Information processing method and device for evaluating and predicting genetic risk, equipment and medium
By using machine learning models to predict genetic risk scores based on genotype data, this approach addresses the problem that existing technologies struggle to comprehensively reflect the multifactorial genetic basis of hepatocellular carcinoma, achieving high efficiency and reliability in genetic risk assessment and optimizing screening strategies.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-11-24
- Publication Date
- 2026-03-10
AI Technical Summary
Existing technologies are insufficient to fully reflect the multifactorial genetic basis of hepatocellular carcinoma (HCC), and in clinical practice, non-invasive fibrosis scoring is insufficient to move individuals with high genetic risk to an appropriate screening starting point and match them with appropriate follow-up intensity.
A machine learning model was used to predict genetic risk scores based on genotype data. A genetic risk assessment model was constructed using Bayesian sparse regression and elastic network regression algorithms. Combined with clinical risk assessment results, a genetic risk level and a recommended screening start age were generated.
It achieves high efficiency and reliability in genetic risk assessment, and the output directly includes risk level and screening start age recommendations, avoiding the subjectivity of manual interpretation and improving the cost-effectiveness and health benefits of screening.
Smart Images
Figure CN121641391A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of information processing technology, and in particular to information processing methods, apparatus, devices and media for assessing and predicting genetic risks. Background Technology
[0002] Hepatocellular carcinoma (HCC) is one of the most common primary liver cancers worldwide. Because it often presents with no specific symptoms in its early stages, clinical practice commonly uses imaging techniques (such as ultrasound, computed tomography, and magnetic resonance imaging) and blood biomarkers for population screening and follow-up. These methods have clear value within established populations and standardized procedures, but their effectiveness is easily affected by factors such as baseline risk, follow-up intervals, the degree of liver fibrosis, and operator differences. Especially when dealing with individuals who "have not yet shown obvious clinical / imaging abnormalities but have a high congenital genetic susceptibility," existing procedures struggle to identify them promptly and provide differentiated starting age and monitoring intensity. Genetic factors, as stable and quantifiable lifelong markers, have long been expected to be used for disease risk assessment. With the development of genomics, genome-wide association studies (GWAS) have systematically revealed a large number of genetic variations associated with hepatic steatosis, metabolic abnormalities, and fibrosis, providing a scientific basis for introducing genetic stratification into early HCC screening. Polygenic Risk Score (PRS) can aggregate small effects from hundreds or thousands of loci into a personalized risk measure, featuring "one-time testing, long-term applicability," providing additional, early, and relatively independent risk information beyond traditional clinical indicators. However, existing PRS applications related to HCC still have two limitations: First, many models focus on a single trait (such as liver fat content), making it difficult to comprehensively reflect the multifactorial genetic basis of HCC; second, the widely used non-invasive fibrosis score in clinical practice mainly characterizes the current liver state, which is a complementary dimension to lifelong genetic risk, but it is difficult to move individuals with high genetic risk to a more appropriate screening starting point and match them with a more suitable follow-up intensity. Summary of the Invention
[0003] In view of this, the purpose of this invention is to provide an information processing method, apparatus, device, and medium for assessing and predicting genetic risks, capable of directly predicting genetic risk scores based on genotype data and outputting corresponding genetic risk assessment results. The specific solution is as follows:
[0004] In a first aspect, this application discloses an information processing method for assessing and predicting genetic risks, applied to a computer device, comprising:
[0005] Receive the genotype data of the user to be predicted and input the genotype data to the preset genetic risk assessment model;
[0006] The target genotype data is calculated using the preset genetic risk assessment model to output a target comprehensive risk score; wherein, the genetic risk assessment model is a machine learning model that has learned the correlation between the genetic effects of multiple target traits and the risk of the target state;
[0007] Based on the target comprehensive risk score, a genetic risk prediction result is generated, which includes the genetic risk level, a suggested starting age for risk screening for the user to be predicted, and a joint risk warning after integrating with the clinical risk assessment results.
[0008] Optionally, before inputting the genotype data to be predicted into the preset genetic risk assessment model, the method further includes:
[0009] Obtain genome-wide association analysis data for each phenotype associated with multiple target traits;
[0010] The set of single nucleotide polymorphism (SNP) site weights corresponding to each target trait in the genome-wide association analysis data is calculated using a preset Bayesian sparse regression algorithm; the set of SNP site weights is a set of site weights used to construct a single-trait multigene risk score for the target trait.
[0011] Historical genotype data of each training user in the training samples are obtained, and the target single trait polygenic risk score for each training user corresponding to each target trait is calculated based on the weight set of multiple single nucleotide polymorphism sites.
[0012] The target single trait polygenic risk scores are jointly modeled using an elastic network regression algorithm to learn the combined weights of each target single trait polygenic risk score when predicting the target state, thereby obtaining a preset genetic risk assessment model; wherein, the preset genetic risk assessment model includes the set of single nucleotide polymorphism site weights and the combined weights.
