CN-122024860-A - Screening method of skeletal muscle aging markers, clock construction and kit

CN122024860ACN 122024860 ACN122024860 ACN 122024860ACN-122024860-A

Abstract

The invention discloses a screening method of skeletal muscle aging markers, clock construction and a kit, and belongs to the technical field of biomedical engineering. The method comprises the steps of obtaining multi-mode biomarker data of an individual, wherein the multi-mode biomarker data comprise an inflammation marker, a metabolic marker, a hormone marker and a muscle function marker, inputting the data into a trained skeletal muscle organism age prediction model to obtain skeletal muscle organism age (SMBA), screening specific methylation site sets related to skeletal muscle aging states from genome-wide methylation sites based on the skeletal muscle organism age (SMBA), constructing a methylation aging clock model by using the methylation data of the site sets, and finally accurately predicting and evaluating the skeletal muscle aging states of the individual by using the model. The invention realizes the conversion from multi-modal clinical indexes to specific single-dimensional molecular markers, and the constructed aging clock model and derived products can be used for personalized evaluation and early risk early warning of skeletal muscle aging.

Inventors

CHEN CHENG
LIU YOUWEI
HE JUN
ZHOU PENG
Wen Zhaoxiang

Assignees

苏州大学附属第四医院(苏州市独墅湖医院)

Dates

Publication Date: 20260512
Application Date: 20260113

Claims (11)

1. A screening method of skeletal muscle aging markers is characterized by comprising the following steps: s1, acquiring multi-mode biomarker data and whole genome methylation sites of healthy individuals, wherein the multi-mode biomarker data comprises inflammatory markers, metabolic markers, hormone markers and muscle function markers; S2, inputting the multi-mode biomarker data into a trained skeletal muscle organism age prediction model to obtain skeletal muscle organism ages (SMBA) of the individuals; S3, screening specific methylation site sets related to the aging state of skeletal muscle from the whole genome methylation sites based on the age (SMBA) of the skeletal muscle organisms.
2. The method according to claim 1, wherein in step S2, the skeletal muscle biological age prediction model is a biologically a priori constrained adapted transducer architecture, wherein different biomarker classes use independent and unshared Q/K/V projection matrices.
3. The method according to claim 1, wherein in step S3, the screening comprises: Calculating a measure of correlation between the methylation level of each candidate methylation site and the skeletal muscle organism age (SMBA); screening out a subset of candidate sites based on the correlation metric; Quantifying the contribution of each site in the subset of candidate sites to the skeletal muscle organism age (SMBA) using an interpretive artificial intelligence method; Determining the set of specific methylation sites based on the contribution.
4. The method of claim 3, wherein the correlation metric comprises a pearson correlation coefficient, a linear regression coefficient, or a mutual information value, wherein the criterion for the saliency screening is |r| > 0.6 and the FDR corrected q value < 0.05 when the correlation metric is pearson correlation coefficient (r), or wherein the criterion for the saliency screening is normalized regression coefficient (β) absolute value greater than 0.6 and the FDR corrected q value < 0.05 when the correlation metric is linear regression coefficient, or wherein the criterion for the saliency screening is MI > 0.1 and the FDR corrected q value < 0.05 when the correlation metric is mutual information value (MI).
5. The method of claim 3, wherein the interpretive artificial intelligence method is SHAP value analysis.
6. The method of claim 1, wherein the genes associated with the sites in the set of specific methylation sites are functionally enriched for core biological processes directly related to skeletal muscle aging such as PI3K-Akt signaling pathway, hippo signaling pathway, transcriptional regulation, and muscle tissue development.
7. A clock construction method based on a screening method of skeletal muscle aging markers for use as a methylation aging clock model for predicting skeletal muscle biological age, comprising the steps of: S4, constructing a methylation aging clock model for predicting Skeletal Muscle Clock Age (SMCA) based on methylation data of the specific methylation site collection and corresponding skeletal muscle organism age (SMBA) of any one of claims 1-6.
8. A methylation detection kit for skeletal muscle aging status assessment, comprising a detection component for detecting the set of specific methylation sites of claim 1 or 6.
9. The kit of claim 8, wherein the detection assembly comprises reagents for bisulfite conversion of genomic DNA, and specific amplification primers and/or detection probes designed for the specific set of methylation sites.
10. The kit of claim 8, wherein the detection component is configured to effect detection by either multiplex fluorescent quantitative PCR or a microarray chip.
11. The kit of claim 8, further comprising a quality control component comprising a reference gene locus, a negative control, and/or a standard of known methylation level.

