Age determination method and device

Abstract

The present invention relates to the determination of age. In particular, the present invention relates to a method for determining an age indicator, and a method for determining the age of an individual. The method is based on data comprising the DNA methylation level of a set of genomic DNA sequences. Preferably, the age indicator is determined by applying a regression method comprising a least absolute shrinkage and selection operator (LASSO) on the data, preferably by combining the method with a subsequent stepwise regression. In addition, the present invention relates to a set of genomic DNA sequences and a gene set, and their use in diagnosing the health status and / or physical condition of an individual and identifying molecules that affect aging. In other aspects, the present invention relates to a chip or a kit, in particular a chip or a kit that can be used to detect the DNA methylation level of the set of genomic DNA sequences.

Description

[0001] The present invention relates to the determination of age. In particular, the present invention relates to a method for determining an age indicator, and a method for determining the age of an individual. The method is based on data comprising DNA methylation levels of a set of genomic DNA sequences. Preferably, the age indicator is determined by applying a regression method comprising a least absolute shrinkage and selection operator (LASSO) on the data, preferably by combining the method with a subsequent stepwise regression. In addition, the present invention relates to an ensemble of genomic DNA sequences and a gene set, and their use in diagnosing the health status and / or physical condition of an individual and identifying molecules that affect aging. In other aspects, the present invention relates to a chip or a kit, in particular a chip or a kit that can be used to detect the DNA methylation level of the set of genomic DNA sequences. Background Art

[0002] When people get older, their bodies change in many ways, such as teeth wearing down, joints wearing out, muscles weakening, and thinking skills declining. But while a person's health may generally decline as they age, there is still a lot of variation in health between individuals, even for people born on the same date. As a result, some people age faster than others.

[0003] Furthermore, in one study that looked at the age of twins, it was found that only about 25% of average lifespan was determined by genetic inheritance, while lifestyle and environmental factors accounted for the remaining 75% of the variability in lifespan.

[0004] It has been found that some diseases are more common in humans as chronological age increases. However, chronological age is not an ideal indicator of an individual's age-related health status (commonly referred to as "biological age"). Determining an age that is more similar to biological age may help assess whether an individual is at higher risk for aging-related diseases (such as Alzheimer's disease). If the determined age is greater than the chronological age, it may indicate the need to take preventive measures, such as lifestyle changes, to prevent or slow the progression of aging-related diseases. Determining an alternative age may also help improve diagnosis, such as helping to evaluate whether aging-related diseases should be of concern.

[0005] Furthermore, if the chronological age of an individual is unknown, a surrogate age (albeit different) can be used as an indicator of chronological age. If the determination of the surrogate age is based on a biological sample, it can also be used, for example, in forensics when traces of a criminal's blood are found at a crime scene.

[0006] It has further been proposed that certain groups of individuals age slower than others, for example people in certain countries with particular local habits in terms of nutrition, etc. Determining the age of individuals in different groups may help identify factors that influence biological age. See Alegria-Torres et al., Epigenomics, 2011 Jun;3(3):267-277.

[0007] It should be noted that in the case where both the chronological age and the age different from the chronological age are known, the difference from the chronological age may be indicated instead of the absolute value.

[0008] It has been proposed to determine the age of a human individual based on the methylation level of genomic DNA sequences found in that individual. In particular, reference is made to WO2012 / 162139. In WO2012 / 162139, it has been proposed to observe the cytosine methylation of genomic DNA selected from one or more of a large set of CG loci names.

[0009] Reference may also be made to WO2015 / 48665, which lists other CpG sites.

[0010] It is also proposed in document WO 2012 / 162139 that a reference (training) data set of, for example, 100 individuals of different chronological ages can be collected using a specific technical platform and organization, and then a specific multivariate linear model for fitting the reference data set is designed, the reference data set including the methylation level of the CpG loci obtained for each individual. In order to estimate the coefficient, for example, least squares regression has been proposed. The unknown alternative age of these individuals not included in the training data set is then determined using the coefficient assigned to each CpG locus. It has been proposed to use "leave-one-out analysis" to determine these coefficients. In this "leave-one-out analysis", all other objects except one object in the reference data set are fitted with a multivariate regression model, and then the predicted value is compared with the chronological age of the object left. In addition, WO 2012 / 162139 has proposed to test to screen out the best predictor to improve the accuracy of the model.

[0011] Nevertheless, despite the fact that an indicator of age is derived from a very large number of corresponding methylation level measurements using a very large number of CpG loci and extensive experimental and computational work, the average accuracy achieved in WO 2012 / 162139 is still said to be only in the range of 3 to 5 years. This suggests that the accuracy and / or efficiency of current age determination methods are not optimal.

[0012] Furthermore, measuring and evaluating a large number of methylation levels is costly.

[0013] In this regard, it is noted that approximately 28 million CpG loci can be found in the human genome. Even if it is taken into account that the methylation levels of some of these CpG loci may not be affected by aging, there are still a large number of CpG loci whose methylation levels are affected by age. Although it is believed that the detection methods for determining methylation levels may improve over time, so that the methylation levels of more and more CpG loci can be determined, the methylation levels of at least about 800,000 (800000) CpG loci can be determined using commercially available instruments and methods. Nevertheless, the cost of such measurements is high, and therefore, it would be very expensive to determine age based on measuring a very large number of CpG loci. Therefore, current age determination methods are based on only hundreds of CpG loci. However, the cost, equipment and expertise required to determine age based on hundreds of CpG loci remain obstacles to the promotion and use of current age determination methods.

[0014] Therefore, there is a need for improved age determination methods. In particular, there is a need for improved age determination methods that require less data input while having at least approximately the same accuracy.

[0015] There is also a need for improved means for screening drugs for treating or preventing aging-related diseases or cancer, or phenotypes associated with aging-related diseases or cancer. In particular, similar means for diagnosing the health status or wellness status of an individual are also desired.

[0016] It would also be desirable to determine age in a cost-effective manner.

[0017] It is also desirable to determine age in a manner that, even if not very cost-effective and / or not very accurate, at least allows for independent evaluation of other age determination methods. In other words, there is a need for an alternative age indicator that can be used to validate the age determined by other age indicators. Such cross-validation is very important in diagnostics. Summary of the invention

[0018] The present disclosure provides means for solving the above-mentioned technical problems, and these means are summarized below.

[0019] In its broadest aspect, the invention relates to a method of determining an age indicator, a method of determining the age of an individual, and / or a collection of genomic DNA sequences.

[0020] In particular, the method of determining an age indicator of the present invention provided herein comprises the following steps:

[0021] (a) Provide multiple individual training data sets, each individual training data set includes:

[0022] (i) DNA methylation levels of a set of genomic DNA sequences, and

[0023] (ii) chronological age, and

[0024] (b) applying a regression method including a least absolute shrinkage and selection operator (LASSO) on the training data set to thereby determine an age indicator and a reduced training data set,

[0025] Wherein the independent variable is the methylation level of the genomic DNA sequence, and preferably, where the dependent variable is age, where the age indicator comprises:

[0026] (i) a subset of the set of genomic DNA sequences, and

[0027] (ii) contains at least one coefficient for each genomic DNA sequence in the collection, and

[0028] The reduced training data set includes all data in the training data set except the DNA methylation levels of the genomic DNA sequences eliminated by LASSO.

[0029] In particular, the method for determining the age of an individual comprises the following steps:

[0030] (a) Provide multiple individual training data sets, each individual training data set includes:

[0031] (i) DNA methylation levels of a set of genomic DNA sequences, and

[0032] (ii) chronological age, and

[0033] (b) applying a regression method including a least absolute shrinkage and selection operator (LASSO) on the training data set to thereby determine an age indicator and a reduced training data set,

[0034] Wherein the independent variable is the methylation level of the genomic DNA sequence, and preferably, where the dependent variable is age, where the age indicator comprises:

[0035] (i) a subset of the set of genomic DNA sequences, and

[0036] (ii) contains at least one coefficient for each genomic DNA sequence in the collection, and

[0037] The reduced training data set includes all data in the training data set except the DNA methylation levels of the genomic DNA sequences eliminated by LASSO, and

[0038] (c) providing the DNA methylation level of the individual whose age is to be determined, the DNA methylation level being the DNA methylation level of at least 80%, preferably 100%, of the genomic DNA sequences included in the age indicator, and

[0039] (d) determining the age of the individual based on the DNA methylation level of the individual and an age indicator,

[0040] Preferably, the determined age may be different from the chronological age of the individual.

[0041] In particular, the set of genomic DNA sequences includes at least one of the following, preferably at least 10, preferably at least 50, preferably at least 70, preferably all: cg11330075, cg25845463, cg22519947, cg21807065, cg09001642, cg18815943, cg06335143, cg01636910, cg10501210, cg03324695, cg19432688, cg22540792, cg11176990, cg00097800, cg27320127, cg09805798, cg03526652, cg 09460489, cg18737844, cg07802350, cg10522765, cg12548216, cg0087634 5. cg15761531, cg05990274, cg05972734, cg03680898, cg16593468, cg193 01963, cg12732998, cg02536625, cg24088134, cg24319133, cg03388189, c g05106770, cg08686931, cg25606723, cg07782620, cg16781885, cg1423156 5. cg18339380, cg25642673, cg10240079, cg19851481, cg17665505, cg133 33913, cg07291317, cg12238343, cg08478427, cg07625177, cg03230469, c g13154327, cg16456442, cg26430984, cg16867657, cg24724428, cg081943 77. cg10543136, cg12650870, cg00087368, cg17760405, cg21628619, cg01 820962, cg16999154, cg22444338, cg00831672, cg08044253, cg08960065, cg07529089, cg11607603, cg08097417, cg07955995, cg03473532, cg06186 727, cg04733826, cg20425444, cg07513002, cg14305139, cg13759931, cg1 4756158, cg08662753, cg13206721, cg04287203, cg18768299, cg05812299,cg04028695, cg07120630, cg17343879, cg07766948, cg08856941, cg16950671, cg01520297, cg27540719, cg 24954665, cg05211227, cg06831571, cg19112204, cg12804730, cg08224787, cg13973351, cg21165089, cg05 087008, cg05396610, cg23677767, cg21962791, cg04320377, cg16245716, cg21460868, cg09275691, cg1921 5678, cg08118942, cg16322747, cg12333719, cg23128025, cg27173374, cg02032962, cg18506897, cg0529201 6. cg16673857, cg04875128, cg22101188, cg07381960, cg06279276, cg22077936, cg08457029, cg20576243, cg09965557, cg03741619, cg04525002, cg15008041, cg16465695, cg16677512, cg12658720, cg27394136, cg1 4681176, cg07494888, cg14911690, cg06161948, cg15609017, cg10321869, cg15743533, cg19702785, cg16267121, cg13460409, cg19810954, cg06945504, cg06153788 and cg20088545, or a fragment thereof comprising at least 70%, preferably at least 90% of the contiguous nucleotide sequence.

[0042] Preferably, said collection of genomic DNA sequences is comprised in a reduced training data set and / or in the age indicator obtained by said method of determining an age indicator.

