Fast prediction method of reference genetic distance between different mammals based on infrared spectrum of milk

Abstract

This invention belongs to the field of spectral applications and mammalian evolution, specifically relating to a rapid prediction method for reference genetic distances between different mammals based on the mid-infrared spectroscopy (MIRS) of milk. The applicant has, for the first time, used milk MIRS to predict reference genetic distances between different mammalian populations, establishing a regression model that can predict the genetic distances between new populations of known mammals, and even new populations of unknown mammalian species, and known mammalian populations, thereby providing a preliminary assessment of evolutionary differences between mammalian populations. The method exhibits high accuracy and clear trends, approaching the results obtained from proteomics or other omics analyses.

Description

Technical Field

[0001] This invention belongs to the field of spectral applications and mammalian evolution, specifically involving a rapid prediction method for reference genetic distances between different mammals based on the mid-infrared spectrum of milk. Background Technology

[0002] To meet the need for timely and rapid testing of animal milk and dairy products, infrared spectroscopy (IRS) was first used in 1967 to detect the fat percentage (FP), protein percentage (PP), and lactose percentage (LP) in milk. [1] Mid-infrared spectroscopy (MIRS) is the fundamental absorption band of organic matter and inorganic ions. The position, shape, number, and intensity of characteristic MIRS peaks in animal milk can reflect the composition and content of milk, making it particularly suitable for the analysis of liquid milk. [2-3] MIRS testing technology is widely used in milk composition testing and global dairy herd improvement (DHI) testing due to its advantages such as low cost, simplicity, speed, batch processing, accuracy, and non-destructive nature.

[0003] The composition of milk varies considerably among different species of mammals (collectively referred to as animals). Horse and donkey milk, belonging to the order Perissodactylus, has low FP (1.21%) and high LP (6.37%). [4-6] The FP and PP (6% and 4%) in camel milk (even-toed ungulates) are lower than those in pigs (9.97% and 5.25%), while the PP in pig colostrum is as high as 17%. [7-8] The FP and PP (5.3-9.3% and 4.5-6.6%) of goat and sheep milk are both higher than those of dairy cows. [9-10] Buffalo milk has higher FP and PP (8.5% and 4.5%) than Holstein milk, and also has a higher fatty acid content. [11-12] The concentrations of FP, PP, and calcium in yak milk are 1.68-2 times higher than those in Holstein milk.

[13] Jersey and Simmental milk had higher percentages of total protein (PP) (3.32%-3.61%) than Holstein milk, while Holstein milk had percentages of total fat (FP) and PP of no less than 3.4% and 2.9%, respectively. [14-15] .

[0004] The MIRS (Mild Intervals and Reductions) also varies considerably among different animals. The MIRS for goat milk, sheep milk, and cow milk is around 29-27 cm⁻¹. -1 There are differences in the CH2 absorption bands at different locations; the MIRS absorption peak increases with increasing FP. Sheep milk PP (1654 cm⁻¹) -1 Nearby amide I and 1544cm -1Nearby amide II) and LP (1159cm) -1 and 1076cm -1 The associated MIRS absorption peak is higher than that of cow's milk and goat's milk.

[16] Currently, only the MIRS bands corresponding to the main milk components have been studied, and more research on characteristic MIRS bands and their applications has not yet been carried out.

[0005] The wave points of the MIRS (Milk Injection Reduction Syndrome) are heritable. The heritability of the 1060 wave points of the MIRS ranges from 0.003 to 0.42, similar to the estimated values ​​for milk-related traits, indicating that milk MIRS, like milk components, is heritable.

[17] The MIRS polation pattern exhibits moderate to high heritability (0-0.63), suggesting that genes influence MIRS polation by regulating milk composition.

[18] The heritability of MIR polation points in Holstein and Jersey cattle ranged from 0 to 0.31, and MIR polation points were significantly correlated with the genome.

[19] Our team's previous research indicated that the heritability of the 1060 spots in MIRS averaged 0.04, and was similar to the inheritance patterns of FP, PP, and LP.

[20] .

