Rapid evaluation method and system for rice flavor based on high-throughput sequencing

Abstract

This invention discloses a rapid digital evaluation method and system for rice flavor based on high-throughput sequencing, belonging to the field of agricultural product quality testing technology. The method involves rice sample collection and homogenization, total RNA extraction and magnetic bead purification, RNA quality detection, cDNA synthesis, sequencing library construction and quality control; high-throughput sequencing and data processing to construct a flavor evaluation model; establishing a digital flavor scoring system based on the model; evaluating the overall flavor score of rice using the digital flavor scoring system; and outputting a digital evaluation report. This invention combines transcriptome sequencing technology with sensory evaluation, achieving objective and digital evaluation of rice flavor with high detection accuracy and a model cross-validation accuracy ≥85%. It can be widely applied in rice variety breeding, quality identification, and origin traceability.

Description

Technical Field

[0001] This invention belongs to the field of agricultural product quality testing technology, and in particular relates to a rapid digital evaluation method for the sensory characteristics of rice based on high-throughput sequencing. Background Technology

[0002] Rice is one of my country's most important food crops, and its flavor and quality are key indicators determining its edible and market value. Traditional rice flavor evaluation primarily relies on sensory assessment, with professional evaluators subjectively scoring indicators such as aroma, taste, and texture. However, traditional sensory evaluation methods have the following problems: Highly subjective: Different evaluators have different scoring standards, resulting in poor repeatability of results; Inefficient: It requires organizing a professional evaluation team, which is time-consuming and labor-intensive; High cost: Requires trained evaluators, resulting in high labor costs; Difficult to quantify: The scoring results are difficult to store and analyze digitally; High sample consumption: A large number of samples need to be cooked for evaluation.

[0003] In recent years, with the development of molecular biology techniques, transcriptome sequencing technology has been widely used in plant gene expression analysis. The synthesis of flavor compounds in rice is related to multiple metabolic pathways, including fatty acid metabolism, amino acid metabolism, and the synthesis of volatile aromatic compounds. The expression levels of key genes in these pathways are significantly correlated with flavor quality.

[0004] However, no published method has yet combined high-throughput sequencing technology with sensory evaluation of rice flavor. Existing technologies include an electronic nose-based method for rice flavor detection, but this method only detects volatile substances and cannot reflect the flavor formation mechanism at the gene expression level; another method discloses a near-infrared spectroscopy-based method for rice quality detection, but this method has low accuracy in predicting flavor indicators.

[0005] Therefore, developing a rapid digital evaluation method for rice flavor sensory perception based on high-throughput sequencing technology to achieve objective and digital evaluation of flavor quality has significant application value. Summary of the Invention

[0006] To address the aforementioned technical problems and overcome the shortcomings of existing rice flavor evaluation methods, such as strong subjectivity, low efficiency, and difficulty in quantification, this invention provides a rapid digital evaluation method for rice flavor sensory perception based on high-throughput sequencing, achieving objective, digital, and rapid evaluation of flavor quality.

[0007] The objective of this invention is achieved through the following technical solution: This invention provides a rapid digital evaluation method for the sensory characteristics of rice based on high-throughput sequencing, comprising: Rice samples from multiple production areas or varieties were collected and then processed by crushing and homogenization to obtain rice flour. Total RNA was extracted from rice flour using TRIzol reagent and purified using magnetic beads to obtain purified RNA. The purified RNA was subjected to quality testing, and RNA samples that met the set quality standards were retained. Using the quality-tested RNA sample as a template, the first cDNA strand was synthesized through reverse transcription. After degrading the RNA-DNA hybrid strand, double-stranded cDNA was synthesized. The double-stranded cDNA was fragmented, and then successively subjected to end repair, A-tail addition, adapter ligation, magnetic bead purification, and PCR enrichment to construct a high-throughput sequencing library. The fragment size and concentration of the high-throughput sequencing library are detected, and libraries that meet the set quality control standards are retained as qualified libraries; The qualified library was loaded into a high-throughput sequencing platform for paired-end sequencing to obtain paired-end sequences. The obtained paired-end sequences were subjected to quality control to filter out low-quality sequences. Valid sequences were distinguished based on the characteristic sequences at both ends of the paired-end sequences and primer sequences. The sequence orientation was then corrected to obtain flavor-related gene sequence data. A flavor evaluation model is constructed based on the obtained flavor-related gene sequence data. A flavor digital scoring system is established based on the flavor evaluation model. The comprehensive score of rice flavor is evaluated through the flavor digital scoring system, and a digital evaluation report is output.

