AISNP marker combinations and multi-dimensional ancestral and geo-inference models

Abstract

The invention provides an AISPN marker combination and a multi-dimensional ancestral and geographic inference model, and relates to the technical field of forensic medicine. According to the method, firstly, seven groups of ancestral inference SNP (Single Nucleotide Polymorphism) sites (AISNPs, AISNPs) of East Asia and Southeast Asia crowds are screened, the screened sites are respectively subjected to efficiency evaluation based on principal component analysis, ADMIXTURE and t-SNE methods, and a set of multi-dimensional ancestral and geographic source inference model is constructed in combination with six different machine learning methods. The AISPN marker combination provided by the invention comprises at least 50 AISPN sites, has relatively high differentiation in East Asia and Southeast Asia crowds, can distinguish nine crowds based on relatively few sites, provides site information for preparation of kits, and provides valuable information for research of forensic medicine, anthropology and population genetics.

Description

Technical Field

[0001] This invention belongs to the fields of bioinformatics and forensic medicine, specifically involving AISNP marker combinations and multi-dimensional ancestry and geographic inference models. Background Technology

[0002] The genetic characteristics of East Asian and Southeast Asian populations are highly diverse and intertwined with historical migrations, making it difficult for traditional ancestry inference methods to accurately capture subtle genetic differences among these groups. How to efficiently and accurately infer the ancestry of East Asian and Southeast Asian populations has become a major challenge in genetic research.

[0003] Ethnic origin inference refers to the use of ancestral information markers (AIMs) that exhibit population-specific differences to infer the biogeographical ancestry of unknown individuals through statistical methods, thereby aiding in criminal investigations and litigation. AIMs are molecular markers that exhibit genetic differences between populations, and the accuracy of ethnic origin inference can be greatly improved, especially when multiple related loci are combined.

[0004] Although some ancestry inference methods and techniques have achieved certain results in practical applications, limitations remain when inferring ancestry within subgroups, especially East Asian and Southeast Asian subgroups. Because different studies focus on different target groups, the chosen AIMs and inference models also vary, often resulting in limited information when previous ancestry inference findings are applied to East Asian and Southeast Asian groups. Therefore, further exploration is needed for ancestry inference within subregions, particularly for East Asian and Southeast Asian groups within Asia. Summary of the Invention

[0005] To address the aforementioned issues, this invention screened seven groups of AIMs and conducted performance evaluations, providing an AISNP tag combination and a multi-dimensional ancestry and geographic inference model.

[0006] On one hand, the present invention provides an AISNP marker combination comprising at least 50 sites of the AISNPs described in Table 1.

[0007] On the other hand, the present invention provides the application of the above-mentioned AISNP tag combination in constructing a multi-dimensional ancestry and geographic origin inference model.

[0008] On the other hand, this invention provides a method for constructing a multi-dimensional ancestry and geographical origin inference model.

[0009] Specifically, the construction method includes the following steps: (1) Collect sample data; (2) Use unsupervised methods to analyze sample data and identify the ancestral components of the samples; (3) Based on the results of step (2), conduct supervised analysis to determine the population grouping; (4) Select AISNPs sites, form AIMs groups based on the ranking of AISNPs, and further refine the AISNPs groups based on their ranking. I n The values ​​are ranked to generate seven groups of AIMs, each group of AIMs comprising at least 50 sites from the AISNPs of claim 1; (5) Use machine learning algorithms to process AIMs groups and build diversified machine learning models.

[0010] More specifically, in step (1), the collection of sample data includes the following steps: (a) Extract modern human sample data from the Human Origins Project dataset, excluding samples from third-degree relatives; (b) Filter the sites and retain sites that meet the following conditions: deletion rate less than 0.1, Hardy-Weinberg equilibrium p-value greater than 0.001, and minimum allele frequency greater than 0.01; (c) Filter the samples and retain those that meet the following condition: missing rate is less than 0.1.

[0011] More specifically, in step (3), the supervised analysis is performed based on the random forest method.

[0012] Specifically, in step (4), after screening out AISNPs sites, the method of evaluating the effectiveness of the screened sites is also included using PCA, ADMIXTURE, t-SNE and other methods.

[0013] Specifically, in step (4), the AIMs group includes at least 50 AISNP sites.

[0014] More specifically, in step (4), the AIMs group includes at least 100 AISNP sites.

[0015] More specifically, in step (4), the AIMs group includes at least 250 AISNP sites.

[0016] More specifically, in step (4), the AIMs group includes at least 500 AISNP sites.

[0017] More specifically, in step (4), the AIMs group includes at least 1000 AISNP sites.

[0018] Preferably, in step (4), the AIMs group includes at least 1500 AISNP sites.

[0019] More preferably, in step (4), the AIMs group includes at least 2000 AISNP sites.

[0020] Specifically, in step (5), the machine learning algorithms include: convolutional neural network, k-nearest neighbor algorithm, logistic regression analysis, random forest, support vector machine, extreme gradient boosting, decision tree, linear regression and / or recurrent neural network.

[0021] Specifically, the construction method further includes the following steps: (6) Construct a deep neural network model.

[0022] More specifically, the deep neural network model is constructed using the Locator method. Based on the genotype information of the seven selected AISNPs, the model is trained to learn the complex nonlinear relationship between genotype data and geographic coordinates, and to predict the location of the sample in geographic space.

[0023] In certain specific embodiments of the present invention, the construction of a deep neural network model using the Locator method includes the following steps: S1: Enter the locator environment, the code is as follows: conda activate locator; S2: Input the genotype data and latitude / longitude coordinates of the training set samples, train the neural network to approximate the relationship between genotype and geographical location, and build a deep neural network model. The code is as follows: locator --vcf data / test_genotypes.vcf.gz --sample_data data / test_sample_data.txt --out out / test / test; S3: After model training, based on the selected 7 groups of AISNPs and 597,569 SNPs, the geographic coordinates of 9 samples in the validation set were randomly predicted. The error between the predicted coordinates and the true coordinates (such as kernel density peak error and centroid error) was calculated using the plot_locator.R script in R software. The accuracy of the model predictions was evaluated, and a map was drawn. The code is as follows: Rscript . / plot_locator_gai.R --infile out / test_windows / --sample_datasample_data.txt --out 9samples --samples CA_Turkic_ALT-093, NC_Sino_Tibetan_Altaic_Han1243, NEA_Tungusic_2BT20220100009, SC_Tai-Kadai_Sinitic_BoY184,SEA_Austroasiatic_BRU101, SEA_Tai-Kadai_Sino-Tibetan_RM129, SECC_ISEA_Austronesian_NA13597, SWA_Hmong_Mien_HM104, WS_Turkic_BAS-661 --error. On the other hand, the present invention provides the application of the accurate ancestry inference model constructed by the above construction method in predicting the ethnic background, genetic characteristics and / or geographical location of samples.

