We screened a population of CSSLs (BC4F3) derived from the cross of R498 (as recipient parent) with 19 different donor parents using marker-assisted selection and chalkiness-related trait investigation. The parental line R498 and CSSLs were planted in Wenjiang in 2022. Nitrogen fertilizer treatment experiments were carried out in Wenjiang in 2021, consistent with a previous work (Li et al., 2023). Three varying levels of nitrogen fertilizer were applied: 0, 10, and 16 kg/666.7 m². The nitrogen fertilizer was applied in two stages, with 70% applied as a base fertilizer and the remaining 30% at the tillering stage. For planting, the crops were organized into four rows, each consisting of 10 plants spaced at intervals of 26.7 × 16.7 cm. Field management adhered to established local agricultural practices.

The agronomic traits were collected consistent with previous research (Hu et al., 2020; Tu et al., 2020). Five plants in the middle of each row were harvested and air-dried after maturation. The agronomic traits including plant height (PH), panicle length (PL), grain width (GW), grain length (GL), and 1,000-grain weight (KGW) were measured. GW, GL, and KGW were measured using SC-G seed counting and a grain weighing device (Wanshen Ltd., Hangzhou, China). Mean phenotypic values were compared using the Student’s t-test, and the data were analyzed using Microsoft Excel 2013.

For the determination of amylose content and amylopectin chain-length distribution, the milled rice was ground to rice flour with a grinder and sieved through a 0.5-mm mesh. The amylose content of the milled rice was measured using the standard iodine colorimetry method described in ISO 6647-2-2011 (Deng et al., 2018). The detailed operation is as follows: 0.01 g of sieved rice powder was placed into a 10-mL centrifuge tube. After 0.1 mL 95% ethanol was added, it was then gently shaken and mixed. Next, 0.9 mL 1 mol/L sodium hydroxide solution was added along the tube wall and mixed in. The centrifugal tube cover was tightly covered and placed at 30°C for 12 h for digestion. After digestion, 9 mL of distilled water was added to the centrifuge tube and mixed thoroughly. A 0.5-mL sample solution was added to a new 10-mL centrifuge tube, combined with 9 mL distilled water, 0.1 mL acetic acid, and 0.2 mL iodine, fully shaken and mixed, and then rested for 20 min. The blank solution was configured and used to adjust the wavelength of 620 nm to zero to measure the absorbance value of the colored sample solution.

Another portion of milled rice was used for the analysis of the amylopectin chain-length distribution, which was measured using a capillary electrophoresis fluorescent detection system (PA800 plus; Beckman-Coulter, CA, USA). From the original electropherogram scan, the peak area for each degree of polymerization (DP) was obtained, which is proportional to the number of chains. The DP was classified into four types following the method of Cameron et al.: A (6 ≤ DP ≤ 12), B1 (13 ≤ DP ≤ 24), B2 (25 ≤ DP ≤ 36), and B3+ (37 ≤ DP ≤ 60) (Cameron et al., 2007).

Protein content was measured from the total nitrogen content of head rice with a conversion index of 5.95 following the protocol of Deng et al. (2021). The crude fatty acid content in the rice grain power was measured by using the Soxhlet extraction method. In brief, 1 ± 0.05 g (m) of sieved rice flour in a filter paper bag was dried at 105°C for 3 h to a constant weight (m1). The crude fatty acids were extracted with petroleum ether for 90 min with ANKOMXT15i. Then, the filter paper bag containing the rice grain power was evaporated to dryness in an oven until a constant weight was obtained (m2). The fat content in the sample was calculated according to the following formula: (m1 - m2)/m × 100.

The dried seeds were processed into polished rice, and 100–200 polished rice grains from each plant were randomly selected to measure chalkiness. Images of head rice were captured, and chalkiness traits were measured with a Microtek ScanWizard EZ scanner and rice quality analyzer SC-E software (Hangzhou Wanshen Detection Technology Co., Ltd., Hangzhou, China; www.wseen.com). Viscosity was tested with Rapid Visco Analyzer (RVA) (Newport Scientific, Australia), and taste quality was estimated using Rice Taste Meter (KETT, Japan) following the method of Tu et al. (2022). Excel 2013 and SPSS v19.0 were used for data collation and analysis.