[0006] Regression models based on MIRS can be established through machine learning, including feature selection for MIRS and training, testing, and optimization of a preliminary model. MIRS features are selected using methods such as Principal Component Analysis (PCA) and Uninformative Variable Elimination (UVE), and regression models are built using methods such as Partial Least Squares Regression (PLSR). The model is then determined based on the coefficient of determination (R²). 2 The optimal model is selected using methods such as root mean square error (RMSE).

[0007] In summary, the composition and MIRS (Mutual Reference Distance) of milk vary among different animals. MIRS not only reflects milk composition characteristics but also possesses heritability, thus suggesting that using MIRS to construct animal evolutionary relationships is feasible. Currently, animal evolution is mainly analyzed using genome sequencing, while research on MIRS in animal evolutionary relationships is limited and lacks literature reports. MIRS technology has many advantages, including low cost, speed, batch processing, accuracy, simplicity, and non-destructive nature. Therefore, the main objective of this invention is to break through the traditional thinking of relying on genetic material to study animal evolutionary relationships and to create a new rapid prediction technology for reference genetic distances between different mammals based on MIRS. Summary of the Invention

[0008] The purpose of this invention is to provide a rapid prediction method for reference genetic distances between different mammals based on the mid-infrared spectrum of milk. This method is simple and highly accurate.

[0009] Another objective of this invention is to provide a rapid prediction method for reference genetic distances between different mammals based on the mid-infrared spectrum of milk.

[0010] To achieve the above objectives, the present invention adopts the following technical measures:

[0011] A rapid prediction method for reference genetic distances between different mammals based on the mid-infrared spectrum of milk includes the following steps:

[0012] 1) Analyze each milk sample from the N animal species to be tested to obtain MIRS data, and obtain the average MIRS for each animal species. The 62 effective bands selected in the MIRS values ​​are:

[0013] The 62 MIR bands mentioned are: 929.778cm -1 —949.068cm -1 983.79cm -1 —999.222cm -1 1022.37cm -1 —1033.944cm -1 1057.092cm -1 —1084.098cm -1 1122.678cm -1 —1126.536cm -1 1138.11cm -1 —1149.684cm -1 1157.4cm -1 —1168.974cm -1 1184.406cm -1 —1192.122cm -1 1207.554cm -1 —1211.412cm -1 1234.56cm -1 —1242.276cm -1 1276.998cm -1 —1288.572cm -1 1300.146cm -1 —1304.004cm -1 1331.01cm -1 —1334.868cm -1、1338.726cm -1 —1381.164cm -1 、1408.17cm -1 —1427.46cm -1 、1462.182cm -1 —1469.898cm -1 、1512.336cm -1 —1531.626cm -1 、1701.378cm -1 —1705.236cm -1 、1766.964cm -1 —1790.112cm -1 、1813.26cm -1 —1817.118cm -1 、1820.976cm -1 —1840.266cm -1 、1859.556cm -1 —1901.994cm -1 、1959.864cm -1 —1975.296cm -1 、2033.166cm -1 —2052.456cm -1 、2118.042cm -1 —2148.906cm -1 、2164.338cm -1 —2179.77cm -1 、2195.202cm -1 —2214.492cm -1 、2249.214cm -1 —2260.788cm -1 、2295.51cm -1 —2322.516cm -1 、2330.232cm -1 —2334.09cm -1 、2349.522cm -1 —2361.096cm -1 、2372.67cm -1 —2388.102cm -1 、2415.108cm -1 —2445.972cm -1 、2465.262cm -1 —2469.12cm -1、2484.552cm -1 —2488.41cm -1 、2511.558cm -1 —2538.564cm -1 、2569.428cm -1 —2581.002cm -1 、2592.576cm -1 —2596.434cm -1 、2627.298cm -1 —2646.588cm -1 、2704.458cm -1 —2708.316cm -1 、2719.89cm -1 —2723.748cm -1 、2739.18cm -1 —2750.754cm -1 、2758.47cm -1 —2773.902cm -1 、2797.05cm -1 —2812.482cm -1 、2820.198cm -1 —2827.914cm -1 、2835.63cm -1 —2885.784cm -1 、2908.932cm -1 —2916.648cm -1 、2928.222cm -1 —2932.08cm -1 、2943.654cm -1 —2955.228cm -1 、3645.81cm -1 —3668.958cm -1 、3676.674cm -1 —3680.532cm -1 、3684.39cm -1 —3703.68cm -1 、3726.828cm -1 —3738.402cm -1 、3784.698cm -1 —3792.414cm -1 、3819.42cm -1 —3854.142cm -13877.29cm -1 —3881.148cm -1 3892.722cm -1 —3896.58cm -1 3904.296cm -1 —3908.154cm -1 3935.16cm -1 —3939.018cm -1 3942.876cm -1 —3973.74cm -1 3981.456cm -1 —3989.172cm -1 3996.888cm -1 —4000.746cm -1 ;