[0024] Specifically, the samples were sourced from East Asia and Southeast Asia.

[0025] Compared with the prior art, the present invention has the following advantages: This invention utilizes the random forest method to group populations and screen for ancestry inference loci. It can differentiate nine population groups with different backgrounds and genetic characteristics in East and Southeast Asia based on a limited number of loci, providing locus information for reagent kit preparation. Six different machine learning methods are used to build models that can infer the population origin of unknown individuals. Simultaneously, the deep learning method Locator is used to construct a deep neural network model using genotype data and latitude / longitude coordinates of training samples, achieving the goal of predicting the geographical origin of samples. This invention provides SNP loci for the preparation of reagent kits for inferring the biogeographical origin of East and Southeast Asian populations, which can help narrow down the scope of investigation in criminal cases, provide a method for building artificial intelligence models, and provide valuable information for forensic medicine, anthropology, and population genetics research. Attached Figure Description

[0026] Figure 1 This is a density distribution map of AISNPs on the 22 chromosomes.

[0027] Figure 2 Principal component analysis plots of 7 AIMs and full test data in 9 populations.

[0028] Figure 3The figures are PCA and ADMIXTURE (K=10) analysis diagrams based on 2000 AISNPs. A is the PCA analysis diagram based on 2000 AISNPs, and B is the ADMIXTURE analysis diagram based on 2000 AISNPs.

[0029] Figure 4 t-SNE analysis plots of 7 AIMs and full test data in 9 populations.

[0030] Figure 5 This is a graph showing the average classification accuracy and the number of SNPs for a multivariate machine learning algorithm.

[0031] Figure 6 To predict the geographical location map of samples using a deep neural network Locator model based on 2000 AISNPs.

[0032] Figure 7 This is a geolocation prediction map of the same individual in NEA_Tungusic based on different locations.

[0033] Figure 8 Kernel density peak error and centroid error of the Locator model constructed for 7 sets of AIMs and full measurement data. Detailed Implementation

[0034] The present invention will be further described in detail below with reference to specific embodiments. The following embodiments are not intended to limit the present invention, but only to illustrate the present invention. Unless otherwise specified, the experimental methods used in the following embodiments are generally performed under conventional conditions. Unless otherwise specified, the materials and reagents used in the following embodiments are commercially available.

[0035] Example 1: Construction of a diversified machine learning model 1. Sample collection and data quality control Based on a Linux system, 1,703 modern human samples were extracted from the Human Origin dataset (HO). Samples within third-degree relatives were excluded using the KING software. Quality control was performed on the merged data using the following code: "--mind 0.1, --geno 0.1, --hwe 0.001, and --maf 0.01".

[0036] Loci were filtered to retain those that met the following criteria: deletion rate less than 0.1 (--mind 0.1), Hardy-Weinberg equilibrium p-value greater than 0.001 (--hwe 0.001), and minimum allele frequency greater than 0.01 (--maf 0.01). The samples were filtered to retain those that met the following condition: missing rate less than 0.1 (--geno 0.1), resulting in a full test dataset containing 597,569 SNPs.

[0037] 2. ADMIXTURE Analysis The ancestral components of whole-genome sequencing data were evaluated using an unsupervised algorithm in the ADMIXTURE software.

[0038] 3. Grouping the population based on the random forest method Based on the ADMIXTUER results (optimal K=9) of the full test data (597,569 SNPs), 1028 samples with an ancestry proportion greater than 50% in the Q file were defined as the definitive group population. Using the results of K=2~9 as input data, the remaining 675 samples of uncertain group population were predicted. Using a supervised method of random forest, all 1703 research samples were divided into nine groups, named according to geographical location and language family: CA_Turkic (115 people), NC_Sino-Tibetan_Altaic (254 people), NEA_Tungusic (285 people), SC_Tai-Kadai_Sinitic (302 people), SEA_Austroasiatic (234 people), SEA_Tai-Kadai_Sino-Tibetan (159 people), SECC_ISEA_Austronesian (60 people), SWA_Hmong-Mien (54 people), and WS_Turkic (240 people).

[0039] 4. Screening for ancestral information genetic loci (1) The selection criteria are as follows: ① The locus is located in an intron region; ② The selected loci are located on different chromosomes or on the same chromosome at a distance of more than 1 Mb, to reduce the mutual influence between SNPs; ③ There is a high degree of genetic distribution difference between the paired populations (F... st , I n ); ④ The population studied conforms to Hardy-weinberg equilibrium (HWE).

[0040] (2) Screening steps: ① Calculate the allele frequencies for each subpopulation. Based on the PLINK software, the allele frequencies were calculated using the following code: "plink --bfile pop --double-id --snps-only--freq --make-bed --keep-allele-order --out pop".

[0041] ② Calculate the squared correlation coefficient (r) of the overall linkage disequilibrium values. 2 ) Based on the PLINK software, the following code was used to calculate the squared correlation coefficient (r) of the overall linkage disequilibrium values: `plink --vcf 9pop.vcf.gz --double-id --snps-only --r2 --ld-window-kb 2000 --ld-window 99999999 --ld-window-r2 0.2 --make-bed --out 9pop`. 2 ).

[0042] ③ Configure and run AIM-generator Calculate the inter-population ratio using the AIM-generator package I n Value, based on I n Locus selection was performed based on value size; the density distribution map on the 22 chromosomes is shown below. Figure 1 The configuration file parameters are set as follows: ldthresh=0.2, hetthresh=0.01, distances=1000000, nraims=50 / 100 / 250 / 500 / 1000 / 1500 / 2000. According to... I n The values ​​will be sorted by the selected sites and then filtered accordingly. I n The SNP loci with the highest values ​​of 50, 100, 250, 500, 1000, 1500, and 2000 are shown in Table 1.

[0043] Table 1. Top 50, 100, 250, 500, 1000, 1500, and 2000 AISNPs

[0044] Note: In the table, serial numbers 1-2000 represent rankings 1-2000.

[0045] The above SNP sites are divided into 7 groups, namely: The 50AIMs group consists of SNP loci numbered 1 to 50, totaling 50 SNP loci.

[0046] Group 100AIMs: Composed of SNP loci numbered 1 to 100, totaling 100 SNP loci.

[0047] Group 250AIMs: Composed of SNP loci numbered 1 to 250, totaling 250 SNP loci.

[0048] The 500AIMs group consists of SNP loci numbered 1 to 500, totaling 500 SNP loci.

