Fermentation type identification method of yeast for making hard liquor, computer program product, equipment and medium

Abstract

The invention relates to a yeast fermentation type identification method, a computer program product, equipment and a medium. The yeast fermentation type identification method comprises the following steps: acquiring key microorganism data in yeast to be identified; and inputting the key microorganism data into a pre-trained fermentation type identification model to obtain a fermentation type identification result of the to-be-identified yeast. According to the method, the fermentation type of the yeast for making hard liquor is put forward and defined by focusing on the fermentation force of microorganisms, and the fermentation type identification result of the yeast for making hard liquor to be identified is automatically output by adopting the trained model capable of identifying the fermentation type of the yeast for making hard liquor according to the obtained key microorganism data of the yeast for making hard liquor to be identified, so that the classification of the yeast for making hard liquor is more scientific and objective; and the method has production significance.

Description

Technical Field

[0001] This application relates to the field of Daqu quality identification technology, and in particular to a fermentation type identification method, computer program product, equipment and medium for Daqu. Background Technology

[0002] Making koji is the foundation of brewing, and high-quality high-temperature koji is an important source and guarantee of the style characteristics of sauce-flavored baijiu.

[0003] "Unpacking" refers to the final stage of fermentation in the koji-making process. It mainly involves removing the fermented koji blocks from the koji storage after a certain period of fermentation. Different types of koji blocks play different roles in the brewing process, influencing the taste and flavor of the baijiu.

[0004] Traditionally, koji (fermented grain starter) is classified into three types based on color: white koji, yellow koji, and black koji. This classification primarily relies on the physicochemical properties of the koji, neglecting the fermentation power of the different microorganisms within it. Even when using two pieces of the same type of koji to brew baijiu, such as two pieces of black koji, the different microbial compositions, functions, and fermentation powers of the two pieces will result in significant differences in the final taste and flavor of the baijiu. Summary of the Invention

[0005] To address the aforementioned technical problems, this application provides a method, computer program product, device, and medium for identifying the fermentation type of Daqu (a type of starter culture).

[0006] In a first aspect, this application provides a fermentation type identification method, comprising:

[0007] Acquire key microbial data from the koji (fermented liquor) to be identified;

[0008] The key microbial data is input into a pre-trained fermentation type identification model to obtain the fermentation type identification result of the Daqu to be identified.

[0009] In one embodiment, the training process of the fermentation identification model includes:

[0010] Obtain microbial composition data samples from multiple Daqu (fermented koji) samples;

[0011] Based on the microbial composition data sample, determine the corresponding fermentation type label for the Daqu sample;

[0012] Based on the microbial composition data sample, the key microbial genera were determined, and the key microbial data sample of the Daqu sample was obtained.

[0013] The fermentation type identification model is obtained by training a model based on the key microbial data samples and fermentation type labels of multiple Daqu samples.

[0014] In one embodiment, determining the fermentation type label of the corresponding Daqu sample based on the microbial composition data sample includes:

[0015] Based on the microbial composition data sample, determine the dissimilarity of microbial composition between every two Daqu samples;

[0016] Clustering is performed based on the dissimilarity of the microbial composition to determine the optimal number of clusters;

[0017] The fermentation type label for each of the Daqu samples is determined based on the optimal cluster number;

[0018] and / or,

[0019] The step of determining key microbial genera based on the microbial composition data sample includes:

[0020] The microbial composition data samples of multiple samples are analyzed to obtain analysis results, which include evaluation indicators for distinguishing different fermentation types based on microbial genera.

[0021] The key microbial genera are determined from the microbial genera according to the evaluation indicators.

[0022] In one embodiment, the Daqu sample is prepared from dismantled Daqu;

[0023] The fermentation type identification result is used to characterize whether the Daqu to be identified is a first fermentation type or a second fermentation type. The first fermentation type has a stronger starch degradation ability and saccharification power than the second fermentation type.

[0024] In one embodiment, the key microbial genus includes a first key microbial genus corresponding to the first fermentation type and a second key microbial genus corresponding to the second fermentation type;

[0025] The first key microbial genus includes Bacillus;

[0026] The second key microbial genus includes the genus *Bacillus*.

[0027] In one embodiment, the first key microbial genus further includes at least one of the following: Pantotheca, Phyllanthae, Pediococcus, Micrococcus, Enterococcus, and Lactobacillus.

[0028] and / or,

[0029] The second key microbial genus also includes at least one of the genera *Croppensteadella*, *Ciminovicella*, and *Penicillium*.

[0030] In one embodiment, obtaining microbial composition data samples from multiple Daqu (a type of starter culture) samples includes:

[0031] Obtain multiple large-scale samples;

[0032] Deoxyribonucleic acid (DNA) was extracted and sequenced from each of the Daqu samples to obtain the corresponding metagenomic sequencing data samples.

[0033] Based on the metagenomic sequencing data samples, the microbial composition data samples of each Daqu sample are obtained, and the microbial composition data samples include the microbial genera in the Daqu sample and the relative abundance among the microbial genera;

[0034] And / or,

[0035] The acquisition of key microbial data in the Daqu (a type of starter culture) to be identified includes:

[0036] The deoxyribonucleic acid of the Daqu to be identified was extracted and sequenced to obtain the corresponding metagenomic sequencing data;

[0037] Based on the metagenomic sequencing data, the microbial composition data of the Daqu to be identified is obtained, and the microbial composition data includes the microbial genera present in the Daqu sample to be identified and the relative abundance among the microbial genera.

[0038] The key microbial genera and their relative abundance are obtained from the microbial composition data and used as the key microbial data.

[0039] In a second aspect, this application provides a computer program product, including a computer program / instructions, which, when executed by a processor, implement the steps of the method described in the first aspect.

[0040] Thirdly, this application provides a computer device including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the method described in the first aspect.