[0008] Furthermore, the quality testing of the purified RNA involves detecting the concentration and purity of the purified RNA using nanodrop spectrophotometry and detecting the integrity of the purified RNA using agarose gel electrophoresis; the acceptable conditions are: (1) RNA concentration ≥200 ng / μL; (2) The A260 / A280 ratio in the range of 1.8-2.2 reflects the degree of protein contamination; (3) An A260 / A230 ratio > 2.0 reflects the degree of organic solvent pollution; (4) The ratio of fluorescence intensity of 28S rRNA to 18S rRNA bands is ≥1.8.

[0009] Furthermore, the TRIzol reagent is used to extract total RNA from rice flour, using the following method: (1) Sample lysis: Add TRIzol reagent, homogenize thoroughly, and let stand at room temperature for 5-10 minutes; (2) Chloroform extraction: Add chloroform, shake to mix, and centrifuge at 4°C and 12000-15000g for 15 minutes; (3) Isopropanol precipitation: Take the supernatant, add an equal volume of isopropanol, and precipitate at -20℃ for 30-60 minutes; (4) Ethanol washing: Wash the precipitate 2-3 times with 75% ethanol; (5) DEPC water dissolution: After air drying, dissolve the RNA precipitate with DEPC water.

[0010] Furthermore, the method for synthesizing the double-stranded cDNA is as follows: (1) Primer binding: Take the quality-tested RNA sample, add random primers or Oligo(dT) primers, incubate at 65℃ and then on ice to allow the primers to bind to the RNA; (2) First strand synthesis: Add reverse transcription buffer, dNTP mixture, reverse transcriptase and RNase inhibitor, and incubate at 42°C to synthesize the first cDNA strand; (3) RNA degradation: Add RNase H to degrade the RNA in the RNA-DNA hybrid chain; (4) Second strand synthesis: DNA polymerase buffer, dNTP mixture and DNA polymerase are added to synthesize double-stranded cDNA.

[0011] Furthermore, the double-stranded cDNA is fragmented by means of fragmentation enzyme treatment or ultrasonic disruption, and the target fragment size after fragmentation is 200-500bp. The quality control standards for the qualified document library are as follows: (1) The effective fragment size distribution of the library is 200-300bp; (2) Library concentration ≥ 10 nmol / L; (3) In the Agilent 2100 detection spectrum, no characteristic peak corresponding to the connector dimer was detected below 150bp.

[0012] Furthermore, the quality indicators for the paired-end sequencing are: Q30 value ≥ 85%, base error rate ≤ 0.1%, and effective data volume ≥ 6 Gb / sample.

[0013] Furthermore, the criteria for filtering low-quality sequences are as follows: (1) The percentage of bases with sequencing quality value Q < 20 is > 5%; (2) The proportion of N bases in the sequence is >5%; (3) Sequence length < 50 bp; (4) Contains connector contamination sequence; Flavor-related genes include those involved in fatty acid metabolism, amino acid metabolism, synthesis of volatile aromatic substances, carbohydrate metabolism, and terpene synthesis pathways.