[0049] The 1000AIMs group consists of SNP loci numbered 1 to 1000, totaling 1000 SNP loci.

[0050] Group 1500AIMs: Composed of SNP loci numbered 1 to 1500, totaling 1500 SNP loci.

[0051] Group 2000AIMs: Composed of SNP loci numbered 1 to 2000, totaling 2000 SNP loci.

[0052] Example 2 Screening site efficacy evaluation (1) Principal component analysis Principal component analysis was performed on the selected AISNPs combinations using the smartPCA program in the EIGENSOFT package. The first two components, PC1 and PC2, were visualized as two-dimensional scatter plots using R software. The nine study groups were classified using different colors. (See below) Figure 2 .

[0053] (2) ADMIXTURE analysis The ancestor proportions of the selected AIMs set were evaluated using an unsupervised algorithm in the ADMIXTURE software, and the results were visualized using R software. PCA and ADMIXTURE analyses of 2000 SNPs are shown below. Figure 3 .

[0054] (3) t-distributed random neighborhood embedding (t-SNE) Using the results of principal component analysis and the Q-file of the optimal K as input files, t-distributed random neighborhood embedding analysis was performed using the Rtsne package in R software, and the results were visualized. Figure 4 .

[0055] (4) Constructing diversified machine learning models 80% of the samples were randomly selected as the training set and 20% as the test set. The PC1-PC10 data from principal component analysis and the Q file of the optimal K in ADMIXTURE were used as input data. Six machine learning methods, namely Convolutional Neural Network (CNN), K-Nearest Neighbors (K-NN), Logistic Regression, Random Forest (RF), Support Vector Machine (SVM), and Extreme Gradient Boosting (XGBoost), were applied to construct machine learning models for the selected AISNPs combinations (50 AIMs, 100 AIMs, 250 AIMs, 500 AIMs, 1000 AIMs, 1500 AIMs, and 2000 AIMs) and their performance was evaluated.

[0056] This includes the average classification accuracy (AAC) for nine population groups. For each AISNP group, the sensitivity, specificity, negative predictive value, and positive predictive value for each population group were systematically evaluated. The formulas for calculating the sensitivity, specificity, negative predictive value, and positive predictive value are as follows: Accuracy =

[0057] Sensitivity = ; Specificity = ; Positive predictive value (PPV) = ; Negative predictive value (NPV) = ; In the above calculation formula, TP represents true positive, which successfully predicts a positive sample as positive; TN represents true negative, which successfully predicts a negative sample as negative; FN represents false negative, which incorrectly predicts a positive sample as negative; and FP represents false positive, which incorrectly predicts a negative sample as positive.

[0058] The experimental results are shown in Table 2-3 and Figure 5 , Class 1: NC_Sino-Tibetan_Altaic; Class 2: NEA_Tungusic; Class 3: SC_Tai-Kadai_Sinitic; Class 4: SWA_Hmong-Mien; Class 5: SEA_Tai-Kadai_Sino-Tibetan; Class 6: SEA_Austroasiatic; Class 7: SECC_ISEA_Austronesian; Class 8: CA_Turkic; Class 9: WS_Turkic.

[0059] Table 2. Population prediction accuracy based on 2000 AISNPs using six machine learning methods

[0060] Table 3. Sensitivity, specificity, negative predictive value, and positive predictive value based on 2000 AISNPs

[0061] Example 3: Predicting the geographical location of samples based on a deep neural network model (1) Data preprocessing The longitude and latitude information of 1703 research samples were collected and the file was named sample_data.txt. Using Locator software, the input data (genotype data of the selected AISNPs and the full data corresponding to the research population) were converted into allele count vectors for each individual. Using R software, 425 samples (25% of all samples) were randomly selected, their true geographical coordinates were hidden, and their longitude and latitude information was named "NA" to form the validation set. The remaining 1278 samples were used as the training set for model training.

[0062] (2) Constructing a deep neural network model Using the Locator method, a supervised machine learning approach based on deep neural networks, this study analyzed the genotype data of individuals corresponding to seven different AISNP loci and 597,569 SNPs across the entire genome. A model was built by training the model to learn the complex nonlinear relationship between genotype data and geographic coordinates, and then used to predict the latitude and longitude geographic coordinates of samples in the validation set. The steps are as follows: ① Enter the locator environment using the following code: conda activate locator.

[0063] ② Input the genotype data and latitude / longitude coordinates of the input samples, train a neural network to approximate the relationship between genotype and geographical location, and construct a deep neural network model. The code is as follows, based on Python software: cd ~ / locator;mkdir out / test;locator.py --vcf data / test_genotypes.vcf.gz --sample_data data / test_sample_data.txt --out out / test / test ③ Predicting geographic coordinates: After model training, the geographic coordinates of samples in the validation set are randomly predicted based on the selected 7 groups of AISNPs and 597,569 SNPs. The errors between the predicted and true coordinates (such as kernel density peak error and centroid error) are calculated using the plot_locator.R script in R software. The accuracy of the model's predictions is evaluated, and a map is drawn. The code is as follows: Rscript scripts / plot_locator.R --infile out / test_windows / --sample_data data / test_sample_data.txt --out out / test_ --map F.

[0064] The results of a deep neural network Locator model, constructed based on 2000 AISNPs, randomly predicting the geographical locations of nine samples from nine population groups are shown below. Figure 6 The results of the deep neural network Locator model predicting geographical locations for the same NEA_Tungusic sample based on different SNPs are shown below. Figure 7 The experimental results show that the 2000 selected AISNPs have good predictive efficacy for geographic origin in the nine population groups. The predictive efficacy further improves with the increase in the number of selected AISNPs. In the NEA_Tungusic geographic location prediction of the same individual, the Locator model based on 2000 AISNPs showed even higher predictive efficacy than the model using 597,569 SNPs, indicating that the AISNPs selected in this invention have good biogeographical prediction application value. The error between the predicted coordinates and the actual coordinates in the model is calculated as follows: Figure 8 The study found that the deep learning model built based on our selected sites has good performance in predicting geographic origins. In particular, the deep neural network Locator model based on 2000 AISNPs achieved the same predictive performance as the deep neural network Locator model based on 597,569 SNPs. This shows that the screening process of the present invention not only extracts key information, but also ensures the stability of the model under various conditions, effectively distinguishes people from different geographic regions, improves the practicality of the model, and reduces the demand for computing resources.

[0065] Finally, it should be noted that the above content is only used to illustrate the technical solution of the present invention, and is not intended to limit the scope of protection of the present invention. Simple modifications or equivalent substitutions made by those skilled in the art to the technical solution of the present invention do not depart from the essence and scope of the technical solution of the present invention.