[0041] Fourthly, this application provides a computer-readable storage medium having a computer program / instructions stored thereon, which, when executed by a processor, implement the steps of the method described in the first aspect.

[0042] Based on common knowledge in the field, the above-mentioned preferred conditions can be combined arbitrarily to obtain various preferred embodiments of this application.

[0043] The beneficial effects that the above-mentioned fermentation type identification method, computer program product, equipment, and medium for Daqu (a type of starter culture) can achieve include:

[0044] Focusing on the core characteristic of Daqu (a type of starter culture) – the fermentation power of microorganisms – different fermentation types have unique microbial species composition, functional potential, microbial interactions, metabolic capabilities, etc. This paper proposes and defines the "fermentation type" of Daqu for the first time. Based on this, a fermentation type identification method for Daqu is proposed. The method uses a trained model that can identify the fermentation type of Daqu and automatically outputs the fermentation type identification result of the Daqu to be identified based on the key microbial data of the Daqu to be identified.

[0045] The classification scheme based on "fermentation-type" Daqu proposed in this application, especially for Daqu after dismantling, is more scientific, objective, and of greater significance to production than the traditional classification scheme that divides Daqu into white, yellow, and black Daqu according to color. Attached Figure Description

[0046] Figure 1 This is a principal component analysis diagram of Daqu samples classified by fermentation type in the embodiments of this application;

[0047] Figure 2 This is a graph showing the test results of the physicochemical properties of the Daqu sample in the embodiments of this application;

[0048] Figure 3 This is a bar chart showing the LDA fraction distribution of microbial genera in the Daqu samples in this application embodiment;

[0049] Figure 4 This is a schematic diagram illustrating the changes in the accuracy of the model in the embodiments of this application;

[0050] Figure 5 This is a flowchart illustrating the fermentation type identification method for Daqu (a type of starter culture) in an embodiment of this application.

[0051] Figure 6 This is a flowchart illustrating the fermentation identification model for dismantled yeast in this embodiment of the application.

[0052] Figure 7 This is a schematic diagram of the fermentation type identification system for Daqu (a type of starter culture) in an embodiment of this application.

[0053] Figure 8 This is a first internal structure diagram of the computer device in an embodiment of this application;

[0054] Figure 9 This is a second internal structure diagram of the computer device in an embodiment of this application. Detailed Implementation

[0055] In order to make the purpose, technical solutions and advantages of this application more clear, the following further describes this application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0056] It should be noted that the illustrations provided in this embodiment are merely schematic representations of the basic concept of this application. The figures only show components relevant to this application and are not drawn according to the actual number, shape, and size of components in implementation. In actual implementation, the form, quantity, and proportion of each component can be arbitrarily changed, and the component layout may also be more complex. The structures, proportions, sizes, etc., shown in the accompanying drawings are only used to complement the content disclosed in the specification for those skilled in the art to understand and read, and are not intended to limit the implementation conditions of this application. Therefore, they have no substantial technical significance. Any modification to the structure, change in the proportional relationship, or adjustment of the size, without affecting the effect and purpose that this application can produce, should still fall within the scope of the technical content disclosed in this application. At the same time, the terms such as "upper," "lower," "left," "right," "middle," and "one" used in this specification are only for clarity of description and are not intended to limit the scope of implementation of this application. Changes or adjustments in their relative relationships, without substantially altering the technical content, should also be considered within the scope of implementation of this application.

[0057] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the document does not necessarily refer to the same embodiment, nor is it a mutually exclusive, independent, or alternative embodiment. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0058] As illustrated herein, unless the context clearly indicates otherwise, words such as “a,” “an,” “an,” and / or “the” do not specifically refer to the singular and may also include the plural. Generally speaking, the terms “comprising” and “including” only indicate the inclusion of explicitly identified steps and elements, which do not constitute an exclusive list, and the method or apparatus may also include other steps or elements.

[0059] The definitions used herein, such as the terms “having,” “may have,” “comprising,” or “may include,” indicate the presence of the corresponding function, operation, element, etc., and do not limit the presence of one or more other functions, operations, elements, etc. Furthermore, it should be understood that the terms “comprising” or “having” as used herein mean the presence of the features, figures, steps, operations, elements, components, or combinations thereof described in the specification, without excluding the presence or addition of one or more other features, figures, steps, operations, elements, components, or combinations thereof.

[0060] The prefixes such as "first" and "second" used in this application embodiment are merely for distinguishing different descriptive objects and do not limit the position, order, priority, quantity, or content of the described objects. The use of ordinal numbers and other prefixes used to distinguish descriptive objects in this application embodiment does not constitute a limitation on the described objects. The description of the described objects is given in the claims or the context of the embodiments, and should not constitute unnecessary limitations due to the use of such prefixes. Furthermore, in the description of this embodiment, unless otherwise stated, "multiple" means two or more.

[0061] This application aims to classify Daqu (a type of starter culture) by assessing the overall state of the microbial community within it, and trains a fermentation-type identification model.

[0062] In one embodiment, the training process of the above-mentioned fermentation identification model includes steps S101 to S104:

[0063] S101. Obtain microbial composition data samples from multiple Daqu (a type of Chinese liquor) samples.

[0064] In one embodiment, step S101 includes:

[0065] S1011, Obtain multiple large-scale samples;

[0066] S1012. Deoxyribonucleic acid was extracted and sequenced from each Daqu sample to obtain the corresponding metagenomic sequencing data sample.

[0067] S1013. Based on metagenomic sequencing data samples, microbial composition data samples were obtained for each Daqu sample.

[0068] For example, in step S1011, 172 large koji blocks (opened koji) from a soy sauce-flavored brewery are collected, including a total of 59 white koji blocks, 53 yellow koji blocks and 60 black koji blocks.

