Methods for identification and functional assessment of novel rare variant-driven Alzheimer's disease genes

Abstract

The present invention discloses a method for identifying and functionally evaluating new genes for Alzheimer's disease driven by rare variants, relates to the field of molecular biology, and solves the problems of strong racial specificity of rare variants and limited research in non-European populations in the genetic research of AD in the prior art; and low statistical power of rare variants; the present invention obtains DNA from peripheral blood samples of target subjects and performs whole genome sequencing on the DNA in the peripheral blood samples to obtain a cohort-level VCF file; obtains rare and harmful genomic variants in the VCF file by designing a variant priority pipeline for screening; performs optimal sequence-nuclear association tests on genomic variants on a gene basis, and performs repeated and meta-analyses in combination with replication cohorts; evaluates the biological plausibility of candidate genes through network analysis, pathway enrichment, and endophenotype association, and explains the functional consequences of top candidate variants through in vitro experiments, thereby solving the above problems.

Description

Technical Field

[0001] The present invention belongs to the field of molecular biology, specifically to a method for identifying and functionally evaluating new genes for Alzheimer's disease driven by rare mutations. Background Art

[0002] Alzheimer's disease (AD) is a progressive neurodegenerative disease characterized by a gradual decline in cognitive function, ultimately leading to dementia and death. With the aging of the global population, the incidence of AD has increased rapidly, which not only seriously affects the quality of life of patients and their families, but also becomes one of the serious social public health issues.

[0003] Existing technologies mainly explore genomic variations in Alzheimer's disease (AD) through genome-wide association studies (GWAS) and exome sequencing (WES). GWAS uses large-scale population data to identify AD risk sites through case-control analysis. For example, in 2013, the International Alzheimer's Genomics Project (IGAP) conducted a meta-analysis using data from more than 70,000 participants and identified 19 AD susceptibility sites in people of European descent. In contrast, exome sequencing focuses on coding regions and identifies rare, functionally relevant AD coding region mutations through deep sequencing. For example, in a 2017 exome microarray study of 37,022 LOAD (late-onset Alzheimer's disease) cases and 48,402 controls, Sims and other scientists found rare variants in TREM2, PLCG2, and ABI3 that were significantly associated with the entire AD genome. Subsequently, exome sequencing and gene-based testing further discovered rare coding variation signals in ATP8B4 and ABCA16. This discovery is of great significance for understanding the genetic mechanisms of AD and developing potential treatments.

[0004] However, most of the existing AD genetics research focuses on European and American populations, while there are fewer studies on AD genetic factors in East Asian populations. Considering that rare variants have strong ethnic specificity, the applicability of existing risk genes in different races is limited; at the same time, the frequency of rare variants in the population is very low, making it difficult to meet the statistical power of GWAS, and the limited coverage of WES will miss important genomic regions such as non-coding variants and splicing sites, making it difficult to fully capture rare variants.

[0005] Therefore, the present invention proposes a method for identifying and functionally evaluating new genes for Alzheimer's disease driven by rare mutations to solve the above problems. Summary of the invention

[0006] The present invention aims to solve at least one of the technical problems existing in the prior art; to this end, the present invention proposes a method for identifying and functionally evaluating new Alzheimer's disease genes driven by rare variants, which is used to solve the technical problems that most of the existing AD genetics research focuses on European and American populations, while there are fewer studies on AD genetic factors in East Asian populations, resulting in the limited applicability of the discovered risk genes in different races; and the statistical power of rare variants is weak, and it is difficult to obtain significant signals through traditional statistical methods with limited sample sizes.

[0007] To achieve the above objectives, the first aspect of the present invention provides a method for identifying and functionally evaluating new genes for Alzheimer's disease driven by rare mutations, comprising:

[0008] Step 1: Obtain DNA from peripheral blood samples of target subjects and perform whole genome sequencing on the DNA in the peripheral blood samples to obtain a cohort-level VCF file;

[0009] Step 2: Screen rare and harmful genomic variants in VCF files by designing a variant prioritization pipeline;

[0010] Step 3: Perform optimal sequence-nuclear association tests on genomic variants on a gene-by-gene basis and repeat on replication cohorts;

[0011] Step 4: Candidate genes are evaluated through network analysis, pathway enrichment, and endophenotype associations, and the functional consequences of top candidate variants are explained through in vitro experiments.