Claims

1. An AISNP tag combination, characterized in that, Includes at least 50 of the following AISNPs: rs12441154, rs73374393, rs174550, rs4655956, rs1431403, rs79508014, rs966770, rs56757009, rs922452, rs1551748, rs5751117, rs75256421, rs193218545, rs5750871, rs73783845, rs1505066, rs2927599, snp_7_138602682, Affx-50071660, rs114470475, rs2148359, rs1 2424204, rs9808255, rs62318988, rs181331688, rs75122574, rs3181202, rs6792783, rs2441727, rs10420646, snp_2_8695382, rs112416191, rs111 877248, Affx-50012601, rs11981433, rs57347846, rs72664675, rs17356935, rs7898005, Affx-50079265, rs57607537, rs9382713, rs1346327, rs757 43048, rs12292304, rs13072132, rs969528, rs56379200, rs16864187, rs144367093, rs28393696, rs59820740, rs16915333, rs4646649, rs3125759, rs3858150, rs6500380, rs114953206, rs1562789, Affx-50073691, rs1045574, rs867767, rs111312034, rs144965950, rs1418042, rs11174549, rs111 09878, snp_13_27105872, rs7127397, rs1881600, rs820335, rs11000335, rs10480067, rs4868356, rs2403222, rs11156408, rs189695372, rs171989 50. rs1613747, rs6437783, rs140057503, rs2179043, rs12495572, rs10124155, rs2851027, rs73494, rs3173798, rs1571596, rs7100854, rs3099862,rs12549790、rs111963872、rs416867、rs12457129、rs859069、rs2646149、rs2542364、rs4872318、rs10762771、rs1792326、rs9513543、rs112654272、rs1240747、Affx-50125044、rs7399591、Affx-50053608、rs1886081、rs12042017、rs12227163、rs9401955、rs28745273、rs17147258、rs6569926、rs12149763、rs4845116、rs1864500、rs12675271、rs149490、rs10983767、rs76030351、Affx-50114236、rs4246627、rs2717176、rs61560679、rs2150696、rs4814142、rs58954514、rs10167200、Affx-50031467、rs142357133、rs1467028、rs142485592、rs11780720、rs8049660、rs2118390、rs2564988、rs10860946、rs72945511、rs7700048、Affx-50018709、Affx-24730811、rs6975395、rs186570802、rs4411464、rs2601881、rs59024601、rs9575385、rs35339929、rs1363713、rs16952025、rs7133116、rs734226、rs7688824、rs295573、rs2793833、snp_1_43657952、rs77959195、rs35277342、rs6992898、rs10842369、rs74837625、snp_4_93445674、rs6595252、rs73123088、rs11097631、rs403906、Affx-50016773、rs13285837、rs10498043、rs117593437、rs145908113、rs3813987、rs1913653、rs12092484、rs9576332、rs1749553、Affx-50106164、rs4925810、Affx-50051429、rs16853338、rs78296377、rs12672026、rs7738608、rs10515625、rs6112263、snp_6_53916393、rs8053831、rs10916209、rs7573044、rs7112255、rs12327410、rs9994192、rs17675813、rs6842124、rs3133787、rs17068421、rs2785189、snp_13_48717981、rs17106242、rs11220787、rs982997、rs2205895、rs2803266、rs12709959、rs73920949、rs141863656、rs55804216、rs284200、rs13397985、rs10183431、rs4762126、rs10401590、rs7128458、rs4759347、rs1192658、rs1467845、rs2574951、rs10971939、rs6445264、rs140469033、rs12638947、rs10875746、rs789982、rs4976398、rs1000313、rs1706824、rs11217798、rs4710978、rs1436797、rs55768113、rs7525145、Affx-50068479、rs34125126、rs3904725、rs934882、rs10173678、rs77922840、rs8042807、rs512888、rs28661325、rs3843024、rs58740148、rs17127329、rs115574267、rs2398832、rs7170961、rs113577965、rs12598094、rs1789158、rs7202124、rs1474249、rs801180、rs2067676、rs4810486、rs7123066、rs7016764、rs10819283、snp_3_15110989、Affx-50072027、rs1275977、Affx-50013997、rs4909598、rs380649、rs9439530、rs10946813、rs2851060、rs660361、rs12691687、rs1765855、rs6470034、rs138703586、snp_6_29730185、rs1288367、rs339973、rs35930291、rs77468152、rs12639387、rs502919、rs2290830、rs759656、rs16828910、rs944717、rs10514547、rs9653392、rs11904650、rs114753298、rs9386049、rs4519530、rs1426117、rs12604439、rs4658883、rs41070、snp_2_223420928、rs75618811、rs4408369、rs6739469、rs13327928、rs1484744、rs10220285、rs4920184、rs6867956、snp_4_155596649、rs61775129、rs115035006、rs17551079、rs7965102、rs36069987、rs11187300、rs10474479、rs9981369、rs6571768、rs677066、rs10814248、rs9329332、rs2634982、rs12173156、rs12291531、rs6592836、rs7713630、rs3796689、rs59987069、Affx-50115263、rs114489298、rs9356128、rs10975141、rs11191514、rs17168818、rs371318、Affx-50079415、rs8041084、rs1316874、rs61069609、rs4731513、rs4846567、rs6443746、snp_2_20328992、rs3758306、rs2753971、Affx-50042480、rs1362785、rs13032261、rs7172774、rs7978717、rs116867118、rs1033732、snp_21_22800279、snp_11_65208431、rs4976606、rs2609081、rs17172574、rs6074798、rs7667144、rs7838079、rs7863573、rs6672776、rs12711244、rs989430、rs4791876、rs10483857、rs574923、rs190651901、Affx-50013170、rs1349187、rs2886778、rs4680899、rs13435651、rs11263456、rs7627506、rs931145、rs148179324、rs58298719、rs10832820、rs10458187、rs322622、rs6680653、rs890266、rs6706465、rs464273、rs74784266、rs4686934、rs6984296、rs1561092、rs6049874、rs8121977、rs484696、rs11064160、Affx-50062468、rs2785245、rs12206022、snp_7_19506454、rs3756381、rs6021721、rs9311793、rs763136、rs1382579、rs660871、rs7165650、rs11189628、rs116805161、rs11081234、rs9991725、rs1915237、rs10207020、rs16931669、rs4810161、rs741245、rs2920879、rs77168561、rs184546912、rs1038236、rs7937519、Affx-23499836、rs560018、rs770267、rs4324551、rs12423629、rs281446、rs1571504、rs2763979、rs17168765、rs17073870、rs847851、rs1016188、rs1964868、rs10752073、rs117257765、rs7227637、rs1204777、rs803318、rs137900317、Affx-50119336、rs2982561、rs11207325、rs114368717、rs111257980、rs1508753、rs17531839、rs13217627、rs6055528、rs332115、rs3847554、rs7129118、rs11213022、rs1349642、rs1011276、rs4810108、rs4410663、rs34488193、rs7221792、rs11073121、rs932768、rs12210974、rs12122195、snp_3_176303559、rs11121615、rs3942395、rs12231509、rs11675180、rs8001778、rs2216145、rs6715575、rs219435、rs9981599、rs2798379、rs117417101、rs266805、rs10100760、rs6829669、rs836483、rs111568371、rs115428306、rs6600362、rs3751163、rs76453343、rs181181105、rs17603472、rs57068357、rs9847154、rs12108640、rs2403731、rs12888084、rs6427062、rs2131774、rs10814680、rs