[0069] The outer skin of the Daqu (a type of koji) block was softened by immersing it in 5 mL of ultrapure water for 2 minutes. A small handheld drill equipped with a 5 cm radius and 2.2 cm height drilled into the Daqu block at three randomly selected points on its surface, drilling to a depth of 0-2.2 cm from the surface. The samples from the three points were then mixed thoroughly to obtain Daqu samples. These Daqu samples were obtained from Daqu blocks unpacked from different years and stored at -80℃.

[0070] In the brewing of Maotai-flavor liquor, the koji (fermentation starter) needs to be made into blocks and fermented in a warehouse for several days under high temperature. First, ultrapure water is used to soften the outer shell of the koji blocks, preventing excessively high temperatures from affecting the state of the microorganisms within.

[0071] In step S1012, a kit for extracting genomic DNA can be used to extract DNA (Deoxyribonucleic acid) from all Daqu samples. For example, the Magnetic Bead Fecal and Soil Genome Extraction Kit (MAGEN, Guangzhou, China) can be used, and extraction can be performed according to the corresponding manual instructions.

[0072] For example, take 100-200 mg of sample from Daqu (a type of Chinese liquor) into a centrifuge tube containing grinding beads, add 1 mL of Buffer ATL (a tissue lysis buffer) / PVP-10 (a polyvinylpyrrolidone), grind the sample using a high-speed grinder (Shanghai Jingxin Technology, China), incubate at 65°C for lysis for 20 min, then centrifuge at 14000 g for 5 min in an Eppendorf centrifuge (Germany), transfer the supernatant to a new centrifuge tube, add 0.6 mL of Buffer PCI (a reagent for nucleic acid purification), vortex mix for 15 s, centrifuge at 18213 g for 10 min, and then transfer the supernatant to a deep-well plate containing magnetic bead binding solution.

[0073] The deep-well plate can use Elution Buffer. The magnetic bead binding solution includes 600 μL of magnetic bead binding buffer, 20 μL of proteinase K, 5 μL of Ribonuclease A, 700 μL of Wash 1 (a high-salt wash buffer), 700 μL of Wash 2 (a low-salt wash buffer), 700 μL of Wash 3 (a wash buffer containing deionized water), and 100 μL of Elution Buffer (an elution buffer).

[0074] Launch the Kingfisher (Kingfisher, Thermo Fisher, USA) and select the corresponding program. Place each deep-well plate in its designated position on the instrument and run the program. After the program finishes, transfer the DNA solution from the deep-well plates to 1.5 mL centrifuge tubes for storage.

[0075] The extracted DNA solution is then subjected to metagenomic sequencing. Library preparation can be performed using the MGI Easy Universal DNA Preparation Kit (MGI-Shenzhen, China).

[0076] For example, a certain amount of genomic DNA is taken from a DNA solution and fragmented. The fragmented sample is then subjected to magnetic bead fragment selection. A reaction system is prepared and a reaction program is set to repair DNA ends and add an A base to the 3' end; an adapter ligation reaction system is prepared and a reaction program is set to ligate the adapter to the DNA. A PCR reaction system is prepared and a reaction program is set to amplify the product. The quality-checked library undergoes denaturation to obtain single-stranded DNA library molecules. Then, a circularization reaction system is used to obtain single-stranded circular products, and uncirculated linear DNA molecules are digested. The single-stranded circular DNA molecules undergo phi29 and rolling circle replication reactions to form a DNA nanosphere (DNB) containing multiple copies. The obtained DNBs are added to the mesh wells on a high-density DNA nanochip using high-density DNA nanochip technology, and PE100 sequencing is performed on the DNBSEQ-G400 sequencing platform using combined probe anchoring polymerization (cPAS). The sequencing results are used as metagenomic sequencing data samples.

[0077] In step S1013, Kraken2 and Bracken2 are used to perform species annotation and abundance calculation on the reads of metagenomic sequencing data samples to obtain microbial composition data samples.

[0078] That is, the microbial composition data sample includes the genera of microorganisms in the Daqu sample and the relative abundance among the genera.

[0079] S102. Based on the microbial composition data sample, determine the fermentation type label of the corresponding Daqu sample. The fermentation type label is used to characterize whether the Daqu sample is a first fermentation type or a second fermentation type.

[0080] In one embodiment, step S102 includes:

[0081] S1021. Based on the microbial composition data samples, determine the dissimilarity of microbial composition between every two Daqu samples;

[0082] S1022. Clustering based on the dissimilarity of microbial composition to determine the optimal number of clusters;

[0083] S1023. Determine the fermentation type label for each Daqu sample based on the optimal cluster number.

[0084] For example, in step S1021, the Bray-Curtis distance between every two Daqu samples can be calculated to measure the dissimilarity of microbial composition between Daqu samples and generate a distance matrix between Daqu samples.

[0085] The Bray-Curtis distance is a commonly used dissimilarity index in ecology and environmental science, used to measure the degree of difference between two samples in terms of species composition or other attributes. A larger Bray-Curtis distance between two samples indicates greater dissimilarity in their microbial composition.

[0086] In step S1022, the distance matrix is ​​clustered using the Partitioning around medoids (PAM) algorithm, and the optimal number of clusters is determined using the CH index (Calinski-Harabasz index) and / or the SI index (Silhouette index).

[0087] PAM is a classic clustering algorithm that aims to divide data into k clusters, each represented by a centroid, such that the sum of the distances from all data points to their respective cluster centroids is minimized. PAM is characterized by using actual data points as centroids, making it more robust to noise and outliers, and it is computationally efficient when the amount of data is small. The CH index and SI index are important indicators for evaluating clustering quality. They can be used to quantify the clustering performance under different numbers of clusters (k values) to help determine the optimal k value, thus obtaining the optimal number of clusters.

[0088] The PAM clustering algorithm and the calculation of the CH index and SI index can all be implemented using the toolkits "cluster" and "clusterSim" in the R language.