[0014] 2) Subtract the absolute values ​​of the average MIRS of N animals to obtain the subtracted MIRS, resulting in a total of N*(N-1) / 2 subtracted spectra. Substitute these N*(N-1) / 2 subtracted spectra into the model D1+SNV+PLSR(n_components=14) to predict the corresponding genetic distance; where N≥2.

[0015] In the methods described above, preferably, the animal is a horse, camel, cow, goat, donkey, or pig.

[0016] In the methods described above, preferably, the cattle are buffalo, yaks, Holstein cows, Simmental cows, Jersey cows, or Xinjiang brown cattle; and the sheep are goats or sheep.

[0017] The scope of protection of this invention also includes the application of the method provided by this invention in predicting the aforementioned genetic distances among mammals. Compared with the prior art, the advantages of this invention are:

[0018] 1. For the first time, milk MIRS was used to predict reference genetic distances between different mammalian populations. The results showed high accuracy and clear trends, and were close to those obtained from proteomics or other omics analyses.

[0019] 2. Based on the reference genetic distance established by the whole genome, a regression model was established to predict the genetic distance between new populations of known mammals in nature, or even new populations of unknown mammal species, and known mammal populations, thereby making a preliminary judgment on the evolutionary differences between mammal populations.

[0020] 3. The optimal preprocessing and algorithm combination for establishing the RMGD model was selected, the optimal parameters were determined, and the accuracy of the model was improved. 1. Regarding the selection of feature bands, a combination of algorithmic feature selection and manual adjustment was used to finally select the wave points or bands for modeling.

[0021] 4. It enables rapid, accurate, and low-cost prediction of genetic distance in mammals, and allows for rapid batch testing of animal populations. Each animal test takes only 10-15 seconds, and milk samples from about 30 animals can accurately predict their genetic distance from other mammals. The total prediction time for each species does not exceed 10 minutes, and it can be widely applied in the field of mammalian evolution. Attached Figure Description

[0022] Figure 1 The average MIRS of 12 animals and the MIRS absorption regions of 3 major milk components are shown in (a), and the TSNE dimensionality reduction clustering diagram of the MIRS of 12 animals is shown in (b).

[0023] Figure 2 This refers to the evolutionary hierarchy (phylogenetic tree) of animals based on the whole genome.

[0024] Figure 3 The correlation between RMGD-based genetic distances of eight animal species and genome-wide genetic distances of eight animal species: (a) The correlation between the genetic distances of the eight animal species that did not participate in the modeling and the genome-wide genetic distances of the eight animal species; (b) The genetic distances of four animal species and their correlation with genome-wide genetic distances; (c) Detailed implementation method:

[0025] Unless otherwise specified, the technical solutions described in this invention are all conventional solutions in the field; unless otherwise specified, the reagents or materials described are all from commercial sources.

[0026] 1. Experimental Materials

[0027] Milk samples (all regular milk) and porcine colostrum (Table 1) from 12 animal species were collected from 15 provinces and regions in China, totaling 10,284 samples. Following the DHI milk collection and testing technical procedures, 30-50 ml of milk was collected using a flow meter or manually (ensuring thorough mixing throughout the milking process). A small amount of bromonitrobenzene glycol preservative was added to the collected milk sample, dissolved, and mixed thoroughly. The milk samples were then rapidly transported to the DHI laboratory at 2-4°C.