[0012] Preferably, obtaining DNA from a peripheral blood sample of a target subject and performing whole genome sequencing on the DNA in the peripheral blood sample comprises:

[0013] S1: DNA is collected from peripheral blood samples of target subjects; target subjects refer to elderly subjects in China, including patients with Alzheimer's disease, mild cognitive impairment, subjective cognitive decline, mild cognitive symptoms and healthy controls;

[0014] S2: Perform whole genome sequencing of DNA in peripheral blood samples through a high-throughput sequencing platform to obtain DNA sequences; wherein the high-throughput sequencing platform includes: Illumina platform or MGI platform;

[0015] S3: Reads were aligned to the GRCh38 human reference genome using the BWA-MEM algorithm, followed by removal of duplicate reads and basic quality score recalibration;

[0016] S4: Single nucleotide variants and small insertions and deletions were identified in each sample using Sentieon software and jointly tested to generate cohort-level VCF files.

[0017] It should be noted that "read segment" refers to a data fragment that represents a small sequence of the original DNA or cDNA molecule generated after the sequencing process. When performing whole genome sequencing, the sequencer will read short sequences on the DNA molecule. These short sequences are called "read segments". The length of the read segment depends on the sequencing technology and platform used. For example, on the Illumina sequencing platform, the length of the read segment can be 75bp, 100bp, 150bp or even longer.

[0018] VCF is a standard format for storing sequence variation data, including the location, type, quality score, and occurrence of the variation in different samples.

[0019] Preferably, the method of screening rare and harmful genomic variations in VCF files by designing a variation prioritization pipeline includes:

[0020] The nucleotide variants and small insertion and deletion variants were annotated using VEP v109 to obtain annotation results, including the minor allele frequency of the genome aggregation database, the minor allele frequency of subgroups of the same ethnicity in the gnomAD database, the functional consequences of the variant on the protein, and the potential pathogenicity prediction score of the variant.

[0021] Identify rare and deleterious genomic variants based on annotation results.

[0022] It should be noted that VEP v109 (Variant Effect Predictor version 109) is a specific version of a bioinformatics tool provided by Ensembl for annotating genomic variations. VEP can provide comprehensive annotations for genomic variations such as single nucleotide polymorphisms, small insertions and deletions, and structural variations.

[0023] Preferably, the identification of rare and harmful genomic variations based on the annotation results comprises:

[0024] Based on the annotation results, rare and deleterious genomic variants were identified by setting the following criteria:

[0025] A: The minor allele frequency in the genome aggregation database and the minor allele frequency of the subpopulation of the same ethnicity in gnomAD are both less than 1%;

[0026] B: Minor allele frequency in each study cohort was less than 5%;

[0027] C: Variants that affect protein function annotated by VEP v109, including nonsense variants, splice acceptor / donor variants, frameshift variants, and missense variants;

[0028] D: For loss-of-function variants, they need to be marked as “HC” by the LOFTEE tool. For missense variants predicted to be harmful, the REVEL score needs to be greater than 0.75, or they need to be predicted to be harmful by at least three tools including CADD, SIFT, PolyPhen2, LRT, M-CAP, and MutationTaster.

[0029] E: The variant is located in a gene with a GDI lower than 13.34 or a known AD pathogenic gene.

[0030] It should be noted that gnomAD (Genome Aggregation Database) is a large-scale human genome variation database that integrates variation data from multiple large-scale genome and exome sequencing projects, aiming to provide a comprehensive resource to explore the variation frequency and pattern of the human genome; the Genome Aggregation Database (gnomADv3.1.1) is a specific version of the gnomAD database;

[0031] A "study cohort" refers to a group of samples participating in a specific genetic study, such as a group of Alzheimer's patients or a group of healthy controls. MAF (Minor Allele Frequency) refers to the frequency of the allele with a lower frequency in a specific population. Setting a threshold of MAF below 1% or 5% is to screen out variants with a lower frequency in a specific study cohort and avoid systematic errors;

[0032] The Gene Damage Index (GDI) is an indicator used to evaluate the cumulative effects of mutations in genes. The lower the GDI, the fewer harmful mutations the gene has accumulated in healthy people, which usually means that the gene is more sensitive to the occurrence of mutations and is therefore more likely to be associated with the disease.

[0033] Preferably, the optimal sequence-nuclear association test for genome variation is performed on a gene-by-gene basis, comprising:

[0034] SKAT-O analysis was performed on a gene-by-gene basis in the discovery cohort to identify genes that reached the nominal significance threshold in the discovery cohort; the discovery cohort refers to the collection of peripheral blood samples obtained from the target subjects.