[0089] The CH index can measure the separation between clusters and the cohesion within clusters based on the ratio of inter-cluster variance to intra-cluster variance. For example, if the value of k is in the range of [2, 10], PAM clustering is performed based on the above distance matrix to obtain the clustering results corresponding to each k value. Based on the clustering results, the CH index corresponding to each k value can be calculated, and the k value with the largest CH index is selected as the optimal number of clusters.

[0090] The SI index can measure the reasonableness of clustering a single sample based on the distance difference from the sample to its own cluster and nearest neighbor cluster. Similarly, if the value of k is in the range of [2,10], the average SI index corresponding to each k value is calculated, and the k value with the largest average SI index is selected as the optimal number of clusters.

[0091] However, a single CH index or SI index may be biased due to data characteristics. For example, the CH index may overestimate the k value in high-dimensional data, while the SI index may fail in non-convex clustering. Therefore, the two can be used together to reduce the risk of misjudgment.

[0092] For example, if the peak values ​​of the CH index and the SI index occur at different k values, such as the peak value of the CH index occurring at k value 4 and the peak value of the SI index occurring at k value 3, the k value at which the sum of the CH index and the SI index is maximized can be selected as the optimal number of clusters.

[0093] The CH index assesses cluster structure from the perspective of overall variance, while the SI index provides a more detailed analysis from the perspective of individual samples. Combining the two can provide a more comprehensive assessment of cluster quality.

[0094] Based on the CH index and SI index, the optimal number of clusters is 2. Therefore, Daqu can be divided into two fermentation types according to its microbial composition. The two fermentation types can be named the first fermentation type Fet_1 and the second fermentation type Fet_2, respectively.

[0095] Table 1 shows the number of white koji, yellow koji, and black koji samples belonging to the first fermentation type and the second fermentation type, respectively.

[0096] Table 1. Distribution differences of white koji, yellow koji, and black koji with different fermentation types in Daqu samples.

[0097] Types of dismantling songs First fermentation type Fet_1 Second fermentation type Fet_2 Baiqu 39 20 Huangqu 10 43 Black Song 9 51

[0098] As shown in Table 1, the proportion of the first fermentation type Fet_1 is higher in white koji, at 66.1%, while the proportion of the second fermentation type Fet_2 is higher in yellow koji and black koji, at 81.1% and 85% respectively.

[0099] It is evident that the results of classifying Daqu (a type of starter culture) according to color are significantly different from the results of classifying it according to the two fermentation types proposed in this application.

[0100] In addition, to clarify the characteristics of the two fermentation types and the specific differences between them, this application also performed principal component analysis (PCoA) on the 172 Daqu samples classified by fermentation type, obtaining the following results: Figure 1 The principal component coordinate diagram is shown, and the physicochemical properties were tested according to national standards, yielding the following results: Figure 2 The test results are shown.

[0101] exist Figure 1 In the middle, the box plots at the top and right represent the distribution of Daqu samples along the PCo 1 and PCo 2 axes, respectively. Dark gray sample points represent Daqu samples from the first fermentation stage, and light gray sample points represent Daqu samples from the second fermentation stage. The solid and dashed ellipses represent the sample ranges of the 80% and 95% confidence intervals in the Daqu samples, respectively. The box plot below shows the distribution of white, yellow, and black Daqu along the PCo 1 axis.

[0102] from Figure 1 It is evident that the proportions of the two fermentation types proposed in this application in white koji, yellow koji, and black koji are significantly different.

[0103] exist Figure 2 The results show the differences in acidity, sugar (dry weight), moisture, starch (dry weight) content, and saccharification power between the first and second fermentation types of Daqu samples.

[0104] from Figure 2 It is evident that the Daqu samples belonging to the first fermentation type and the second fermentation type, respectively, showed significant differences in moisture content and saccharification power.

[0105] Furthermore, this application also performs functional annotation on the metagenomic sequencing data samples obtained above based on the KEGG and CAZy databases, and sums up the basic abundance of the annotated functions as functional abundance.

[0106] Based on functional annotations and functional abundance statistics, the carbohydrate-active enzymes (CAZymes) enriched in the Daqu samples of the first fermentation type include: CBM21, CBM34, CBM66, CE1, GH10, GH104, GH128, GH130, GH131, GH148, GH152, GH171, GH22, GH32, GH4, GH52, GH64, GH65, GH70, GH77, GH84, GH85, GT24, GT27, GT3, GT92, PL0, and PL3.

[0107] The carbohydrate-active enzymes enriched in the second fermentation type of Daqu sample include: AA2, AA4, CBM12, CBM26, CBM3, CBM32, CBM37, CBM43, CBM54, CBM6, CBM63, CBM87, CE11, CE12, CE8, GH102, GH126, GH15, GH18, GH25, GH26, GH3, GH38, GH39, GH46, GH48, GH72, GH8, GH81, GH94, GT112, GT14, GT21, GT47, GT50, GT51, GT59, GT60, GT61, GT62, GT71, GT84, GT90, PL42, PL7, and PL9.

[0108] Because the Daqu samples of the first fermentation type are enriched with starch hydrolytic enzymes (such as GH10, GH104, GH128, GH130, GH131) and some glycosyltransferases (such as GT3, GT24, GT27, GT92), they have strong starch degradation potential. In particular, GH10 (amylase) can effectively hydrolyze the large starch molecules in sorghum to produce fermentable small sugar molecules, providing an important carbon source for fermentation. It is also enriched with a variety of enzymes related to sugar conversion (such as GH4, GH52, GH64, etc.). These enzymes can further process the starch after conversion to generate fermentable sugars or other glycosylation products, which helps to improve the fermentation efficiency. This makes the Daqu of the first fermentation type have strong starch degradation ability and saccharification power.

[0109] In the second fermentation type of Daqu sample, starch hydrolases such as GH102, GH126, and GH15 were enriched. Although these enzymes have a slightly weaker ability to degrade starch compared to GH10 and GH104 in Fet_1, other CAZymes enriched there, such as CBM12, CBM26, and CBM3, can also help microorganisms to better bind and hydrolyze polysaccharides in sorghum, especially on more complex polysaccharide substrates (such as cellulose and hemicellulose), where they may show better performance.