Description

Screening method of skeletal muscle aging markers, clock construction and kit Technical Field The invention belongs to the technical field of biomedical engineering and aging evaluation, and particularly relates to a method for constructing a skeletal muscle biological age (SkeletalMuscleBiologicalAge, SMBA) model and a skeletal muscle aging clock (SkeletalMuscleClock, SMC) by processing a multi-mode biomarker through an artificial intelligent model based on a Transformer framework, and a detection kit for developing specific methylation sites screened by the method. The invention can be used for accurate assessment of skeletal muscle aging state and early warning of sarcopenia risk. Background Skeletal muscle aging significantly increases the risk of falls, fractures, disability, and total cause death. As the global population ages, it has become a significant public health problem affecting the health, quality of life, and individual ability of the elderly. The clinically advanced phenotype of skeletal muscle aging is sarcopenia, and clinical diagnosis is mainly based on measurement of muscle mass (as measured by dual energy X-ray absorption method DXA, bioelectrical impedance method BIA) and muscle function (as grip strength, pace). However, these methods generally only identify "disease states" in which significant muscle loss or decline in function has occurred, lack the ability to evaluate early, continuous and quantitative skeletal muscle decline in aging processes, and have late early warning windows, which are detrimental to early intervention. In recent years, biomarker-based aging assessment studies have been increasingly emphasized. Studies have shown that inflammatory markers (e.g., CRP, IL-6), metabolic markers (e.g., serum albumin, creatinine), hormonal levels (e.g., IGF-1, testosterone), and muscle function indicators are all closely related to the development and progression of sarcopenia. However, the single biomarker has limited prediction efficacy, and integration of multi-modal data is required to obtain the ideal prediction effect. The traditional method has the defects in fusing heterogeneous multi-mode data and extracting intersystem interaction characteristics with clear biological explanation, and lacks an interpretable and transformable technical path from clinical macroscopic indicators to high-specificity molecular markers. The artificial intelligent model represented by a transducer architecture has great potential in the biomedical field by virtue of the strong characteristic fusion and pattern recognition capability. However, there is no report on the systematic application of the same to skeletal muscle aging-specific assessment and skeletal muscle biological age construction. On the other hand, DNA methylation has been demonstrated to accurately predict aging processes as a key epigenetic regulatory mechanism. An "epigenetic clock" (e.g., horvath clock) constructed based on whole genome methylation data has made significant progress in predicting overall biological age. However, the existing methylation clocks are mostly "pan tissue" senescence clocks, lacking specific predictive power for the risk of senescence and sarcopenia in skeletal muscle, a specific tissue. In addition, the direct use of hundreds or thousands of methylation sites to construct a clock results in high detection cost and complex operation, which is not beneficial to the industrial application and popularization of the technology. In view of the foregoing, there is a great need to develop a aging evaluation technique that has both high prediction accuracy and skeletal muscle specificity, and is convenient for large-scale popularization and application. In particular, it is necessary to establish an interpretable technical transformation path from macroscopic, clinically accessible, multimodal biomarkers to microscopic, specific DNA methylation markers, and develop a corresponding detection kit to achieve convenient, accurate assessment of skeletal muscle aging status and early intervention early warning. Disclosure of Invention The present invention aims to overcome the above-mentioned drawbacks of the prior art, solving the following key problems: 1. the early warning window period is late, and the existing sarcopenia assessment method is mostly intervened after obvious loss or functional decline of muscles occurs, so that early warning and intervention are difficult to realize. 2. The model integration and prediction precision is insufficient, the traditional model is insufficient in integration and utilization of multi-mode biomarkers (especially core indexes of skeletal muscle quality and functions), deep feature interaction is difficult to mine, and the prediction precision is limited. 3. The lack of tissue-specific assessment tools is that the existing high-precision methylation aging clocks are mostly universal tissue clocks, and the specific assessment capability for skeletal muscle aging states