[0014] Furthermore, the method for establishing the digital flavor evaluation system is as follows: Quantitative descriptive analysis was used to score the flavor of rice samples from qualified libraries, and a sensory evaluation benchmark database was established. Flavor-related gene sequence data obtained from high-throughput sequencing and rice volatile flavor fingerprints obtained from GC-MS detection of qualified libraries were subjected to Pearson correlation analysis with a sensory evaluation benchmark database to screen out the gene set of flavor markers and the set of volatile flavor markers that are related to sensory scores in the sensory evaluation benchmark database. The screening criteria were: P < 0.05, |r| > 0.6, where P is the restriction interval and r is the correlation coefficient. A random forest machine learning algorithm was used to construct a rice flavor evaluation model, with the selected flavor biomarker gene set and volatile flavor compound biomarker set as core feature variables and sensory scores from the sensory evaluation benchmark database as response variables. 70% of the rice sample data in the qualified library was used for training the flavor evaluation model, and 30% of the rice sample data was used for validation, ensuring that the cross-validation accuracy of the flavor evaluation model was ≥85%. Based on the trained flavor evaluation model, a digital flavor scoring system of 0-100 points is established, with built-in flavor level classification standards: 85-100 points is excellent, 70-84 points is level one, 55-69 points is level two, and <55 points is level three.

[0015] Furthermore, the digital evaluation report includes: overall flavor score, flavor grade determination, description of main flavor characteristics, comparative analysis with benchmark samples, and gene expression profile of flavor biomarkers.

[0016] This invention discloses a rapid digital evaluation system for the sensory characteristics of rice based on high-throughput sequencing, used to execute the method described in this invention, comprising: The sample processing module is used to collect rice samples from multiple production areas or varieties, and then process them into rice flour through crushing and homogenization. The RNA extraction and purification module is used to extract total RNA from rice flour and purify the total RNA using magnetic beads to obtain purified RNA. The RNA quality testing module is used to test the quality of purified RNA and retain RNA samples that meet the set quality qualification conditions. The cDNA synthesis module is used to synthesize the first cDNA strand through reverse transcription using the quality-tested RNA sample as a template, and to synthesize double-stranded cDNA after degrading the RNA-DNA hybrid chain. The sequencing library construction module is used to fragment double-stranded cDNA, sequentially performing end repair, A-tail addition, adapter ligation, magnetic bead purification, and PCR enrichment to construct a high-throughput sequencing library. The library quality control module is used to detect the fragment size and concentration of the high-throughput sequencing library and retain libraries that meet the set quality control standards as qualified libraries. The sequencing module is used to load qualified libraries into a high-throughput sequencing platform for paired-end sequencing to obtain paired-end sequence data. The sequencing data processing module is used to distinguish samples from effective sequences based on the characteristic sequences at both ends of the obtained paired-end sequences and primer sequences, and to correct the sequence orientation to obtain flavor-related gene sequence data. The flavor evaluation module is used to construct a flavor evaluation model based on the obtained flavor-related gene sequence data, establish a flavor digital scoring system based on the flavor evaluation model, evaluate the comprehensive flavor score of rice through the flavor digital scoring system, and output a digital evaluation report.

[0017] The beneficial effects of this invention are as follows: 1. This invention enables objective evaluation: flavor evaluation is based on gene expression data, avoiding the influence of subjective human factors, and the evaluation results are highly reproducible; Digital output: Flavor evaluation results are presented in digital scoring format, which facilitates data storage, analysis, and comparison; High detection accuracy: The model cross-validation accuracy is ≥85%, which is significantly higher than that of traditional instrument detection methods; High throughput: ≥50 samples can be tested at a time, suitable for batch sample testing.

[0018] 2. This invention has a wide range of applications: it can be widely used in multiple fields such as rice variety selection, quality identification, and origin traceability.

[0019] 3. The present invention consumes less sample: only about 100mg of sample is needed to complete the test, which is far lower than that of traditional sensory evaluation.

[0020] 4. The invention provides a wealth of information: in addition to flavor scores, it can also obtain gene expression profiles of flavor markers, providing molecular evidence for breeding research. Attached Figure Description

[0021] Figure 1 This is a system block diagram of the present invention; Figure 2 This is a nanotiter detection spectrum for RNA quality detection in this invention; Figure 3 This is an agarose gel electrophoresis image used in RNA quality detection according to the present invention; Figure 4This is a schematic diagram of the high-throughput sequencing library quality control detection pattern in the method of the present invention, namely: Agilent 2100 Bioanalyzer detection pattern; Figure 5 This is a schematic diagram illustrating the grading system of the flavor digital scoring system of the present invention; Figure 6 This is a correlation analysis diagram between Embodiment 1 of the present invention and traditional sensory evaluation methods. Detailed Implementation