[0110] In addition, Fet2 also has advantages in the hydrolysis and fermentation of non-starch sugars (such as mannose, glucose, etc.).

[0111] Therefore, Fet_1's advantage mainly lies in its stronger starch degradation ability, especially the enrichment of amylases such as GH10, GH104, and GH128, which enables it to decompose starch more efficiently during sorghum fermentation and produce the small-molecule sugars required for fermentation. Fet_2, on the other hand, performs better in a wider range of sugar substrate conversions and non-starch polysaccharide hydrolysis, making it suitable for more complex fermentation substrate treatments.

[0112] Based on the results of the physicochemical property tests, it can be seen that Fet_1 has a stronger starch degradation capacity and saccharification power than Fet_2, and is especially suitable for starch conversion and alcoholic fermentation of sorghum.

[0113] Furthermore, Fet_1 and Fet_2 koji each have their own advantages and complement each other, which can jointly promote the fermentation process of sorghum. In addition, the two fermentation types of koji can be reasonably matched according to production needs, thereby improving production efficiency.

[0114] In step S1023, after determining the optimal number of clusters to be 2, the fermentation type of each Daqu sample can be obtained based on the corresponding clustering results, and then the fermentation type label of each Daqu sample can be obtained.

[0115] S103. Determine the key microbial genera based on the microbial composition data sample, and obtain the key microbial data sample of the Daqu sample.

[0116] In one embodiment, identifying key microbial genera based on a microbial composition data sample includes:

[0117] S1031. Analyze multiple microbial composition data samples to obtain analysis results, including evaluation indicators for different fermentation types based on microbial genera.

[0118] S1032. Based on the evaluation indicators, determine the key microbial genera from the microbial genera. The key microbial genera include the first key microbial genera corresponding to the first fermentation type and the second key microbial genera corresponding to the second fermentation type.

[0119] For example, in step S1031, all 172 large-batch samples can be analyzed using LEfSe (Linear Discriminant Analysis Effect Size).

[0120] LEfSe is a statistical analysis method commonly used in fields such as microbiome and ecology. It is mainly used to identify features that show significant differences between different groups, assess the size of their effects, and then discover and interpret the biomarkers.

[0121] LEfSe analysis mainly consists of three steps:

[0122] (1) Kruskal-Wallis rank-sum test: to detect species with significant differences in abundance among different groups.

[0123] (2) Wilcoxon rank-sum test: further examine whether species with significant differences converge between subgroups.

[0124] (3) Linear discriminant analysis (LDA): Dimensionality reduction of the data and assessment of the influence of species with significant differences, i.e., the evaluation index LDA score.

[0125] See also Figure 3 The LDA score distribution histogram shows the species with significant differences in LDA scores greater than 3.5 between the first fermentation type Fet_1 and the second fermentation type Fet_2. The length of the histogram represents the size of the species' influence.

[0126] Therefore, the biomarkers between the first fermentation type Fet_1 and the second fermentation type Fet_2 can be identified as key microbial genera based on the LDA scores.

[0127] from Figure 3It can be seen that the main enriched microorganisms in Fet_1 include Bacillus, Pantoea, Duffyella, Pediococcus, Kosakonia, Enterococcus, and Lactiplanti Bacillus, while the main enriched microorganisms in Fet_2 include LentiBacillus, Kroppenstedtia, Siminovitchia, and Penicillium.

[0128] In step S1032, the microbial genera with the highest evaluation index can be selected as the corresponding key microbial genera for the first fermentation type and the second fermentation type, respectively.

[0129] For example, in one embodiment, the key microbial genus includes a first key microbial genus corresponding to a first fermentation type and a second key microbial genus corresponding to a second fermentation type. The first key microbial genus includes Bacillus, and the second key microbial genus includes LentiBacillus.

[0130] Alternatively, microbial genera with high evaluation indicators can be selected as key microbial genera. For example, in another embodiment, the first key microbial genera includes, in addition to Bacillus, at least one of Pantoea, Duffyella, Pediococcus, Kosakonia, Enterococcus, and Lactiplanti Bacillus. The second key microbial genera includes, in addition to Lenti Bacillus, at least one of Kroppenstedtia, Siminovitchia, and Penicillium.

[0131] S104. Based on key microbial data samples and fermentation type labels from multiple Daqu samples, a fermentation type identification model is trained to obtain the model.

[0132] The key microbial data samples of the aforementioned Daqu (a type of fermented bean curd) samples can be divided into training, validation, and test sets. The training set is used for model training and learning, the validation set is used to adjust parameters and monitor model performance during training, and the test set is used to evaluate the final performance of the model.

[0133] For example, a neural network model (NeuralNetTorch) built using the PyTorch framework is trained using a training set. Key microbial data samples of Daqu (a type of starter culture) are input into the model to obtain the model's prediction results. The prediction results are then compared with the corresponding fermentation type labels to calculate the prediction loss, and the model parameters are adjusted based on the prediction loss.

[0134] The model's predictions can be summarized into a confusion matrix to visually reflect the model's correct and incorrect classifications. Various performance metrics, such as accuracy, precision, recall, F1 score, and area under the curve (AUC), can be calculated to reflect the model's fitting ability on the training set.

[0135] Then, the validation set is used to evaluate the model's generalization ability, adjust the model parameters, and select the optimal hyperparameters. The validation set evaluation helps detect whether the model is overfitting or underfitting.

[0136] Finally, the model was evaluated on unseen data using a test set. The results of the test set reflect the model's performance in real-world applications, especially its generalization ability.

[0137] Based on the results from the training set, validation set, and test set, the performance of the model can be comprehensively evaluated, ensuring the consistency and stability of the model across different datasets.