[0013] Optionally, the phenotype includes any one or more of the following: liver fat content-related traits, body mass index, glycated hemoglobin, blood pressure-related indicators, blood lipid-related indicators, liver function enzyme indicators, and platelet count.
[0014] Optionally, generating a genetic risk prediction result based on the target comprehensive risk score, including a genetic risk level, a suggested starting age for risk screening for the user to be predicted, and a joint risk warning fused with clinical risk assessment results, includes:
[0015] The target comprehensive risk score is compared with a preset risk stratification threshold to determine the genetic risk level of the user to be predicted; wherein, the preset risk stratification threshold is a set risk score threshold for preset low genetic risk level, preset medium genetic risk level, and preset high genetic risk level.
[0016] Genetic risk prediction results containing the genetic risk level are generated.
[0017] Optionally, generating a genetic risk prediction result based on the target comprehensive risk score, including a genetic risk level, a suggested starting age for risk screening for the user to be predicted, and a joint risk warning fused with clinical risk assessment results, includes:
[0018] Based on the genetic risk level, a suggested screening start age for the target condition is generated by using a preset mapping relationship between the risk level and the screening start age. The preset mapping relationship between the risk level and the screening start age includes a preset mapping relationship between low genetic risk level and the first screening start age, a preset mapping relationship between medium genetic risk level and the second screening start age, and a preset mapping relationship between high genetic risk level and the third screening start age, with the first screening start age, the second screening start age, and the third screening start age decreasing sequentially.
[0019] Genetic risk prediction results are generated, which include the genetic risk level and the corresponding recommended screening start age.
[0020] Optionally, generating a genetic risk prediction result based on the target comprehensive risk score, including a genetic risk level, a suggested starting age for risk screening for the user to be predicted, and a joint risk warning fused with clinical risk assessment results, includes:
[0021] Obtain the clinical biomarker data of the user to be predicted, and calculate the non-invasive fibrosis score based on the clinical biomarker data to obtain the clinical risk assessment result;
[0022] The target comprehensive risk score is integrated with the clinical risk assessment results to generate a joint risk alert;
[0023] Genetic risk prediction results are generated, which include the target comprehensive risk score and joint risk indication.
[0024] Optionally, the step of fusing the target comprehensive risk score with the clinical risk assessment results to generate a joint risk alert includes:
[0025] The target comprehensive risk score is compared with a preset genetic risk threshold to obtain a first comparison result;
[0026] The non-invasive fibrosis score is compared with a preset clinical risk threshold to obtain a second comparison result;
[0027] When the first comparison result is that the target comprehensive risk score is greater than the preset genetic risk threshold and the non-invasive fibrosis score is greater than the preset clinical risk threshold, a joint risk warning indicating that enhanced screening is required is generated.
[0028] Secondly, this application discloses an information processing apparatus for assessing and predicting genetic risks, applied to a computer device, comprising:
[0029] The data receiving module is used to receive the genotype data to be predicted from the user to be predicted, and input the genotype data to be predicted into the preset genetic risk assessment model;
[0030] The scoring prediction module is used to calculate the genotype data to be predicted using the preset genetic risk assessment model, so as to output the target comprehensive risk score; wherein, the genetic risk assessment model is a machine learning model that has learned the correlation between the genetic effects of multiple target traits and the risk of the target state;
[0031] The risk output module is used to generate a genetic risk prediction result based on the target comprehensive risk score, which includes the genetic risk level, a suggested starting age for risk screening for the user to be predicted, and a joint risk warning after merging with the clinical risk assessment results.
[0032] Thirdly, this application discloses an electronic device, including:
[0033] Memory, used to store computer programs;
[0034] A processor is configured to execute the computer program to implement the steps of the aforementioned disclosed information processing method for assessing and predicting genetic risks.
[0035] Fourthly, this application discloses a computer-readable storage medium for storing a computer program; wherein, when the computer program is executed by a processor, it implements the steps of the aforementioned disclosed information processing method for assessing and predicting genetic risks.
[0036] Therefore, this application discloses receiving genotype data of users to be predicted and inputting the genotype data into a preset genetic risk assessment model; calculating the genotype data using the preset genetic risk assessment model to output a target comprehensive risk score; wherein, the genetic risk assessment model is a machine learning model that has learned the correlation between the genetic effects of multiple target traits and the risk of the target state; based on the target comprehensive risk score, generating a genetic risk prediction result that includes the genetic risk level, a suggested age for starting risk screening for the user to be predicted, and a joint risk warning after fusion with the clinical risk assessment result. It is evident that directly processing genotype data and outputting a comprehensive score through a preset genetic risk assessment model avoids the subjectivity and inconsistency of manual interpretation, improving assessment efficiency and reliability. Furthermore, the output genetic risk prediction result directly includes a comprehensive prediction result of risk level, screening start age, and joint risk warning, directly mapping genetic information into specific clinical actions without the need for secondary analysis and decision-making. By providing a comprehensive, stratified, timed, and integrated conclusion in a single output, the system lays the foundation for reducing unnecessary screening (for low-risk individuals) and strengthening surveillance in advance (for high-risk individuals), thereby improving the overall cost-effectiveness of screening and the health benefits to the population. Attached Figure Description
[0037] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.
