Performance of starch RVA profiles of parents and CSSLs in different years
The starch RVA profiles of two parents, 9311 and Nipponbare, differed significantly over 4 years of the study. Most of the RVA parameters varied greatly among 4 years, except PeT and PaT which were less affected by environment (Table 1). For indica rice 9311, almost all parameters were much smaller than those in Nipponbare, with the exception of SB and PeT.
The starch RVA profiles were continually distributed in the CSSLs over the different years except for PaT, which exhibited a double-peak distribution (Figure 1). Among the various RVA parameters, the maximum value of peak viscosity (PV) was greater than the other parameters as consequence of greater kurtosis. The starch RVA profiles of CSSLs showed partial separation for some parameters in 2016 and 2017, such as PV, final viscosity (FV) and pasting temperature (PaT). Among the CSSLs, the PV, FV and PaT phenotypic values exhibited greater distribution than the other traits over the 4 years, whereas breakdown viscosity (BD), setback viscosity (SB) and peak time (PeT) were stable and varied little. The mean values for the various RVA parameters of the CSSLs in each year were near the mid-parent value, but some values were not (Table 1). In addition, the phenotypic values for all starch RVA profiles had bidirectional ultra-parental genetic types in the CSSL population, suggesting that those traits were of quantitative traits and controlled by polygenes.
Correlation analysis using the average value of four years as variable revealed that PV was positively correlated with trough viscosity (TV), BD and FV, but negatively correlated with SB and PaT. TV was positively correlated with FV and PeT. FV was positively correlated with PeT and PaT but negatively correlated with SB. BD was negatively correlated with all other RVA parameters. SB was positively correlated with PeT and PaT. These correlations were highly significant (P < 0.01). Among the 4 years, the relationship between SB to TV and FV was changed (Supplementary table). SB correlated with FV except in 2013. TV was significantly positively correlated with SB in 2016 and 2017.
Microclimate
Because of the variation in daily maximum temperature and average temperature, mean of daily maximum temperature during recorded periods ranged from 28.7 ◦C to 34.3 ◦C among four years in our research. The daily average temperatures during observational periods in 2013, 2016 and 2017 were about 28.0 ◦C, higher than 25.1 ◦C in 2014 (Figure 2). The average of daily maximum temperature showed the same tendency as the daily average temperatures among those four years. In 2017, it had the highest recorded temperature among those years. The relative humidity (RH) of those four years showed big difference with the temperature. The average daily RH of observational periods in 2014 and 2017 were about 91%, higher than 80.5% and 82.5%, those in 2013 and 2016.
QTL analysis
To elucidate the interference of environmental factors on starch viscosity among the CSSLs, we mapped QTLs that influenced the starch RVA profile characteristics over the 4 years of the study. In total, 310 QTLs were detected for all seven RVA characteristics. We detected 44 PV-related QTLs, 42 TV and 43 FV QTLs. 43, 49, 34 and 55 QTLs were identified that controlled BD, SB, PeT and PaT, respectively.
Among 310 QTLs, 22 loci for PV, 15 for TV, 25 for FV and 11 for BD were detected in more than one year, and 50%, 35.7%, 58.1% and 25.6% of those QTLs were detected in different years that control those four starch RVA profile characters. For SB, PeT and PaT, 28, 8 and 27 QTLs, respectively, were mapped in different years. Among all 310 QTLs, 81 and 50 QTLs were mapped in 2 and 3 years, respectively. Five constant QTLs that control TV, FV and PaT, were identified in all 4 years (Table 2). The other 174 QTLs appeared in only 1 year. These results implied that, although the QTLs showed large variation in different years, there still exits some constant QTLs related to starch RVA profile characters. Such as the area located on 16.9–19 Mb of Chr. 2, which was found to affect both FV and TV each year. qPaT7 and qTV7, which are located near RM432 on Chr. 7, and qFV9, located near RM219 on Chr. 9, were also mapped over the 4 years. A QTL near RM3827 on Chr.6, which was associated with SB, and a QTL near RM1812 on Chr.11 associated with BD, is not been previously identified in those particular locations (Table 3). Those 6 QTLs were not affected by environmental factors in our research, there are very important for the RVA characteristics of rice starch.
We observed pleiotropism of QTLs for starch RVA profile characteristics. In total, 26 marker intervals corresponding to 55 QTLs were identified in more than 3 years in which QTL were associated with starch RVA profiles. Among those constant QTLs, 13 chromosomal intervals showed pleiotropy with more than one starch RVA profile characteristic (Table 3). In particular, the region near RM219 on Chr. 9 contained a QTL that was associated with all starch RVA profile characteristics except TV. Moreover, marker interval RM469 to RM587 on Chr. 6 was found to be associated with five parameters, and the interval near marker RM3795 on Chr. 2 as well as markers RM289 and RM178 on Chr. 5 and marker 10-1.63 on Chr. 10 were found to be associated with four RVA parameters.
Several studies have reported on QTLs that corresponded to starch RVA profiles of rice [17, 19, 20, 22, 33]. In our study present here, those 26 intervals detected more than 3 times were compared with QTL mapped previously. Except for two intervals, all the other intervals were found to overlap with QTLs identified for starch RVA profiles among different mapping populations and different environments (Table 3). Certain hot spots were identified such as region RM469 to RM587 on Chr. 6 and RM1375 on Chr. 10. In the region RM469 to RM587 on Chr. 6, we mapped QTLs associated with SB, PV, BD, PeT and PaT. Many QTLs associated with PeT, TV, FV, BD and SB were identified in the same genomic regions reported previously [1, 3, 6, 19, 22]. qBD10, located near RM1375 (15.9–18 Mb) on Chr. 10, was in a similar chromosomal region as certain starch RVA profile QTLs which related with FV, SB, PeT, TV, PaT and PV [3, 22, 33].
Our stable QTL clusters shows overlapping to some of the known SSRGs compared with previous studies (Table 4). All of these QTL clusters were found in regions where some of the SSRGs are, such as GBSSI (Wx), SSIII-1, SSIV-2 and SBE3. Besides the QTL in the interval 0.56–2.86 Mb on Chr. 6, which overlapped with Wx, the region 3.24–5.38 Mb and 30.45–32.65 Mb on Chr. 4, 24.59–26.37 Mb on Chr.5, and 16.9–20.95 Mb on Chr.2 contained PUL, SSIII-1, SSIV-2, and SBE3, respectively. ADPlar, ADPsma, SSII-1, and ISA were near our starch RVA QTLs. For ADPlar, ADPiso, GBSSII, SSII-2, SSIII-2, SSIV-1 and PUL, the QTLs near or overlapped these genes were detected in only 1 or 2 years, implying that those QTLs were unstable and easily affected by environment. No QTL associated with starch RVA traits was detected in regions that contained SSI,SSII-3, SBE1 and SBE4. The region between RM6748 to RM5473 on Chr. 4 where the gene SSIII-2 situated is related to SB. SSIII-2 is one of the essential gene controlling Pat and PeT [5]. However, the effort of this region on Pat and PeT did not tested in our research.
The clusters closed to SSII-1 and SSII-3 were reported that had major effects on the PT, and minor effects on gel consistency (GC), AC, PV, CPV, BD, and SB [2]. In this research, we found that the QTL near SSII-1 was associated with BD. Yan reported the QTL clusters near the SBE3 locus on the Chr. 2, which had 4 major QTLs associated with HPV, consistency viscosity, viscosity at 95℃ and BD were detected in 2 years [25]. We also found QTL clusters overlapped with SBE3 gene, which was associated with FV, TV, SB and PaT.