[0138] By comparing performance metrics such as accuracy, F1 score, AUC, precision, and recall, one can gain a comprehensive understanding of the model's strengths and weaknesses in classification tasks.

[0139] Alternatively, different numbers of biomarkers can be selected as key microbial genera for model training, and the group with the best performance can be selected as the key microbial genera.

[0140] In order to compare the impact of different biomarkers on the model, this application uses the top 1, top 2, top 3... top 10 biomarkers with the highest LDA scores as key microbial genera. Based on the identified key microbial genera, the relative abundance of key microbial genera is obtained from the microbial composition data, and key microbial genera data samples are obtained for model training.

[0141] The key microbial data samples include the key microbial genera present in the Daqu samples and their relative abundance.

[0142] When the top 1-10 biomarkers with the highest LDA scores are used as key microbial genera, the model's accuracy varies across the training, validation, and test sets as follows: Figure 4 As shown.

[0143] from Figure 4 It is evident that when the top 3, 7, and 8 biomarkers with the highest LDA scores are used as key microbial genera, the accuracy of the trained model is good in the training, validation, and test sets.

[0144] Therefore, in one embodiment, the first key microbial genus includes Bacillus, and the second key microbial genus includes LentiBacillus and Kroppenstedtia.

[0145] In another embodiment, the first key microbial genera include Bacillus, Pantoea, Duffyella, Pediococcus, and Kosakonia, and the second key microbial genera include LentiBacillus and Kroppenstedtia.

[0146] The performance evaluation results of the fermentation identification model trained using the above 7 key microbial genera and their relative abundance as key microbial data samples are shown in Table 2.

[0147] Table 2 Performance evaluation results of the fermentation type identification model

[0148] Dataset accuracy F1 score AUC Accuracy Recall rate training set 0.989247312 0.984615385 0.984848485 1 0.96969697 Validation set 0.975609756 0.965517241 0.981481481 0.933333333 1 test set 0.973684211 0.952380952 0.954545455 1 0.909090909

[0149] According to Table 2, the performance of the trained fermentation identification model on the training set is as follows:

[0150] Accuracy: 0.989, demonstrating extremely high classification accuracy.

[0151] The F1 score of 0.985 indicates that the model achieves a good balance between precision and recall.

[0152] The AUC of 0.985 indicates that the model is very strong in distinguishing between different categories (Fet_1 and Fet_2).

[0153] Precision: 1 indicates that all samples predicted as Fet_1 on the training set by the model are true classes and there are no misclassifications.

[0154] Recall: 0.970, which means that the model has a very high ability to identify Fet_1 category, but the recall rate drops slightly when recalling Fet_2 category samples.

[0155] Overall, the model demonstrates extremely high accuracy on the training set, indicating that the model fits the training data very well.

[0156] As shown in Table 2, the trained fermentation identification model performs as follows on the validation set:

[0157] Accuracy: 0.976, indicating that the model still performs very well on unseen data.

[0158] The F1 score is 0.966, indicating that the model maintains a good balance in performance on the validation set.

[0159] AUC: 0.981, which validates the model's ability to distinguish between different fermentation types (Fet_1 and Fet_2).

[0160] Precision: 0.933. The precision has decreased, indicating that some Fet_1 samples may have been misclassified as Fet_2.

[0161] Recall: 1. The recall rate for the Fet_1 category on the validation set is perfect, indicating that the model can effectively identify all samples of the Fet_1 category.

[0162] The performance on the validation set was slightly lower than that on the training set, especially in terms of precision, but the recall remained perfect, indicating that the model has good generalization ability on different datasets.

[0163] According to Table 2, the performance of the trained fermentation identification model on the test set is as follows:

[0164] Accuracy: 0.974. The model's performance on the test set is similar to that on the training and validation sets, demonstrating its good generalization ability.

[0165] F1 score: 0.952, slightly lower than the validation set, but still at a high level.

[0166] The AUC of 0.955 indicates that the model can still effectively distinguish Fet_1 from Fet_2 on the test set.

[0167] Precision: 1, all Fet_1 predictions on the test set are Fet_1, demonstrating that the model has high accuracy in this task.

[0168] Recall: 0.909. The recall on the test set is low, mainly in the Fet_2 category, indicating that the model is slightly inadequate in detecting Fet_2 category samples.

[0169] The test set results show that the model has strong robustness and wide applicability. Although the recall rate has decreased, the accuracy remains at a high level.

[0170] Table 3 shows the confusion matrix of the model on the training, validation, and test sets, which is used to further analyze the model's classification performance.

[0171] Table 3. Confusion matrix of the model on the training, validation and test sets.

[0172]

[0173] From the confusion matrix of the training set, we can see that:

[0174] On the training set, among the samples that were actually Fet_1, all 60 samples were correctly predicted as Fet_1, and among the 32 samples that were actually Fet_2, only 1 was mispredicted as Fet_1, while the other 31 samples were accurately predicted as Fet_2. This indicates that the model can perfectly distinguish between the Fet_1 and Fet_2 categories on the training set, with an extremely low misclassification rate.

[0175] From the confusion matrix of the validation set, we can see that:

[0176] On the validation set, among the samples that were actually Fet_1, 26 samples were correctly predicted as Fet_1, and 1 sample was mispredicted as Fet_2; among the samples that were actually Fet_2, 14 samples were correctly predicted as Fet_2, with no misclassification. The validation set performance remains good, and the model has strong reliability in distinguishing between Fet_1 and Fet_2 categories.

[0177] From the confusion matrix of the test set, we can see that:

[0178] On the test set, among the samples that were actually Fet_1, 27 samples were correctly predicted as Fet_1, and 1 sample was mispredicted as Fet_2; among the samples that were actually Fet_2, 10 samples were correctly predicted as Fet_2, and 1 sample was mispredicted as Fet_1. The confusion matrix results on the test set indicate that the model is very accurate in predicting the Fet_1 category, but still has a few misclassifications in the Fet_2 category.