[0038] Figure 1 This is a flowchart of an information processing method for assessing and predicting genetic risks disclosed in this application;
[0039] Figure 2 This is a schematic diagram of an information processing device for assessing and predicting genetic risks disclosed in this application.
[0040] Figure 3 This is a structural diagram of an electronic device disclosed in this application. Detailed Implementation
[0041] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.
[0042] Hepatocellular carcinoma (HCC) is one of the most common primary liver cancers worldwide. Because it often presents with no specific symptoms in its early stages, clinical practice commonly uses imaging techniques (such as ultrasound, CT, and MRI) and blood biomarkers for population screening and follow-up. These methods have clear value within established populations and standardized procedures, but their effectiveness is easily affected by factors such as baseline risk, follow-up intervals, the degree of liver fibrosis, and operator differences. Especially when dealing with individuals who "have not yet shown obvious clinical / imaging abnormalities but have a high congenital genetic susceptibility," existing procedures struggle to identify them promptly and provide differentiated starting age and monitoring intensity. Genetic factors, as stable and quantifiable lifelong markers, have long been expected to be used for disease risk assessment. With the development of genomics, genome-wide association studies (GWAS) have systematically revealed a large number of genetic variations associated with hepatic steatosis, metabolic abnormalities, and fibrosis, providing a scientific basis for introducing genetic stratification into early HCC screening. Polygenic risk scores (PRS) can accumulate small effects from hundreds or thousands of loci into an individualized risk measure, featuring "one-time testing, long-term applicability," providing additional, early, and relatively independent risk information beyond traditional clinical indicators. However, existing PRS applications related to HCC still have two limitations: first, many models focus on a single trait (such as liver fat content), making it difficult to comprehensively reflect the multifactorial genetic basis of HCC occurrence; second, the widely used non-invasive fibrosis score in clinical practice mainly characterizes the current liver state, which is a complementary dimension to lifelong genetic risk, but it is difficult to move individuals with high genetic risk to a more appropriate screening starting point and match them with a more appropriate follow-up intensity.
[0043] Therefore, the present invention provides an information processing scheme for assessing and predicting genetic risks, which can directly predict genetic risk scores based on genotype data and output corresponding genetic risk assessment results.
[0044] Reference Figure 1 As shown, the present invention provides an information processing method for assessing and predicting genetic risks, applied to a computer device, comprising:
[0045] Step S11: Receive the genotype data of the user to be predicted and input the genotype data to be predicted into the preset genetic risk assessment model.
[0046] In this embodiment, the genotype data to be predicted specifically refers to genotype information of single nucleotide polymorphism sites covering the entire genome, obtained through gene chip or gene sequencing technology, after quality control processing, such as GWAS summary statistics and individual genotype data, specifically SNP (Single Nucleotide Polymorphism) sites.
[0047] In this embodiment, the obtained genotype data to be predicted is input into a preset risk assessment model to predict the genetic risk of hepatocellular carcinoma. Specifically, the preset risk assessment model in this scheme is for predicting the genetic risk of hepatocellular carcinoma. Therefore, the genotype data to be predicted can specifically be the genotype data of all single nucleotide polymorphism sites related to the following target traits of the user to be predicted. The target traits specifically include: liver proton density fat fraction, body mass index, glycated hemoglobin, systolic and diastolic blood pressure, triglycerides, high-density lipoprotein cholesterol, aspartate aminotransferase, alanine aminotransferase, and platelet count.
[0048] Step S12: Calculate the genotype data to be predicted using the preset genetic risk assessment model to output a target comprehensive risk score; wherein, the genetic risk assessment model is a machine learning model that has learned the correlation between the genetic effects of multiple target traits and the risk of the target state.
[0049] In this embodiment, before inputting the genotype data to be predicted into the preset genetic risk assessment model, the method further includes: acquiring genome-wide association analysis (GWIA) data of each phenotype associated with multiple target traits; calculating the set of single nucleotide polymorphism (SNP) site weights for each target trait in the GWIA data using a preset Bayesian sparse regression algorithm; the SNP site weight set is a set of site weights used to construct a single-trait polygenic risk score for the target trait; acquiring the historical genotype data of each training user in the training samples and calculating the target single-trait polygenic risk score for each training user corresponding to each target trait based on the multiple SNP site weight sets; jointly modeling the target single-trait polygenic risk scores using an elastic network regression algorithm to learn the combined weights of each target single-trait polygenic risk score when predicting the target state, thereby obtaining the preset genetic risk assessment model; wherein, the preset genetic risk assessment model includes the set of SNP site weights and the combined weights. The phenotypes include any one or more of the following: liver fat content-related traits, body mass index, glycated hemoglobin, blood pressure-related indicators, blood lipid-related indicators, liver function enzyme indicators, and platelet count.