[0035] It should be noted that SKAT-O (Sequence Kernel Association Test-Optimal) is a statistical method mainly used in the analysis of whole genome sequencing data or exome sequencing data to detect the association between rare variants and specific phenotypes (such as diseases); unlike traditional genome-wide association studies (GWAS), which usually focus on common genetic variants, SKAT-O aggregates rare variants onto genes and performs statistical analysis on a gene basis to overcome the problem of low statistical power of single low-frequency variants;

[0036] In statistics, reaching the nominal significance threshold means that the p-value of a test result is less than or equal to the pre-set significance level, which is usually used to determine whether an observed effect is unlikely to be caused by random variation or chance. The nominal significance threshold is the significance standard when multiple comparison correction is not taken into account.

[0037] Preferably, the repeating on the replication queue comprises:

[0038] S1: In the replication cohort, the effect size and direction of genes were evaluated by SKAT-O analysis of the same genes that reached the nominal significance threshold as in the discovery cohort; the replication cohort included: Alzheimer's Disease Neuroimaging Initiative (ADNI) or Alzheimer's Disease Sequencing Project (ADSP);

[0039] S2: The discovery cohort and replication cohort are integrated to form a summarized dataset;

[0040] S3: Meta-analysis of genes that reached the nominal significance threshold was performed using MetaSKAT software.

[0041] It should be noted that the replication cohort is used to verify whether the association observed in the discovery cohort is real. The use of replication cohorts is a key step to ensure the reliability of research results and avoid false positives. If similar association patterns are observed in the replication cohort, then these findings are more likely to be real rather than caused by accidental statistical fluctuations or research biases.

[0042] The effect size is the magnitude of the effect that a variant or set of variants has on a particular phenotype (e.g., disease risk). The direction is the direction of the effect of a variant or set of variants on the phenotype, i.e., increasing or decreasing risk. For example, in a genetic study, if a gene variant is associated with an increased risk of Alzheimer's disease, the effect direction is positive; conversely, if a variant is associated with a decreased risk, the effect direction is negative.

[0043] The purpose of performing a meta-analysis of genes that reached the nominal significance threshold in the data set using MetaSKAT software is to expand the sample size, increase statistical power, and reduce false positive results.

[0044] Preferably, the evaluation of candidate genes by network analysis, pathway enrichment and endophenotype association comprises:

[0045] S1: Evaluation of the distance between candidate genes and core genes of Alzheimer's disease in the protein-protein interaction network based on the human gene connectome database;

[0046] S2: Hierarchical clustering is used to form gene modules with candidate genes and core genes of Alzheimer's disease. If the gene module is enriched in known Alzheimer's disease-related pathways, it means that the candidate gene captures the biological mechanism of Alzheimer's disease.

[0047] S3: SKAT-O analysis was used to evaluate the association of candidate genes with changes in cognitive function scores and with changes in the volume of specific brain regions;

[0048] S4: The differential expression of candidate genes in different brain tissues was evaluated using transcriptome data in the AMP-AD database.

[0049] It should be noted that the known Alzheimer's disease pathways include: Aβ metabolism, lipid metabolism, and immune neuroinflammation; specific brain regions include: hippocampus, amygdala, nucleus accumbens, caudate nucleus, putamen, globus pallidus, thalamus, etc.;

[0050] Candidate genes are genes that show significant association with Alzheimer's disease (AD) in the SKAT-O test;

[0051] AD core genes refer to those genes that have been extensively studied and confirmed to be closely related to the occurrence and development of Alzheimer's disease (AD). These genes have been identified as playing a key role in the pathophysiological process of AD through various research methods, including genome-wide association studies (GWAS), sequencing studies, in vitro experiments, and functional validation. AD core genes are usually involved in typical pathological features of AD (such as β-amyloid deposition and tau protein hyperphosphorylation), and are associated with disease risk, progression, or susceptibility;

[0052] Cognitive function scores are standardized tests used to assess an individual's cognitive abilities and mental status, such as the Mini-Mental State Examination (MMSE). These tests are designed to detect and quantify various aspects of cognitive function, including memory, attention, language skills, executive function, and spatial orientation.

[0053] Preferably, the functional consequences of the top candidate variants are explained by in vitro experiments, including:

[0054] Build cell models carrying top candidate variants;

[0055] The effects of top candidate variants on cell function were assessed by in vitro assays to measure cell status and cell function in cell models.

[0056] It should be noted that the cell status includes: cell proliferation ability, apoptosis and cell cycle arrest, etc.; the cell functions include: expression of key endosomal proteins such as Rab5, early endosomal volume and Aβ secretion, etc.

[0057] Top candidate variants refer to specific genetic variants in candidate genes that have the greatest impact on the onset of AD.