[0179] As shown above, the model performs exceptionally well on the training, validation, and test sets, especially on the training and test sets, where accuracy exceeds 97% and precision is 1, indicating almost no error in predicting these two fermentation types. This demonstrates that the model can effectively distinguish between different fermentation types of dismantled yeast.

[0180] Recall: Recall varies across the training, validation, and test sets, with the recall for the Fet_2 class being slightly insufficient, particularly on the test set. This may be related to the relatively small sample size for the Fet_2 class and data imbalance, suggesting that future model optimizations could improve recall for the Fet_2 class through data augmentation or adjusting the classification threshold.

[0181] Generalization ability: Performance on the validation and test sets demonstrates that the model has strong generalization ability and can adapt to classification tasks with unseen data. In particular, the model maintains high levels of AUC and F1 scores across different datasets, proving its good stability and robustness in practical applications.

[0182] In summary, the model proposed in this application demonstrates high classification performance on the training, validation, and test sets, effectively distinguishing between the fermentation types of *Fet_1* and *Fet_2*. The model exhibits excellent precision, recall, and F1 score, and possesses strong generalization ability. Although the recall rate slightly decreased, the overall performance remains excellent, indicating that this model has broad application potential in practical production and scientific research.

[0183] In another embodiment, the first key microbial genera include Bacillus, Pantoea, Duffyella, Pediococcus, and Kosakonia, and the second key microbial genera include LentiBacillus, Kroppenstedtia, and Siminovitchia.

[0184] This application focuses on the core characteristic of Daqu—the fermentation power of microorganisms. Different fermentation types have unique microbial species composition, functional potential, microbial interaction relationships, metabolic capacity, etc. This application proposes and defines the "fermentation type" of Daqu for the first time, and divides the high-latitude complex composition of Daqu into two fermentation types.

[0185] "Fermentation type" is based entirely on the composition of microbial species, representing a unique microbial species structure and functional potential. In contrast, the traditional classification scheme based on color is a description of the physicochemical properties of Daqu itself. According to Table 1 mentioned above in this application, it can be found that even if two pieces of Daqu are white (or yellow or black), they may correspond to two completely different microbial compositions and functional characteristics.

[0186] Therefore, the classification scheme based on "fermentation type" Daqu proposed in this application is entirely based on the microbial species composition structure, reflecting the overall microbial species composition, functional potential, microbial community cooperation network, and microbial community metabolic characteristics of Daqu. It can be applied to different fermentation types of Daqu according to different needs.

[0187] In particular, for Daqu (a type of starter culture) dismantled from storage, the classification scheme is more scientific, objective, and of greater significance to production compared to the traditional method of classifying Daqu into white, yellow, and black Daqu based on color.

[0188] Furthermore, the concept of fermentation type in Daqu (a type of starter culture) applies not only to Daqu after its initial fermentation but also to Daqu at other stages, and can be widely used in any study involving comparisons between groups. For example, comparisons between different seasons, locations, and temperature zones. By describing the proportional changes in the fermentation types of Daqu under different groups, the differences in Daqu microorganisms can be more intuitively demonstrated, thereby achieving a description of the overall microbial community differences based on Daqu.

[0189] Furthermore, with in-depth exploration of the biomarkers, functions, networks, and metabolic characteristics of Daqu fermentation, it will be possible to conduct more in-depth and focused analyses of the microbial differences in Daqu, thereby providing new theoretical basis and practical guidance for the optimization and control of the Daqu fermentation process.

[0190] Based on the above definition and concept of "fermentation type", this application also provides a method for identifying the fermentation type of Daqu, which uses a trained model that can identify the fermentation type of Daqu, and automatically outputs the fermentation type of Daqu to be identified by acquiring key microbial data of Daqu to be identified.

[0191] In one embodiment, the fermentation type identification method of Daqu is as follows: Figure 5 As shown, steps S201 to S202 are included:

[0192] S201. Obtain key microbial data in the Daqu (a type of Chinese liquor) to be identified;

[0193] S202. Input the key microbial data into the pre-trained fermentation type identification model to obtain the fermentation type identification results of the Daqu to be identified.

[0194] In one embodiment, step S201 includes:

[0195] S2011. Deoxyribonucleic acid (DNA) was extracted and sequenced from the Daqu (a type of Chinese liquor) to be identified, and the corresponding metagenomic sequencing data was obtained.

[0196] S2022. Based on metagenomic sequencing data, obtain the microbial composition data of the Daqu to be identified. The microbial composition data includes the genera of microorganisms present in the Daqu sample to be identified and the relative abundance among the genera.

[0197] S2023. Obtain key microbial genera and their relative abundance from the microbial composition data, as key microbial data.

[0198] The process of deoxyribonucleic acid extraction and sequencing is described in the previous text and will not be repeated here.

[0199] If the Daqu to be identified is a type of dismantled Daqu, see [link to relevant documentation]. Figure 6 Similar to the training process of the fermentation identification model described above, the key microbial data samples of the Daqu sample prepared by dismantling the Daqu can be obtained and the fermentation identification model can be trained.

[0200] The key microbial data of the Daqu to be identified are input into the trained fermentation identification model to obtain the fermentation type identification result of the Daqu to be identified. The fermentation type identification result is used to characterize whether the Daqu to be identified is a first fermentation type or a second fermentation type. The first fermentation type has a stronger starch degradation ability and saccharification power than the second fermentation type.

[0201] It should be understood that, although Figure 5 The steps in the flowchart are shown sequentially as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise explicitly stated herein, there is no strict order in which these steps are executed, and they can be performed in other orders. Figure 5 At least some of the steps in the process may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least some of the sub-steps or stages of other steps.

[0202] To implement the above-mentioned method for identifying the fermentation type of Daqu (a type of starter culture), this application also provides a computer program product, including a computer program / instructions, which, when executed by a processor, implement the steps of the above-described method for identifying the fermentation type of Daqu.