[0050] Understandably, aggregated statistical data related to ten hepatocellular carcinoma phenotypes were obtained from a large genome-wide association study database. These phenotypes included: liver proton density fat fraction (the gold standard for quantitative assessment of liver fat content via MRI imaging), body mass index, glycated hemoglobin, systolic and diastolic blood pressure, triglycerides, high-density lipoprotein cholesterol, aspartate aminotransferase, alanine aminotransferase, and platelet count. These phenotypes cover the complete pathophysiological process from fat deposition and metabolic abnormalities to hepatocellular damage. For each phenotype, the PRS-CS (auto) Bayesian sparse regression algorithm was used for calculation. This algorithm utilizes the linkage disequilibrium reference panel provided by the 1000 Genomes Project, automatically estimates global contraction parameters through Gibbs sampling, and calculates the optimal set of single nucleotide polymorphism (SNP) site weights for each phenotype without requiring manual setting of P-value thresholds. Compared with traditional methods, it effectively preserves the contribution of small-effect sites, reduces redundancy caused by linkage disequilibrium, and thus improves the accuracy of genetic signal capture.
[0051] Furthermore, quality-controlled genotype data for each training user in the training sample population was obtained. Based on the aforementioned set of single nucleotide polymorphism (SNP) site weights for each phenotype, a single-trait polygenic risk score for each training user across ten phenotypes was calculated. These single-trait scores were then standardized to ensure they had the same scale across the training set. On this basis, a resilient network regression algorithm was used to jointly model these standardized single-trait scores. The resilient network algorithm combines the advantages of L1 (Lasso) and L2 (Ridge) regularization. By optimizing model parameters on the training set, it automatically learns the optimal combination weights for each single-trait score in predicting hepatocellular carcinoma risk, while also addressing multicollinearity by shrinking low-contribution or redundant phenotypic weights to near zero, ultimately forming a sparse and robust risk assessment model. The final preset genetic risk assessment model includes the set of SNP site weights for each phenotype and cross-phenotype combination weights learned through the resilient network. This ensures that the model can comprehensively capture information from multiple genetic pathways involved in the development and progression of hepatocellular carcinoma.
[0052] Specifically, PRS-CS (auto) is a Bayesian sparse continuous prior multi-SNP regression framework that utilizes an external linkage disequilibrium reference to perform Gibbs sampling on a block linkage disequilibrium matrix, automatically estimating the global shrinkage parameter (without requiring a grid search p-value threshold). Its advantages include preserving small effect loci information, reducing linkage disequilibrium redundancy, and lowering the risk of overfitting. The software used is PRS-CS (Python); LD reference: a 1,000-person genome reference panel (matched to the target population), using the official block linkage disequilibrium file; Input: GWAS summary statistics for the target trait (unified by allele direction); Key parameters: global shrinkage parameter φ is estimated using auto mode (no manual setting required); other parameters are default; Output: a SNP weight file for each trait (effect size * shrinkage estimate), subsequently multiplied position-by-position by the individual genotype and summed to obtain the single-trait PRS; Standardization: each PRS is standardized on the training set for cross-trait comparability and subsequent weighting. Furthermore, a multi-trait polygenic risk score (metaPRS) is constructed. In the aforementioned steps, a polygenic risk score (PRS) for each trait (such as liver proton density fat fraction PDFF, body mass index BMI, aspartate aminotransferase AST, alanine aminotransferase ALT, etc.) has been obtained using the PRS-CS (auto) method. To achieve a holistic, stable, and interpretable genetic risk profile for hepatocellular carcinoma (HCC), cross-trait weights are learned using Elastic-Net regression in the training set, thereby constructing the multi-trait polygenic risk score (metaPRS). To optimize the contribution of each trait polygenic risk score (PRS) in HCC risk assessment, this invention uses Elastic-Net regression to jointly model and learn the weights of the standardized trait PRS (such as PDFF, BMI, AST, ALT, etc.) in the training set. Elastic-Net incorporates both L1 (Lasso) and L2 (Ridge) regularization, enabling stable estimation even in the presence of multicollinearity. It automatically reduces redundant or weakly correlated traits to low weights or even zero, thus obtaining robust and interpretable cross-trait weights (β). Subsequently, in the validation and application phase, the PRS for each trait is first standardized for new individuals using the mean and standard deviation of the training set. Then, a comprehensive score (metaPRS) is constructed by linearly combining the coefficients obtained during training. This score integrates multidimensional genetic information such as fat deposition, metabolic abnormalities, and hepatocellular damage on the same scale, reflecting the individual's overall genetic susceptibility. When the independent contribution of certain traits to HCC is insufficient, their coefficients are reduced to zero by the algorithm without affecting the consistency of the overall framework. Low / medium / high genetic risk stratification can be achieved based on population quantile thresholds (e.g., P20 / P80), and it can be used in parallel with clinical non-invasive fibrosis scores for subsequent triage and follow-up strategy development.
