL12.1 were only identified at LD and MD. For qFLSL5.1, the positive genotypes came from ‘9311’, while for the other two QTL (qFLSL1.1 and qSDW12.1), the positive genotypes came from ‘NIP’. For first leaf length, all the five QTLs (qFLL2.1, qFLL2.2, qFLL6.1, qFLL9.1 and qFLL9.2) were identified in at least two seeding densities, in which qFLL6.1 explained maximum phenotypic variances of 17.98%, 18.67% and 15.08% at LD, MD and HD in Ex1, respectively. Furthermore, the alleles from ‘9311’ for all QTLs increased first leaf length. The QTLs for second leaf length were detected on chromosome 2 (qSLL2.1), chromosome 3 (qSLL3.1) and chromosome 6 (qSLL6.1). qSLL6.1 explained the largest phenotypic variation and existed at all three seeding densities, while qSLL3.1 was only detected at MD and HD. All the three QTLs of SLL functioned in the same direction, with the allele from ‘9311’ increasing the phenotypic value. Three QTLs were identified for RDW on chromosomes 1 (qRDW1.1, qRDW1.2) and chromosomes 3 (qRDW3.1). Interestingly, qRDW1.1 and qRDW3.1 were detected in LD and MD, while qRDW1.2 was detected only in HD. Besides, the allele of ‘9311’ increased RDW for all three QTLs. Five QTLs for shoot dry weight (qSDW2.1, qSDW4.1, qSDW5.1, qSDW5.2 and qSDW11.1) were identified on chromosomes 2, 4, 5(2), and 11. For qSDW2.1, the positive genotype came from ‘NIP’, while for the other four QTL (qSDW4.1, qSDW5.1, qSDW5.2 and qSDW11.1), the positive genotype came from ‘9311’. Two QTLs (qSDW4.1 and qSDW11.1) were detected at both three seeding densities, while qSDW5.1 and qSDW5.2 were detected only at LD and MD.
QTL hotspot is a region where multiple traits were co-located. There were five QTL hotspots (bin311, bin411, bin1135, bin1565 and bin2518) identified clustered within genomic regions on chromosomes 1, 2, 4, 6 and 11 (Table 2 and Figure 3). However, it was not clear whether these five QTL hotspots were single locus with pleiotropic effects on the multiple traits, or a group of tightly linked loci. Further analyses revealed some noteworthy features about these regions. Three of these hotspots, bin311, bin1135 and bin2518 showed almost consistent effects across the three seeding densities, while bin411 mainly detected at MD and HD for FLL and SLL, with minor effects at LD. However, bin411 functioned at LD for SDW. The QTL hotspot at bin1565 on chromosome 6 were simultaneously detected for FLL and SLL at all three seeding densities, however, this hotspot was only detected for FLSL at LD and MD. These results suggest that bin311, bin1135 and bin2518 were not sensitive to seeding density, while bin411 and bin1565 appeared to be QTL-specific for seedling density. Another interesting finding was that the effects of QTL hotspots for the multiple seedling traits showed diversity. For bin311, the allele from ‘NIP’ increased trait values, with the additive effects in the same direction for SH and FLSL. For the other three QTL hotspots (bin1135, bin1565 and bin2518), the allele from ‘9311’ increased trait values, and consistent among those multiple traits. However, the additive effect of bin411 showed the opposite direction, with positive effects for SDW and negative effects for FLL and SLL. All the five QTL hotspots produced considerable individual effects on the seedling traits at the three seedling densities. These results were consistent with the observation that all the traits displayed continuous variation in the RIL population, implying that seedling traits in rice were contributed by many loci with small effects. Furthermore, the existence of specific QTLs at different seeding densities also revealed the differences in seedling traits of indica rice and japonica rice in response to seeding densities.
Genome-wide additive and epistatic effects
Although the main-effect QTLs of young seedlings traits related to machine transplanting under different seeding densities had been fully explored. We did not have a complete understanding of the contributions of the ‘9311’ and ‘Nip’ genomes to these traits at whole-genome level. Genome-wide additive and epistatic effects for each trait at three seeding densities were estimated based on the high-density bin map. The genome-wide additive and epistatic effects and the distribution of QTLs in Ex1 were detailed showed in Figure 4.