[0058] Compared with the prior art, the present invention has the following beneficial effects:

[0059] 1. Most of the existing AD genetics research focuses on European and American populations, while there are fewer studies on AD genetic factors in East Asian populations, resulting in limited applicability of the discovered risk genes in different races. The present invention systematically explores rare variants of AD in the Chinese population by performing whole genome sequencing on patients with neurodegenerative diseases in China, thereby identifying risk genes unique to the East Asian population, enhancing the understanding of the genetic mechanism of AD in the East Asian population, and providing more representative AD genetics research data.

[0060] 2. Due to the large variety of rare variants and their low frequency in the population, it is difficult to identify a small number of truly pathogenic variants; the present invention obtains all variants including coding and non-coding regions by whole-genome sequencing of DNA in peripheral blood samples, and at the same time, based on comprehensive annotation information and a designed variant priority pipeline, screens rare and harmful genomic variants to exclude potential benign mutations and reduce data noise.

[0061] 3. The statistical power of rare variants is weak, and it is difficult to obtain significant signals through traditional statistical methods (such as GWAS) with limited sample sizes. The present invention solves the problem of low statistical efficiency of rare variants by performing optimal sequence-nuclear association tests on genomic variants on a gene-by-gene basis and repeating the test on a replication cohort, thereby ensuring the statistical reliability of the results. At the same time, candidate genes are evaluated through network analysis, pathway enrichment, and endophenotype association, and the functional consequences of top candidate variants are explained through in vitro experiments, providing additional evidence for the functional analysis of variants. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0063] Figure 1 A schematic diagram of the method steps of an embodiment of the present invention;

[0064] Figure 2 Schematic diagram of the burden of LoF mutations in AD core genes according to an embodiment of the present invention;

[0065] Figure 3 4 is a bar graph of AD candidate genes according to an embodiment of the present invention. DETAILED DESCRIPTION

[0066] The technical solution of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0067] See also Figure 1 The first embodiment of the present invention provides a method for identifying and functionally evaluating new genes for Alzheimer's disease driven by rare mutations, comprising:

[0068] Step 1: Obtain DNA from peripheral blood samples of target subjects and perform whole genome sequencing on the DNA in the peripheral blood samples to obtain a cohort-level VCF file;

[0069] Step 2: Screen rare and harmful genomic variants in VCF files by designing a variant prioritization pipeline;

[0070] Step 3: Perform optimal sequence-nuclear association tests on genomic variants on a gene-by-gene basis and repeat on replication cohorts;

[0071] Step 4: Candidate genes are evaluated through network analysis, pathway enrichment, and endophenotype associations, and the functional consequences of top candidate variants are explained through in vitro experiments.

[0072] Obtain DNA from a peripheral blood sample of the target subject and perform whole genome sequencing on the DNA in the peripheral blood sample, including:

[0073] S1: DNA is collected from peripheral blood samples of target subjects; target subjects refer to elderly subjects in China, including patients with Alzheimer's disease, mild cognitive impairment, subjective cognitive decline, mild cognitive symptoms and healthy controls;

[0074] S2: Perform whole genome sequencing of DNA in peripheral blood samples through a high-throughput sequencing platform to obtain DNA sequences; wherein the high-throughput sequencing platform includes: Illumina platform or MGI platform;

[0075] S3: Reads were aligned to the GRCh38 human reference genome using the BWA-MEM algorithm, followed by removal of duplicate reads and basic quality score recalibration;

[0076] S4: Single nucleotide variants and small insertions and deletions were identified in each sample using Sentieon software and jointly tested to generate cohort-level VCF files.

[0077] The VCF files were screened for rare and harmful genomic variants by designing a variant prioritization pipeline, including:

[0078] The nucleotide variants and small insertion and deletion variants were annotated using VEP v109 to obtain annotation results, including the minor allele frequency of the genome aggregation database, the minor allele frequency of subgroups of the same ethnicity in the gnomAD database, variants affecting protein function, and variant predictions.

[0079] Identify rare and deleterious genomic variants based on annotation results.

[0080] Identify rare and deleterious genomic variants based on annotation results, including:

[0081] Based on the annotation results, rare and deleterious genomic variants were identified by setting the following criteria:

[0082] A: The minor allele frequency in the genome aggregation database and the minor allele frequency of the subpopulation of the same ethnicity in gnomAD are both less than 1%;

[0083] B: Minor allele frequency in each study cohort was less than 5%;

[0084] C: Variants that affect protein function annotated by VEP v109, including nonsense variants, splice acceptor / donor variants, frameshift variants, and missense variants;

[0085] D: For loss-of-function variants, they need to be marked as “HC” by the LOFTEE tool. For missense variants predicted to be harmful, the REVEL score needs to be greater than 0.75, or they need to be predicted to be harmful by at least three tools including CADD, SIFT, PolyPhen2, LRT, M-CAP, and MutationTaster.