12465401、rs10027346、rs7175977、rs11748320、rs9920619、rs6740653、rs4687307、rs4954799、rs1287044、rs9308836、Affx-50016187、rs9753377、rs4405339、Affx-50091621、rs9471185、Affx-50031591、rs4889799、rs11763980、rs17293010、rs10146962、rs1535593、rs7100934、rs56138618、rs9579248、rs111901154、rs7093309、rs182374068、rs4237219、rs17101013、rs4657237、rs73859870、rs2444769、rs4142748、rs12514007、rs878175、rs79285903、rs16973908、rs35501269、rs904750、rs1465327、rs7614497、rs4657449、rs1534596、rs542788、rs2467232、rs62180777、rs9311137、rs12120232、rs10943076、rs1482718、rs57766590、snp_1_32452296、rs7220048、rs2649734、rs2825529、rs16962840、rs6429540、rs77225203、rs10837839、snp_10_7920069、Affx-50092652、rs10139750、rs9583210、rs10782785、rs11259027、rs4781256、rs4364954、snp_11_111576194、rs4233905、rs4653336、rs2304401、rs1026490、rs10764585、rs61536198、rs11118607、rs12056053、rs11155510、rs62333881、rs7649709、rs7592457、rs139072471、rs12547701、rs150420267、rs1025934、rs55806384、rs72620522、rs7727027、rs12741210、rs2814771、rs111986994、rs62336256、rs749588、rs10489599、snp_7_67054546、rs28802716、rs55687305、rs12590451、rs2687656、Affx-50031669、rs10868851、rs875727、rs7183105、rs9284989、rs13016652、rs74468693、rs2160766、rs12188494、rs6754561、rs16945349、rs4514962、rs11139192、rs4701734、rs11151771、rs9954415、rs10810611、rs59021370、rs10496354、rs1327406、rs13162983、Affx-50034204、rs7925049、rs2429943、Affx-12508200、rs1472095、rs2286540、rs661669、rs201508858、rs363976、rs233955、rs56050759、rs73935273、rs12589424、rs177698、rs10510107、rs12831686、rs34052907、rs6991944、rs75114913、rs80324685、rs7836608、rs79065746、rs111678717、rs28712159、rs555008、rs13262465、rs2193587、rs7143714、rs1465894、rs4688621、rs12340764、rs76839785、snp_9_136623450、rs1365280、snp_15_26219876、rs12052388、rs10917067、rs4793014、rs7045635、rs67938508、rs2068741、rs146878137、rs9478713、rs58619341、rs16839958、rs1922233、rs12898636、rs2424306、rs11240503、snp_19_12893216、rs75635066、rs6971685、rs62286210、snp_13_109691348、rs17299467、rs4581803、rs4777901、rs2236946、rs74659258、rs73159573、rs76773209、rs9871022、rs6062363、rs3860903、rs17097380、rs62211087、rs16927458、rs60869409、rs3846452、rs1078502、rs13393294、rs7331661、rs738276、Affx-50036013、rs1216491、rs1060570、rs7679696、rs1568869、rs1319584、rs2372888、rs5011906、rs73633268、rs638006、rs45474193、rs3000678、rs1897589、rs1883610、rs28407613、rs10487384、rs8008779、rs148871550、rs6730185、rs4964156、rs148987454、rs3825386、rs207701、rs1889189、rs12868344、rs76752998、Affx-50011576、rs2038338、rs74488994、rs35891645、rs488679、rs12587866、rs2893262、rs1415415、rs1991967、rs7698004、rs2066881、rs2906308、rs55872795、rs73079246、rs115213347、rs4902628、rs10813882、rs57965597、rs2026171、rs780553、rs778556、rs498911、rs11642665、snp_1_241560279、rs10471436、rs980944、rs7545014、rs979375、rs6986584、rs1548029、rs13079040、rs147918451、rs11074660、rs6475963、rs76022828、rs60649507、rs2278290、rs2279566、rs78708040、rs968102、Affx-50079904、rs16850728、rs12079477、rs1519998、rs2418082、rs684463、rs2868148、rs79200319、rs1950093、rs10002376、rs595422、rs151378、rs544135、rs185229608、rs56004992、rs6860250、rs1196069、rs6548685、rs9343914、rs6583007、rs6995650、rs1360948、rs8012328、rs11160364、rs8086918、rs850913、rs10766635、rs10935101、rs2255749、rs394244、rs73987343、rs17132007、rs12921440、rs1495530、rs1873494、rs78041038、rs4770482、rs902399、rs115011552、rs10997727、rs11629508、rs7579930、rs117673505、rs9505225、rs13278931、rs252677、rs13326163、rs11014358、rs11921230、rs9553464、Affx-50011189、rs1840815、rs4521564、rs7005532、rs8097422、rs7697509、rs74907998、rs2088170、rs11265780、rs1361595、rs113936369、rs7671283、rs12555385、rs35170334、rs7192202、rs55980496、rs4917141、rs73047487、rs74062874、rs6761395、rs668988、rs10843219、rs7577695、rs9874843、rs34109658、rs6061358、rs73231771、rs9847948、rs6852609、rs149392467、rs4823006、rs7072133、rs1821242、rs8091698、rs7808830、Affx-50028987、rs7609450、rs1509049、rs1881731、rs16932382、rs6917207、rs883472、rs1924373、rs34068146、rs5756391、rs13059753、rs710819、rs9354724、rs79816186、rs28458412、Affx-50121902、rs111364139、rs12188310、rs10424168、rs9603541、rs6799270、rs1353406、rs5877、rs60660419、rs6424152、rs12380416、rs4135231、rs9466625、rs12186190、rs3909711、rs4908594、rs2521685、snp_6_27967302、rs6664618、rs1492485、rs1537418、snp_6_83248548、rs9391292、rs1570592、rs3821762、rs7287595、rs2283019、rs58424360、rs6593507、rs6735307、rs9310645、rs2441805、rs7192185、rs2206902、rs113069001、rs4903860、Affx-50040022、rs114091486、rs12502383、rs33950471、rs6547234、rs2791997、rs2915131、rs72830425、rs9540125、rs4376434、rs9475655、rs76395338、rs142071844、rs28430294、rs34428905、snp_15_75791640、rs1369804、rs7196251、rs7553708、Affx-50059027、rs1830981、rs4711760、rs8044612、rs7070379、rs114281043、rs11590319、rs59745495、rs2604296、rs2080858、rs4237941、rs3755438、rs10025443、rs4806340、rs17147002、rs2590949、rs855918、rs1289219、rs11649230、rs664584、rs798329、rs6859806、rs8067664、rs4949261、rs9316969、rs1004579、rs4240872、rs3947733、rs11055784、rs868822、rs7566741、rs17826588、rs13270721、rs11581692、rs628419、rs6732518、rs13270872、rs2527212、rs16860970、rs7123981、rs12121971、rs149189378、rs79969342、snp_18_33986543、rs12679359、rs231947、rs6435293、rs928831、rs4648485、rs1861288、rs399649、rs2242152、rs7222696、rs2991228、rs16943463、rs201183576、rs12141654、rs10006747、snp_18_26943421、rs6774578、rs9590359、rs7482329、rs2345976、rs2643195、rs62246991、rs4895189、rs3821883、rs12036832、rs3773364、rs4335057、rs2250648、rs114539812、rs11973064、rs16968313、rs7019837、rs59755163、rs112757136、rs12332207、rs35825018、rs7981942、rs12979014、rs6953542、rs1926306、rs7531679、rs67520393、rs17812521、rs6556883、rs326057、rs286554、rs1024043、rs17810799、rs17796669、rs3848278、rs12692903、rs2112606、