[0053] In addition, the PRS-CS (auto) method can be replaced with any multi-gene modeling algorithm, such as LDpred2, SBayesR, lassosum, or clumping+thresholding (C+T). The above methods construct individual-level PRS using different principles for subsequent cross-trait weighting and stratification. The GWAS source can be replaced with external publicly available GWAS, multi-center Meta-GWAS, local cohort GWAS, or summary statistics from cross-phenotype joint analysis, etc., which will not be elaborated here.
[0054] In this embodiment, based on the preset genetic risk assessment model after the above training, weighted calculation is performed on the genotype data to be predicted under the single nucleotide polymorphism site weight set and the combined weight to output the final target comprehensive risk score.
[0055] Step S13: Generate a genetic risk prediction result including a genetic risk level, a recommended starting age for risk screening for the user to be predicted, and a combined risk prompt after fusion with the clinical risk assessment result based on the target comprehensive risk score.
[0056] In this embodiment, the target comprehensive risk score is compared with a preset risk stratification threshold to determine the genetic risk level of the user to be predicted; wherein, the preset risk stratification threshold is the corresponding risk score threshold set for the preset low genetic risk level, preset medium genetic risk level, and preset high genetic risk level; a genetic risk prediction result including the genetic risk level is generated. It can be understood that first is risk stratification and threshold setting. Specifically, in the training set or representative population, metaPRS is standardized, and corresponding risk score thresholds are set according to the <20% / 20% - 80% / >80% quantiles to obtain low / medium / high three-level genetic risk groups. Specifically, low risk: metaPRS ≤ 20%; medium risk: 20% < metaPRS ≤ 80%; high risk: metaPRS > 80%. Based on the above preset risk stratification threshold, the genetic risk level corresponding to the current target comprehensive risk score can be obtained.
[0057] In addition, in addition to P20 / P80, P25 / P75, tertiles (T1 / T2), deciles (D1 / D9), or programmed cut-off points for fixed population acceptable NNS thresholds can be used. Different threshold rules all achieve the goal of mapping metaPRS to low / medium / high risk layers to drive different screening strategies, which will not be elaborated here.
[0058] In this embodiment, based on the genetic risk level, a suggested screening start age for the target state is generated through a preset mapping relationship between risk level and screening start age. This preset mapping relationship includes a preset mapping relationship between low genetic risk level and the first screening start age, a preset mapping relationship between medium genetic risk level and the second screening start age, and a preset mapping relationship between high genetic risk level and the third screening start age, with the first, second, and third screening start ages decreasing sequentially. A genetic risk prediction result containing the genetic risk level and the corresponding suggested screening start age is generated. It is understood that, based on the results of metaPRS stratification, personalized screening start age suggestions are provided for individuals in different risk groups: Low-risk group: According to metaPRS calculations, the suggested screening start age for individuals in the low-risk group is 65 years old. This group has a lower risk of hepatocellular carcinoma, so screening can be postponed; Medium-risk group: For individuals in the medium-risk group, screening is recommended to begin at age 50. Individuals in the medium-risk group may have potential metabolic problems or liver damage and need regular screening; High-risk group: For individuals in the high-risk group, screening is recommended to begin at age 43. High-risk individuals should undergo regular checkups and monitoring as early as possible to ensure early detection of hepatocellular carcinoma. Therefore, based on the screening start age for the different risk groups mentioned above, the screening age corresponding to the genetic risk level under the target comprehensive risk score is output.
[0059] In this embodiment, clinical biomarker data of the user to be predicted is obtained, and a non-invasive fibrosis score is calculated based on the clinical biomarker data to obtain a clinical risk assessment result; the target comprehensive risk score is fused with the clinical risk assessment result to generate a joint risk warning; and a genetic risk prediction result containing the target comprehensive risk score and the joint risk warning is generated.
[0060] Specifically, the target comprehensive risk score is compared with a preset genetic risk threshold to obtain a first comparison result; the non-invasive fibrosis score is compared with a preset clinical risk threshold to obtain a second comparison result; when the first comparison result is that the target comprehensive risk score is greater than the preset genetic risk threshold and the non-invasive fibrosis score is greater than the preset clinical risk threshold, a combined risk warning indicating the need for enhanced screening is generated. It is understood that combining the target comprehensive risk score (metaPRS) with the clinical non-invasive fibrosis score improves the accuracy and efficiency of early screening for hepatocellular carcinoma (HCC). Specifically, metaPRS is combined with common non-invasive fibrosis scoring tools (such as the Forns score, FIB-4, APRI, etc.) to form a more comprehensive and accurate screening system. The non-invasive fibrosis score is a diagnostic tool that assesses the degree of liver fibrosis through blood biomarkers (such as ALT, AST, platelet count, etc.) or imaging examinations. Compared with traditional liver biopsy methods, non-invasive fibrosis scoring has the advantages of low risk, non-invasiveness, and ease of operation, and is therefore widely used in clinical practice. Common non-invasive fibrosis scoring tools include: Forns score: Based on factors such as ALT, AST, platelet count, and age, it assesses the risk of liver fibrosis using a mathematical formula. A high Forns score effectively indicates the severity of liver fibrosis. FIB-4 index: Combines age, AST, ALT, and platelet count to calculate an index of liver fibrosis, with high predictive accuracy. APRI score: Combines AST and platelet count to assess the degree of liver fibrosis and is widely used for screening patients with chronic liver disease. Therefore, combining metaPRS with non-invasive fibrosis scoring can effectively combine genetic factors with clinical indicators, thereby improving the sensitivity and accuracy of screening. The specific steps are as follows:
[0061] Data Collection: Clinical data for each individual was collected, including common liver health indicators such as alanine aminotransferase (ALT), aspartate aminotransferase (AST), and platelet count (PLT). This clinical data reflects an individual's liver function and health status, particularly the degree of liver fibrosis.