[0086] E: The variant is located in a gene with a GDI lower than 13.34 or a known AD pathogenic gene.

[0087] Perform optimal sequence-nuclear association tests on genomic variations on a gene-by-gene basis, including:

[0088] SKAT-O analysis was performed on a gene-by-gene basis in the discovery cohort to identify genes that reached the nominal significance threshold in the discovery cohort; the discovery cohort refers to the collection of peripheral blood samples obtained from the target subjects.

[0089] Repeat on the replication queue, including:

[0090] S1: In the replication cohort, the effect size and direction of genes were evaluated by SKAT-O analysis of the same genes that reached the nominal significance threshold as in the discovery cohort; the replication cohort included: Alzheimer's Disease Neuroimaging Initiative (ADNI) or Alzheimer's Disease Sequencing Project (ADSP);

[0091] S2: The discovery cohort and replication cohort are integrated to form a summarized dataset;

[0092] S3: Meta-analysis of genes that reached the nominal significance threshold was performed using MetaSKAT software.

[0093] Candidate genes were evaluated through network analysis, pathway enrichment, and endophenotype associations, including:

[0094] S1: Evaluation of the distance between candidate genes and core genes of Alzheimer's disease in the protein-protein interaction network based on the human gene connectome database;

[0095] S2: Hierarchical clustering is used to form gene modules with candidate genes and core genes of Alzheimer's disease. If the gene module is enriched in known Alzheimer's disease-related pathways, it means that the candidate gene captures the biological mechanism of Alzheimer's disease.

[0096] S3: SKAT-O analysis was used to evaluate the association of candidate genes with changes in cognitive function scores and with changes in the volume of specific brain regions;

[0097] S4: The differential expression of candidate genes in different brain tissues was evaluated using transcriptome data in the AMP-AD database.

[0098] Functional consequences of top candidate variants explained through in vitro experiments, including:

[0099] Build cell models carrying top candidate variants;

[0100] The effects of top candidate variants on cell function were assessed by in vitro assays to measure cell status and cell function in cell models.

[0101] For example: This embodiment is called the ZIB_AD cohort, with a total of 988 subjects, including 239 Alzheimer's disease (AD) patients, 230 mild cognitive impairment (MCI) patients, 206 subjective cognitive decline (SCD) patients, 177 mild cognitive symptoms (SCS) patients and 136 normal controls; among them, the diagnosis of AD patients followed the guidelines of the National Association on Aging Alzheimer's Association (NIA-AA); MCI patients were diagnosed according to the Jak / Bondi criteria, and the classification of SCD patients was based on the criteria proposed by Jessen et al.

[0102] In this example, individuals with MCI, SCD or SCS are collectively referred to as cognitively impaired patients (CI). Additional inclusion criteria for normal controls (NC) are as follows:

[0103] (1) No medical or family history of other neurological or psychiatric diseases, such as Parkinson's disease, depression, epilepsy, and neurodevelopmental delay;

[0104] (2) No serious physical illness;

[0105] (3) Have adequate vision and hearing.

[0106] All participants underwent a battery of cognitive assessments, including the Mini-Mental State Examination (MMSE), the Montreal Cognitive Assessment-Basic (MoCA_B), and the Addenbrooke's Cognitive Assessment-III (ACE-III). Some participants also underwent blood RNA sequencing, plasma metabolic measures, and T1-weighted brain MRI scans.

[0107] Step 1: Obtain DNA from peripheral blood samples of target subjects and perform whole genome sequencing on the DNA in the peripheral blood samples to obtain a cohort-level VCF file;

[0108] For example, genomic DNA was extracted from peripheral blood samples of each subject by using the QIAamp DNA Blood Kit (Qiagen), and libraries were constructed using the TruSeq Nano DNA Sample Preparation Kit.

[0109] Whole-genome sequencing was performed on the Illumina Novaseq 6000 platform, generating DNA sequences with a read length of 150 bp and an average sequencing depth of 33.46 times.

[0110] Sequence alignment and variant detection were performed using Sentieon Genomics software (v202001.02) according to the best practices given by the Broad Institute. In brief, the reads were aligned to the GRCh38 human reference genome using the Sentieon BWA-MEM algorithm, followed by removal of duplicate reads and basic quality score recalibration, identification of single nucleotide variants (SNVs) and small insertions and deletions (INDELs) in each sample, and finally joint detection to generate cohort-level VCF files.