Affx-50075582、rs6793434、rs2124959、rs76436438、rs76847913、rs4445429、rs7520936、snp_18_63816579、rs17014647、rs1153292、rs12624012、rs72667119、rs10766495、rs4073471、rs4891302、rs182253764、rs12486781、rs7211602、rs337875、rs1684150、rs62160261、rs7663941、rs7926165、rs72837527、rs1835828、rs3846985、rs2795467、rs9321662、rs728939、rs11181885、rs17196820、rs16924096、rs2577717、rs5995269、rs8025217、rs2404233、rs11601217、rs61183526、rs224291、rs16952766、rs187756888、rs685851、rs10830732、rs10492497、rs73956925、rs7587717、rs6725131、rs16882350、rs4764043、rs2176324、rs17628078、rs2790937、rs7868836、snp_4_4682472、rs9487255、rs10059367、rs7597355、snp_8_53668704、rs10186264、rs7355975、rs2816316、Affx-50081407、rs6499631、rs2359893、rs7546440、rs11661737、rs1761023、rs2393579、rs10865077、rs117820746、Affx-50007730、rs17052053、rs4483185、Affx-50127727、rs12607851、rs9309587、rs13260843、rs9860974、rs334724、rs13056344、rs7095441、rs10425630、rs11737706、rs422130、rs2816030、rs7020549、rs927570、rs2833400、rs12365021、rs12406019、rs4753885、rs472296、rs7832751、rs9932757、rs1435434、rs12587950、rs4893916、rs16974449、rs10250329、rs13098990、snp_11_67792769、rs4760487、rs73651573、rs722896、Affx-50090134、rs11926052、rs7215687、rs421817、rs2687864、rs10257184、rs112583642、rs16862432、rs4611026、rs1896799、rs6494306、rs12538565、rs742600、rs57538712、rs4696917、Affx-50066935、rs6984094、rs132424、rs8046298、rs9429593、rs10867195、rs113211971、rs314346、rs60581100、rs4574502、snp_6_78143760、rs58582728、rs11653646、rs11028125、rs512750、rs12513897、rs35551795、rs136037、Affx-50074792、rs3828535、rs348473、rs7172168、rs6968341、rs10858485、rs2535957、Affx-50012381、rs6430487、rs17724104、rs4321390、rs1567675、rs12910991、rs73943017、snp_3_110660860、rs186248548、rs4731117、rs74877090、rs6462990、rs1781655、rs524534、rs12740394、rs77028522、rs7551916、rs6650546、rs12670760、Affx-50028190、rs6814036、rs7934284、rs1643469、rs6954592、Affx-50091663、Affx-50063178、rs10491042、rs511141、rs70156、snp_19_42942810、rs6770068、rs10789108、rs3763095、rs7616596、rs79532113、rs2761089、rs6657893、rs58761390、rs62334724、rs72732414、rs12915634、rs28542550、rs4312931、rs13344592、rs7629904、rs11205672、rs1409778、rs6010138、rs113517583、rs462904、rs7126715、rs9491372、rs160380、rs1419383、rs2324716、rs7246719、rs6560783、rs187744437、rs1028778、rs12462004、snp_2_201272536、rs1533236、rs8050946、rs1321100、rs1394050、rs9668964、rs2456251、rs205390、rs57579458、rs62428470、rs2823104、rs746517、snp_4_106820913、rs1810467、rs2030383、rs3810113、rs12648715、rs12461381、rs7966949、rs1777610、rs437866、Affx-50039813、rs10133203、rs77600020、rs17666614、rs10814589、rs12703864、rs142439678、Affx-50040986、rs9645819、rs604706、rs9639877、rs740746、rs11105584、rs8008963、rs7535810、rs321237、rs187609、rs9308819、rs726131、rs12035809、rs10785649、rs2978128、rs17811686、rs2074157、rs78194438、rs10519830、rs6892956、rs9544843、rs73488406、rs1994464、rs6567241、rs3861665、rs1467073、rs5998505、rs13324409、rs2665120、rs10008952、rs2384457、rs12756245、rs7726664、rs498573、rs17547641、rs219822、rs2215023、rs219464、rs6574978、rs4552489、rs6885189、rs2252267、rs11803200、rs3812604、rs12880675、rs709332、Affx-50124258、rs3786163、rs10209121、rs276727、rs10842366、rs76640514、rs160143、rs13120400、rs12232294、rs7459565、rs2182921、rs10024084、rs16851032、rs4435670、rs66735056、snp_13_69636922、rs6489415、rs2816290、rs13008832、rs925111、rs2131525、rs1887628、Affx-50109911、rs7756087、rs1334962、rs1759976、rs28410235、rs4451148、rs16873291、Affx-50075238、rs11133452、rs343079、rs4787651、snp_12_91754917、rs1799999、rs304051、rs2390992、rs7386365、rs6105508、rs300489、rs6842950、rs187871467、rs734147、rs10768586、rs3773211、rs72769011、rs9563119、rs11044995、rs7523381、rs993283、rs3912201、rs58560083、rs1532451、rs10429149、rs4861039、Affx-50032203、rs1293391、rs12729871、rs16879934、rs60153055、rs115569073、snp_7_53251789、rs1564595、rs4948530、rs11679490、rs6580376、rs6959431、rs6658481、rs6555701、rs17666091、rs2500503、rs196074、rs956685、rs939500、rs76218592、rs2877122、rs5754739、rs10404868、rs73633714、rs12806537、rs73943624、rs9895150、rs1023934、rs4670838、rs9964594、rs10151608、rs7239592、rs111418263、rs10277950、rs1881251、rs3847656、rs17252350、rs66497055、rs2293600、rs7795932、rs1033218、rs9305252、rs2154366、rs7774406、rs78830764、rs4320330、rs776207、rs17445853、rs394961、rs113109879、rs10811928、rs77166563、rs12539396、rs183356218、rs7223919、rs695589、rs9538755、rs6490126、rs3111458、rs7005270、rs1562481、Affx-50120328、rs145806101、rs482202、rs7796370、rs342457、Affx-50105777、rs9586455、rs4452041、rs1396146、rs12076873、rs2339934、rs73128832、Affx-19125492、rs1790256、rs61443775、rs12137590、rs1375693、rs7573751、rs74461924、rs2726385、rs603495、rs74895607、rs3802873、rs56149054、rs338185、rs10895595、rs79911293、rs9957011、rs9566846、rs76752148、rs1781347、rs10832164、rs11159131、rs12901668、rs4434311、rs6794782、snp_3_164511166、rs1204572、rs17554747、snp_2_237031037、Affx-50092037、rs2571128、rs6452827、rs6928860、rs731710、rs7862967、rs7429951、rs16961570、rs2383671、rs4688116、rs10188397、rs11720817、rs995148、rs2029093、rs6948019、rs75465853、rs7562024、rs1536578、rs145338707、rs62006055、rs3818706、rs963536、rs1341211、rs2118836、rs2894895、rs1990293、rs78930755、rs13093165、rs6427832、rs11634609、rs9386159、rs2267820、rs17737524、rs11157247、rs347668、rs7804857、rs1919413、rs1009458、rs9599719、rs4077193、rs2823730、rs79569648、rs11651301、rs11014033、rs57276252、rs1495274、rs6073623、rs758489、rs12892055、rs4711900、rs12211885、rs1857608、rs7761363、rs6745401、rs56007971、rs16954281、rs12575698、rs1325202、rs3755076、rs12433253、rs352557、rs12195751、rs11100163、rs4973473、rs7738507、rs58759374、snp_15_42620915、rs4373347、rs182943、rs62176863、rs1527556、rs7791649、rs4731227、rs12538516、rs9844423、rs4144