[0043] In another preferred aspect, the present invention relates to a gene set comprising at least 1, preferably at least 10, preferably at least 30, preferably at least 50, preferably at least 70, preferably all of the following: SIM bHLH transcription factor 1 (SIM1), microtubule-associated protein 4 (MAP4), protein kinase C ζ (PRKCZ), glutamate ion receptor AMPA-type subunit 4 (GRIA4), BCL10, immune signal binder (BCL10), 5' nucleotidase domain-containing protein 1 (NT5DC1), suppressor of tumorigenicity 7 (ST7), protein kinase Cη (PRKCH), glial cell line-derived neurotrophic factor (GDNF), muskelin 1 protein (MKLN1), exocyst complex component 6B (EXOC6B), protein S (PROS1), calcium voltage-gated channel subunit α1D (CACNA1D), kelch-like family member 42 (KLHL42), OTU deubiquitinase 7A (OTUD7A), death-associated protein (DAP), coiled-coil domain-containing protein 179 (CCDC179), iodothyronine deiodinase 2 (DIO2), transient receptor potential cation channel subfamily V member 3 (TRPV3), MT-RNR2-like protein 5 (MTRNR2L5), filament protein B (FLNB), furin, paired alkaline amino acid lyase (FURIN), solute carrier family 25 member 17 (SLC25A17), G-patch structure domain-containing protein 1 (GPATCH1), UDP-GlcNAc:βGalβ-1,3-N-acetylglucosamine transferase 9 (B3GNT9), zyg-11 family member A, cell cycle regulator (ZYG11A), epilepsy-associated protein 6 (SEZ6L), myosin X (MYO10), acetyl-CoA carboxylase alpha (ACACA), G protein subunit alphai1 (GNAI1), CUE domain-containing protein 2 (CUEDC2), homeobox D13 (HOXD13), Kruppel-like factor 14 (KLF14), solute carrier family 1 member 2 (SLC1A2), acetoacetyl-CoA synthetase (AACS), ankyrin repeat and sterile alpha motif domain-containing protein 1A (ANKS1A), microRNA 7641-2 (MIR7641-2), collagen V α1 chain (COL5A1), arsenite methyltransferase (AS3MT), solute carrier family 26 member 5 (SLC26A5), nucleoporin 107 (NUP107), intergenic nonprotein long coding RNA 1797 (LINC01797), myosin IC (MYO1C), ankyrin repeat domain 37 (ANKRD37), phosphodiesterase 4C (PDE4C), EF-hand domain-containing protein 1 (EFHC1),uncharacterized LOC375196 (LOC375196), ELOVL fatty acid elongase 2 (ELOVL2), WAS protein family member 3 (WASF3), chromosome 17 open reading frame 82 (C17orf82), G protein-coupled receptor 158 (GPR158), F-box and leucine-rich repeat protein 7 (FBXL7), rippled transcription repressor 3 (RIPPLY3), VPS37C subunit of ESCRT-1 (VPS37C), polypeptide N-acetylgalactosaminyltransferase-like protein 6 (GALNTL6), DENN domain-containing protein 3 (DENND3), nuclear receptor co-inhibitor 2 (NCOR2), endothelial PAS domain-containing protein 1 (EPAS1), PBX homeobox protein 4 (PBX4), intergenic nonprotein long coding RNA 1531 (LINC01531), family with sequence similarity 110 member A (FAM110A), glycosyltransferase 8 domain-containing protein 1 (GLT8D1), G protein subunit gamma 2 (GNG2), MT-RNR2-like protein 3 (MTRNR2L3), zinc finger protein 140 (ZNF140), kinase inhibitor 1 (KSR1), protein disulfide isomerase family A member PDIA5, spermatogenesis-associated protein 7 (SPATA7), pantothenate kinase 1 (PANK1), ubiquitin-specific protease 4 (USP4), G protein subunit αq (GNAQ), potassium voltage-gated channel modifier subfamily S member 1 (KCNS1), DNA polymerase γ2, auxiliary subunit (POLG2), forkhead box 2 (STOX2), neuronal axon protein 3 (NRXN3), BMS1, ribosome biogenesis factor (BMS1), stork head box E3 (FOXE3), NADH:ubiquinone oxidoreductase subunit A10 (NDUFA10), relaxin family peptide receptor 3 (RXFP 3), GATA binding protein 2 (GATA2), isoprenoid synthase domain-containing protein (ISPD), adenosine deaminase, RNA-specific B1 (ADARB1), Wnt family member 7B (WNT7B), pleckstrin and Sec7 domain-containing protein 3 (PSD3), membrane-anchored junction protein (MAJIN), pyridine nucleotide-disulfide oxidoreductase domain 1 (PYROXD1), cingulin-like protein 1 (CGNL1), chromosome 7 open reading frame 50 (C7orf50), MORN repeat-containing protein 1 (MORN1), atlastin GTPase protein 2 (ATL2), WD repeat and FYVE domain-containing protein 2 (WDFY2), transmembrane protein 136 (TMEM136), inositol polyphosphate-5-phosphatase A (INPP5A), TBC1 domain family member 9 (TBC1D9), interferon regulatory factor 2 (IRF2), sirtuin 7 (SIRT7),Type XXIII collagen alpha 1 chain (COL23A1), guanine monophosphate synthase (GMPS), potassium double pore domain potassium channel subfamily K member 12 (KCNK12), SIN3-HDAC complex-associated factor (SINHCAF), hemoglobin subunit ε1 (HBE1), and tudor domain-containing protein 1 (TDRD1).

[0044] Preferably, the gene set is obtained by selecting genomic DNA sequences encoding proteins or microRNAs or long non-coding RNAs from a collection of genomic DNA sequences.

[0045] In other preferred aspects, the present invention relates to the use of a collection of genomic DNA sequences or a gene set according to the present invention in diagnosing the health status and / or health status of an individual.

[0046] In other preferred aspects, the present invention relates to an in silico and / or in vitro screening method for identifying molecules that affect aging, comprising the step of providing a collection of genomic DNA sequences or a gene set according to the present invention, wherein when administered to an individual, the molecules improve, prevent and / or reverse at least one aging-associated disease, at least one phenotype associated with at least one aging-associated disease, and / or cancer.

[0047] In other preferred aspects, the present invention relates to a chip or a kit, in particular a chip or a kit that can be used to detect the DNA methylation level of a set of genomic DNA sequences or a gene set according to the present invention.

[0048] In particular, the chip comprises genomic DNA sequences or gene sets according to the invention, wherein each sequence is contained in a separate spot.

[0049] In particular, the kit comprises:

[0050] (a) at least one unique primer pair,

[0051] wherein in each primer pair, one primer is a forward primer that binds to the reverse strand of a genomic DNA sequence included in the set of genomic DNA sequences, or to the reverse strand of a gene included in the gene set according to the present invention, and the other primer is a reverse primer that binds to the forward strand of the genomic DNA sequence, or to the forward strand of the gene,

[0052] and wherein the two nucleotides complementary to the 3' ends of the forward primer and the reverse primer are more than 30 and less than 3000, preferably less than 1000 nucleotides apart, or

[0053] (b) at least one probe which is complementary to a genomic DNA sequence included in the set of genomic DNA sequences or to a gene included in the gene set according to the present invention.

[0054] The present invention is based at least in part on the following surprising discovery, that can be determined by using the regression method comprising the least absolute shrinkage and selection operator (LASSO) comprising an age indicator comprising a further reduced genomic DNA sequence set, but still with acceptable quality, wherein the independent variable is the methylation level of the genomic DNA sequence, and the dependent variable is age. This is particularly surprising, because the ridge regression (Ridge regression) (L2 parameter) required in the previous method is omitted. Further surprisingly, the overlap between the genomic DNA sequence set determined in the present invention and the previously determined genomic DNA sequence set is minimal. Therefore, further surprisingly, a kind of age indicator can be found, which comprises very different genomic DNA sequences but good performance compared with known age indicators.

[0055] Reducing the number of genomic DNA sequences while ensuring the accuracy of age determination has many advantages. One advantage is to reduce the cost, workload and / or necessary expertise of determining the DNA methylation level of the genomic DNA sequence, particularly because it allows the use of simpler laboratory methods. Another advantage is to narrow the scope of candidate drug targets encoded by the set of reduced genomic DNA sequences. Another advantage is to provide an alternative or improved tool for diagnosing individual health status. Therefore, a method for determining an alternative or improved age indicator can also be used to verify the results obtained by other methods (i.e. diagnosis or candidate drugs).

[0056] General terms

[0057] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which the invention belongs. Preferred methods and materials are described, however, any methods and materials similar or equivalent to those described herein may be used in the practice or testing of the present invention. For purposes of the present invention, the following terms are defined as follows.

[0058] As used herein, the articles "a" and "an" refer to one or more (ie, at least one) of the grammatical object of the article. By way of example, "an element" means one element or more than one element.

[0059] As used herein, "and / or" refers to and encompasses any and all possible combinations of one or more of the associated listed items and excludes combinations when interpreting the alternative (or).

[0060] The object of the present invention is to provide novelties for industrial applications.

[0061] This object is achieved by what is claimed in the present disclosure.

[0062] The present disclosure also describes some preferred embodiments. It will be apparent to those skilled in the art that preferred embodiments that are not currently claimed can be found in the specification. In addition, it should be noted that although certain aspects of the present invention are not currently claimed in the independent claims, these aspects can still be found in the specification and may be referenced later. DETAILED DESCRIPTION

[0063] The following describes embodiments and definitions related to the method for determining an age indicator according to the present invention, the age indicator obtained by the method, the set of genomic DNA sequences included in the age indicator, and the method for determining the age of an individual according to the present invention.

[0064] As used herein, an age indicator refers to a statistical model that can be used to determine the age of an individual based on the DNA methylation levels of certain genomic DNA sequences of the individual.

[0065] As used herein, the determined age of an individual is not necessarily the same as the actual age of the individual. Typically, the determined age and actual age of an individual are different, and the situation where the two are the same is just a coincidence. The determined age is also referred to as an "alternative age" in this article. Any age can be measured in "years", and / or preferably in "days". As used herein, the determined age of an individual is more indicative of the biological age of the individual than its actual age. The actual age of an individual refers to the time that has passed since the individual was born. As used herein, biological age is related to the health status of an individual. Preferably, the health status relates to the state of at least one aging-related disease, at least one phenotype associated with at least one aging-related disease, and / or cancer, wherein the state indicates whether the disease or the phenotype associated with the disease exists, or the stage at which it is located. Therefore, the age indicator of the present invention can be used to diagnose the health status of an individual.

[0066] In particular, the age indicator as used herein refers to a linear model including an independent variable. Herein, the age indicator or the independent variable included in the linear model for generating the age indicator refers to the DNA methylation level of a certain genomic DNA sequence.

[0067] Preferably, the age indicator of the present invention and / or the dependent variable of the linear model used to generate the age indicator of the present invention is age.

[0068] In a linear model, the ages of multiple individuals are predicted by a set of independent variables (methylation levels of certain genomic DNA sequences), each of which has at least one coefficient. Preferably, the predicted age and the actual age are well correlated, or in other words, are preferably very similar on average. However, the predicted age (also referred to herein as the "determined age") of an individual may differ from his actual age, for example by several years.

[0069] Specifically, the methylation level as used herein refers to a β value. As used herein, a β value describes a ratio, which is the ratio of methylated cytosine to the sum of methylated and unmethylated cytosine in all relevant cytosines in a specific part of the genomic DNA of all alleles of all cells contained in the sample. The methylation state of a specific cytosine molecule is binary, either an unmethylated state (0; 0%) or a methylated state (1; 100%). Methylated cytosine is also referred to as "5'mC". Thus, the β value of cytosine at a specific position of a single cell genomic DNA with two alleles is therefore typically 0, 0.5 or 1. Therefore, the β value at a specific CpG position in the genomic DNA of a cell population (regardless of the number of alleles) can take a value between 0 and 1. In addition, when considering all CpGs in a certain genomic DNA sequence of a single allele, the β value can take a value between 0 and 1. Preferably herein, only one CpG is considered in a certain genomic DNA sequence. In this article, the sample preferably contains more than one cell, which may contain more than one allele. Therefore, it is obvious that the beta value of the genomic DNA sequence as used herein can actually take any value between 0 and 1. In this article, the methylation level of CpG is defined by the cytosine contained in the CpG, rather than guanine.

[0070] Preferably, CG / CpG corresponds to the CG cluster number (Illumina TM Methylation probe ID number) specified by Illumina TM Probes. Use Illumina TM DNA methylation arrays can measure the methylation levels of a preselected set of CpGs. To quantify the methylation levels of CpGs, software can be used to calculate the beta value of methylation. Illumina TM The ID of the methylation probe is characterized by the term "cg" followed by a number, such as cg11330075 or cg25845463. The terms "CG", "cg", "CpG", "CpG locus", "CpG site" and "cg site" are used interchangeably herein. TMDNA methylation arrays for determining DNA methylation levels are well known and accepted and can be used in the present invention, however, other methods will also be described and may be preferred for the reasons indicated. Therefore, alternatively or additionally, the methylation levels of CpGs can also be quantified using other methods known in the art. Nonetheless, unless otherwise specified, the CG / CpGs identified in the present invention correspond to Illumina TM Methylation probe ID.

[0071] In addition, although it is possible, it is not necessary to determine the methylation of cytosine at a single nucleotide resolution for the determination of the methylation level of a genomic DNA sequence, and the average methylation signal of the relevant cytosine in the sequence is sufficient. Preferably, only the cytosine (CpG dinucleotide) followed by guanine is considered relevant in this article. The common names of bases and nucleotides (such as cytosine and cytosine nucleoside) are used interchangeably herein, respectively, and refer to the specific nucleotides comprising the corresponding bases. In this article, the terms "methylation level" and "DNA methylation level" are used interchangeably. When referring to the methylation level, ranges of 0% to 100% and 0 to 1 are used interchangeably herein.

[0072] As used herein, a genomic DNA sequence refers to a contiguous portion of an individual's genomic DNA. In this article, a certain genomic DNA sequence is not necessarily identical to the reference sequence of the genomic DNA sequence to which it is involved, but can be a variant thereof. Preferably, the genomic DNA sequence is a unique sequence. By querying databases such as "GenBank" or "EMBL-NAR" and using common sense, a technician can easily determine whether a sequence is a variant of a specific reference genomic DNA sequence.

[0073] Herein, the methylation level of a genomic DNA sequence refers to the methylation level of at least one cytosine in at least one CpG dinucleotide contained in the genomic DNA sequence.

[0074] Preferably herein, the methylation level of a genomic DNA sequence refers to the methylation level of only one cytosine in only one CpG dinucleotide contained in the genomic DNA sequence. Preferably, the genomic DNA sequence comprises other nucleotides, the methylation levels of these nucleotides are not considered, but they can be used to identify the CpG dinucleotides. Therefore, herein, the genomic DNA sequence can be defined by the CpG locus.