[0111] To ensure variant quality, variant quality score recalibration (VQSR) was performed, and further quality control was performed at the variant and individual levels. A variant was considered low quality if it met at least one of the following criteria:

[0112] (1) The VQSR filter status is not PASS;

[0113] (2) the missingness rate in the cohort exceeded 5%;

[0114] (3) Violation of Hardy-Weinberg equilibrium.

[0115] At the same time, the following criteria were applied to exclude low-quality individuals:

[0116] (1) The mutation loss rate is higher than 5%;

[0117] (2) the inferred gender does not match the record;

[0118] (3) abnormal heterozygosity (3 standard deviations from the mean);

[0119] (4) Abnormal transition / transversion ratio. The above filtering was implemented based on PLINK (v1.9). In addition, King software was used to identify close relatives within three levels, and only one person was retained in each identified group.

[0120] Variant quality score recalibration (VQSR) allows the quality of variants to be assessed and corrected, ensuring that only high-quality variants are used for subsequent analysis.

[0121] Step 2: Screen rare and harmful genomic variants in VCF files by designing a variant prioritization pipeline;

[0122] For example: This example designs a variant annotation and prioritization pipeline to assist in screening rare and high-impact genomic variants.

[0123] First, SNVs and INDELs were annotated using VEP v109 (based on Ensembl 109), and only variants located on autosomes were considered;

[0124] Second, the following criteria were defined to identify rare, potentially deleterious variants:

[0125] (1) The minor allele frequency (MAF) in the Genome Aggregation Database (gnomAD v2.1.1) and the MAF of subpopulations of the same ethnicity (e.g., East Asian population) in gnomAD were all lower than 1%;

[0126] (2) the MAF in each study cohort was less than 5% to avoid systematic errors;

[0127] (3) VEP-annotated variants that may affect protein function, including nonsense variants, splice acceptor / donor variants, frameshift variants, and missense variants;

[0128] (4) For loss-of-function variants (LoF), they must be marked as “HC” by the LOFTEE tool; for predicted harmful missense variants (DMis), the REVEL score must be greater than 0.75, or they must be predicted as harmful by at least three tools including CADD, SIFT, PolyPhen2, LRT, M-CAP, and MutationTaster;

[0129] (5) It must be located in genes with a Gene Damage Index (GDI) lower than 13.34 or known AD pathogenic genes, because genes with high variation in healthy individuals are unlikely to be associated with the disease.

[0130] See also Figure 2As a validation of previous knowledge, this example first focuses on rare possible pathogenic variants in known AD pathogenic genes. Through literature review, 25 AD core genes that have been widely confirmed by genome-wide association studies, sequencing or in vitro experiments were compiled. At the same time, benign variants in ClinVar were excluded, and the focus was on more destructive and convincing LoF variants. The results of this example confirmed that the incidence of rare LoF variants in AD core genes was higher in AD and CI groups than in normal controls, supporting the involvement of rare and high-impact variants in AD pathology. At the same time, after excluding variants present in the normal control group, 10 rare LoF variants were identified in AD core genes, including 4 frameshifts, 4 termination gains, and 2 splicing LoF variants. They were carried by 10 AD cases (4.2%) in the ZIB_AD cohort, of which more than half (6) were APOEε4 negative, indicating that the contribution of these variants was independent of the known APOEε4 allele. In addition, 4 variants were not present in the gnomAD non-East Asian population, indicating that they are ultra-rare and may be unique to the East Asian population. In summary, the variant prioritization pipeline of this embodiment can effectively identify and screen rare AD variants that may be pathogenic.

[0131] Step 3: Perform optimal sequence-nuclear association tests on genomic variants on a gene-by-gene basis and repeat on replication cohorts;

[0132] For example: This example aims to identify potential AD risk genes that are susceptible to rare and harmful variants to overcome the challenge of low statistical power of rare variants:

[0133] First, SKAT-O analysis was performed on a gene-by-gene basis in the ZIB_AD cohort to identify genes that reached the nominal significance threshold (P < 0.01);

[0134] Then, SKAT-O analysis was performed in the replication cohorts (ADNI and ADSP cohorts) on the same genes that reached the nominal significance threshold in the ZIB_AD cohort to assess the effect size and direction of these genes;

[0135] The ZIB_AD cohort was integrated with the ADNI and ADSP cohorts to form a summary dataset;

[0136] Meta-analysis was performed on genes that reached the nominal significance threshold in the dataset using the MetaSKAT R package.

[0137] Separate models were established for LoF and DMis variants, and genes with minor allele counts less than 2 were excluded. Gender, age, years of education (if applicable), and number of APOEε4 alleles were included as covariates. Specific analyses were performed with the help of R software packages SKAT (v2.2.5) and MetaSKAT (v0.82).