[0028] Table 1 Information on the samples used in this invention

[0029]

[0030]

[0031] 2. Mid-infrared spectroscopy measurement and acquisition

[0032] The sample was poured into a cylindrical sample tube with a diameter of 3.5 cm and a height of 9 cm, and incubated in a water bath at 42℃ for 15-20 min. Sheep milk was measured using a Bentley NexGen-400 milk composition analyzer (Minnesota, USA), and 12 kinds of animal milk were measured using a FOSS Milkoscan FT+ milk composition analyzer (Hilleroed, Denmark). (Due to the high dry matter content of pig colostrum, it was diluted with water at a ratio of 1:1 before measurement. The solid fiber optic probe was inserted into the liquid, and the sample was mixed and scanned.)

[0033] 3. Selection of Valid Samples

[0034] Sixty abnormal samples were removed, including those with missing or incomplete MIRS data, Mahalanobis distance (GH) ≥ 3 based on MIRS, and DHI-determined contents of the four main milk components below 0%, FP > 25%, PP > 19%, LP > 7.5%, and TS > 38%. From the 10,284 milk samples collected from 12 animal species, 10,224 valid milk samples were selected. Due to the large number of Holstein, Simmental, and buffalo samples, a small number of samples were randomly selected based on sampling time for the construction and validation of the model (Table 1) to ensure sample balance during the modeling process.

[0035] Example 1:

[0036] Standard genetic distance based on genome-wide prediction:

[0037] To effectively reflect the evolutionary relationships of animals, a phylogenetic tree was constructed using whole genome sequences. Reference genomes of 12 animals were selected from NCBI, including Pig (GCF_000003025.6_Sscrofa11.1), Horse (GCF_002863925.1_EquCab3.0), Ass (GCF_016077325.2_ASM1607732v2), Camel (GCF_000767855.1_Ca_bactrianus_MBC_1.0), Goat (GCA_001704415.1_ARS1), Sheep (GCA_016772045.1_ARS-UI_Ramb_v2.0), Yak (GCA_005887515.2_BosGru3.0), and Buffalo (GCA_003121395.1_UOA_WB_1);

[0038] Four cattle breeds—Holstein (GCA_021347905.1_ARS-LIC_NZ_Holstein-Friesian_1), Jersey (GCA_021234555.1_ARS-LIC_NZ_Jersey), and Simmental (GCA_018282465.1_ARS_Simm1.0)—and Xinjiang Brown Cattle (XJB; since there was no XJB reference genome, the more closely related Brown Swiss Cattle (GCA_914753205.1_Brown_Swiss_cow) was chosen as a substitute)—were used. The genetic distances of the 12 animals were estimated using MASH software with default parameters.

[24] The obtained genetic distances are then uploaded to ITOL (Interactive Tree of Life) to construct an evolutionary tree.

[25] .

[0039] Based on the method of constructing phylogenetic trees from whole genome sequences, the standard genetic distances among the 12 animals are shown in Table 2:

[0040] Table 2 Genetic distances among 12 animals based on whole-genome calculations.

[0041]

[0042] Example 2:

[0043] A rapid prediction method for reference genetic distances between different mammals based on the mid-infrared spectra of milk:

[0044] Branch length (genetic distance) in an animal evolutionary tree (relationship) is a major indicator of the degree of animal evolution. The distance between branches of two species represents the genetic distance; therefore, the regression model of genetic distance (RMGD) is used to determine the degree of animal evolution. 973 samples (Table 1) from eight animal species (Holstein cow, Simmental cow, Jersey cow, yak, goat, buffalo, camel, and horse) were used for modeling.

[0045] RMGD construction approach:

[0046] (1) Using whole genome data, a standard animal evolutionary tree was constructed and reference values ​​of genetic distance among mammals were calculated, as shown in Table 2;

[0047] (2) The average MIRS of milk from different mammals is used as the standard MIRS of the animal. Genetic distance regression models of different animals (with the same evolutionary level discrimination model) are established using MIRS data to predict the genetic distance between species and to determine the genetic distance between the animal and other animals.

[0048] RMGD construction process:

[0049] (1) 579 MIR points (925.92-1543.2 cm⁻¹) outside the water region were selected. -1 1720.67-2970.66cm -1 3634.24-3998.59cm -1 ) as candidate wave points for RMGD;

[0050] (2) 973 samples from 8 species were selected as the model construction objects. The average MIRS of each animal was used as the standard MIRS of that animal. The average MIRS of the 8 species were subtracted from each other and the absolute value was taken to obtain a total of 28 subtracted MIRS as RMGD samples. The reference genetic distance between species calculated using whole genome data was used as the target value of the model.