[0075] In this article, it is very preferred to use Illumina TM Define genomic DNA sequence. The term "Illumina TM Methylation Probe ID, "IlluminaTM CpG Cluster ID", "Illumina TM Cluster CG number", "Illumina TM Probes, Illumina TM Methylation probe ID numbers and their equivalents (with or without the term “Illumina TM "or" TM ”) are used interchangeably herein.

[0076] As used herein, multiple individuals refer to more than one individual. As used herein, an individual refers to an organism with 5'-methylated cytosine (5'-mc) in its genomic DNA. Preferably, the organism is a vertebrate, more preferably a mammal, most preferably a human. Preferably, the methylation level of at least one genomic DNA sequence of an individual can be associated with the aging and / or health status of the individual. As used herein, an individual can have any gender, for example, can be male, female, hermaphrodite, etc. Therefore, in the context of an individual herein, the terms "he", "she", "it" or "his", "her", "its" can be used interchangeably.

[0077] Usually, the identity of the individual is known, but this is not necessary. In particular, even if the identity and / or actual age of the individual is unknown, the age of the individual can be determined by the method of the present invention. Therefore, the method for determining the age of an individual according to the present invention allows the prediction of the actual age of an individual when only individual biological samples are available. Such biological samples include, for example, hair cells, cheek cells, saliva, blood and / or sperm. Therefore, the method for determining the age of the individual can be used to estimate the actual age of an individual who appears at the crime scene and whose biological materials are left there. In addition, when the data on the actual age of an individual have not yet been recorded or are not available, the method for determining the age of the individual can be used to estimate the actual age of the individual.

[0078] As used herein, regression methods refer to statistical procedures for estimating relationships between variables, particularly between a dependent variable and one or more independent variables. Regression analysis is also used to understand which of the independent variables are related to the dependent variable, and to explore the form of these relationships. Preferably, the regression method comprises linear regression. Preferably, the regression method comprises linear regression using shrinkage. Shrinkage refers to the shrinkage of data values ​​to a central point (such as a mean). In this article, the regression method comprises the least absolute shrinkage and selection operator (LASSO).

[0079] LASSO encourages the use of simple, sparse models (i.e., models with fewer parameters). This particular type of regression is well suited for models with high levels of multicollinearity or for models where some parts of model selection (such as variable selection and / or parameter elimination) are desired to be automated. LASSO regression performs L1 regularization, which adds a penalty term equal to the absolute value of the coefficient magnitude. This type of regularization produces a sparse model with fewer coefficients; some coefficients can be set to zero and eliminated from the model. The larger the penalty, the closer the coefficient value is to zero, which is ideal for producing a simpler model. In other words, LASSO can be used to reduce the number of independent variables in a linear model. In this article, the terms "LASSO", "lasso", and "Lasso regression" are used synonymously.

[0080] In a preferred embodiment, LASSO is performed using the biglasso R package, preferably by applying the "cv.biglasso" command. Preferably, "nfold" is 20.

[0081] In a preferred embodiment, the LASSO L1 regularization parameter / α parameter is 1.

[0082] Preferably, the regression method of the present invention does not include ridge regression (L2 regularization) or the L2 regularization parameter / λ parameter is 0.

[0083] In contrast, in the Elastic Net method, the L1 regularization parameter or α parameter is not 1, but is approximately 0.1 to 0.9. Furthermore, the Elastic Net method includes Ridge Regression. Therefore, preferably, the regression method of the present invention does not include the Elastic Net method. Furthermore, preferably, the age indicator of the present invention is not determined by applying the Elastic Net method.

[0084] Preferably, the regression method of the present invention further comprises applying stepwise regression after LASSO.Preferably, stepwise regression is applied to a reduced training data set.

[0085] Therefore, in a particularly preferred embodiment, the method of determining an age indicator comprises the following steps:

[0086] (a) Provide multiple individual training data sets, each individual training data set includes:

[0087] (i) DNA methylation levels of a set of genomic DNA sequences, and

[0088] (ii) chronological age, and

[0089] (b) Apply regression methods including the following on the training dataset:

[0090] (i) Least Absolute Shrinkage and Selection Operator (LASSO), thereby determining the reduced training dataset, and

[0091] (ii) a subsequent stepwise regression to determine an age indicator, preferably wherein said stepwise regression is applied to said reduced training data set,

[0092] Wherein the independent variable is the methylation level of the genomic DNA sequence, and preferably, where the dependent variable is age, where the age indicator comprises:

[0093] (i) a subset of the set of genomic DNA sequences, and

[0094] (ii) contains at least one coefficient for each genomic DNA sequence in the collection, and

[0095] The reduced training data set includes all data in the training data set except the DNA methylation levels of the genomic DNA sequences eliminated by LASSO.

[0096] Therefore, in a particularly preferred embodiment, the method of determining the age of an individual comprises the following steps:

[0097] (a) Provide multiple individual training data sets, each individual training data set includes:

[0098] (i) DNA methylation levels of a set of genomic DNA sequences, and

[0099] (ii) chronological age, and

[0100] (b) Apply regression methods including the following on the training dataset:

[0101] (i) Least Absolute Shrinkage and Selection Operator (LASSO), thereby determining the reduced training dataset, and

[0102] (ii) a subsequent stepwise regression to determine an age indicator, preferably wherein said stepwise regression is applied to said reduced training data set,

[0103] Wherein the independent variable is the methylation level of the genomic DNA sequence, and preferably, where the dependent variable is age, where the age indicator comprises:

[0104] (i) a subset of the set of genomic DNA sequences, and

[0105] (ii) contains at least one coefficient for each genomic DNA sequence in the collection, and

[0106] The reduced training data set includes all data in the training data set except the DNA methylation levels of the genomic DNA sequences eliminated by LASSO, and

[0107] (c) providing the DNA methylation level of the individual whose age is to be determined, the DNA methylation level being the DNA methylation level of at least 80%, preferably 100%, of the genomic DNA sequences included in the age indicator, and

[0108] (d) determining the age of the individual based on the DNA methylation level of the individual and an age indicator,

[0109] Preferably, the age determined may be different from the chronological age of the individual.

[0110] As used herein, stepwise regression is a method for fitting a regression model in which predictors are selected by an automatic procedure. In each step, a variable is considered to be added or subtracted from the explanatory variable set based on certain predetermined criteria. This can take the form of a series of F tests or t tests, but other techniques are also feasible, such as adjusted R2, Akaike information criterion (Akaike information criterion, AIC), Bayesian information criterion, Mallows's Cp (Mallows's Cp), PRESS or false discovery rate. The main methods are forward selection, backward elimination and bidirectional elimination. Forward selection includes: testing the addition of each variable using the selected model fitting criterion; adding variables (if any), the addition of which can bring the most significant improvement to the fit statistically; and repeating the process until no variable can statistically significantly improve the model.

[0111] Backward elimination includes: testing the deletion of each variable using the selected model fit criteria; deleting the variable (if any) whose loss brings the least statistically significant degradation to the model fit; and repeating the process until the deletion of any variable does not cause a statistically significant loss. Two-way elimination is a combination of forward selection and backward elimination, testing the variables to be added or excluded in each step. In this article, preferably, the variables considered by stepwise regression are those selected by LASSO regression.

[0112] In a preferred embodiment, the stepwise regression is a two-way elimination. Preferably, when applying the stepwise regression, statistically insignificant independent variables are removed. Preferably, the significance level for determining whether to add / include or remove / exclude a variable is 0.05.

[0113] In order to determine the age indicator according to the invention, the set of genomic DNA sequences is reduced in at least one step, preferably in two steps, by means of the regression method according to the invention. Preferably, the starting set of genomic DNA sequences is preselected from genomic DNA sequences whose methylation levels can be correlated with chronological age. For example, such a preselected set is Illumina TM DNA methylation arrays. LASSO was then applied to determine age indicators and a reduced training data set, both of which consisted of a collection of genomic DNA sequences.

[0114] In certain embodiments, the set of genomic DNA sequences included in the training data set is preselected from genomic DNA sequences whose methylation levels can be associated with chronological age. Preferably, the preselected set includes at least 400,000, preferably at least 800,000 genomic DNA sequences. Particularly suitable are sequences determined by the Infinium MethylationEPIC BeadChip kit.

[0115] In certain embodiments, the genomic DNA sequences included in the training data set do not overlap with each other and / or each allele only occurs once. This is particularly preferred when only a relatively small set of genomic DNA sequences (i.e., less than 10,000 genomic DNA sequences) is preselected.

[0116] In a preferred embodiment, stepwise regression is applied to a reduced training data set to determine an age indicator for a collection of genomic DNA sequences.

[0117] It has further surprisingly been found that the set of genomic DNA sequences determined by applying LASSO and subsequent stepwise regression is smaller and the corresponding age indicator performs better than the set of genomic DNA sequences or the age indicator determined by applying only LASSO without applying stepwise regression.

[0118] It was further surprisingly found that despite the reduced variables, the age indicator determined by applying the LASSO and subsequent stepwise regression has at least as high or even higher accuracy than prior art methods such as in Horvath, Genome Biology, 2013, 14: R115.

[0119] In this context, the subset included in the age indicator is also referred to as a "set" or a "set of genomic DNA sequences." As used herein, a subset (set) of genomic DNA sequences is at most as large as the set of genomic DNA sequences.

[0120] Preferably, the collection comprised in the age indicator of the present invention is smaller than the collection of genomic DNA sequences used to determine said age indicator.

[0121] Preferably, the set included in the age indicator of the invention is smaller than the set of genomic DNA sequences included in the reduced training data set used to determine the age indicator.

[0122] In certain embodiments, the reduced training data set includes at least 90, preferably at least 100, and preferably at least 140 genomic DNA sequences.

[0123] In certain embodiments, the reduced training data set comprises fewer than 5000, preferably fewer than 2000, preferably fewer than 500, preferably fewer than 350, preferably fewer than 300 genomic DNA sequences.

[0124] Compared with the genomic DNA sequence set of preselection, the genomic DNA sequence set included in the reduced training data set preferably significantly reduces, preferably reduces and reduces by 90%, preferably reduces and is greater than 99%, preferably reduces and is greater than 99.9%.But, the genomic DNA sequence set must be large enough, thereby avoids prematurely limiting the optimization potential of follow-up stepwise regression and / or avoids obtaining the weaker age indicator of performance.In this article, it is contemplated that, compared with the age indicator that comprises at least 30, preferably at least 50, preferably at least 60, preferably at least 80 genomic DNA sequences, the age indicator that comprises being less than 30 genomic DNA sequences has weaker performance.But, preferably comprise the age indicator of the fewest genomic DNA sequences.

[0125] Thus, in certain embodiments, the age indicator comprises at least 30, preferably at least 50, preferably at least 60, preferably at least 80 genomic DNA sequences.

[0126] In a preferred embodiment, the age indicator comprises less than 300, preferably less than 150, preferably less than 110, preferably less than 100, preferably less than 90 genomic DNA sequences.

[0127] In a very preferred embodiment, the age indicator comprises 80 to 100, preferably 80 to 90, preferably 88 genomic DNA sequences.

[0128] In addition, the age indicator of the present invention comprises at least one coefficient for each genomic DNA sequence included in the set. Since one coefficient is sufficient, preferably, for each genomic DNA sequence included in the set, the age indicator comprises only one coefficient.

[0129] As used herein, coefficient refers to the weight of the independent variable (in this article, the methylation level of a certain genomic DNA sequence). To predict or determine the age of an individual, the coefficient is multiplied by the methylation level of the genomic DNA sequence, or in other words, a weight is applied to each genomic DNA sequence and its methylation level; then all weighted methylation levels are summed. Preferably, the methylation level is between 0 and 1 (unmethylated and fully methylated, respectively).

[0130] In this article, the data set used to generate the age indicator is also referred to as a "training data set". As used herein, a reduced training data set refers to a training data set in which data of certain genomic DNA sequences are eliminated or not considered. In this article, the reduced training data set is determined by applying a regression method including LASSO on the training data set.

[0131] In a preferred embodiment, the training data set comprises a matrix, the columns of which are the methylation levels of the genomic DNA sequences contained in the genomic DNA sequence set, and the rows of which are multiple individuals. Preferably, the chronological age of the individual is contained in another column of the matrix.

[0132] In certain embodiments, the age indicator of the present invention is iteratively updated, the iterative updating comprising adding data of at least one other individual to the training data in each iteration, thereby iteratively expanding the training data set.

[0133] It is expected that the iterative updating of the age indicator iteratively improves the performance of the age indicator, in particular its accuracy.

[0134] As used herein, iterative updating refers to multiple consecutive rounds of updating the age indicator. As used herein, a round of updating or an updating round is referred to in certain or preferred embodiments of the present invention. As used herein, different rounds of updating may refer to the same or different embodiments. Preferably, each iterative updating round is referred to by the same embodiment of the present invention. Additional individuals with respect to updating the age indicator refer to individuals that have not yet contributed data to the training data set, but whose data is added to an updating round. As used herein, extension of a training data set refers to adding data of at least one additional individual to the training data set.