[0062] Calculating Noninvasive Fibrosis Scores: Based on collected clinical data, an individual's liver fibrosis score is calculated using commonly used noninvasive fibrosis scoring tools (such as the Forns score, FIB-4, APRI, etc.). Each noninvasive fibrosis score reflects an individual's liver health status, especially the degree of progression of liver fibrosis, which is crucial for assessing the risk of hepatocellular carcinoma.
[0063] Calculating metaPRS: First, a metaPRS is calculated for each individual using a genetic risk assessment model. This score comprehensively considers the impact of multiple genetic risk factors (such as liver proton density fat fraction, body mass index, aspartate aminotransferase, alanine aminotransferase, etc.) on the risk of hepatocellular carcinoma. metaPRS provides an individual's overall risk assessment at the genetic level.
[0064] Combining with Non-invasive Fibrosis Score: The calculated metaPRS is used in conjunction with non-invasive fibrosis scores (such as the Forns score, FIB-4, etc.). The non-invasive fibrosis score provides important clinical information about an individual's liver health status. Combined with genetic risk (metaPRS), a comprehensive assessment model is formed. This model can more comprehensively assess an individual's HCC risk, integrating genetic risk and clinical liver health status, providing more accurate risk predictions.
[0065] Risk Grouping: Individuals can be grouped into different risk groups based on a combination of metaPRS and noninvasive fibrosis score. If an individual has a high metaPRS value and a high noninvasive fibrosis score, it indicates a high-risk status both genetically and clinically. These individuals should be classified as high-risk and require enhanced screening and monitoring. Conversely, if an individual has a low metaPRS value and a low noninvasive fibrosis score, they are in the low-risk group and screening can be postponed.
[0066] Optimize screening strategies: Based on individual risk groups, high-risk groups should be prioritized for more frequent screening, which may include earlier screening start age and higher monitoring frequency. Low-risk groups can have screening appropriately delayed based on their genetic and clinical characteristics to avoid wasting resources.
[0067] In addition to Forns, FIB-4, and APRI clinical tools, NAFLD Fibrosis Score, ELF detection, transient elastography (FibroScan LSM / CAP), magnetic resonance elastography, or liver cancer-related biomarkers / combined scores (AFP, PIVKA-II, GALAD variable combination) can be used as alternatives, which will not be elaborated further. The reference age for risk estimation can be replaced by 45 years, 60 years, or as specified by regional guidelines, or an absolute risk threshold (e.g., 10-year risk ≥ X%) can be directly used to map the starting age. The substitution of the starting age anchor will not be elaborated further here.
[0068] Therefore, this application discloses receiving genotype data of users to be predicted and inputting the genotype data into a preset genetic risk assessment model; calculating the genotype data using the preset genetic risk assessment model to output a target comprehensive risk score; wherein, the genetic risk assessment model is a machine learning model that has learned the correlation between the genetic effects of multiple target traits and the risk of the target state; based on the target comprehensive risk score, generating a genetic risk prediction result that includes the genetic risk level, a suggested age for starting risk screening for the user to be predicted, and a joint risk warning after fusion with the clinical risk assessment result. It is evident that directly processing genotype data and outputting a comprehensive score through a preset genetic risk assessment model avoids the subjectivity and inconsistency of manual interpretation, improving assessment efficiency and reliability. Furthermore, the output genetic risk prediction result directly includes a comprehensive prediction result of risk level, screening start age, and joint risk warning, directly mapping genetic information into specific clinical actions without the need for secondary analysis and decision-making. By providing a comprehensive, stratified, timed, and integrated conclusion in a single output, the system lays the foundation for reducing unnecessary screening (for low-risk individuals) and strengthening surveillance in advance (for high-risk individuals), thereby improving the overall cost-effectiveness of screening and the health benefits to the population.