To prepare the rice for analysis, 30 g of polished rice grains was washed three times and soaked for 10 min. Then, 1.4 times the resulting volume of distilled water was added, cooked for 20 min, and then stewed for 10 min. The cooked rice was then cooled to ambient room temperature and dried using hot air at 80°C for 3 h. The instant rice was prepared by adding 1.2 times the volume of boiled water and incubated for 10 min. The taste quality of the freshly cooked and rehydrated instant rice was detected as described in “Measurement of grain qualities”.

In order to explore the effects of amylose content, ASV, and grain chalkiness on the ECQ of rice, we screened a CSSL population previously constructed in our laboratory to identify CSSLs carrying different alleles of Wx and ALK genes and CSSLs with high grain chalkiness. The process of selecting segment substitution lines involved careful screening based on specific criteria. Agronomic traits such as plant height, growth rate, yield potential, disease resistance, and stress tolerance were assessed in the field. This allowed for the identification of lines that exhibited desirable traits for further analysis. In total, we obtained five CSSLs containing Wx^lv , Wx^a , Wxⁱⁿ , and wx alleles and two CSSLs carrying ALK^G-GC allele with low ASV. In addition, we identified four CSSLs with high chalkiness, including one CSSL with high WCR and three CSSLs with high WBR, which share identical Wx and ALK genotype as the receptor parent, R498 ( Figures 1A, B ; Supplementary Table S1 ).

We investigated the main agronomic traits including plant height (PH), panicle length (PL), tiller number (TN), seed setting rate (SSR), 1,000 grain weight (KGW), grain length (GL), and grain width (GW) of R498 and each CSSL. Our data reveal multiple variations of main agronomic traits between the CSSLs and R498, among which the differences in grain-shape-related traits, including KGW, GL, and GW, were the most obvious, which may be caused by the imported chromosome fragments containing the major genes of these agronomic traits ( Supplementary Table S1 ; Figure 1A ). Generally, the plant morphology of each CSSL was similar to that of 498 ( Supplementary Figure S1 ), consistent with the results that the genetic backgrounds of all the CSSLs are relatively clear. Gui Chao 2 is a high-yield rice variety with wide adaptability, but its appearance quality and ECQ are poor and used as a negative control in this study ( Figure 1B ; Supplementary Table S2 ).

Genomic sequencing analysis was conducted to unravel the genetic makeup of the CSSLs. High-throughput sequencing technologies were employed to obtain comprehensive genomic data, with R498 as the reference genome. The fragment substitutions of CSSLs on each chromosome are shown in Figure 1A , in which each genome is divided into 0.01-Mb fragments. The position information of the mutation site and the number of SNPs that differ between each fragment and R498 were calculated. The genome infiltration of each CSSL was assessed, and it revealed that the backgrounds of all the CSSLs are relatively pure ( Supplementary Figure S1 ). Both Wx and ALK genes located on chromosome 6 can be excluded from the four CSSLs with high chalkiness ( Figure 1A ).

Earlier studies have revealed that ALK is the major gene controlling ASV variation between indica and japonica rice. ALK encodes starch synthase IIa (SSIIa) and plays a key role in the synthesis of long B1 chains by extending the short A and B1 chains of amylopectin in the endosperm (Gao et al., 2003, 2011; Nakamura et al., 2005; Yu et al., 2011). The indica allele ALKⁱ (G-GC) controlled lower ASV, and inactive japonica alleles ALK^j (A-GC or G-TT) controlled high ASV (Fan et al., 2017; Gao et al., 2011; Wang et al., 2015).