[0135] In certain embodiments, the added data for each additional individual in an updating round includes the following DNA methylation levels for the individual:

[0136] (i) DNA methylation levels of at least 5%, preferably 50%, and more preferably 100% of the genomic DNA sequence set included in the initial training dataset or any extended training dataset, and / or

[0137] (ii) DNA methylation levels of genomic DNA sequences contained in the reduced training dataset.

[0138] In particular, the selection scheme (i) refers to a starting set of genomic DNA sequences, in particular a pre-selected set of genomic DNA sequences. Typically, the starting set of genomic DNA sequences is large and includes, for example, at least 800,000 genomic DNA sequences. Therefore, adding at least 5% of the methylation levels of the starting set to the training data set provides a sufficiently large training data set that can be used to determine age indicators. Preferably, the training data set is limited to such genomic DNA sequences: the DNA methylation levels of the genomic DNA sequences of all individuals included in the set are present.

[0139] Thus, in a preferred embodiment, all genomic DNA sequences (independent variables) that are not present in all individuals contributing data to the extended training data set are removed from the extended training data set. Preferably, the updating of the age indicator according to said option (i) comprises adding at least 50%, preferably 100% of the set of genomic DNA sequences contained in the initial or any extended training data set, in particular when several or more rounds of updating are completed.

[0140] In a preferred embodiment, in one update round, for each of the other individuals, the set of genomic DNA sequences to which the added methylation levels belong is the same. This is particularly useful for avoiding excessive removal of genomic DNA sequences in one round of update.

[0141] In this article, the updating of the age indicator can change the set of the genomic DNA sequence (independent variable) included therein and / or the coefficient of each described genomic DNA sequence. It is noteworthy that the selection scheme (i) allows the set of the described genomic DNA sequence to be expanded, restricted and / or changed, while the selection scheme (ii) only allows the set of the described genomic DNA sequence to be restricted. Both the selection schemes (i) and (ii) allow the described coefficient to change. However, the advantage of the selection scheme (ii) is that it is only necessary to provide the methylation level of the genomic DNA sequence set of the reduction of at least one other individual. In addition, the selection scheme (ii) is particularly useful for further reducing the size of the set of the described genomic DNA sequence. In other words, the particularly useful aspect of the selection scheme (i) is that different age indicators (such as the age indicators of some individual groups) for different purposes are generated, or different age indicators are determined as the basis for further improvement; the selection scheme (ii) is particularly useful for fine-tuning and optimizing the universal age indicator (i.e., for further reducing the number of independent variables), such as its non-personalized off-the-shelf use. Both options (i) and (ii) can be combined to combine the flexibility of option (i) with the simplicity of option (ii).

[0142] In certain embodiments, an updating round comprises applying LASSO on an expanded training data set to thereby determine an updated age indicator and / or an updated reduced training data set.

[0143] In some embodiments, the training data set to which the data of at least one other individual is added is a reduced training data set, which may be the initial or any updated reduced training data set. Preferably, the reduced training data set is a previous reduced training data set in an iteration.

[0144] Therefore, by applying LASSO to the expanded training dataset and / or by adding data of at least one other individual to the reduced training dataset, an updated reduced training dataset may be obtained.

[0145] In a preferred embodiment, one updating round comprises applying stepwise regression to the reduced training data set to determine updated age indicators.

[0146] In certain embodiments, during an updating round, data of at least one individual is removed from the training dataset and / or the reduced training dataset.

[0147] In certain embodiments, the training data set, reduced training data set and / or added data further comprises at least one factor related to a lifestyle or risk pattern that may be associated with the individual and / or a characteristic of the individual. Preferably, the factor is selected from the group consisting of medication consumption, environmental pollutants, shift work and stress.

[0148] In certain embodiments, the preselection of genomic DNA sequence and / or the addition and / or removal of individual data depend on at least one feature of the individual. In this article, the feature of the individual is, for example, nationality, sex, chronological age, residence, place of birth, at least one disease and / or at least one lifestyle factor. As used herein, lifestyle factors are selected from drug consumption, exposure to environmental pollutants, shift work or pressure.

[0149] In certain embodiments, the training data set and / or reduced training data set is restricted to genomic DNA sequences whose DNA methylation levels and / or activity / levels of encoded proteins are associated with at least one of the traits and / or lifestyle factors.

[0150] Based on lifestyle factors and / or characteristics of the individual, selecting data from the training data set and / or the reduced training data set in any step (i.e. at the beginning during preselection of genomic DNA sequences and / or during updating of said data set and / or age indicators) may allow determining an age indicator that is particularly suitable for determining the age of an individual or a group of individuals having a certain combination of the above characteristics and / or lifestyle factors. Furthermore, applying different age indicators for age determination may help to determine certain tendencies of an individual or group of individuals, such as the main effects of stress or drug consumption. For example, if the age of an individual determined using an age indicator optimized for aging associated with smoking is much higher than expected relative to using an age indicator optimized for aging associated with shift work, this may indicate that smoking is a more important factor for the individual's aging-related health state than shift work.

[0151] In certain embodiments, the mass of the age indicator is determined, wherein determining the mass comprises the following steps:

[0152] (a) providing a test data set of multiple individuals who have not contributed data to the training data set, wherein the test data set of each individual includes:

[0153] (i) DNA methylation levels of the set of genomic DNA sequences included in the age indicator, and

[0154] (ii) chronological age; and

[0155] (b) determining the quality of the age indicator by statistical evaluation and / or domain boundary evaluation,

[0156] The statistical evaluation includes:

[0157] (i) determine the ages of the individuals included in the test dataset,

[0158] (ii) correlating the determined age with the chronological age of the individual and determining at least one statistical parameter describing the correlation, and

[0159] (iii) determining whether the statistical parameter indicates that the age indicator has acceptable quality, preferably, wherein the statistical parameter is selected from the coefficient of determination (R 2 ) and mean absolute error (MAE), where R 2 A value greater than 0.50, preferably greater than 0.70, preferably greater than 0.90, preferably greater than 0.98 and / or a MAE less than 6 years, preferably less than 4 years, preferably at most 1 year indicates acceptable quality, and

[0160] The domain boundary evaluation includes:

[0161] (iv) determine the domain boundaries of the age indicator,

[0162] wherein the domain boundaries are the minimum and maximum DNA methylation levels for each genomic DNA sequence included in the age indicator, and

[0163] wherein the minimum and maximum DNA methylation levels are present in a training data set that has been used to determine an indicator of age, and

[0164] (v) Determine whether the test data set exceeds the domain boundary, where not exceeding the domain boundary indicates acceptable quality.

[0165] As used herein, a test data set is a data set that can be used to evaluate an age indicator that has been determined based on a training data set. Typically, the training data set and the test data set have the same structure. In particular, the test data set and the training data set include the same set of genomic DNA sequences. However, the essential difference is that the test data set only contains data from individuals that have not yet contributed data to the corresponding training data set.

[0166] As used herein, evaluation of age indicators includes statistical evaluation and / or domain boundary evaluation.

[0167] For statistical evaluation, the age of the individuals in the test data set is determined and compared to the chronological age of the individuals. Any statistical measure or parameter commonly used to describe the correlation of two variables can be applied. Preferably, the statistical parameter is selected from the coefficient of determination (R 2 ) and mean absolute error (MAE). Preferably, R 2Greater than 0.50, preferably greater than 0.70, preferably greater than 0.90, preferably greater than 0.98 and / or MAE less than 6 years, preferably less than 4 years, preferably at most 1 year indicates acceptable quality. If not specified herein, the skilled person can evaluate the measurement results or parameters based on common sense. If there is doubt, the quality should be judged as unacceptable.

[0168] If the test data set is not completely contained within the domain boundaries of the age indicator, the age indicator is judged to be of unacceptable quality. As used herein, the domain boundaries of the age indicator refer to the minimum and maximum DNA methylation levels of each genomic DNA sequence included in the age indicator. More specifically, the minimum and maximum DNA methylation levels are present in the training data set that has been used to determine the age indicator.

[0169] The test data set should be of a reasonable size. In particular for statistical evaluation it should not be too small, but should include at least 10, preferably at least 30, preferably at least 200 individuals. Furthermore, for the determination of the domain boundaries the test data set should not be too large either, thus including at most 1000 individuals, preferably at most 200 individuals. If larger, some violations of the domain boundaries can be allowed, for example for 5%, preferably 1%, of the test data set individuals.

[0170] In certain embodiments, the training data set and / or the test data set comprises at least 10, preferably at least 30 individuals, preferably at least 200 individuals. Preferably, the training data set comprises at least 200 individuals, and the test data set comprises at least 30 individuals.

[0171] It is noteworthy that an age indicator that has been judged to be of unacceptable quality may still be used to determine the age of an individual. As used herein, the term "acceptable quality" refers to the best age indicator determined, in particular the best age indicator determined by updating. Therefore, as used herein, the acceptable or unacceptable quality of an age indicator has nothing to do with the absolute quality of the age indicator, but rather with its relative quality compared to other age indicators, in particular to age indicators determined in different rounds of updating according to the method of the invention.

[0172] In a preferred embodiment, the age indicator is updated when the quality of the age indicator is unacceptable.The quality is judged as acceptable or unacceptable as explained above in the context of evaluating the age indicator.

[0173] In some embodiments, when the number of individuals included in the data has reached a predetermined value and / or a predetermined time has passed since the last update, the age indicator is not further updated. The predetermined time may also refer to the number of quality evaluations for a potential update round.

[0174] For example, if the age indicator already contains data for thousands or even millions of individuals, or the last 10 or even 100 evaluations using a new test dataset have shown that the quality is acceptable, then no further optimization of the age indicator is needed and updating may be stopped.

[0175] In certain embodiments, the DNA methylation level of a genomic DNA sequence of an individual is measured in a sample of biological material of said individual comprising said genomic DNA sequence.Preferably, the sample comprises buccal cells.

[0176] Suitable methods for determining DNA methylation levels are, for example, methylation sequencing, bisulfate sequencing, PCR (polymerase chain reaction, PCR), high resolution melting (HRM), methylation-sensitive single-nucleotide primer extension (MS-SnuPE), methylation-sensitive single-strand conformation analysis, methyl-sensitive cut counting (MSCC), base-specific cleavage / MALDI-TOF, combined bisulfate restriction analysis (COBRA), methylated DNA immunoprecipitation (MeDIP), microarray-based methods, bead array-based methods, pyrosequencing and / or direct sequencing without bisulfate treatment (nanopore technology).

[0177] In a preferred embodiment, the DNA methylation level of an individual is measured using a DNA methylation array, such as the Illumina TM DNA methylation arrays, preferably the Infinium MethylationEPIC BeadChip kit. DNA methylation arrays are particularly suitable when the DNA methylation levels of a very large number of genomic DNA sequences are to be measured, especially for starting and / or preselected genomic DNA sequences.

[0178] In a preferred embodiment, the DNA methylation level of the genomic DNA sequence of the individual is measured by base-specific cleavage / MALDI-TOF and / or PCR methods, preferably, wherein the base-specific cleavage / MALDI-TOF is Agena technology and the PCR method is methylation-specific PCR. The base-specific cleavage / MALDI-TOF and / or PCR methods are particularly suitable when the DNA methylation level of a reduced set of genomic DNA sequences is to be measured, in particular when the data is added to a reduced training data set and / or when providing the methylation level of an individual whose age is to be determined using the age indicator of the present invention.

[0179] Additional details regarding the determination of DNA methylation levels are further explained in the other aspects and embodiments of the invention below.

[0180] In certain embodiments, the method for determining an age indicator and / or the method for determining the age of an individual according to the present invention further comprises the step of obtaining a sample of biological material from the individual. The biological material may be from any part of the individual, but preferably, the sample is obtained non-invasively. Preferably, the individual is not an embryo.

[0181] In a preferred embodiment, the sample is obtained by buccal swab.

[0182] In this context, the age indicator of the present invention can be used as a tool for determining the age of an individual. Therefore, the method for determining the age of an individual according to the present invention comprises all the steps of the method for determining the age indicator of the present invention, or comprises the step of providing the age indicator according to the present invention. In addition, the method for determining the age of an individual comprises the following steps: providing the DNA methylation level of the individual whose age is to be determined, the DNA methylation level being the DNA methylation level of at least 80%, preferably 100% of the genomic DNA sequences included in the age indicator; and determining the age of the individual based on the DNA methylation level of the individual and the age indicator.

[0183] In other words, the methylation levels of at least 80%, preferably 100%, of the genomic DNA sequences contained in the provided age indicator must be provided for the individual whose age is to be determined. The methylation levels of the missing genomic DNA sequences of the individual can be deduced, for example by using the median or average of the provided methylation levels.