[0069] Reference Figure 2 As shown, the present invention also discloses an information processing apparatus for assessing and predicting genetic risks, applied to a computer device, comprising:
[0070] Data receiving module 11 is used to receive the genotype data to be predicted of the user to be predicted, and input the genotype data to be predicted into the preset genetic risk assessment model;
[0071] The scoring prediction module 12 is used to calculate the genotype data to be predicted using the preset genetic risk assessment model, so as to output the target comprehensive risk score; wherein, the genetic risk assessment model is a machine learning model that has learned the correlation between the genetic effects of multiple target traits and the risk of the target state;
[0072] The risk output module 13 is used to generate a genetic risk prediction result based on the target comprehensive risk score, which includes the genetic risk level, a suggested starting age for risk screening for the user to be predicted, and a joint risk warning after merging with the clinical risk assessment results.
[0073] Therefore, the synergy of the three modules—data reception, scoring prediction, and risk output—significantly improves the efficiency and standardization of genetic information processing, avoiding the subjective bias and inconsistencies introduced by reliance on manual interpretation in traditional methods. Secondly, the risk output module provides comprehensive decision support information integrating genetic risk levels, recommended screening start ages, and combined risk warnings. This directly transforms genetic data into actionable clinical management guidelines, enabling early and intensive monitoring of high-risk individuals while avoiding over-screening of low-risk groups. This significantly improves the overall effectiveness and cost-effectiveness of early hepatocellular carcinoma screening.
[0074] Furthermore, embodiments of this application also disclose an electronic device, Figure 3 This is a structural diagram of an electronic device 20 according to an exemplary embodiment. The content of the diagram should not be construed as limiting the scope of this application.
[0075] Figure 3 This is a schematic diagram of the structure of an electronic device 20 provided in an embodiment of this application. Specifically, the electronic device 20 may include: at least one processor 21, at least one memory 22, a power supply 23, a communication interface 24, an input / output interface 25, and a communication bus 26. The memory 22 stores a computer program, which is loaded and executed by the processor 21 to implement the relevant steps in the information processing method for assessing and predicting genetic risks disclosed in any of the foregoing embodiments. Alternatively, the electronic device 20 in this embodiment may specifically be an electronic computer.
[0076] In this embodiment, the power supply 23 is used to provide operating voltage for each hardware device on the electronic device 20; the communication interface 24 can create a data transmission channel between the electronic device 20 and external devices, and the communication protocol it follows can be any communication protocol applicable to the technical solution of this application, and is not specifically limited here; the input / output interface 25 is used to acquire external input data or output data to the outside world, and its specific interface type can be selected according to specific application needs, and is not specifically limited here.
[0077] The processor 21 may include one or more processing cores, such as a quad-core processor or an octa-core processor. The processor 21 may be implemented using at least one hardware form selected from DSP (Digital Signal Processing), FPGA (Field-Programmable Gate Array), and PLA (Programmable Logic Array). The processor 21 may also include a main processor and a coprocessor. The main processor, also known as a CPU (Central Processing Unit), is used to process data in the wake-up state; the coprocessor is a low-power processor used to process data in the standby state. In some embodiments, the processor 21 may integrate a GPU (Graphics Processing Unit), which is responsible for rendering and drawing the content to be displayed on the screen. In some embodiments, the processor 21 may also include an AI (Artificial Intelligence) processor, which is used to handle computational operations related to machine learning.
[0078] In addition, the memory 22, as a carrier for resource storage, can be a read-only memory, random access memory, disk or optical disk, etc. The resources stored thereon can include operating system 221, computer program 222, etc., and the storage method can be temporary storage or permanent storage.
[0079] The operating system 221 manages and controls the various hardware devices and computer programs 222 on the electronic device 20 to enable the processor 21 to perform calculations and processing on the massive amounts of data 223 in the memory 22. The operating system 221 can be Windows Server, Netware, Unix, Linux, etc. The computer program 222, in addition to including a computer program capable of performing the information processing method for assessing and predicting genetic risks as disclosed in any of the foregoing embodiments, may further include computer programs capable of performing other specific tasks. The data 223 may include data received by the electronic device from external devices, as well as data collected by its own input / output interface 25.
[0080] Furthermore, this application also discloses a computer-readable storage medium for storing a computer program; wherein, when the computer program is executed by a processor, it implements the aforementioned disclosed information processing method for assessing and predicting genetic risks. Specific steps of this method can be found in the corresponding content disclosed in the foregoing embodiments, and will not be repeated here.
[0081] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since it corresponds to the method disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to in the method section.
[0082] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application. The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein can be implemented directly in hardware, software modules executed by a processor, or a combination of both. The software module may be located in random access memory (RAM), memory, read-only memory (ROM), electrically programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), registers, hard disks, removable disks, CD-ROMs (Compact Disc-Read Only Memory), or any other form of storage medium known in the art.
[0083] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.
[0084] The solution provided by the present invention has been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only for helping to understand the method and core ideas of the present invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.
Claims
1. An information processing method for assessing a predictive genetic risk, characterized by, The application is applied to a computer device, comprising: receiving to-be-predicted genotype data of a to-be-predicted user, and inputting the to-be-predicted genotype data into a preset genetic risk assessment model; calculating the to-be-predicted genotype data by the preset genetic risk assessment model to output a target comprehensive risk score; wherein the genetic risk assessment model is a machine learning model that learns the correlation between the genetic effects of multiple target traits and target state risks; generating a genetic risk prediction result comprising a genetic risk level, a risk screening starting age suggestion for the to-be-predicted user, and a joint risk prompt after fusing a clinical risk assessment result according to the target comprehensive risk score.