To investigate the impact of introducing the ALK^GGC allele into R498 (ALK^GTT ) on the physicochemical properties of amylopectin, the XRD pattern and amylopectin branching chain-length distribution of these two CSSLs were detected in collaboration with Sanshubio (http://www.sanshubio.com/). Both of the CSSLs displayed typical A-type diffraction patterns ( Figure 2A ). Compared to R498, these two substitution lines showed a decrease in the number of A chains and B2 chains, while the number of B1 chains increased ( Figures 2B, C ). This could be attributed to the promotion of A chains into B1 chains by SSIIa, resulting in an increase in rice gelatinization temperature (Nakamura et al., 2005).

To test the accuracy of the rice taste meter and hardness viscosity analyzer, we selected five rice varieties for testing: Yixiangyou 2115 (YX2115), Huihe 5A/R584 (HY584), Quanxiang 3A/R612 (QXY612), Fyou 498 (FY498), and Gui Chao 2. These rice varieties were chosen to represent a sequential decrease in taste quality as determined by sensory evaluation results. The ECQ of these five rice varieties was detected by the rice taste meter and hardness viscosity analyzer, and the comprehensive taste scores were revealed to be consistent with the results of the sensory evaluation ( Supplementary Table S2 ). The rice varieties with high comprehensive scores had better color, higher appearance, viscosity, balance, and taste scores and lower hardness. A correlation analysis of the measured hardness, viscosity, balance, elasticity and comprehensive taste score was carried out, and the results showed that the comprehensive taste core was negatively correlated with hardness and positively correlated with viscosity and balance ( Supplementary Table S3 ). In short, the rice taste meter and hardness viscosity analyzer can accurately determine the ECQ of rice.

Our analysis of appearance quality and the main nutrient content revealed significant findings. Specifically, CSSLs containing Wx^lv , Wx^a , and Wxⁱⁿ exhibited a notably higher amylose content than R498, and their chalky-grain percentage (CGP) and chalky-grain grade (CGG) were significantly higher than those of R498 ( Table 1 ), consistent with recent findings that the Wx genotype may be associated with grain chalkiness formation and rice varieties with Wx^a and Wxⁱⁿ alleles produced more chalky grains (Li et al., 2023; Xia et al., 2022). CSSL-AC-5 carrying the wx allele is a glutinous rice with a milky-white grain. The CSSLs with high chalkiness showed a decrease in amylose content overall, while the protein and lipid contents tended to increase ( Table 1 ), perhaps due to the lack of tight filling of nutrient materials (Lin et al., 2014; Liu et al., 2011). In contrast, CSSLs with lower ASV exhibited lower CGP than R498 ( Table 1 ).

Throughout the rice preparation process, the cooked rice grains from the four CSSLs with high chalkiness, two CSSLs with lower ASV, and R498 under varying fertilization conditions all exhibited excellent grain integrity and color ( Figure 3A ). Rice with low to medium levels of amylose, such as CSSL-AC-5 (wx), a glutinous rice, and R498 (Wx^b ), attained the highest taste quality scores ( Table 1 ). Conversely, CSSLs with higher amylose levels, such as CSSL-AC-1 (Wx^lv ), CSSL-AC-2 (Wx^a ), CSSL-AC-3 (Wxⁱⁿ ), and CSSL-AC-4 (Wxⁱⁿ ), exhibited lower scores across categories of appearance, taste, and comprehensive taste quality ( Table 1 ). Furthermore, we observed that cooked rice grains from CSSLs with elevated amylose content displayed greater dispersion, akin to Gui Chao 2, a rice variety known for its lower taste quality ( Figure 3A ). In conclusion, the amylose content of rice has a substantial impact on both its appearance and eating and cooking quality (ECQ). The comprehension and manipulation of amylose content can yield valuable insights to enhance the overall quality and consumer acceptance of rice varieties (Tian et al., 2009; Xu et al., 2021; Zhang et al., 2021).