[0138] See also Figure 3 In the ZIB_AD cohort, a total of 90 genes reached the suggestive significance threshold (SKAT-O P < 0.01). Among them, 11 genes passed the significance threshold in at least one replication cohort (MetaSKAT-O FDR < 0.05 and the direction was consistent with ZIB_AD), so they were identified as candidate genes. Among them, the HNMT gene ranked highest in the meta-analysis (MetaSKAT-O ZIB_AD-ADNI FDR = 1.08 × 10 -5 ), which encodes a histamine-metabolizing enzyme that is essential for the regulation of the histamine system in the central nervous system and is also the inhibitory target of the early AD drug tacrine. Two DMis variants were identified in this gene, both of which were carried by normal controls, supporting the role of neuronal histamine in AD pathology and indicating possible protective sites in HNMT.

[0139] Another important gene, RABEP1, was the risk gene with the strongest signal (MetaSKAT-O ZIB_AD-ADSP FDR = 1.18 × 10 -3 ), whose rare DMis variants have an increased burden in the AD group. This gene encodes the Rab5 effector protein, which plays an important role in endocytic membrane docking and fusion and is widely expressed in various tissues such as the nervous system.

[0140] Previous studies have nominated RABEP1 as an AD-related gene, but its functional mechanism in AD is still unclear, and this invention provides a new perspective for understanding the role of this gene in AD. Overall, this invention identified 11 potential AD candidate genes through cross-cohort SKAT-O analysis, among which HNMT and RABEP1 are the most important protective genes and pathogenic genes, respectively.

[0141] Step 4: Evaluate candidate genes through network analysis, pathway enrichment, and endophenotype associations, and interpret the functional consequences of top candidate variants through in vitro experiments:

[0142] For example: To further explore the functional rationality of candidate genes in the AD mechanism, this example uses the human gene connectome (HGC) to evaluate their biological connection with AD core genes. The results showed that the average distance between candidate genes and core genes was significantly shorter than random sampling (10,000 permutation tests, P = 2.19 × 10-2). In addition, the candidate genes and core genes formed three gene modules through hierarchical clustering, and were enriched in known AD pathways such as Aβ metabolism (module 1, module 2, module 3), lipid metabolism (module 2), and immune neuroinflammation (module 1, module 3). The above results show that rare variants reproduce known AD mechanisms and strengthen the biological association between candidate genes and AD.

[0143] This embodiment also studies the association of 11 candidate genes with AD-related endophenotypic characteristics such as cognition and brain volume. SKAT-O analysis showed that four candidate genes were significantly correlated with changes in MMSE scores (P<0.05), exceeding random expectations. In addition, RABEP1 and VIPR1 were significantly correlated with the volume reduction of multiple AD marker brain regions (such as hippocampus and amygdala). Consistent with existing studies, individuals carrying rare harmful variants of RABEP1 showed severe atrophy of the putamen and thalamus.

[0144] This example evaluates the differential expression of candidate genes in different brain tissues by using transcriptome data from the AMP-AD database. The results show that 9 candidate genes are significantly dysregulated in at least one brain region, especially in the parahippocampal gyrus (PHG), temporal cortex (TCX), and superior temporal gyrus (STG) related to learning and memory, supporting the potential role of these candidate genes in the pathogenesis of AD.

[0145] This example further explored individual sites with large effects, focusing on the RABEP1 p.R845W mutation that occurs most frequently in AD, and constructed a HEK293T cell model carrying the RABEP1 p.R845W mutation.

[0146] In vitro experiments showed that compared with wild-type cells, the proliferation ability of RABEP1 p.R845W mutant cells was significantly reduced, early and late apoptosis were significantly increased, and the cell cycle was arrested in the G2 phase. In addition, the expression of Rab5 and Rabex-5 proteins in RABEP1p.R845W mutant cells was significantly upregulated, and EEA1 immunofluorescence showed an increase in the diameter of early endosomes, indicating that there was a disorder in endosomal function. Further ELISA results showed that the level of Aβ42 in RABEP1p.R845W mutant cells was increased, the level of Aβ40 was moderately reduced, and the ratio of Aβ42 / Aβ40 was significantly increased.

[0147] The above phenomena indicate that the RABEP1 p.R845W mutation leads to the accumulation of neurotoxic Aβ by interfering with endosomal processes, thereby promoting the pathological progression of AD. In summary, the present invention proves the biological rationality of the candidate gene from multiple angles and emphasizes the importance of RABEP1 as a potential intervention target.