[0203] In one embodiment, this computer program product presents itself as a fermentation-type identification system for koji (a type of Chinese liquor).

[0204] like Figure 7 As shown, the fermentation type identification system for Daqu (a type of starter culture) includes:

[0205] Data acquisition module 301 is used to acquire key microbial data in the Daqu (a type of starter culture) to be identified;

[0206] The fermentation type identification module 302 is used to input key microbial data into a pre-trained fermentation type identification model to obtain the fermentation type identification result of the Daqu to be identified.

[0207] Specific limitations regarding the fermentation type identification system can be found in the limitations of the fermentation type identification method described above, and will not be repeated here. Each module in the aforementioned fermentation type identification system can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the corresponding operations of each module.

[0208] This application also provides a computer device. In one embodiment, the computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of the fermentation identification method described in the above embodiments.

[0209] In one embodiment, the computer device may be a server, and its internal structure diagram may be as follows: Figure 8 As shown, the computer device includes a processor, memory, network interface, and database connected via a system bus. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The database stores relevant data for fermentation pattern identification. The network interface communicates with external terminals via a network connection. When the computer program is executed by the processor, it implements the steps of the fermentation pattern identification method described in the above embodiments.

[0210] In one embodiment, the computer device may be a terminal, and its internal structure diagram may be as follows: Figure 9 As shown, the computer device includes a processor, memory, network interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The network interface is used to communicate with external terminals via a network connection. When the computer program is executed by the processor, it implements the steps of the fermentation identification method described in the above embodiments. The display screen can be a liquid crystal display (LCD) or an electronic ink display. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad mounted on the computer device casing, or an external keyboard, touchpad, or mouse.

[0211] Those skilled in the art will understand that Figure 8and Figure 9 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0212] This application also provides a computer-readable storage medium storing a computer program / instructions thereon, which, when executed by a processor, implements the steps of the fermentation identification method in the above embodiments.

[0213] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0214] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0215] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims

1. A method for identifying the fermentation type of Daqu (a type of starter culture), characterized in that, The fermentation type identification method includes: Acquire key microbial data from the koji (fermented liquor) to be identified; The key microbial data is input into a pre-trained fermentation type identification model to obtain the fermentation type identification result of the Daqu to be identified.

2. The fermentation type identification method as described in claim 1, characterized in that, The training process of the fermentation-type recognition model includes: Obtain microbial composition data samples from multiple Daqu (fermented koji) samples; Based on the microbial composition data sample, determine the corresponding fermentation type label for the Daqu sample; Based on the microbial composition data sample, the key microbial genera were determined, and the key microbial data sample of the Daqu sample was obtained. The fermentation type identification model is obtained by training a model based on the key microbial data samples and fermentation type labels of multiple Daqu samples.

3. The fermentation type identification method as described in claim 2, characterized in that, The step of determining the fermentation type label of the corresponding Daqu sample based on the microbial composition data sample includes: Based on the microbial composition data sample, determine the dissimilarity of microbial composition between every two Daqu samples; Clustering is performed based on the dissimilarity of the microbial composition to determine the optimal number of clusters; The fermentation type label for each of the Daqu samples is determined based on the optimal cluster number; and / or, The step of determining key microbial genera based on the microbial composition data sample includes: The microbial composition data samples of multiple samples are analyzed to obtain analysis results, which include evaluation indicators for distinguishing different fermentation types based on microbial genera. The key microbial genera are determined from the microbial genera according to the evaluation indicators.

4. The fermentation type identification method as described in claim 2 or 3, characterized in that, The Daqu sample was prepared from dismantled Daqu. The fermentation type identification result is used to characterize whether the Daqu to be identified is a first fermentation type or a second fermentation type. The first fermentation type has a stronger starch degradation ability and saccharification power than the second fermentation type.

5. The fermentation type identification method as described in claim 4, characterized in that, The key microbial genera include the first key microbial genera corresponding to the first fermentation type and the second key microbial genera corresponding to the second fermentation type; The first key microbial genus includes Bacillus; The second key microbial genus includes the genus *Bacillus*.

6. The fermentation type identification method as described in claim 5, characterized in that, The first key microbial genus also includes at least one of the following: Pantotheca, Phyllostachys, Pediococcus, Micrococcus, Enterococcus, and Lactobacillus. and / or, The second key microbial genus also includes at least one of the genera *Croppensteadella*, *Ciminovicella*, and *Penicillium*.

7. The fermentation type identification method as described in claim 2, characterized in that, The acquisition of microbial composition data samples from multiple Daqu (fermented koji) samples includes: Obtain multiple large-scale samples; Deoxyribonucleic acid (DNA) was extracted and sequenced from each of the Daqu samples to obtain the corresponding metagenomic sequencing data samples. Based on the metagenomic sequencing data samples, the microbial composition data samples of each Daqu sample are obtained, and the microbial composition data samples include the microbial genera in the Daqu sample and the relative abundance among the microbial genera; and / or, The acquisition of key microbial data in the Daqu (a type of starter culture) to be identified includes: The deoxyribonucleic acid of the Daqu to be identified was extracted and sequenced to obtain the corresponding metagenomic sequencing data; Based on the metagenomic sequencing data, the microbial composition data of the Daqu to be identified is obtained, and the microbial composition data includes the microbial genera present in the Daqu sample to be identified and the relative abundance among the microbial genera. The key microbial genera and their relative abundance are obtained from the microbial composition data and used as the key microbial data.

8. A computer program product comprising a computer program / instructions, characterized in that, When the computer program / instructions are executed by the processor, they implement the steps of the method according to any one of claims 1-7.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1-7.

10. A computer-readable storage medium having a computer program / instructions stored thereon, characterized in that, When the computer program / instructions are executed by the processor, they implement the steps of the method described in any one of claims 1-7.

Images