[0184] In certain embodiments, the age of an individual is determined based on the individual's DNA methylation level and an updated age indicator. In particular, when the quality of the age indicator initially provided is unacceptable, the age is determined based on the updated age indicator.

[0185] In a preferred embodiment, the age indicator is used to determine the age of an individual only if the individual has not contributed data to a training data set that is or has been used to generate the age indicator.

[0186] In certain embodiments, the method of determining the age of an individual further comprises the step of determining at least one lifestyle factor associated with a difference between the determined age and the chronological age of the individual.

[0187] In a preferred embodiment, the collection of genomic DNA sequences according to the present invention does not include cg27320127.

[0188] In certain embodiments, the collection of genomic DNA sequences according to the present invention comprises at least one, preferably at least four, preferably at least ten, preferably at least thirty, preferably at least seventy, preferably all of the following: cg11330075, cg00831672, cg27320127, cg27173374, cg14681176, cg06161948, cg08224787, cg05396610, cg15609017, cg09805798, cg19215678, cg12333719, cg03741619, cg16677512, cg03230469, cg1985148 1. cg10543136, cg07291317, cg26430984, cg16950671, cg16867657, cg220 77936, cg08044253, cg12548216, cg05211227, cg13759931, cg08686931, cg 07955995, cg07529089, cg01520297, cg00087368, cg05087008, cg2472442 8. cg19112204, cg04525002, cg08856941, cg16465695, cg08097417, cg2162 8619, cg09460489, cg13460409, cg25642673, cg19702785, cg18506897, cg 21165089, cg27540719, cg21807065, cg18815943, cg23677767, cg07802350 , cg11176990, cg10321869, cg17343879, cg08662753, cg14911690, cg1280 4730, cg16322747, cg14231565, cg10501210, cg09275691, cg15008041, cg0 5812299, cg24319133, cg12658720, cg20576243, cg03473532, cg07381960 , cg05106770, cg04320377, cg19432688, cg22519947, cg06831571, cg08194 377, cg01636910, cg14305139, cg04028695, cg15743533, cg03680898, cg2 0088545, cg13333913, cg19301963, cg13973351, cg16781885, cg04287203,cg27394136, cg10240079, cg02536625 and cg23128025, or a fragment thereof comprising at least 70%, preferably at least 90% of the contiguous nucleotide sequence.

[0189] Preferably, said collection of genomic DNA sequences is comprised in an age indicator obtained by an age indicator determination method, wherein said method comprises applying stepwise regression after LASSO.

[0190] In this article, gene refers to a genomic DNA sequence encoding a protein (coding sequence; CDS), or a microRNA or a long non-coding RNA. In this article, the genomic DNA sequence encoding a protein also encodes the mRNA for translating the protein. MicroRNA (miRNA) is a small non-coding RNA molecule (comprising about 22 nucleotides) that plays a role in the post-transcriptional regulation of RNA silencing and gene expression. Long non-coding RNA (long ncRNA, lncRNA) is a transcript, typically having more than 200 nucleotides, which are not translated into proteins (but may be translated into peptides). However, most long non-coding RNAs may be functional, i.e., have a function in transcriptional regulation.

[0191] In a preferred embodiment, the gene set of the present invention does not include KCNK12.

[0192] In certain embodiments, the gene set of the present invention includes at least one, preferably at least 5, preferably at least 10, preferably at least 30, and preferably all of the following: ISPD, KCNK12, GNG2, SIRT7, GPATCH1, GRIA4, LINC01531, LOC101927577, NCOR2, WASF3, TRPV3, ACACA, GDNF, EFHC1, MYO10, COL23A1, TDRD1, ELOVL2, GNAI1, MAP4, CCDC179, KLF14, ST7, INPP5A, SIM1, SLC1A2, AS3MT, KSR1, DSCR6, IRF2, KCNS1, NRXN3, C11orf85, HBE1, FOXE3, TMEM136, HOXD13, LOC375196, PANK1, MIR107, COL5A1, PBX4, ZNF140, GALNTL6, NUP107, LOC100507250, MTRNR2L5 , C17orf82, MKLN1, FURIN, KLHL42, MORN1, ANKS1A, BCL10, DENND3, FAM110A, PROS1, WNT7B, FBXL7, GATA2, VPS37C, NRP1, POLG2, ANKRD37, GMPS, and WDFY2.

[0193] Preferably, the gene set is obtained by selecting those genomic DNA sequences encoding proteins, or microRNAs or long non-coding RNAs from a collection of genomic DNA sequences, wherein the collection of genomic DNA sequences is contained in an age indicator obtained by an age indicator determination method, wherein the method comprises applying stepwise regression after LASSO.

[0194] In a preferred embodiment, the collection of genomic DNA sequences according to the present invention includes at least one, preferably at least four, preferably at least ten, preferably all of the following: cg11330075, cg00831672, cg27320127, cg27173374, cg14681176, cg06161948, cg08224787, cg05396610, cg15609017, cg09805798, cg1921567 8. cg12333719, cg03741619, cg03230469, cg19851481, cg10543136, cg07291317, cg26430984, cg16950671, cg16867657, cg13973351, cg16781885, cg04287203, cg27394136, cg10240079, cg02536625, and cg23128025.

[0195] Preferably, the set of genomic DNA sequences is contained in an age indicator obtained by an age indicator determination method, wherein the method comprises applying stepwise regression after LASSO, and wherein the absolute value of the coefficient of each of the genomic DNA sequences contained in the age indicator is greater than 20.

[0196] In a very preferred embodiment, the collection of genomic DNA sequences according to the invention comprises at least one, preferably at least four, preferably all of the following: cg11330075, cg00831672, cg27320127, cg10240079, cg02536625 and cg23128025.

[0197] Preferably, the set of genomic DNA sequences is contained in an age indicator obtained by an age indicator determination method, wherein the method comprises applying stepwise regression after LASSO, and wherein the absolute value of the coefficient of each of the genomic DNA sequences contained in the age indicator is greater than 40.

[0198] In a preferred embodiment, the genomic DNA sequences contained in the collection of genomic DNA sequences according to the present invention are complete sequences, rather than fragments thereof.

[0199] In a preferred embodiment, the set of genomic DNA sequences according to the present invention comprises complementary sequences in addition to and / or in place of the set of genomic DNA sequences. In this article, a genomic DNA sequence refers to the described sequence and / or its reverse complementary sequence. The technician can easily determine whether the described sequence or its reverse complementary sequence should be used. By default, and for most applications, the described sequence will be used; however, for some applications, for example, when a probe is used to determine the methylation level of the sequence, its complementary sequence will be used for the probe.

[0200] In a preferred embodiment, the gene set of the present invention includes at least 1, preferably at least 5, preferably at least 10, preferably at least 20, preferably all of the following:

[0201] Microtubule-associated protein 4 (MAP4), protein kinase Cζ (PRKCZ), glutamate ion receptor AMPA type subunit 4 (GRIA4), inhibitor of tumorigenicity 7 (ST7), protein kinase Cη (PRKCH), calcium voltage-gated channel subunit α1D (CACNA1D), death-associated protein (DAP), transient receptor potential cation channel subfamily V member 3 (TRPV3), furin, paired alkaline amino acid lyase (FURIN), acetyl-CoA carboxylase α (ACACA), G protein subunit αi1 (GNAI1), solute carrier family 1 member 2 (SLC1A2), phosphodiesterase 4C (PDE4C), ELOVL fatty acid elongase 2 (ELOVL2), nuclear receptor co-inhibitor 2 (NCOR2), endothelial PAS structure domain protein 1 (EPAS1), G protein subunit gamma 2 (GNG2), pantothenate kinase 1 (PANK1), ubiquitin-specific protease 4 (USP4), G protein subunit alpha q (GNAQ), potassium voltage-gated channel modifier subfamily S member 1 (KCNS1), DNA polymerase gamma 2, auxiliary subunit (POLG2), NADH:ubiquinone oxidoreductase subunit A10 (NDUFA10), relaxin family peptide receptor 3 (RXFP3), isoprenoid synthase domain-containing protein (ISPD), inositol polyphosphate-5-phosphatase A (INPP5A), sirtuin 7 (SIRT7), guanine monophosphate synthase (GMPS), SIN3-HDAC complex-associated factor (SINHCAF), tudor domain-containing protein 1 (TDRD1).

[0202] Preferably, the gene set is obtained by further filtering the gene set of the invention on genes encoding proteins whose levels and / or activities can be determined by available assays. In other words, the gene set is further enriched for candidate drug targets.

[0203] Generally speaking, the method for determining age indicators according to the present invention and the set of genomic DNA sequences according to the present invention are closely linked together, and are based on common inventive concepts. Therefore, the description and definition of the set of genomic DNA sequences according to the present invention herein can be used to further limit the age indicators of the present invention and / or the reduced training data sets, and both the age indicators and the reduced training data sets include the set of genomic DNA sequences. In addition, the age indicators and / or the reduced training data sets can be used to further limit the method for determining age indicators and / or the method for determining individual age. Similarly, preferably, by the method for determining age indicators according to the present invention, the set of genomic DNA sequences according to the present invention that can be included in the age indicators of the present invention is obtained. And, this also applies to gene sets of the present invention, which are preferably selected from the set of genomic DNA sequences according to the present invention.

[0204] In other preferred aspects, the present invention relates to an age indicator obtained by the age indicator determination method according to the present invention, and / or a collection of genomic DNA sequences contained in the age indicator obtained by the method.

[0205] In other preferred aspects, the invention relates to an age indicator as described in the examples herein.

[0206] Regarding the use of the age indicators described in the embodiments, the age indicators obtained by the age indicator determination method and / or the collection of genomic DNA sequences contained therein, they are the same as the uses of the collection of genomic DNA sequences and / or gene sets according to the present invention described herein, in particular, the same as the uses in methods for diagnosing the health status and / or physical condition of an individual and / or in computer simulations and / or in vitro screening methods for identifying molecules that affect aging.

[0207] In other preferred aspects, the present invention relates to a method for diagnosing an individual's health status and / or physical condition, which comprises the step of providing a collection of genomic DNA sequences according to the present invention or a gene set according to the present invention.

[0208] Preferably herein, a health state includes the state of at least one aging-related disease, at least one phenotype associated with at least one aging-related disease, and / or cancer, wherein the state indicates whether the disease or the phenotype associated with the disease is present, or the stage it is in. Therefore, as used herein, a health state is preferably associated with aging.

[0209] In this article, the phenotype associated with an aging-related disease preferably refers to at least one symptom of an aging-related disease. In addition, an aging-related disease or cancer or a phenotype associated therewith is usually at a certain stage of development. Therefore, in this article, an aging-related disease or cancer or a phenotype associated therewith may not exist or exist, or be at a certain stage.

[0210] In preferred embodiments, the aging-related disease is Alzheimer's disease, Parkinson's disease, atherosclerosis, cardiovascular disease, cancer, arthritis, cataracts, osteoporosis, type 2 diabetes, hypertension, age-related macular degeneration and / or benign prostatic hyperplasia.

[0211] Preferably herein, the physical condition includes blood pressure, body weight, immune cell level, inflammation level and / or cognitive function of the individual.

[0212] Preferably in this article, the health status and / or physical condition of an individual is related to its biological age. Moreover, compared with the chronological age of the individual, the individual age determined according to the present invention better describes the biological age and / or the health status and / or physical condition.

[0213] In particular, diagnosing the health status and / or physical condition of an individual is complementary to diagnosing a specific disease and / or health / physique parameter. First, diagnosing the health status and / or physical condition state can provide an overall or comprehensive view of the individual. For example, in the case of a negative diagnosis, this may indicate that the individual has changed his lifestyle and / or his environment. Moreover, diagnosing the health status and / or physical condition is particularly useful for evaluating whether a certain medical treatment or a change in lifestyle or environment of an individual has improved the overall health status and / or physical condition of the individual. Obviously, the overall health status and / or physical condition of an individual, particularly the overall health status and / or physical condition when related to aging, is a key factor in the well-being of the individual. In other words, the diagnostic method of the health status and / or physical condition according to the present invention not only determines the disease state of the individual, but also can biologically diagnose whether the individual is young or old.

[0214] In certain embodiments, the method of diagnosing a health status and / or physical condition of an individual further comprises the step of determining the methylation level of a genomic DNA sequence in a biological sample of the individual comprising the genomic DNA sequence.

[0215] Concerning the steps of determining the methylation level of a genomic DNA sequence and a biological sample, it is the same as already described above in the context of the method of determining an age indicator and / or the method of determining the age of an individual of the present invention.

[0216] The method for diagnosing the health status and / or physical condition of an individual according to the present invention includes medical and / or non-medical applications of the method.

[0217] In other preferred aspects, the present invention relates to a computer simulation and / or in vitro screening method for identifying molecules that affect aging, the method comprising the step of providing a collection of genomic DNA sequences according to the present invention or a gene set according to the present invention. Preferably, when administered to an individual, the molecule improves, prevents and / or reverses at least one aging-related disease, at least one phenotype associated with at least one aging-related disease, and / or cancer, preferably, the screening method is an in vitro method.