[0148] The above embodiments are only used to illustrate the technical method of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. A method for identifying and functionally evaluating new genes for Alzheimer's disease driven by rare variants, characterized in that: include: Step 1: Obtain DNA from peripheral blood samples of target subjects and perform whole genome sequencing on the DNA in the peripheral blood samples to obtain a cohort-level VCF file; Step 2: Screen rare and harmful genomic variants in VCF files by designing a variant prioritization pipeline; Step 3: Perform optimal sequence-nuclear association tests on genomic variants on a gene-by-gene basis and repeat on replication cohorts; Step 4: Evaluate candidate genes through network analysis, pathway enrichment, and endophenotype associations, and interpret the functional consequences of top candidate variants through in vitro experiments; The VCF files were screened for rare and harmful genomic variants by designing a variation prioritization pipeline, including: Nucleotide variants and small insertion and deletion variants were annotated using VEP v109 to obtain annotation results; rare and harmful genomic variants were identified based on the annotation results; The annotation-based identification of rare and harmful genomic variations includes: Based on the annotation results, rare and deleterious genomic variants were identified by setting the following criteria: A: The minor allele frequency in the genome aggregation database and the minor allele frequency of the subpopulation of the same ethnicity in gnomAD are both less than 1%; B: Minor allele frequency in each study cohort was less than 5%; C: Variants that affect protein function annotated by VEP v109, including nonsense variants, splice acceptor / donor variants, frameshift variants, and missense variants; D: For loss-of-function variants, they need to be marked as "HC" by the LOFTEE tool. For missense variants predicted to be harmful, the REVEL score needs to be greater than 0.75, or they need to be predicted to be harmful by at least three tools including CADD, SIFT, PolyPhen2, LRT, M-CAP, and MutationTaster. E: The variant is located in a gene with a GDI lower than 13.34 or a known AD pathogenic gene; The optimal sequence-nuclear association test for genome variation is performed on a gene-by-gene basis, including: Perform SKAT-O analysis on a gene-by-gene basis in the discovery cohort to identify genes that reach the nominal significance threshold in the discovery cohort; the discovery cohort refers to the collection of peripheral blood samples obtained from the target subjects; The above is repeated on the replication queue, including: S1: In the replication cohort, the effect size and direction of genes were evaluated by SKAT-O analysis of the same genes that reached the nominal significance threshold as in the discovery cohort; the replication cohort included: Alzheimer's Disease Neuroimaging Initiative (ADNI) or Alzheimer's Disease Sequencing Project (ADSP); S2: The discovery cohort and replication cohort were integrated to form a summarized dataset; S3: Meta-analysis of genes that reached the nominal significance threshold was performed using MetaSKAT software; The candidate genes were evaluated by network analysis, pathway enrichment and endophenotype association, including: S1: Evaluation of the distance between candidate genes and core genes of Alzheimer's disease in the protein-protein interaction network based on the human gene connectome database; S2: Hierarchical clustering is used to form gene modules with candidate genes and core genes of Alzheimer's disease. If the gene module is enriched in known Alzheimer's disease-related pathways, it means that the candidate gene captures the biological mechanism of Alzheimer's disease. S3: SKAT-O analysis was used to evaluate the association of candidate genes with changes in cognitive function scores and with changes in the volume of specific brain regions; S4: The differential expression of candidate genes in different brain tissues was evaluated using transcriptome data in the AMP-AD database.

2. The method for identifying and functionally evaluating new genes for Alzheimer's disease driven by rare mutations according to claim 1, characterized in that: The step of obtaining DNA from a peripheral blood sample of a target subject and performing whole genome sequencing on the DNA in the peripheral blood sample comprises: S1: DNA is collected from peripheral blood samples of target subjects; target subjects refer to elderly subjects in China, including patients with Alzheimer's disease, mild cognitive impairment, subjective cognitive decline, mild cognitive symptoms and healthy controls; S2: Perform whole genome sequencing of DNA in peripheral blood samples through a high-throughput sequencing platform to obtain DNA sequences; wherein the high-throughput sequencing platform includes: Illumina platform or MGI platform; S3: Reads were aligned to the GRCh38 human reference genome using the BWA-MEM algorithm, followed by removal of duplicate reads and basic quality score recalibration; S4: Single nucleotide variants and small insertions and deletions were identified in each sample using Sentieon software and jointly tested to generate cohort-level VCF files.

3. The method for identifying and functionally evaluating new genes for rare variant-driven Alzheimer's disease according to claim 1, characterized in that: The functional consequences of the top candidate variants are explained by in vitro experiments, including: Build cell models carrying top candidate variants; The effects of top candidate variants on cell function were assessed by in vitro assays to measure cell status and cell function in cell models.

Images