[0022] The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0023] Example: Figures 1-6 As shown, this invention provides a rapid digital evaluation method for the sensory characteristics of rice based on high-throughput sequencing technology, comprising the following steps: Step S1: Sample collection and processing: In this example, 100 rice variety samples were collected from four major rice producing areas, namely Heilongjiang, Jiangsu, Hunan and Guangdong. Each sample weighed 500g. The samples were pulverized to 80 mesh using an ultra-low temperature pulverizer, flash-frozen in liquid nitrogen and stored at -80℃ to obtain rice flour samples. Step S2: Total RNA Extraction and Purification: Total RNA was extracted from rice flour using TRIzol reagent and purified using magnetic beads. Specifically, 100 mg of crushed rice sample was weighed, 1 ml of TRIzol reagent was added, and the mixture was thoroughly homogenized. The mixture was allowed to stand at room temperature for 5 min, then 200 μl of chloroform was added, and the mixture was vigorously shaken for 15 s and allowed to stand at room temperature for 3 min. The mixture was then centrifuged at 12000 g for 15 min at 4 °C, and the supernatant was collected. An equal volume of isopropanol was added, and the mixture was precipitated at -20 °C for 30 min. The mixture was then centrifuged at 12000 g for 10 min at 4 °C, the supernatant was discarded, and the precipitate was washed with 75% ethanol. The mixture was air-dried for 5-10 min, and the RNA was dissolved in 50 μl of DEPC water. The mixture was stored at -80 °C for later use. Step S3: RNA quality detection: Take 1-2 μL of the purified RNA sample obtained in step S2 and use a NanoDrop™ One spectrophotometer to detect the concentration and purity. The qualified criteria are: RNA concentration ≥200 ng / μL, A260 / A280 ratio in the range of 1.8-2.2, and A260 / A230 ratio >2.0. Prepare a 1.2% agarose gel, load 500 ng of the purified RNA sample obtained in step S2, and perform electrophoresis at 120V for 20 min to check the integrity of the purified RNA. Acceptance criteria: 28S / 18S ratio ≥ 1.8, clear bands without obvious tailing (meaning no diffuse weak fluorescent band below the band). Tailing is a typical characteristic of RNA degradation (intact RNA breaks into small fragments of varying sizes, forming diffuse tails). The absence of tails further verifies that the RNA integrity is acceptable; that is, the 28S and 18S rRNA bands have sharp edges and no diffusion, indicating high RNA purity and no contamination from genomic DNA, proteins, or other impurities.

[0024] The RNA samples that passed the quality inspection were retained. In this example, out of 100 samples, 92 samples had RNA quality that passed the inspection and proceeded to the next step. Step S4: cDNA Synthesis: Using a quality-tested RNA sample as a template, the first cDNA strand is synthesized via reverse transcription. After degrading the RNA-DNA hybrid strand, double-stranded cDNA is synthesized, specifically as follows: Take 1 μg of the quality-tested RNA sample obtained in step S3, add 1 μL of random primers (10 μmol / L), and bring the volume to 12 μL with DEPC water. Incubate at 65°C for 5 minutes, then on ice for 3 minutes to allow the primers to bind to the RNA. Add 4 μL of 5× reverse transcription buffer, 1 μL of dNTP mixture (10 mmol / L), 1 μL of reverse transcriptase (200 U / μL), and 1 μL of RNase inhibitor (40 U / μL) sequentially, for a total volume of 20 μL. Incubate at 42°C for 60 minutes to synthesize the first cDNA strand. Add 1 μL of RNase H (10 U / μL), and incubate at 37°C for 20 minutes to degrade the RNA in the RNA-DNA hybrid strand. Add 20 μL of 2× DNA polymerase buffer, 1 μL of dNTP mixture, and 1 μL of DNA polymerase (5 U / μL), and bring the volume to 40 μL with DEPC water. Incubate at 16°C for 2 hours to synthesize double-stranded cDNA.