This investigation revealed a significant correlation between nitrogen application and the protein content in R498 grains. As the nitrogen levels increased, protein content also rose markedly. However, this increase in protein content was accompanied by a gradual reduction in viscosity and balance of appearance, resulting in diminished taste quality and lower comprehensive taste scores. Under the condition of no nitrogen fertilizer application, the comprehensive taste scores, appearance, and taste scores all showed improvements, along with enhanced viscosity and balance. Importantly, there was no significant change in hardness, resulting in the best taste quality. Conversely, under high nitrogen levels, taste quality was at its lowest, with a decrease in viscosity and poorer balance ( Table 1 ). This outcome aligns with the observation that as protein content increases, the taste value of rice decreases. It is worth noting that protein content is a vital indicator to evaluate the nutritional quality of rice, in line with prior research findings (Zhang et al., 2020).

Interestingly, the CSSLs with high chalkiness had high comprehensive taste scores, but their cooked rice grains were more fragile, particularly those from CSSL-Chalk-1, which had a high WCR ( Figure 3A ; Table 1 ). Therefore, chalkiness is one of the factors that determine the appearance quality of rice. The two low-ASV CSSLs had higher appearance and taste scores, and there was an upward trend in the comprehensive taste scores ( Table 1 ). Thus, within the context of R498, ASV appears to exert minimal influence on taste and predominantly affects the gelatinization temperature of rice.

The correlation analysis revealed a significant negative correlation between amylose content, protein content, and taste quality. Amylose content has the strongest correlation with taste quality and the greatest impact on taste, followed by protein content ( Supplementary Table S4 ). On the other hand, chalkiness shows a weak correlation with taste qualities.

The RVA measurement results of materials with different amylose contents are shown in Figure 3B . CSSL-AC-5 with the wx allele exhibits low gelatinization temperature and peak viscosity (PV), and its breakdown value (BDV) was greater than its setback value (SBV), resulting in good taste ( Figure 3B ). The RVA curves of CSSL-AC-1 (Wx^lv ) and CSSL-AC-2 (Wx^a ) with amylose content higher than 25% are similar to that of the negative control Gui Chao 2 ( Figure 3B ), while CSSL-AC-3 and CSSL-AC-4 with the Wxⁱⁿ alleles had amylose content ranging from 18% to 22%. They show high final viscosity (FV) with significantly greater consistency values (CSV) than BDV, indicating a high retrogradation tendency ( Figure 3B ).

The CSSLs with high chalkiness have higher PV compared to the negative control Gui Chao 2 and the control R498. Their FVs are lower than the negative control Gui Chao 2. The gelatinization temperature is lower than that of Gui Chao 2, with negative SBVs and BDVs higher than CSVs, resulting in better taste. Thus, chalkiness appears to have a relatively weak impact on taste ( Figure 3C ; Supplementary Table S4 ).

The CSSLs with lower ASV exhibit lower PV and FV values, along with positive SBVs that do not show significant variation. This pattern indicates a reduced tendency for retrogradation in the rice ( Figure 3D ). The RVA profile of CSSL-ALK ^GGC-2 shows a higher gelatinization temperature compared to CSSL-ALK ^GGC-1, suggesting that the starch granules in CSSL-ALK ^GGC-2 are more difficult to gelatinize. Compared to CSSL-ALK ^GGC-2, CSSL-ALK ^GGC-1 has no significant changes in PV and trough viscosity, but has higher FV, resulting in a larger CSV and SBV and relatively poorer taste ( Figure 3D ).

Among the nitrogen fertilizer treatments, the RVA curves of the three different nitrogen application levels are similar. In comparison to the negative control Gui Chao 2, they generally exhibit lower TV and FV, with the CSV being smaller than the BDV, indicating a relatively poorer taste quality ( Figure 3E ). Of the three nitrogen application levels, the treatment without nitrogen fertilizer stood out with the lowest SBV and the best taste ( Figure 3E ). In the high nitrogen treatment, the PV is low, and the SBV is the highest, resulting in the worst taste quality ( Figure 3E ).

As amylose content increases, the SBV and retrogradation tendency of starch after gelatinization also rise, leading to a decline in taste quality. On the other hand, ASV primarily influences the gelatinization temperature of starch, with higher ASV resulting in lower gelatinization temperature. Different nitrogen application levels primarily affect the PV during starch gelatinization. Rice subjected to high nitrogen levels exhibits the lowest PV, smaller BDV, and larger SBV and CSV, resulting in variations in taste quality. Based on these findings, we propose an ECQ evaluation criterion for indica rice, which includes a direct amylose content of around 13%–17%, a low degree of chalkiness, and low protein content.