Single-locus positive and negative additive effects were extensively distribution among all the seedling traits, however, the contributions of the allele from ‘9311’ and ‘NIP’ differed across the traits and seeding densities (Additional file 1: Table S5). One of the most noticeable results was that the number of negative additive effect bins was more than that of positive additive effect bins for most traits, and the trend was consistent under different seeding densities (Table 3). However, there were a few exceptions, the number of positive and negative additive effect bins was approximately equal for SH and FLSL in Ex2 and Ex3. Besides, there were no bins with significant positive effect for some traits at both three seeding densities, such as FLL in all three experiments, SLL in Ex1, and RDW in Ex2 (Table 3). These results implied that the allele from ‘9311’ contributing to increase seedling traits performance were widely genomic distribution. The comparisons of the distribution of additive effects under three seeding densities showed that the number of bins simultaneously detected under the three seeding densities took the advantage for all the other traits, except for RDW, the number of bins detected at the LD condition dominated. Another noteworthy result was that the bins that detected under HD were all contained under MD for FLL, while the bins detected under MD were contained under LD for RDW, and the bins detected under HD were included under both LD and MD for SLL (Figure 5).
Table 3. Summary of significant additive effects and two-locus interactions under three seeding densities.
Traita
|
Seeding densityb
|
Ex1
|
Ex2
|
Ex3
|
PAEc
|
NAEd
|
PEEe
|
NEEf
|
PAE
|
NAE
|
PEE
|
NEE
|
PAE
|
NAE
|
PEE
|
NEE
|
SH
|
LD
|
36
|
488
|
12
|
42
|
80
|
99
|
0
|
33
|
166
|
363
|
4
|
50
|
MD
|
62
|
529
|
13
|
35
|
82
|
86
|
0
|
41
|
137
|
228
|
2
|
64
|
HD
|
82
|
465
|
6
|
42
|
73
|
109
|
1
|
53
|
166
|
159
|
3
|
79
|
FLSL
|
LD
|
104
|
225
|
5
|
12
|
168
|
294
|
1
|
79
|
173
|
183
|
1
|
88
|
MD
|
85
|
435
|
9
|
35
|
271
|
245
|
0
|
100
|
161
|
257
|
3
|
91
|
HD
|
74
|
382
|
12
|
19
|
351
|
243
|
0
|
118
|
138
|
267
|
2
|
120
|
FLL
|
LD
|
0
|
323
|
22
|
9
|
0
|
335
|
8
|
36
|
0
|
466
|
15
|
25
|
MD
|
0
|
497
|
16
|
9
|
0
|
278
|
25
|
17
|
0
|
553
|
18
|
31
|
HD
|
0
|
535
|
11
|
13
|
0
|
202
|
3
|
53
|
0
|
379
|
6
|
34
|
SLL
|
LD
|
0
|
437
|
4
|
29
|
39
|
321
|
1
|
40
|
21
|
215
|
4
|
62
|
MD
|
0
|
547
|
18
|
27
|
48
|
391
|
0
|
68
|
26
|
252
|
8
|
103
|
HD
|
0
|
491
|
10
|
31
|
19
|
461
|
0
|
68
|
23
|
381
|
9
|
99
|
RDW
|
LD
|
1
|
425
|
1
|
26
|
0
|
650
|
1
|
5
|
6
|
125
|
5
|
11
|
MD
|
0
|
345
|
2
|
4
|
0
|
184
|
0
|
3
|
0
|
99
|
2
|
9
|
HD
|
0
|
471
|
0
|
18
|
0
|
300
|
0
|
12
|
17
|
83
|
0
|
2
|
SDW
|
LD
|
24
|
484
|
8
|
41
|
63
|
334
|
6
|
15
|
74
|
282
|
8
|
29
|
MD
|
1
|
525
|
5
|
18
|
42
|
108
|
1
|
12
|
19
|
559
|
5
|
14
|
HD
|
5
|
416
|
2
|
55
|
21
|
97
|
4
|
14
|
35
|
440
|
7
|
18
|
a Trait, SH, seedling height; FLSL, first leaf sheath length; FLL, first leaf length; SLL, second leaf length; RDW, root dry weight (mg); SDW, shoot dry weight (mg). b Seeding density, LD, MD, and HD represented low density, medium density, and high density, respectively. c, d PAE and NAE, indicated the number of bins with positive additive effect and negative additive effect, respectively. Positive values indicate that alleles from Nip were in the direction of increasing the trait scores, and negative values indicate that alleles from 9311 are in the direction of increasing the score. e, f PEE, NEE indicated the number of two-locus interactions with significant positive and negative epistatic effects, respectively.