[0051] (3) Five wave point selection algorithms (GA, UVE, SPA, Lars, Cars) were used to filter 579 wave points. Among them, the GA algorithm selected 269 wave points, the UVE algorithm selected 1 wave point, the SPA algorithm selected 8 wave points, the Lars algorithm selected 40 wave points, and the Cars algorithm selected 54 wave points. The union of the wave points selected by the five different algorithms initially selected 327 wave points.

[0052] (4) Using the genetic distance calculated from the whole genome as a reference value, 28 MIRS differences and 327 MIRS points were used to model different combinations of 12 preprocessing methods (None, MMS, SS, CT, SNV, MA, SG, MSC, D1, D2, DT and WAVE) and partial least squares regression (PLSR) algorithm. The best combination of preprocessing methods and PLSR algorithm was selected. Then, the manual selection method was used to select 278 MIRS points (distributed in 62 MIRS bands) for modeling and the optimal dimensionality reduction parameter PLSR (n_components=14) of PLSR algorithm from the 327 MIRS points.

[0053] The 62 MIR bands mentioned are: 929.778cm -1 —949.068cm -1 983.79cm -1 —999.222cm -1 1022.37cm -1 —1033.944cm -1 1057.092cm -1 —1084.098cm -1、1122.678cm -1 —1126.536cm -1 、1138.11cm -1 —1149.684cm -1 、1157.4cm -1 —1168.974cm -1 、1184.406cm -1 —1192.122cm -1 、1207.554cm -1 —1211.412cm -1 、1234.56cm -1 —1242.276cm -1 、1276.998cm -1 —1288.572cm -1 、1300.146cm -1 —1304.004cm -1 、1331.01cm -1 —1334.868cm -1 、1338.726cm -1 —1381.164cm -1 、1408.17cm -1 —1427.46cm -1 、1462.182cm -1 —1469.898cm -1 、1512.336cm -1 —1531.626cm -1 、1701.378cm -1 —1705.236cm -1 、1766.964cm -1 —1790.112cm -1 、1813.26cm -1 —1817.118cm -1 、1820.976cm -1 —1840.266cm -1 、1859.556cm -1 —1901.994cm -1 、1959.864cm -1 —1975.296cm -1 、2033.166cm -1 —2052.456cm -1 、2118.042cm -1 —2148.906cm -1、2164.338cm -1 —2179.77cm -1 、2195.202cm -1 —2214.492cm -1 、2249.214cm -1 —2260.788cm -1 、2295.51cm -1 —2322.516cm -1 、2330.232cm -1 —2334.09cm -1 、2349.522cm -1 —2361.096cm -1 、2372.67cm -1 —2388.102cm -1 、2415.108cm -1 —2445.972cm -1 、2465.262cm -1 —2469.12cm -1 、2484.552cm -1 —2488.41cm -1 、2511.558cm -1 —2538.564cm -1 、2569.428cm -1 —2581.002cm -1 、2592.576cm -1 —2596.434cm -1 、2627.298cm -1 —2646.588cm -1 、2704.458cm -1 —2708.316cm -1 、2719.89cm -1 —2723.748cm -1 、2739.18cm -1 —2750.754cm -1 、2758.47cm -1 —2773.902cm -1 、2797.05cm -1 —2812.482cm -1 、2820.198cm -1 —2827.914cm -1 、2835.63cm -1 —2885.784cm -12908.932cm -1 —2916.648cm -1 2928.222cm -1 —2932.08cm -1 2943.654cm -1 —2955.228cm -1 3645.81cm -1 —3668.958cm -1 3676.674cm -1 —3680.532cm -1 3684.39cm -1 —3703.68cm -1 3726.828cm -1 —3738.402cm -1 3784.698cm -1 —3792.414cm -1 3819.42cm -1 —3854.142cm -1 3877.29cm -1 —3881.148cm -1 3892.722cm -1 —3896.58cm -1 3904.296cm -1 —3908.154cm -1 3935.16cm -1 —3939.018cm -1 3942.876cm -1 —3973.74cm -1 3981.456cm -1 —3989.172cm -1 3996.888cm -1 —4000.746cm -1 ;

[0054] (5) Compare the model performance and compare it with the whole genome prediction results to select the optimal RMGD.