874、rs2033031、rs1502763、Affx-50033371、rs10772948、rs11129651、rs1159931、rs4314511、rs1368532、rs1999426、rs1397225、rs1982126、rs9354281、rs413147、rs11750305、rs12452668、rs9917762、rs2175065、rs72671341、rs57610268、rs141931312、rs1386306、rs12597210、rs1553385、rs78262469、rs11018632、snp_13_100836717、rs12022873、rs73148921、rs9939986、rs35559673、rs79920783、rs10508063、rs3746626、rs62418815、rs11606426、rs61620033、rs77334291、rs1343036、rs114098106、rs2514238、rs11835796、rs7335689、rs4309783、rs9346612、rs114511614、rs3114364、rs2861100、rs4853154、rs2144112、rs1472920、rs77707178、rs1852108、rs1679493、rs714235、rs58569787、rs3760780、rs656374、rs2180055、rs42825、rs6935516、rs7359586、rs75548573、rs491959、rs72761017、rs1480082、rs1444401、rs13103429、rs4073507、rs7969452、rs2585356、rs959319、rs7597717、rs62549274、rs10502432、rs10885405、rs12247028、rs9814976、rs17696277、snp_11_123464478、rs2459572、rs62370957、rs7836816、rs1345751、rs1497050、rs59647219、rs59989438、rs5755607、rs2140445、rs74539229、rs10953825、rs4861066、rs2835810、rs2189467、rs6090149、rs10517528、rs10185596、rs4795090、snp_6_81257704、rs1335024、rs12027024、rs9592536、snp_17_44204373、rs183365326、rs1859251、rs78791844、rs12648687、rs61706291、rs9290284、rs6086797、rs12934443、rs11233148、snp_12_25917408、rs9392299、rs139067810、rs7556480、rs71402276、rs7714387、rs2041076、rs11690210、rs650711、rs9891783、rs9830749、rs9456785、rs7517272、rs141202015、rs12356763、rs12765810、rs4865030、rs712273、rs10863005、rs4667287、rs12493149、rs145291420、rs7870323、rs17094802、rs11051199、rs2405971、rs77312220、rs3891553、rs6050858、rs28579160、rs2370771、rs6579788、rs1855478、snp_3_95760837、rs189404485、rs74595052、rs2527161、rs59495531、rs7461441、rs556203、rs6881461、rs1359453、rs1385511、rs12659559、rs7124126、rs1609371、rs17145169、rs12124367、rs34048487、rs13219800、rs9361005、rs1986854、rs7613684、rs9894298、rs11043246、rs2457637、rs10735819、rs10046510、rs4432101、rs11590105、rs6735110、rs117116114、rs2056222、rs4691575、rs17667305、rs3820594、rs4149014、rs2220686、rs11695967、rs9674129、rs4423607、rs72699237、rs11175173、rs473295、rs12096233、rs2279940、rs13187960、rs1996150、rs17709807、rs2055572、rs2412850、rs1561836、rs10956180、rs7740637、rs10065655、rs56143640、rs6082445、rs6907869、rs73121291、rs27729、rs1592495、rs2053204、rs3862837、rs35379293、rs7313455、rs77864643、rs7077070、rs4441451、rs12505943、rs7742134、rs7567588、rs1414803、rs1036755、Affx-50120392、rs6446196、rs3856134、rs6439360、rs1859604、rs41267634、rs5744531、rs2002311、rs10059458、rs9640381、rs71618215、rs470549、rs112047598、rs12140922、rs387812、rs12640664、rs1448、rs17009397、rs4880317、rs1516034、rs17483415、rs3883936、rs10505560、rs10175958、rs55739117、rs1483128、rs690722、rs4904946、rs10073723、rs9932707、rs7895695、rs4808548、rs7215670、rs4586075、rs11688625、rs62442010、rs12523337、rs4819388、rs11154494、rs9547470、rs2607336、rs74925793、Affx-50076034、rs1562058、rs156961、rs2284214、rs72607335、rs4408643、rs7537668、rs76154618、rs2061510、rs10182758、rs10957504、rs2528614、rs7151583、rs74727830、rs7619119、rs13212099、rs306757、rs35734069、Affx-50095743、rs27323、rs934880、rs6769402、rs59356430、rs9316180、rs41758、rs1504357、rs2831488、rs1851214、rs7117365、rs1864672、rs28606003、rs16953619、rs10932481、rs1483552、rs2581969、rs9316109、rs73125226、rs2755213、rs10880627、rs13094850、rs79434533、rs12450654、rs9348124、rs9344698、rs4677503、rs74107410、rs894419、rs75809067、rs35131119、rs2980773、rs1278603、rs12151976、rs116379927、rs62527980、rs41519549、rs11218326、rs11791730、rs4799929、rs776965、rs9913145、Affx-50088834、rs34687、rs6684209、rs35134、rs1037419、rs35003803、rs1858422、rs73315140、rs675497、rs1510186、rs11723615、rs74703074、rs10832869、rs75512663、rs72980706、rs3771943、rs1479300、rs9296776、rs17128651、rs10949426、rs2046172、rs1044717、Affx-50107142、rs17722974、rs1863438、rs11731808、rs9388479、rs1565833、rs17493550、rs4702966、Affx-50067568、rs341078、rs985492、rs116060342、rs78297688、rs12434830、rs11253536、rs17013326、rs9967057、rs2278844、rs7679286、rs12537087、rs4890388、Affx-50074905、rs4368165、rs9668321、rs2718443、rs3843356、rs2345728、rs4751130、rs78312975、rs7915206、rs1168987、rs35469236、rs1616727、rs7957168、rs7635708、rs8050390、rs652612、rs3860173、snp_20_33121001、rs7258236、rs12626191、rs35023505、rs1868045、rs10038007、rs1816116、rs16850914、rs6922708、rs4934644、rs6997338、snp_5_128508737、rs163425、rs2465040、rs300773、rs12151738、rs9971500、rs2616812、rs28398401、rs2545242、rs9582985、rs10509754、rs2664420、rs67742464、rs12505159、rs10200502、rs1100713、rs17157239、rs79375795、rs7237566、rs7549215、rs115125389、rs13134800、rs6033803、rs3787367、rs2868510、rs7711546、rs17376456、rs11164117、rs79639642、rs879045、rs7429731、rs6574575、rs118033085、rs2762893、rs7556785、rs798949、rs74040105、rs6505219、rs36071004、rs11205415、rs10733187、rs10224153、rs72874100、rs7044210、rs77437387、rs9811323、rs1588191、rs4242682、rs11080775、rs79146332、rs13438238、rs12993413、rs345551、rs4624888、rs73160985、rs7186783、rs35521319、rs8066148、rs73408859、rs9827888、rs57578864、rs57218449、rs6583677、rs1005277、rs1932971、rs1229578、rs79380120、rs972297、rs138993466、rs145034195、rs6752818、rs1104868、rs4817640、rs10917334、rs17395202、rs10481651、rs16928390、rs72912007、rs56047396、rs6576634、rs62419701、rs952263、rs2810420、rs9473472、rs7816853、rs1351734、rs1423127、rs12933229、rs1004044、rs9551751、rs17092070、rs4930175、rs12331273、rs4555717、rs6425031、rs4833508、rs17400042、rs79215366、rs7178872、rs4887772、rs12248406、rs73805002、rs62285567、rs10501331、rs2013347、rs12292034、rs1279912、rs298580、rs12723653、rs62005378、rs16956187。、 2. The application of the AISNP tag combination as described in claim 1 in constructing a multi-dimensional ancestry and geographic origin inference model.