Digenetic interactions of each bin pairs across the entire genome at 0.0001 probability level were shown in Additional file 1: Table S6. Less numbers of digenic interactions contributing for seedling traits, explaining lower proportions of phenotypic variance than single-locus analysis (Table 3). However, most of the interaction pairs individually accounted for more than 6% of the genotypic variation, which were not less than some main-effect QTLs. Significant interactions involve many marker loci, most of which were not detected in the single-locus analysis. However, there were still some QTLs found to interact with two or more other locus. The qFLSL12.1 was identified simultaneously interacted with bin1368, bin1845 and bin2499 at HD. The qFLL6.1 was detected to interacted with qFLL2.2 and qFLL9.1 at HD, and they were both showed positive epistasis effects. However, the interactions might change under different seeding densities. For example, the qFLL6.1 was detected to interacted with bin1938 at LD in Ex1. Similarly, qSLL2.1 at QTL hotspot bin411 interacted with bin2640 at LD, while it interacted with bin1290 at HD in Ex1. QTL for SDW at bin90 interacted with bin892 at LD and MD, while it interacted with bin431 and bin1302 at HD in Ex1 (Figure 4). Another result worth noting was that the number of interactions with negative epistasis effects were predominant for SH, FLSL, SLL, RDW and SDW at three seeding densities, except for FLL (Table 3 and Additional file 1: Table S6). Taken together, genome-wide additive and two-locus epistasis suggested a dynamic of the genetic control underlying different seeding densities.
Table 4. List of candidate genes for QTLs identified in specific seeding density.
QTLs
|
Chromosome
|
MSU locus ID
|
Gene annotation
|
Reference
|
qRDW1.1
|
1
|
LOC_Os01g13520
|
OsARF16, auxin response factor 1-like
|
[64, 65]
|
|
1
|
LOC_Os01g13530
|
ABIL3, Abl interactor-like protein 3, expressed
|
[66]
|
|
1
|
LOC_Os01g13540
|
NLP3, Nodule inception protein-like protein 3
|
[67]
|
|
1
|
LOC_Os01g13570
|
Coding for phosphoglycerate mutase, expressed
|
[68]
|
|
1
|
LOC_Os01g13740
|
OsGLK2, probable transcription factor GLK2
|
[69]
|
qFLL2.1
|
2
|
LOC_Os02g14910
|
OsbZIP19, putative bZIP transcription factor RF2b
|
[70]
|
|
2
|
LOC_Os02g15580
|
OsCNGC1, Cyclic nucleotide-gated channels family, Group I
|
[71]
|
|
2
|
LOC_Os02g15640
|
Abscisic acid receptor PYL3, PYR1-like protein 3
|
[72]
|
qFLSL5.1
|
5
|
LOC_Os05g28350
|
OsABI4, Ethylene-responsive transcription factor ABI4
|
[73]
|
|
5
|
LOC_Os05g28500
|
Pentatricopeptide-repeat proteins (PPRs)
|
[74]
|
|
5
|
LOC_Os05g28730
|
Zinc finger, C3HC4 type domain containing protein
|
[75]
|
|
5
|
LOC_Os05g28740
|
Universal stress protein domain-containing protein, putative
|
[76]
|