[0055] RMGD Best Model:

[0056] Based on the accuracy of the cross-validation set, the optimal model was selected, which was a combination of two preprocessing methods (D1+SNV) and the PLSR algorithm. Specifically, the optimal genetic distance model (D1+SNV+PLSR (n_components=14)) was established using MIRS data from eight animal species. The correlation coefficient of this model on the cross-validation set was 0.99. Figure 3 In section a), the ratios of the genetic distances predicted by the MIRS model and the genetic distances calculated based on whole genomes are all close to 1, as shown in Table 3: Table 3 Genetic distances of eight animals predicted by RMGD and their ratios to genetic distances calculated based on whole genomes.

[0057]

[0058] To prove that the band selected in this embodiment is the optimal band, two bands were randomly selected from the first and second halves of the entire band for addition, deletion, and replacement, and the results were observed as follows:

[0059]

[0060]

[0061] The comparison results show that, regardless of the band change, the correlation coefficient is lower than that of the band selected in this embodiment. Therefore, the band selected in this embodiment is the optimal band.

[0062] Example 3:

[0063] MIRS-based animal RMGD validation, application, and model usage strategies:

[0064] To verify the effectiveness of the established animal RMGD, the genetic distances of the non-modeling animal population (first validation population), the four non-modeling animal populations (second validation population), and other known animal populations (modeling population) were verified using the established model.

[0065] 1. The results of RMGD were better for the eight animal species that did not participate in the modeling process.

[0066] Genetic distances were predicted for animals that did not participate in modeling (i.e., the validation group in Table 1, totaling 738 samples) from eight animal species using MIRS and RMGD. The steps are as follows:

[0067] 1) Milk samples from each of the eight animal species were tested to obtain MIRS data, and the average MIRS for each animal species was obtained. The 62 effective bands selected in the MIRS values ​​are as follows:

[0068] The 62 MIR bands mentioned are: 929.778cm -1—949.068cm -1 、983.79cm -1 —999.222cm -1 、1022.37cm -1 —1033.944cm -1 、1057.092cm -1 —1084.098cm -1 、1122.678cm -1 —1126.536cm -1 、1138.11cm -1 —1149.684cm -1 、1157.4cm -1 —1168.974cm -1 、1184.406cm -1 —1192.122cm -1 、1207.554cm -1 —1211.412cm -1 、1234.56cm -1 —1242.276cm -1 、1276.998cm -1 —1288.572cm -1 、1300.146cm -1 -1304.004cm -1 、1331.01cm -1 —1334.868cm -1 、1338.726cm -1 —1381.164cm -1 、1408.17cm -1 —1427.46cm -1 、1462.182cm -1 —1469.898cm -1 、1512.336cm -1 —1531.626cm -1 、1701.378cm -1 —1705.236cm -1 、1766.964cm -1 —1790.112cm -1 、1813.26cm -1 —1817.118cm -1 、1820.976cm -1 —1840.266cm -1 、1859.556cm -1—1901.994cm -1 、1959.864cm -1 —1975.296cm -1 、2033.166cm -1 —2052.456cm -1 、2118.042cm -1 —2148.906cm -1 、2164.338cm -1 —2179.77cm -1 、2195.202cm -1 —2214.492cm -1 、2249.214cm -1 —2260.788cm -1 、2295.51cm -1 —2322.516cm -1 、2330.232cm -1 —2334.09cm -1 、2349.522cm -1 —2361.096cm -1 、2372.67cm -1 —2388.102cm -1 、2415.108cm -1 —2445.972cm -1 、2465.262cm -1 —2469.12cm -1 、2484.552cm -1 —2488.41cm -1 、2511.558cm -1 —2538.564cm -1 、2569.428cm -1 —2581.002cm -1 、2592.576cm -1 —2596.434cm -1 、2627.298cm -1 —2646.588cm -1 、2704.458cm -1 —2708.316cm -1 、2719.89cm -1 —2723.748cm -1 、2739.18cm -1 —2750.754cm -1 、2758.47cm -1—2773.902cm -1 、2797.05cm -1 —2812.482cm -1 、2820.198cm -1 —2827.914cm -1 、2835.63cm -1 —2885.784cm -1 、2908.932cm -1 —2916.648cm -1 、2928.222cm -1 —2932.08cm -1 、2943.654cm -1 —2955.228cm -1 、3645.81cm -1 —3668.958cm -1 、3676.674cm -1 —3680.532cm -1 、3684.39cm -1 —3703.68cm -1 、3726.828cm -1 —3738.402cm -1 、3784.698cm -1 —3792.414cm -1 、3819.42cm -1 —3854.142cm -1 、3877.29cm -1 —3881.148cm -1 、3892.722cm -1 —3896.58cm -1 、3904.296cm -1 —3908.154cm -1 、3935.16cm -1 —3939.018cm -1 、3942.876cm -1 —3973.74cm -1 、3981.456cm -1 —3989.172cm -1 、3996.888cm -1 —4000.746cm -1 ;