[0025] Step S5: Sequencing library construction: The double-stranded cDNA was fragmented using a fragmentation enzyme and incubated at 37°C for 15 minutes. The high-throughput sequencing library was constructed by sequentially performing end repair, adding A-tails, ligating adapters, purifying with magnetic beads, and PCR enrichment (12 cycles). The cDNA fragmentation was performed using fragmentation enzyme treatment or ultrasonic disruption, and the target fragment size after fragmentation was 200-500 bp.

[0026] Step S6: Library Quality Control: The fragment size and concentration of the high-throughput sequencing library were tested. Fragment distribution was detected using an Agilent 2100 Bioanalyzer, and library concentration was quantified using qPCR. Acceptance criteria: fragment size 200-300 bp, concentration ≥10 nmol / L, no adapter dimers. In this example, 89 out of 92 samples met the library quality standards; libraries meeting these criteria were retained as qualified libraries. Step S7: High-throughput sequencing: Load the obtained qualified library into the Illumina NovaSeq 6000 high-throughput sequencing platform for PE150 paired-end sequencing to obtain paired-end sequence data; real-time monitoring of the quality indicators of paired-end sequencing is as follows: Q30 value ≥ 85%, base error rate ≤ 0.1%, effective data volume ≥ 6Gb / sample.

[0027] Step S8: Sequencing Data Processing: FastQC was used to perform quality control on the paired-end sequence data obtained in Step S7. Trimmomatic was used to filter low-quality sequences. Samples were distinguished based on the characteristic sequences (barcodes) at both ends of the sequence and the primer sequences to obtain valid sequences. The sequence orientation was corrected. The filtering criteria were: Q < 20 bases > 5%, N bases > 5%, length < 50 bp, and presence of adapter contamination. Flavor-related gene sequence data were obtained for screening and analysis of potential target differential metabolites. Cleanreads were aligned to the rice reference genome (IRGSP-1.0) using HISAT2, and gene expression levels (TPM values) were calculated using StringTie. Flavor-related genes were screened, including key genes in fatty acid metabolism pathways (LOX, AOS, etc.), amino acid metabolism pathways (BCAT, AAT, etc.), volatile aromatic compound synthesis pathways (DXS, TPS, etc.), carbohydrate metabolism pathways (AMY, SS, etc.), and terpene compound synthesis pathways (HMGR, SQS, etc.), resulting in the identification of 156 flavor-related gene sequence data.

[0028] Step S9: Digital Evaluation of Flavor: Construct a flavor evaluation model based on the obtained flavor-related gene sequence data, evaluate the overall flavor score of rice using the flavor evaluation model, and output a digital evaluation report, specifically as follows: (1) Establish a sensory evaluation benchmark database: A professional sensory evaluation team of 15 people was formed, and the flavor of rice samples from 89 qualified libraries was scored using quantitative descriptive analysis (QDA) to establish a sensory evaluation benchmark database. The flavor score includes aroma intensity, aroma purity, aroma richness, taste harmony, taste freshness, mouthfeel delicacy, aftertaste persistence, and overall acceptability, with a comprehensive score of 0-100. (2) Correlation analysis: The 156 flavor-related gene sequence data screened by high-throughput sequencing obtained in step S8 and the rice volatile flavor fingerprint obtained by GC-MS detection of qualified library rice samples were respectively subjected to Pearson correlation analysis with the sensory evaluation benchmark database. According to the indicators for judging significance: P<0.05, |r|>0.6, R: correlation coefficient, P: confidence interval, 48 genes related to the sensory evaluation benchmark database were selected as flavor marker gene set and volatile flavor marker gene set.

[0029] (3) Constructing a flavor evaluation model: The random forest machine learning algorithm was used to construct a rice flavor evaluation model with the selected flavor biomarker gene set and volatile flavor substance biomarker set as core feature variables and the sensory scores in the sensory evaluation benchmark database as response variables. During the model construction process, 70% of the rice sample data from 89 samples in the qualified library were used for flavor evaluation model training and 30% of the rice sample data were used for flavor evaluation model validation to ensure that the cross-validation accuracy of the flavor evaluation model was ≥85%. The flavor evaluation model in this example achieved an accuracy of 87.3% after cross-validation, which can accurately realize the rapid digital prediction of rice flavor and provide efficient and objective technical support for rice flavor evaluation.