3. A method for constructing a multi-dimensional ancestry and geographical origin inference model, characterized in that, Includes the following steps: (1) Collect sample data; (2) Use unsupervised methods to analyze sample data and identify the ancestral components of the samples; (3) Based on the results of step (2), conduct supervised analysis to determine the population grouping; (4) According to Rosenberg's I n -Statistics filters out AISNPs, and the I of AISNPs is used to filter them. n The values ​​are ranked to generate seven groups of AIMs, each group of AIMs comprising at least 50 sites from the AISNPs of claim 1; (5) Use machine learning algorithms to process AIMs groups and build diversified machine learning models.

4. The construction method according to claim 3, characterized in that, In step (1), the collection of sample data includes the following steps: (a) Extract sample data, excluding samples from third-degree relatives; (b) Filter the sites and retain sites that meet the following conditions: deletion rate less than 0.1, Hardy-Weinberg equilibrium p-value greater than 0.001, and minimum allele frequency greater than 0.01; (c) Filter the samples and retain those that meet the following condition: missing rate is less than 0.

1.

5. The construction method according to claim 3, characterized in that, In step (5), the machine learning algorithms include: convolutional neural network, k-nearest neighbor algorithm, logistic regression analysis, random forest, support vector machine, extreme gradient boosting, decision tree, linear regression and / or recurrent neural network.

6. The construction method according to claim 3, characterized in that, It also includes the following steps: (6) Construct a deep neural network model.

7. The construction method according to claim 6, characterized in that, The deep neural network model is constructed using the Locator deep learning method, which includes the following steps: S1: Enter the locator environment, the code is as follows: conda activate locator; S2: Input the genotype data and latitude / longitude coordinates of the training set samples, train the neural network to approximate the relationship between genotype and geographical location, and build a deep neural network model. The code is as follows: locator--vcf data / test_genotypes.vcf.gz--sample_data data / test_sample_data.txt--out out / test / test; S3: Use R software to randomly predict the location information of samples and visualize the 95th, 50th, and 10th quantiles in geospatial space. The code is as follows: Rscript scripts / plot_locator.R --infile out / test_windows / --sample_data data / test_sample_data.txt --out out / test_ --map F.

8. An ancestral and geographical origin inference model constructed by any one of claims 3-7.

9. The application of the ancestry and geographic origin inference model constructed by the construction method according to any one of claims 3-7 in predicting the ethnic background, genetic characteristics and / or geographic location of samples.

10. The application according to claim 9, characterized in that, The samples were sourced from East Asia and Southeast Asia.

Images