To validate this ECQ evaluation criterion, we conducted laboratory tests on 10 selected rice varieties to assess the impact of various physicochemical characteristics on ECQs, including amylose content, chalkiness, and ASV, as measured by the China National Rice Research Institute ( Supplementary Table S5 ). The results showed that the hybrid rice varieties Jincheng 3A/Shuhui 91 and Jincheng 2A/Shuhui 91, which had higher amylose content compared to other materials, had lower viscosity and poorer balance, resulting in lower taste scores. The hybrid rice varieties Quanxiang 3A/Shuhui 882, Quanxiang 1A/Shuhui 882, and Quanxiang 3A/Shuhui 569, which had higher protein content, exhibited consistent ASV, chalkiness ratios, and amylose content, leading to lower taste scores and lower appearance and texture scores with higher hardness. The hybrid rice variety S28-13/Yuhe had the highest taste score, with an amylose content of around 16% and lower protein content and ASV. This further confirmed that increasing amylose and protein content decreases the taste quality of rice, while the impact of ASV on taste quality was limited.

Firstly, we conducted a comparison between hot air drying and freeze-drying to assess their impact on the rehydration texture of instant rice. It was observed that grains dried by hot air exhibited a slightly yellowish appearance with cracks on the surface, while the freeze-dried grains were paler and plump, although the freeze-drying process took a longer time. Consequently, we selected hot air drying for subsequent experiments.

Secondly, we performed sensory evaluations on three different rehydration ratios (1:1, 1:1.2, and 1:1.4) for instant rice. The results indicated that using a 1:1 water-to-rice ratio did not lead to complete water absorption, resulting in an uneven texture. Conversely, the rice with a 1:1.4 water-to-rice ratio became excessively soft due to an excess of water. In contrast, the rice with a 1:1.2 water-to-rice ratio exhibited a shiny appearance and effectively absorbed the added water. As a result, we recommend a rehydration ratio of 1:1.2 for the preparation of instant rice.

The assessment of the grain integrity after drying and taste quality of rehydrated instant rice for different CSSLs revealed that materials with a higher amylose content, such as those carrying Wx^lv and Wx^a alleles, had better grain integrity but poorer taste quality compared to directly steamed rice ( Table 2 ). On the other hand, R498 and glutinous rice materials with medium to low amylose content exhibited relatively poorer grain integrity after drying, but there was no significant decrease in palatability after rehydration ( Figure 4A ; Table 2 ). The impact of chalkiness on the taste quality of rehydrated instant rice is relatively weak ( Table 2 ), but different types of chalkiness have different effects on the grain integrity after drying. A CSSL with high WCR is more likely to result in incomplete and fragile rice grains ( Figure 4A ). Therefore, it is advisable to select varieties with lower chalkiness, especially lower WCR, for instant rice production.

Rice grains from CSSLs with low ASV dried similarly to R498, exhibiting better grain integrity. After rehydration, the rice grains could regain their original color and integrity ( Figures 4A, B ). Interestingly, instant rice made from CSSLs with low ASV showed a slight increase in appearance and taste scores, a decrease in hardness, an increase in viscosity, an overall improvement in balance, and a trend of increased comprehensive taste score ( Figure 4B ; Table 2 ). This may be attributed to the resistance to cooking exhibited by rice with low ASV. CSSL-ALK ^GGC-1 had lower appearance and taste scores and higher hardness compared to CSSL-ALK ^GGC-2 ( Table 2 ). This finding suggests that the slight difference in taste scores may be attributed to the involvement of SSIIIa in long-chain synthesis, and the increase in B3 chains led to a significant improvement in the overall score of CSSL-ALK ^GGC-2 rice.

In conclusion, we recommend the use of rice with medium to low amylose content, low chalkiness, and low ASV in the production of instant rice.