2. The information processing method for evaluating a predictive genetic risk according to claim 1, characterized by, Before the to-be-predicted genotype data is input into the preset genetic risk assessment model, the method further comprises: obtaining whole genome association analysis data of each phenotype related to multiple target traits; calculating a single nucleotide polymorphism site weight set of the whole genome association analysis data corresponding to each target trait by a preset Bayesian sparse regression algorithm; the single nucleotide polymorphism site weight set is a site weight set used for constructing a single-trait polygenic risk score of a target trait; obtaining historical genotype data of each training user in a training sample and calculating a target single-trait polygenic risk score of each training user corresponding to each target trait based on multiple single nucleotide polymorphism site weight sets; jointly modeling the target single-trait polygenic risk scores by an elastic network regression algorithm to learn the combination weights of each target single-trait polygenic risk score in predicting a target state, so as to obtain a preset genetic risk assessment model; wherein the preset genetic risk assessment model comprises the single nucleotide polymorphism site weight set and the combination weights.
3. The information processing method for evaluating a predictive genetic risk according to claim 2, characterized by, The phenotype comprises any one or several of a liver fat content related trait, a body mass index, a glycated hemoglobin, a blood pressure related index, a blood lipid related index, a liver function enzyme index, and a platelet count.
4. The information processing method for evaluating a predictive genetic risk according to claim 1, characterized by, The genetic risk prediction result comprises: comparing the target comprehensive risk score with a preset risk stratification threshold to determine the genetic risk level of the to-be-predicted user; wherein the preset risk stratification threshold is a corresponding risk score threshold set for a preset low genetic risk level, a preset medium genetic risk level, and a preset high genetic risk level; generating a genetic risk prediction result comprising the genetic risk level.
5. The information processing method for assessing a predictive genetic risk according to claim 4, characterized in that, The genetic risk prediction result comprises: generating a screening starting age suggestion for the target state based on the genetic risk level and a preset risk level-screening starting age mapping relationship; wherein the preset risk level-screening starting age mapping relationship includes a preset low genetic risk level-first screening starting age mapping relationship, a preset medium genetic risk level-second screening starting age mapping relationship, and a preset high genetic risk level-third screening starting age mapping relationship, and the first screening starting age, the second screening starting age, and the third screening starting age decrease in turn; generating a genetic risk prediction result including the genetic risk level and the corresponding screening starting age suggestion.
6. The information processing method for evaluating a predictive genetic risk according to claim 4, characterized by, The genetic risk prediction result generated according to the target comprehensive risk score includes a genetic risk level, a risk screening starting age suggestion for the user to be predicted, and a joint risk prompt after fusion with a clinical risk assessment result. Obtaining clinical biomarker data of the user to be predicted, and calculating a non-invasive fibrosis score based on the clinical biomarker data to obtain a clinical risk assessment result; fusing the target comprehensive risk score and the clinical risk assessment result to generate a joint risk prompt; generating a genetic risk prediction result including the target comprehensive risk score and the joint risk prompt.
7. The information processing method for evaluating a predictive genetic risk according to claim 6, characterized by, The fusion of the target comprehensive risk score and the clinical risk assessment result to generate a joint risk prompt includes: comparing the target comprehensive risk score with a preset genetic risk threshold to obtain a first comparison result; comparing the non-invasive fibrosis score with a preset clinical risk threshold to obtain a second comparison result; when the first comparison result is that the target comprehensive risk score is greater than the preset genetic risk threshold, and the non-invasive fibrosis score is greater than the preset clinical risk threshold, a joint risk prompt indicating that intensive screening is needed is generated.
8. An information processing apparatus for assessing a predictive genetic risk, characterized by, application in a computer device, comprising: a data receiving module for receiving predicted genotype data of a user to be predicted and inputting the predicted genotype data into a preset genetic risk assessment model; a score prediction module for calculating the predicted genotype data by the preset genetic risk assessment model to output a target comprehensive risk score; wherein the genetic risk assessment model is a machine learning model that learns the correlation between the genetic effect of multiple target traits and the risk of the target state; a risk output module for generating a genetic risk prediction result including a genetic risk level, a risk screening starting age suggestion for the user to be predicted, and a joint risk prompt after fusion with a clinical risk assessment result according to the target comprehensive risk score.
9. An electronic device, comprising: including: a memory for saving a computer program; a processor for executing the computer program to implement the steps of the information processing method for evaluating and predicting genetic risk according to any one of claims 1 to 7.
10. A computer-readable storage medium, characterized in that, for storing a computer program; wherein the computer program is executed by a processor to implement the steps of the information processing method for evaluating and predicting genetic risk according to any one of claims 1 to 7.