[0069] 2) Subtract the absolute values ​​of the average MIRS of the 8 animals to obtain the subtracted MIRS, which totals 28 groups. Substitute the 28 subtracted spectra into the model D1+SNV+PLSR(n_components=14) to predict the corresponding genetic distance.

[0070] The correlation coefficient of the final cross-validation set was 0.97. The genetic distance predicted by the MIR model and the genetic distance calculated based on the whole genome were very close (except for yaks, Holstein cows, and goats), and the ratio between the two was also close to 1. Figure 3 (See Table 4, b).

[0071] Table 4. Genetic distances of the eight animal species not included in the modeling based on the MIRS model (RMGD) and their ratios to the genetic distances of the eight animal species calculated based on whole genomes.

[0072]

[0073] 2. RMGD results were better for the four animal species that did not participate in the modeling.

[0074] Genetic distances were predicted for four animal species that did not participate in modeling (i.e., the external validation animals in Table 1) using MIRS and RMGD. The specific steps are as follows:

[0075] 1) MIRS data were obtained by testing each milk sample from the four animals. The average MIRS for each animal was obtained, and the 62 bands selected for the MIRS values ​​are shown above.

[0076] 2) Subtract the absolute values ​​of the average MIRS of the four animals to obtain the subtracted MIRS, which are a total of 6 groups. Substitute the 6 subtracted spectra into the model D1+SNV+PLSR(n_components=14) to predict the corresponding genetic distance.

[0077] The correlation coefficient of the final cross-validation set was 0.95. The genetic distances predicted based on the MIR model and those calculated based on the whole genome were very similar (except for Xinjiang Brown cattle, Jersey cattle, and yak cattle), and their ratios were close to 1. Figure 3 (middle c, Table 5)

[0078] Table 5 shows the genetic distances predicted by the MIRS model for the four animal species, the genetic distances of the four animals and eight other animal species, and their ratios to the genetic distances calculated from the whole genome.

[0079] Prediction results Holstein cows Simmental dairy cows Jersey Cow buffalo yak goat horse camel sheep donkey Xinjiang Brown Cattle pig sheep 0.05968 0.04464 0.05872 0.01961 0.08534 0.02529 0.16453 0.13263 donkey 0.14637 0.13756 0.14933 0.12717 0.14093 0.11548 0.00614 0.13768 0.14899 Xinjiang Brown Cattle 0.00206 0.01774 0 0.07352 0.00179 0.0611 0.14783 0.16224 0.06179 0.13895 pig 0.1383 0.1752 0.14582 0.12288 0.1359 0.15474 0.15046 0.13604 0.1664 0.14391 0.18131

[0080]

[0081] Therefore, the animal RMGD established in this study has been effectively applied not only to the prediction of genetic distance in animals that did not participate in the modeling (or new unknown animals) among the 8 animal species, but also to the prediction of the evolutionary degree of the 4 animal species (or new unknown animals) that did not participate in the modeling.