[0030] Among them, 156 feature variables are gene sequence data related to rice flavor (fitting the scale of conventional gene detection datasets), 89 are sample numbers, and 15 are the number of people in the sensory evaluation group; the cross-validation accuracy is 87.3%, which represents the degree of agreement between the model's prediction results and the actual sensory scores, verifying the reliability of the model.

[0031] (4) Establish a digital flavor scoring system: Based on the trained flavor evaluation model, establish a digital flavor scoring system for rice with a score of 0-100, with built-in flavor grade classification standards: flavor score of 85-100 is excellent, 70-84 is first grade, 55-69 is second grade, and <55 is third grade; The flavor evaluation model establishes a digital scoring system to evaluate the overall flavor score of rice and outputs a digital evaluation report.

[0032] Example 2: Application of flavor evaluation of test samples.

[0033] Three samples of commercially available rice were collected and tested according to the method described in Example 1.

[0034] RNA quality testing results: The A260 / A280 ratios of the three samples were 1.92, 1.88, and 2.05, respectively, and the 28S / 18S ratios were 1.95, 1.87, and 2.10, respectively, all meeting the acceptable standards. After sequencing, the expression data of flavor biomarker genes were substituted into the evaluation model constructed in Example 1 to calculate the flavor score.

[0035] Test results Sample A 88.5 Superior Sample B 76.2 Level 1 Sample C 61.8 Level 2 Output a digital evaluation report, including overall flavor score, flavor grade determination, description of main flavor characteristics, comparative analysis with benchmark samples, and gene expression profile of flavor biomarkers.

[0036] The application of the method described in this invention in rice quality identification includes flavor quality grade determination, comprehensive taste quality score, freshness grade assessment, and variety authenticity identification.

[0037] The rice quality assessment includes the evaluation of the following indicators: flavor quality grade determination; comprehensive taste quality score; freshness grade assessment; and variety authenticity identification.

[0038] The rice quality assessment is applicable to the following scenarios: quality grading during rice procurement; quality inspection of processed rice products; quality sampling inspection of commercially available rice products; and quality monitoring during rice storage.

[0039] The output formats of the rice quality assessment include: digital quality scoring report; flavor characteristic radar chart; comparative analysis with standard samples; and quality grade certificate.

[0040] Example 3: A rapid digital evaluation system for the sensory characteristics of rice based on high-throughput sequencing technology, used to execute the evaluation method described in this invention, includes: The sample processing module is used to collect rice samples from multiple production areas or varieties, and then process them into rice flour through crushing and homogenization. The RNA extraction and purification module is used to extract total RNA from rice flour and purify the total RNA using magnetic beads to obtain purified RNA. The RNA quality testing module is used to test the quality of purified RNA and retain RNA samples that meet the set quality qualification conditions. The cDNA synthesis module is used to synthesize the first cDNA strand through reverse transcription using the quality-tested RNA sample as a template, and to synthesize double-stranded cDNA after degrading the RNA-DNA hybrid chain. The sequencing library construction module is used to fragment double-stranded cDNA, sequentially performing end repair, A-tail addition, adapter ligation, magnetic bead purification, and PCR enrichment to construct a high-throughput sequencing library. The library quality control module is used to detect the fragment size and concentration of the high-throughput sequencing library and retain libraries that meet the set quality control standards as qualified libraries. The sequencing module is used to load qualified libraries into a high-throughput sequencing platform for paired-end sequencing to obtain paired-end sequence data. The sequencing data processing module is used to distinguish samples from effective sequences based on the characteristic sequences at both ends of the obtained paired-end sequences and primer sequences, and to correct the sequence orientation to obtain flavor-related gene sequence data. The flavor evaluation module is used to construct a flavor evaluation model based on the obtained flavor-related gene sequence data, establish a digital flavor scoring system based on the flavor evaluation model, evaluate the comprehensive flavor score of rice through the digital flavor scoring system, and generate a digital flavor evaluation report.

[0041] The present invention provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the method.

[0042] The present invention provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps of the method.

[0043] The parts not described in detail in this application are all existing conventional technologies and will not be elaborated here.