[0218] Regarding aging-related diseases and phenotypes associated therewith, it is the same as what has been described above in the context of the method for diagnosing the health status and / or health status of an individual. In addition, preventing aging-related diseases and / or phenotypes associated therewith involves maintaining the absence of the disease; improving involves delaying the progression of various stages of the disease, maintaining the stage of the disease and / or causing the disease to regress to an earlier stage; reversing involves regressing the disease to an earlier stage, preferably involving regression to the absence of the disease and / or phenotypes associated therewith.

[0219] In this context, cancer is a preferred aging-related disease.

[0220] In certain embodiments, the screening method of the present invention further comprises the step of determining the DNA methylation level of at least one genomic DNA sequence included in the collection of genomic DNA sequences and / or at least one gene included in the gene set.

[0221] In a preferred embodiment, when administered to an individual, the identified molecule increases and / or decreases the DNA methylation level of at least one of the genomic DNA sequences or genes in the individual. Preferably, the DNA methylation level is altered so that it is associated with a younger chronological age than before the alteration.

[0222] Therefore, the set of genomic DNA sequences or gene sets according to the present invention can be used to screen molecules, i.e. candidate drugs, which change the methylation level of the sequence or gene in a way so that it is associated with a younger chronological age than before the change. For example, when the methylation level of a genomic DNA sequence increases with the growth of chronological age, the drug should reduce the methylation level of the genomic DNA sequence. Similarly, when the methylation level of a genomic DNA sequence decreases with the growth of chronological age, the drug should increase the methylation level of the genomic DNA sequence.

[0223] In certain embodiments, where a gene set of the present invention is provided, the screening method of the present invention further comprises a step of determining the activity of at least one protein encoded by the gene set. Preferably, the gene set comprises only genes encoding proteins.

[0224] In a preferred embodiment, the identified molecule inhibits and / or enhances the activity of at least one protein encoded by the gene set. Preferably, the activity of the protein is altered so that it is associated with a younger chronological age than before the alteration. For example, when the protein activity of a protein encoded by a genomic DNA sequence increases with chronological age, the drug should reduce / inhibit the activity of the protein. Similarly, when the protein activity of a protein encoded by a genomic DNA sequence decreases with chronological age, the drug should increase / enhance the activity of the protein.

[0225] As used herein, the activity of a protein also encompasses the level of said protein, particularly the level of the active form of said protein.

[0226] In other preferred aspects, the present invention relates to a chip comprising a collection of genomic DNA sequences according to the present invention, or a gene set according to the present invention as spots, wherein each sequence is contained in a separate spot. Preferably, the chip is a microarray chip.

[0227] In other preferred aspects, the present invention relates to a kit comprising:

[0228] (a) at least one unique primer pair, wherein in each primer pair, one primer is a forward primer that binds to the reverse strand of one genomic DNA sequence included in the set of genomic DNA sequences according to the present invention, or to the reverse strand of one gene included in the gene set of the present invention, and the other primer is a reverse primer that binds to the forward strand of the one genomic DNA sequence, or to the forward strand of the one gene; and wherein the two nucleotides complementary to the 3' ends of the forward primer and the reverse primer are more than 30 and less than 3000, preferably less than 1000 nucleotides apart;

[0229] (b) at least one probe that is complementary to a genomic DNA sequence included in the set of genomic DNA sequences according to the present invention, or complementary to a gene included in the gene set of the present invention; and / or

[0230] (c) A chip according to the present invention.

[0231] Preferably, the primer pair is used for polymerase chain reaction (PCR). The primer may or may not have DNA methylation specificity. Preferably, the primer is used for a methylation-specific PCR method. The DNA methylation level can be determined by measuring the amplified PCR product, for example by sequencing or by comparing the quantity of products obtained by different PCRs performed using primers that bind to methylated or unmethylated sequences. Preferably, the probe is used for hybridization methods, such as in situ hybridization methods or microarray methods.

[0232] In certain embodiments, the primer or probe binds specifically to methylated or unmethylated DNA, wherein unmethylated cytosine has been converted to uracil.

[0233] In this context, the conversion of unmethylated cytosine to uracil is preferably accomplished by bisulfite treatment.

[0234] In certain embodiments, the kit further comprises a container for the biological material and / or the buccal swab material.

[0235] In certain embodiments, the kit further comprises materials for extracting, purifying and / or amplifying genomic DNA from a biological sample, wherein the materials are spin columns and / or enzymes.

[0236] In certain embodiments, the kit further comprises bisulfite.

[0237] In other preferred aspects, the present invention relates to the use of the chip of the present invention and / or the kit of the present invention for determining the DNA methylation level of at least one genomic DNA sequence included in the set of genomic DNA sequences according to the present invention and / or a gene included in the gene set of the present invention.

[0238] In other preferred aspects, the present invention relates to use of the chip of the present invention and / or the kit of the present invention for diagnosing the health status and / or physical condition of an individual.

[0239] In other preferred aspects, the present invention relates to use of the chip of the present invention and / or the kit of the present invention in computer simulation and / or in vitro screening methods for identifying molecules that affect aging.

[0240] Regarding the diagnosis of an individual's health status and / or physical condition and the computer simulation and / or in vitro screening methods for identifying molecules that affect aging, they are the same as those already described above in the context of the diagnosis methods of an individual's health status and / or physical condition and the computer simulation and / or in vitro screening methods for identifying molecules that affect aging.

[0241] In further preferred aspects, the present invention relates to a data carrier comprising the age indicator of the present invention, the set of genomic DNA sequences of the present invention, and / or the gene set of the present invention.

[0242] In certain embodiments, the kits and / or data carriers of the present invention further comprise a questionnaire for the individual whose age is to be determined, wherein the questionnaire may be blank or contain information about the individual.

[0243] The present invention further relates to the following other aspects and embodiments.

[0244] In another aspect, the present invention relates to a method for determining an individual's age based on the methylation level of a genomic DNA sequence found in an individual, the method comprising the following steps: preselecting a set of genomic DNA sequences from genomic DNA sequences having methylation levels that can be associated with the individual's age; determining the methylation levels of the preselected genomic DNA sequences for a plurality of individuals; selecting a collection of genomic DNA sequences from the preselected set so that the number of genomic DNA sequences in the collection is less than the number of genomic DNA sequences in the preselected set, wherein the ages of the plurality of individuals can be calculated based on the methylation levels of the sequences in the collection, and a statistical evaluation of the calculated ages indicates that the quality of the calculated ages is acceptable; determining the methylation levels of the sequences in the collection in a sample of biological material from an individual; calculating the age of the individual based on the methylation levels of the sequences in the collection; determining a measure of the quality of the calculated age; judging whether the determined quality is acceptable; if the quality is judged to be acceptable, outputting the calculated age of the individual; reselecting genomic DNA sequences for the collection in view of the judgment; and modifying the group of individuals to include the individual based on the judgment; reselecting the collection of genomic DNA sequences from the preselected subset based on the determined methylation levels of the individuals of the modified group.

[0245] In some embodiments, a collection of genomic DNA sequences selected from a large number of genomic DNA sequences having methylation levels that can be associated with the age of an individual is initially used; typically, the number of genomic DNA sequences in the collection is less than the number of sequences in the collection from which it comes; then, the methylation levels of the genomic DNA sequences of the collection are obtained, and the age is determined from them. During a series of age determinations, the composition of the collection and / or the manner in which the age is determined based on the methylation levels obtained for the genomic DNA sequences of the collection are repeatedly changed based on additional information generated or obtained during the series of determinations, in particular based on additionally determined methylation levels. Please note that in some embodiments of the present invention, the determination of age will be based on the evaluation of the methylation levels of specific genomic DNA sequences (or CpG loci) from multiple individuals, wherein the multiple individuals include the individual whose age is to be determined, although this is not necessarily the case.

[0246] Surprisingly, it has been found that in this way, significant improvements compared to the prior art can be achieved.

[0247] In general, this adjustment for the set and / or for the (optimal) way of determining age based on the corresponding methylation levels obtained for the genomic DNA sequences of the set currently under consideration can be modified using any other individuals of known methylation levels and preferably chronological age. Sometimes, this may not be done for every individual, but only for some of them.

[0248] The adjustment can only be performed after the methylation levels of the genomic DNA sequences of more than one additional individual (such as 5, 8, 10, 20, 50 or 100 individuals) have been determined. This will be particularly advantageous in the following situations: the statistical evaluation work required for selecting certain genomic DNA sequences into the collection and / or determining the best age determination method based on the methylation levels of certain genomic DNA sequences is very large.

[0249] Therefore, it is not necessary to re-iterate the composition of the set and / or the best age determination method based on methylation levels only to avoid measuring outliers.

[0250] On the contrary, it is possible to judge the quality to be unacceptable based on a (statistical) measure simply because the (statistical) measure indicates that the size of the multiple reference individuals is smaller than a certain number (for example, smaller than the total number of all individuals whose methylation levels have been determined and / or whose methylation levels have been determined for selected genomic DNA sequences and whose chronological ages are known).

[0251] Before determining the age of a particular individual, the composition of the set and / or the optimal age determination method based on methylation levels obtained for the genomic DNA sequence can be re-iterated; and / or the age of another individual can be first calculated and then the set and / or optimal method can be repeatedly iterated.

[0252] In this document, "individuum" and "individual" are used interchangeably.

[0253] If the composition of the set and / or the best way to obtain the age based on the corresponding methylation level is to be implemented after outputting the age of the individual, the methylation level can be stored together with additional information about the individual, such as his or her full age (if known), so that the stored information can be used later in the statistical (re)evaluation. It is thus possible to collect such methylation level information for a plurality of additional individuals before repeatedly iterating the set and / or the best way.

[0254] From the above it is evident that the present invention basically proposes in one embodiment an improvement of the unknown age determination process, which improvement is achieved based on a statistical evaluation of the measurements that produced the unknown result to be determined. Surprisingly, this is not in itself contradictory, since by including such information in the reference set, the reliability of the method can be improved as a whole. Thus, it has been found that a self-learning method can be easily implemented.

[0255] On average, for a large group of individuals, the age determined by this method should correspond to the mean of their chronological ages. Note that the age determined will be a biological age or at least a closer approximation to a biological age, which may differ from chronological age and is generally only useful if it is comparable to chronological age, as it can then be used to determine whether a particular individual is aging faster than average.

[0256] Therefore, any deviation of the age determined based on the best information available from the chronological age is important. This method can be summarized or restated to relate it to methods for establishing age differences between known biological age and chronological age, or to assess differences between biological ages obtained by different measurements and / or methods.

[0257] It has been found that the use of the best information available for such comparisons will generally include the maximum number of individuals available, rather than a predetermined fixed number. In general, changes in the collection and / or optimal age determination based on methylation levels obtained for genomic sequences will alter the age determined from the same sample obtained from an individual.

[0258] Due to such variations, the overall accuracy and / or variability may be affected, but the present invention therefore provides methods for improving the overall accuracy and / or variability.

[0259] Note that by storing the samples in a way that avoids changes in methylation levels, changes in the determined age can be easily detected if the measurements are sufficiently denoised and the changes introduced by the re-iterations are large enough. Thus, it has been found that frequent implementations of the self-learning method of the present invention can be easily detected.

[0260] In the general method, it is not necessary to actively pre-select a set of genomic DNA sequences less than the known approximately 28 million sites from the total 28 million genomic DNA sequences whose methylation levels are known to be associated with the age or adverse health of an individual. Instead, such active pre-selection should be considered to have been completed as long as the selected method only evaluates a limited number of known sites.

[0261] Preselection can be performed by choosing a specific method for determining methylation levels, such as by Illumina TM The provided method, and / or, selects a DNA chip with a limited number of spot sets, wherein each spot can be used to determine a genomic DNA sequence that is present in only one or some, but not all, individuals and whose methylation level can be correlated with age. Therefore, the decision to use a specific detection method is an implicit preselection.

[0262] Likewise, as long as data obtained in this manner is evaluated (ie, data from fewer than all of the approximately 28 million sites that form the basis of the collection and / or preselected set are evaluated), then preselection may be considered to have occurred.

[0263] Typically, the pre-selected set will be significantly smaller than the 28 million different genomic DNA sequences. In particular, although commercially available methods allow for determination of methylation levels of 800,000 (800000) or more different genomic DNA sequences, it will be appreciated that methods using microarrays allow for determination of methylation levels of a very limited number of different genomic DNA sequences using a collection of specific sites or "spots", and therefore determination of the age of an individual using this method is much cheaper.

[0264] For example, in certain methods, a chip can be used that allows the determination of the methylation level of only one or a few thousand different genomic DNA sequences, preferably even fewer, in particular no more than 1000 CpG loci, preferably 500 different genomic DNA sequences or CpG loci, preferably less than 200 different genomic DNA sequences or CpG loci, preferably no more than 150 different genomic DNA sequences or CpG loci.