Claims

1. A method for predicting the reference genetic distance between different mammals based on the infrared spectrum of milk, comprising the following steps: 1) Detecting each milk sample of N animals to be tested to obtain MIRS data, and obtaining the average MIRS of each animal, wherein the selected effective 62 wave bands in the MIRS value are as follows: The 62 MIR bands are: 929.778 cm -1 —949.068 cm -1 , 983.79 cm -1 —999.222 cm -1 , 1022.37 cm -1 —1033.944 cm -1 , 1057.092 cm -1 —1084.098 cm -1 , 1122.678 cm -1 —1126.536 cm -1 , 1138.11 cm -1 —1149.684 cm -1 , 1157.4 cm -1 —1168.974 cm -1 , 1184.406 cm -1 —1192.122 cm -1 , 1207.554 cm -1 —1211.412 cm -1 , 1234.56 cm -1 —1242.276 cm -1 , 1276.998 cm -1 —1288.572 cm -1 , 1300.146 cm -1 — 1304.004 cm -1 1331.01 cm -1 1334.868 cm -1 1338.726 cm -1 1381.164 cm -1 1408.17 cm -1 1427.46 cm -1 1462.182 cm -1 1469.898 cm -1 1512.336 cm -1 1531.626 cm -1 1701.378 cm -1 1705.236 cm -1 1766.964 cm -1 1790.112 cm -1 1813.26 cm -1 1817.118 cm -1 1820.976 cm -1 1840.266 cm -1 1859.556 cm -1 1901.994 cm -1 1959.864 cm -1 1975.296 cm -1 2033.166 cm -1 2052.456 cm -1 2118.042 cm -1 2148.906 cm -1 2164.338 cm -1 2179.77 cm -1 2195.202 cm -1 2214.492 cm -1 2249.214 cm -1 2260.788 cm -1 2295.51 cm -1 2322.516 cm -1 2330.232 cm -1 2334.09 cm -1 2349.522 cm -1 2361.096 cm -1 2372.67 cm -1 2388.102 cm -1 2415.108 cm -1 — 2445.972 cm -1 , 2465.262 cm -1 — 2469.12 cm -1 , 2484.552 cm -1 — 2488.41 cm -1 , 2511.558 cm -1 — 2538.564 cm -1 , 2569.428 cm -1 — 2581.002 cm -1 , 2592.576 cm -1 — 2596.434 cm -1 , 2627.298 cm -1 — 2646.588 cm -1 , 2704.458 cm -1 — 2708.316 cm -1 , 2719.89 cm -1 — 2723.748 cm -1 , 2739.18 cm -1 — 2750.754 cm -1 , 2758.47 cm -1 — 2773.902 cm -1 , 2797.05 cm -1 — 2812.482 cm -1 , 2820.198 cm -1 — 2827.914 cm -1 , 2835.63 cm -1 — 2885.784 cm -1 , 2908.932 cm -1 — 2916.648 cm -1 , 2928.222 cm -1 — 2932.08 cm -1 , 2943.654 cm -1 — 2955.228 cm -1 , 3645.81 cm -1 — 3668.958 cm -1 , 3676.674 cm -1 — 3680.532 cm -1 , 3684.39 cm -1 — 3703.68 cm -1 , 3726.828 cm -1 — 3738.402 cm -1 , 3784.698 cm -1 — 3792.414 cm -1 , 3819.42 cm -1 — 3854.142 cm -1 , 3877.29 cm -1 — 3881.148 cm -1 , 3892.722 cm -1 — 3896.58 cm -1 , 3904.296 cm -1 — 3908.154 cm -1 , 3935.16 cm -1 — 3939.018 cm -1 , 3942.876 cm -1 — 3973.74 cm -1 , 3981.456 cm -1 — 3989.172 cm -1 , 3996.888 cm -1 — 4000.746 cm -1 ; 2) Subtracting the absolute value of the average MIRS of N animals from each other to obtain the subtracted MIRS, a total of N* (N-1) / 2 subtracted spectra, and substituting the N* (N-1) / 2 subtracted spectra into the model D1+SNV+PLSR to predict the corresponding genetic distance; wherein N≥2, and n_components=14 in PLSR.

2. The method of claim 1, wherein the animals are horses, camels, cows, sheep, donkeys and pigs.

3. The method of claim 2, wherein the cows are buffalos, yaks, Holstein cows, Simmental cows, Jersey cows and Xinjiang brown cows; and the sheep are goats and sheep.

4. The use of the method of claim 1 in predicting the genetic distance between mammals.

Images