[0044] It is understood that the above specific description of the present invention is only for illustrating the present invention and is not limited to the technical solutions described in the embodiments of the present invention. Those skilled in the art should understand that modifications or equivalent substitutions can still be made to the present invention to achieve the same technical effect; as long as the use needs are met, they are all within the protection scope of the present invention.

Claims

1. A rapid digital evaluation method for the sensory characteristics of rice flavor based on high-throughput sequencing, characterized in that: include: Rice samples from multiple production areas or varieties were collected and then processed by crushing and homogenization to obtain rice flour. Total RNA was extracted from rice flour using TRIzol reagent and purified using magnetic beads to obtain purified RNA. The purified RNA was subjected to quality testing, and RNA samples that met the set quality standards were retained. Using the quality-tested RNA sample as a template, the first cDNA strand was synthesized through reverse transcription. After degrading the RNA-DNA hybrid strand, double-stranded cDNA was synthesized. The double-stranded cDNA was fragmented, and then successively subjected to end repair, A-tail addition, adapter ligation, magnetic bead purification, and PCR enrichment to construct a high-throughput sequencing library. The fragment size and concentration of the high-throughput sequencing library are detected, and libraries that meet the set quality control standards are retained as qualified libraries; The qualified library was loaded into a high-throughput sequencing platform for paired-end sequencing to obtain paired-end sequences. The obtained paired-end sequences were subjected to quality control to filter out low-quality sequences. Valid sequences were distinguished based on the characteristic sequences at both ends of the paired-end sequences and primer sequences. The sequence orientation was then corrected to obtain flavor-related gene sequence data. A flavor evaluation model is constructed based on the obtained flavor-related gene sequence data. A flavor digital scoring system is established based on the flavor evaluation model. The comprehensive score of rice flavor is evaluated through the flavor digital scoring system, and a digital evaluation report is output.

2. The method according to claim 1, characterized in that, The quality assessment of the purified RNA was performed by detecting the concentration and purity of the purified RNA using nanodrop spectrophotometry and detecting the integrity of the purified RNA using agarose gel electrophoresis; the acceptable criteria were: (1) RNA concentration ≥200 ng / μL; (2) The A260 / A280 ratio in the range of 1.8-2.2 reflects the degree of protein contamination; (3) An A260 / A230 ratio > 2.0 reflects the degree of organic solvent pollution; (4) The ratio of fluorescence intensity of 28S rRNA to 18S rRNA bands is ≥1.

8.

3. The method according to claim 1, characterized in that, The TRIzol reagent was used to extract total RNA from rice flour, and the method was as follows: (1) Sample lysis: Add TRIzol reagent, homogenize thoroughly, and let stand at room temperature for 5-10 minutes; (2) Chloroform extraction: Add chloroform, shake to mix, and centrifuge at 4°C and 12000-15000g for 15 minutes; (3) Isopropanol precipitation: Take the supernatant, add an equal volume of isopropanol, and precipitate at -20℃ for 30-60 minutes; (4) Ethanol washing: Wash the precipitate 2-3 times with 75% ethanol; (5) DEPC water dissolution: After air drying, dissolve the RNA precipitate with DEPC water.

4. The method according to claim 1, characterized in that, The method for synthesizing the double-stranded cDNA is as follows: (1) Primer binding: Take the quality-tested RNA sample, add random primers or Oligo(dT) primers, incubate at 65℃ and then on ice to allow the primers to bind to the RNA; (2) First strand synthesis: Add reverse transcription buffer, dNTP mixture, reverse transcriptase and RNase inhibitor, and incubate at 42°C to synthesize the first cDNA strand; (3) RNA degradation: Add RNase H to degrade the RNA in the RNA-DNA hybrid chain; (4) Second strand synthesis: DNA polymerase buffer, dNTP mixture and DNA polymerase are added to synthesize double-stranded cDNA.

5. The method according to claim 1, characterized in that, The double-stranded cDNA is fragmented by using fragmentation enzyme treatment or ultrasonic disruption, and the target fragment size after fragmentation is 200-500bp. The quality control standards for the qualified document library are as follows: (1) The effective fragment size distribution of the library is 200-300bp; (2) Library concentration ≥ 10 nmol / L; (3) In the Agilent 2100 detection spectrum, no characteristic peak corresponding to the connector dimer was detected below 150bp.