[0265] The methylation levels of only those genomic DNA sequences that form part of the set may be determined. In this case, during repeated iterations, the composition of the set may only be changed in the following aspects: certain genomic DNA sequences that have been considered previously are no longer considered after repeated iterations, and / or the optimal age determination method based on the methylation levels obtained for the genomic sequences of the set itself has been changed, for example, in the case where the regression coefficient obtained from the multivariate (linear) correlation between the methylation levels obtained from the genomic DNA sequences of the set and the chronological age of the individual is changed. Moreover, it will be possible to determine the methylation levels of other sequences before repeated iterations.

[0266] More (or more) genomic DNA sequences (or CpG loci) than those currently forming part of the collection may be included in each determination of the methylation level for an individual, for example, about or at least 10, or about or at least 20, or about or at least 50 additional sequences or CpG loci.

[0267] Nevertheless, the number of genomic DNA sequences that have a methylation level that can be associated with the age of the individual and that is determined for each individual or a number of individuals and that does not form part of the current set will generally be very small. For example, the methylation level of genomic DNA sequences that do not currently form part of the set can be determined for no more than 5 times the number of genomic DNA sequences that currently form part of the set. Thus, for example, in the case where the set includes 100 different genomic DNA sequences, the total number of different genomic sequences will generally be less than 500. Typically, the number of additional genomic DNA sequences or CpG loci is even less.

[0268] In some embodiments, it can be determined that the methylation level can be associated with age in additional genomic DNA sequences, however, each genomic DNA sequence that does not form part of the set of genomic DNA sequences from which the age is determined is less than 400, preferably less than 300, particularly less than 100 and particularly less than 60, 50 or 40 CpG loci. In addition and / or alternatively, the ratio of genomic DNA sequences that do not form part of the set to genomic DNA sequences in the set is preferably less than 5, preferably less than 4, preferably less than 3, preferably less than 2. It should be noted that for those additional sequences that do not currently form part of the set, but are used to provide additional methylation levels only in the case where they may help to repeat the iteration, careful selection is usually also performed. This can be done in the preselection.

[0269] For example, CpG loci can be selected into the set whose methylation levels are highly correlated with the methylation levels of such CpG loci, although the CpG is also selected into the set, but it has a very low overall methylation level or a high variance. Moreover, CpG loci known to indicate specific unhealthy lifestyles can also be included, even if these loci are not dominant in the statistical multivariate analysis. In addition, CpG loci that are associated with a subset of the initial reference group can be additionally selected.

[0270] It will be apparent from the above that the exact number of sets and / or collections will depend on the availability of affordable measurement methods, such as sufficiently cheap chips. Moreover, data processing costs may be very high. It may be preferred to use a chip suitable for determining the methylation level of a genomic DNA sequence with no more than 1000, 500, 200 spots, each spot being adapted to determine the methylation level of a different CpG locus.

[0271] Particularly preferably, the chip comprises at least 1 spot, preferably at least 10, in particular at least 20, 30, 40, 50, 60, 70, 80, 90 or 100, and in particular all spots, to allow determination of the methylation level of one or more, in particular at least 20, 30, 40, 50, 60, 70, 80, 90 or 100, and in particular all of the following genomic DNA sequences or CpG loci: cg11330075, cg25845463, cg22519947, cg21807065, cg09001642, cg18815943, cg06335143, cg01636910, cg10501210, cg03324695, cg19432688, cg22540792, cg11176990, cg00097800, cg09805 798, cg03526652, cg09460489, cg18737844, cg07802350, cg10522765, cg1 2548216, cg00876345, cg15761531, cg05990274, cg05972734, cg03680898 , cg16593468, cg19301963, cg12732998, cg02536625, cg24088134, cg24319 133. cg03388189, cg05106770, cg08686931, cg25606723, cg07782620, cg1 6781885, cg14231565, cg18339380, cg25642673, cg10240079, cg19851481 , cg17665505, cg13333913, cg07291317, cg12238343, cg08478427, cg0762 5177, cg03230469, cg13154327, cg16456442, cg26430984, cg16867657, cg2 4724428, cg08194377, cg10543136, cg12650870, cg00087368, cg17760405 , cg21628619, cg01820962, cg16999154, cg22444338, cg00831672, cg08044 253, cg08960065, cg07529089, cg11607603, cg08097417, cg07955995, cg0 3473532, cg06186727, cg04733826, cg20425444, cg07513002, cg14305139,cg13759931, cg14756158, cg08662753, cg13206721, cg04287203, cg18768299, cg05812299, cg04028695, cg071206 30. cg17343879, cg07766948, cg08856941, cg16950671, cg01520297, cg27540719, cg24954665, cg05211227, cg068 31571, cg19112204, cg12804730, cg08224787, cg13973351, cg21165089, cg05087008, cg05396610, cg23677767, cg 21962791, cg04320377, cg16245716, cg21460868, cg09275691, cg19215678, cg08118942, cg16322747, cg12333719, cg23128025, cg27173374, cg02032962, cg18506897, cg05292016, cg16673857, cg04875128, cg22101188, cg073819 60. cg06279276, cg22077936, cg08457029, cg20576243, cg09965557, cg03741619, cg04525002, cg15008041, cg1646 5695, cg16677512, cg12658720, cg27394136, cg14681176, cg07494888, cg14911690, cg06161948, cg15609017, cg10321869, cg15743533, cg19702785, cg16267121, cg13460409, cg19810954, cg06945504, cg06153788, and cg20088545.

[0272] In particular, each of the genomic DNA sequences or CpG loci is contained in a separate spot of the chip. In other words, one spot of the chip is defined by one of the genomic DNA sequences or CpG loci. It will be apparent that when using a chip to measure methylation levels, it is useful to at least refer to a plurality of these CpG loci. In particular, at least 10, preferably at least 20, preferably at least 50 and particularly preferably all of the CG loci will constitute part of a preselected set of genomic DNA sequences whose methylation levels can be associated with individual age, so that a collection of genomic DNA sequences containing all of the above-mentioned CG loci, or in a preferred embodiment at least a large number or part of the above-mentioned CG loci, can be easily obtained. In a preferred embodiment of the present invention, the chip can be used to determine the DNA methylation level of a genomic DNA sequence set, in particular the DNA methylation level of a genomic DNA sequence included in a reduced training data set according to the present invention.

[0273] In some cases, the CpG locus will additionally comprise cg27320127, which CpG locus is known in particular from WO2012 / 162139. TM The above CpGs can be identified by methylation probe ID.

[0274] In certain embodiments, the chip will comprise a total number of fewer spots allowing determination of the methylation level of the following genomic DNA sequences, particularly fewer than 1600 spots, particularly fewer than 800 spots, particularly fewer than 400 spots, preferably fewer than 200 spots.

[0275] It should be noted that when defining a set of genomic DNA sequences whose methylation levels can be associated with an individual's age, the set is different from the entire set of human genomic DNA sequences whose methylation levels can be associated with an individual's age, because the set may include some or all of the CpG loci known in the art whose methylation levels are known to be associated with an individual's age (such as those listed in WO2012 / 162139A1). However, it is believed that at least 10, preferably 20, particularly preferably 50, 100, and in particular all of the CpG loci listed above that are considered to be novel than those known in the art may constitute part of a preselected set of genomic DNA sequences whose methylation levels may be correlated with the age of an individual, and in particular include no more than 5000, in particular no more than 2000, in particular no more than 1000, in particular no more than 250 genomic DNA sequences or CpG loci, and / or constitute a portion of at least 10%, preferably at least 10%, particularly preferably at least 15%, 20%, 25%, 33%, 50%, 66%, 75%, 80%, 100% of the total number of genomic DNA sequences in the preselected set. Therefore, in preferred cases, the relevant CpG loci listed and newly disclosed herein will constitute an important part of the set.

[0276] It should be noted that the total number of CG loci in the set from which the selected set is drawn will depend on the number of different loci that are easily measured, and according to the corresponding state-of-the-art technology and in a cost-effective manner. For example, the price of a DNA chip with oligonucleotides that bind to a DNA fragment containing each CpG locus during the measurement process varies greatly with the number of different sites, with the cost decreasing significantly from chips with 1000 sites or more to different chips with 500, 384, 192 or 96 sites.

[0277] It should be noted that, without being limited in any way, the numbers 96 or 384 refer to the numbers often used in current laboratory procedures. It has been pointed out that once it has been decided not to use all CpG loci known to man, but only those that are readily available, the preselection step can usually be considered to have been completed. Therefore, such a preselection step can be completed by referring to a data set that only includes a correspondingly small number of methylation levels.

[0278] Moreover, it should also be noted that the methylation levels of sequences in a set can be determined in a sample of biological material from an individual by reference to measurements that have been made on the sample. Therefore, the determination of the methylation levels of certain sequences can be achieved by opening the corresponding data file. The same is true for selecting a set of genomic DNA sequences in a specific manner from a pre-selected set. In the case of having referenced a data basis comprising such a set determined by previously analyzing reference data of an individual, this selection should be considered to have been completed.

[0279] About the calculation of age, it should be noted that most often, the number of genomic DNA sequences in the set and / or multiple is quite large, for example, because it comprises more than 5, particularly more than 10, particularly at least 50 different genomic DNA sequences. Moreover, the number of individuals in the group is also quite large, and the group preferably comprises at least 10, preferably at least 50, particularly at least 100, particularly at least 200, and in a preferred embodiment, at least 1000 individuals. Therefore, mathematical analysis, particularly statistical analysis, is usually required to determine the best way to calculate the age of an individual based on the determined methylation level. It should be noted that the "best" way of this calculation may not be the absolute best way, but may refer to some very good ways. In other words, if a determination method is particularly simple to calculate and / or because it has used the local extreme value of a statistical function rather than the absolute extreme value, then the determination method can also be said to be the "best" way.

[0280] As can be clearly seen from the above, the methylation level calculation of the individual age based on the sequence of the set is usually completed in the following manner: based on also using the regression coefficient from multiple regression, particularly from multiple linear regression, using the value (such as percentage) related to the methylation level to calculate the age. The calculation of the measure (such as a statistical measure) can be achieved in different ways. For example, it can be determined whether the methylation level of the sequence itself should be considered to be reliable. In the case where the methylation level is extremely low, it may not be recommended to use the corresponding sequence and / or methylation level, because for example, a determination error (for example, caused by the noise in the methylation level measurement) may have occurred, so compared with other methylation levels, the measured value should be ignored or the measured value should be weighted using a lower weight. And, in the case where the methylation level is particularly high or particularly low, the assumption during the initial calculation, such as the multiple linear regression assumption that the methylation level is linearly correlated with age, will no longer apply.

[0281] It should be noted that, in general, while the assumption that methylation levels are linearly correlated with individual age is useful, this assumption is not required in cases where extremely high or low methylation levels are observed or where the chronological age of the individual is significantly younger or older than the mean age of the individuals in the reference set. Before determining how to best calculate individual age based on the methylation levels of sequences found in the reference set, it may be very useful to group certain individuals to determine a more significant linear correlation. For example, it may be advisable to distinguish between male and female individuals, children, adolescents, young people, middle-aged people, and the elderly. Moreover, distinctions that may be useful include, for example, distinguishing between smokers and non-smokers, people with specific different nutritional habits (e.g., eating very greasy or not greasy, eating fish regularly versus eating red meat regularly, drinking alcohol or specific alcoholic beverages (e.g., alcoholic beverages such as beer or wine) regularly), people who exercise regularly or do not exercise, and people who work in harsh environments exposed to pollutants or hazardous substances (e.g., radioactive substances and / or certain chemicals).

[0282] Therefore, when calculating a statistical measure of the quality of the calculated age, one can consider whether the known chronological age deviates from the calculated biological age by a significantly greater amount than the overall deviation obtained for the reference set, and / or by a number of additional individuals for whom the methylation levels of the sequences of the set have been measured. A difference can be considered to be significantly large if the significance is at least 2σ, 3σ, 4σ, 5σ, or 6σ.

[0283] Moreover, the statistical measure of quality can be estimated by determining whether the individual reference set is large enough. For example, when the Spearman correlation is less than 0.85, preferably less than 0.90, preferably less than 0.91 and preferably 0.92 and the mean square error (MAE) is greater than 6 years, preferably greater than 5 years, particularly greater than 4 years, and regression is performed, this is not the case. It will be understood that the confidence interval can also be estimated for each individual age calculated, and the calculation of the statistical measure can include the confidence interval for determining the calculated age. However, for example, by determining whether the potential reference group is large enough, the statistical measure of the quality of the calculated age can also be calculated more easily. If this group is considered too small, it can be determined that the quality is not high enough. If the overall individual quantity in the reference group is too few, and / or if the individual quantity in the reference group is too few relative to the genomic DNA sequence or the quantity of the CpG loci in the preselected set of the genomic DNA sequence or the selected set, then this is exactly the case.