6. The method according to claim 1, characterized in that, The quality indicators for the paired-end sequencing are: Q30 value ≥ 85%, base error rate ≤ 0.1%, and effective data volume ≥ 6 Gb / sample.

7. The method according to claim 1, characterized in that, The criteria for filtering low-quality sequences are as follows: (1) The percentage of bases with sequencing quality value Q < 20 is > 5%; (2) The proportion of N bases in the sequence is >5%; (3) Sequence length < 50 bp; (4) Contains connector contamination sequence; Flavor-related genes include those involved in fatty acid metabolism, amino acid metabolism, synthesis of volatile aromatic substances, carbohydrate metabolism, and terpene synthesis pathways.

8. The method according to claim 1, characterized in that, The method for establishing a digital flavor evaluation system is as follows: Quantitative descriptive analysis was used to score the flavor of rice samples from qualified libraries, and a sensory evaluation benchmark database was established. Flavor-related gene sequence data obtained from high-throughput sequencing and rice volatile flavor fingerprints obtained from GC-MS detection of qualified libraries were subjected to Pearson correlation analysis with a sensory evaluation benchmark database to screen out the gene set of flavor markers and the set of volatile flavor markers that are related to sensory scores in the sensory evaluation benchmark database. The screening criteria were: P < 0.05, |r| > 0.6, where P is the restriction interval and r is the correlation coefficient. A random forest machine learning algorithm was used to construct a rice flavor evaluation model, with the selected flavor biomarker gene set and volatile flavor compound biomarker set as core feature variables and sensory scores from the sensory evaluation benchmark database as response variables. 70% of the rice sample data in the qualified library was used for training the flavor evaluation model, and 30% of the rice sample data was used for validation, ensuring that the cross-validation accuracy of the flavor evaluation model was ≥85%. Based on the trained flavor evaluation model, a digital flavor scoring system of 0-100 points is established, with built-in flavor level classification standards: 85-100 points is excellent, 70-84 points is level one, 55-69 points is level two, and <55 points is level three.

9. The method according to claim 1, characterized in that, The digital evaluation report includes: overall flavor score, flavor grade determination, description of main flavor characteristics, comparative analysis with benchmark samples, and gene expression profile of flavor biomarkers.

10. A rapid digital sensory evaluation system for rice flavor based on high-throughput sequencing, used to perform the method according to any one of claims 1-9, characterized in that, include: The sample processing module is used to collect rice samples from multiple production areas or varieties, and then process them into rice flour through crushing and homogenization. The RNA extraction and purification module is used to extract total RNA from rice flour and purify the total RNA using magnetic beads to obtain purified RNA. The RNA quality testing module is used to test the quality of purified RNA and retain RNA samples that meet the set quality qualification conditions. The cDNA synthesis module is used to synthesize the first cDNA strand through reverse transcription using the quality-tested RNA sample as a template, and to synthesize double-stranded cDNA after degrading the RNA-DNA hybrid chain. The sequencing library construction module is used to fragment double-stranded cDNA, sequentially performing end repair, A-tail addition, adapter ligation, magnetic bead purification, and PCR enrichment to construct a high-throughput sequencing library. The library quality control module is used to detect the fragment size and concentration of the high-throughput sequencing library and retain libraries that meet the set quality control standards as qualified libraries. The sequencing module is used to load qualified libraries into a high-throughput sequencing platform for paired-end sequencing to obtain paired-end sequence data. The sequencing data processing module is used to distinguish samples from effective sequences based on the characteristic sequences at both ends of the obtained paired-end sequences and primer sequences, and to correct the sequence orientation to obtain flavor-related gene sequence data. The flavor evaluation module is used to construct a flavor evaluation model based on the obtained flavor-related gene sequence data, establish a flavor digital scoring system based on the flavor evaluation model, evaluate the comprehensive flavor score of rice through the flavor digital scoring system, and output a digital evaluation report.

Images