[0284] When determining the best estimate and calculating age, it may also be determined that the quality of the calculated age will not be high enough if not all individuals whose methylation levels have been determined are referenced, or all individuals with a certain attribute, such as smokers or women, etc. In this case, the (statistical) measure would be the number of members of the reference group, relative to the number of individuals for whom methylation levels are available and other information is available (if applicable).

[0285] By determining that even before calculating the age of the individual, the new data has not yet been input into the reference group, the calculation of this difference can be completed.Then, it should be noted that although at least the age of the individual is output when the individual quality is judged to be acceptable, the age can also be output even when the quality of the calculation is considered to be suspicious or insufficient.For example, in any case, outputting the age of the individual may still be beneficial, because this allows the operator to check whether any specific problem can be detected, which easily explains why the quality of the calculated age is considered to be below the standard.For example, it may be because of individual grouping errors, resulting in the use of the set and regression coefficient of the genomic DNA sequence obtained for young male heavy smokers (however, this is an elderly woman who does not smoke) to determine the age.In some embodiments, multiple sets of CpG loci are defined based on the set, and one of these sets is selected based on specific information, which is derived from the specific information of one or more specific methylation levels of the CpG loci analyzed and / or is independent of the additional information provided.Some CpG loci in the pre-selected set can be selected, so that a specific set can be selected.This process of providing and / or selecting a specific set from multiple sets itself should be considered to be creative. Furthermore, the calculated age can be output and explained to the individual, and a refund can be provided if the standard or guaranteed quality is not met.

[0286] Furthermore, it is possible that once repeated iterations over the set have been completed and / or an optimal way of deriving age based on the methylation levels of the sequence has been obtained, the age may be recalculated and then the individual age calculated in the modified way may be outputted.

[0287] Usually by comprising one or more individuals whose individual methylation levels have been determined in addition, thus completing the modification of the individual group. However, it is also possible to exchange individuals or modify the individual group by splitting groups, etc. For example, such a situation may occur: the initial individual group is very small, so it is not recommended to distinguish between smokers and non-smokers, between men and women, between the old and the young, between drinkers and non-drinkers, or this distinction is not feasible or unreasonable. Then, a large number of measurements will have been completed after a period of time, and in some cases, other attributes (such as, whether the individual is a smoker) will have been determined, therefore, subsequently, one or some individuals can be added based on its characteristics, and the group can be modified according to these characteristics split groups.

[0288] It should be noted that in many organisms, methylation levels will vary in a manner that is related to the age of the organism. However, age determination methods will often be used to determine the age of mammals, particularly primates, and particularly humans. And, for example, for animals that are particularly expensive to buy and sell, at least a rough estimate of age may still be helpful.

[0289] In a preferred embodiment, the individual is a person. Of course, this will also be applicable to the individual of the reference group. It has been shown above that the many steps listed above will require a large amount of calculations. Therefore, it is crucial to implement these steps in an automated manner performed by a computer. It should be noted that for the situation of considering at least 20 different genomic dna sequences in a set or collection, and at least 20, preferably 100 individuals constitute the situation of the reference group, the calculation performed without a computer is particularly prone to error, so the whole result must be considered as completely useless and unreliable. And, in view of the cost of the human being to complete the calculation, such calculation can neither be afforded, nor can the time for the result be accepted by any individual. Therefore, it is considered crucial to perform at least one and preferably all calculation and evaluation steps by a computer.

[0290] Regarding the manner in which the methylation level of a genomic DNA sequence present in an individual is determined, reference is made to the following methods known in the art: methylation sequencing / bisulfite sequencing; PCR methods, in particular at least one of methylation-specific PCR (MSP), real-time methylation-specific PCR, quantitative methylation-specific PCR (QMSP), COLD-PCR, PCR using a methylated DNA-specific binding protein, targeted multiplex PCR, real-time PCR, and microarray-based PCR; high-resolution melting curve analysis (HRM); methylation-sensitive single nucleotide primer extension (MS-SnuPE); methylation-sensitive single-stranded conformation analysis; methylation-sensitive enzyme digestion counting (MSCC); base-specific cleavage / MALDI-TOF (e.g., Agena); combined bisulfate restriction enzyme digestion analysis (COBRA); methylated DNA co-immunoprecipitation (MeDIP); microarray-based methods; magnetic bead array-based methods; pyrophosphate sequencing; direct sequencing without bisulfate treatment (nanopore technology).

[0291] The inventors anticipate that other methods of determining methylation levels will become available through the use of future technologies or technologies that are known to date but rarely used or not marketed. Therefore, the list of methods given is not exclusive. Moreover, different methods can be used to determine the methylation levels of different CpGs. Moreover, different methods can be used to determine the methylation levels for pre-selection and selection.

[0292] Among those methods for detecting methylation levels in a manner that can be used in the present invention, the following methods are currently particularly preferred: methylation sequencing / bisulfite sequencing, methylation-specific PCR (MSP), real-time methylation-specific PCR, quantitative methylation-specific PCR (QMSP), COLD-PCR, base-specific cleavage / MALDI-TOF (e.g., Agena), microarray-based methods, bead array-based methods, and pyrophosphate sequencing.

[0293] In some embodiments, the group of individuals whose methylation levels are first determined is large enough so that the calculated age obtained thereby remains stable even when self-learning still causes a large number of processes. In other words, although the initial training of the process by repeatedly iterating the selection of the set and / or the best way to obtain the result should involve at least 50 individuals in order to have sufficiently stable values ​​for initial reference, it is generally preferred to use a larger number, such as 100 or 200 individuals, in the reference group before starting the actual measurement. As described above, after sufficient additional individuals can be additionally considered or added to the reference group, the composition of the set and all the best ways to calculate the age therefrom can be postponed.

[0294] In some embodiments, the number of genomic DNA sequences in the pre-selected set can be made quite small while still allowing the set to be modified in a beneficial manner.

[0295] In some embodiments, the preselected set will include, on the one hand, at least 90 CpG loci, preferably at least 100 CpG loci, particularly preferably at least 140 CpG loci, in particular at least 150 CpG loci.

[0296] It should be noted that in cases where a broad category of individuals is to be examined, it is advisable to use a larger number of CpG loci in the preselected set, whereas measuring methylation levels in a clearly specified and well-defined group may rely on a smaller number of CpG loci in the preselected set, sometimes even requiring 90 CpG loci or less. On the other hand, the preselected set should not be oversized for various reasons. First, if more CpG loci are examined for methylation levels, the process of determining the methylation levels of the CpG loci will be more costly and more complicated.

[0297] Therefore, the method that relies on a large number of CpG loci is costly, and reducing the number of CpG loci in a set or pre-selected set does significantly reduce costs. Moreover, if fewer CpG loci need to be considered, data processing can be greatly simplified. This is true for the best way to iterate the CpG loci in the set and for processing the methylation obtained for these loci. It should be noted that, in general, the calculation involved in expanding the specific details of the iterative set and the best way should be considered to increase with the number of genomic DNA sequences considered in a highly nonlinear manner. Therefore, from the perspective of data analysis, it is also preferred to reduce the number of CpG loci considered. However, even if only 350, 170, 150 or even 100 CpG loci are considered in the pre-selected set, if computer implementation is not adopted, the overall computational workload of multivariate analysis (such as multilinear regression, principal component analysis, partial least squares analysis, etc.) used to determine the most important CpG loci methylation level will be unmanageable at any rate without overdetermining the system.

[0298] It is considered necessary to provide the determined methylation levels in an electronic, automated manner, for example by establishing an electronic record or file of the methylation levels for use when processing the data, even if such processing is not performed immediately after the methylation levels have been determined; the transfer of data without the use of a computer interface between the final stage for obtaining the methylation levels from the sample and the stage for data analysis would introduce a source of error that would necessarily be considered unacceptable.

[0299] Therefore, it should be noted that the method is typically a computer-implemented method having computer-implemented steps, and at least some of the steps must be performed using a computer.

[0300] In some embodiments, the selected set will have a relatively small number of CpG loci, in particular including less than 150 CpG loci, in particular less than 110 CpG loci, in particular less than 100 CpG loci, in particular less than 90 CpG loci, in particular less than 80 CpG loci, and in particular less than 70 CpG loci. It has been found that this relatively small number of CpG loci considered still allows for the inclusion of many different influencing factors, such as those from lifestyle, such as those caused by food, folic acid and vitamin intake (such as vitamin B12 intake), polyphenol and selenium intake, obesity and / or physical activity, cigarette smoke, alcohol consumption, environmental pollutants such as arsenic and air pollution, aromatic hydrocarbons and other organic pollutants, psychological stress, shift work, etc. In this regard, reference is made to the following paper: "Alegria-Torres et al., Epigenomics, June 2011; 3(3): 267-277". These authors have shown that lifestyle has a major impact on epigenetics among many factors and that DNA methylation is influenced by lifestyle.

[0301] On the other hand, although a relatively small number of CpG loci may be considered in the set, the set should not be too small. Otherwise, there is a risk that the deviation between the age or chronological age and the determined age may be affected by measurement errors, insufficient database in the reference set, etc. Therefore, in some embodiments, it is recommended to include at least 30 CpG loci, preferably at least 50 CpG loci, and in particular at least 60 CpG loci in the set.

[0302] It should be noted that the above numbers applied to a set are valid only after one or more iterations of determining the best way to determine the data based on the methylation levels of the visible CpG loci in the set.

[0303] In some embodiments, when a set is iterated repeatedly, the number of members in the set after the repeated iterations may be different from the number of members in the set before the repeated iterations.

[0304] However, in some embodiments, through such repeated iterations, the number of CpG loci in the set may optionally vary, i.e., merely replacing one or more CpG loci in the set with one or more CpG loci in the set is not prohibited.

[0305] As noted above, the best way to determine age by the methylation levels of the CpG sites of the set may generally depend on the coefficients obtained by multiple regression (preferably: multiple linear regression) performed by the methylation levels of individuals in the group relative to the known chronological age. In some embodiments, the methylation level is used by a numerical value varying between 0% of the minimum methylation of a given CpG locus and 100% of a given CpG locus, and the latter numerical value 100% is used when the methylation level corresponds to the possible maximum methylation of a given CpG locus. In other words, the value of the methylation level is centered and normalized. Of course, a value between 0 and 1 can also be used instead of a percentage between 0% and 100%. Although other ranges of values ​​can be used, it is particularly intuitive to use a value between 0 and 1 or 0% to 100% when evaluating results, etc.

[0306] As described above, in some embodiments, independent of the determination of the quality of the measurements, the calculated age of the individuals is output prior to reselecting the set.

[0307] Furthermore, there is an embodiment in which if the calculated age of an individual is judged to be unacceptable; and only after the set of genomic DNA sequences has been reselected and the age has been recalculated for the reselected set, the age is output.

[0308] Regarding the statistical analysis of methylation levels or values ​​related to methylation levels, in principle different methods can be used. However, it has been found that it is appropriate to use at least one regression method for statistical analysis, for example, principal component analysis for finding the main components that cause the deviation of the calculated age; least squares regression, partial least squares regression, LASSO / elastic network regression and / or XPG Boost methods for identifying related CpGs. It is worth noting that, as explained above, LASSO and elastic network are different regression methods, at least because LASSO does not contain ridge regression, and / or in elastic network, the L1 regularization parameter is not 1.

[0309] It should be noted that protection is sought not only for the method itself, but also for a kit which can be used when the method according to the invention is to be carried out, which kit is a kit for such a method.

[0310] In particular, such a kit will include at least a container for biological material for an individual, the biological material being obtained and / or prepared in a manner that allows the determination of age according to the methods disclosed herein; the kit also includes an information carrier that carries information related to the identification of the patient or individual; the kit further includes instructions for performing the method of the present invention and / or instructions on how to perform this method of the present invention (for example, by sending the probe to a specific laboratory using a voucher), and / or instructions on how to provide data for the production of a data carrier, the data carrier including age-related data determined according to the method of the present invention, and / or instructions on how to provide a data carrier, the data carrier including age-related data determined according to the method of the present invention.

[0311] As noted above, while it is often necessary to determine the absolute age of an individual, for example where DNA-containing biological material has been sampled from a criminal at a crime scene in order to estimate the criminal's chronological age, it is also often preferred to compare the determined age with a known chronological age.

[0312] Furthermore, given the methylation levels that have been determined for specific CpG loci, it may be beneficial to assess the difference between chronological age and biological age. It should be noted that these specific CpG loci do not have to form part of the set. For example, the methylation levels of certain CpG loci may be highly dependent on whether the person smokes, and whether the smoker is a particularly heavy smoker.

[0313] Including such methylation levels in a panel may not be advisable when calculating a standard biological age for an individual, but may be beneficial when indicating to an individual that certain methylation levels may be indicative of environmental or other stresses on the individual.

[0314] For example, the biological age of an individual can be determined by using a set of CpG loci that are particularly beneficial for non-smokers; this may be beneficial in cases where the individual has been shown not to be a smoker. However, it may happen that a non-smoker has been forced to passively smoke for a long time, for example due to growing up with smoking parents. In that case, the methylation levels of specific CpG loci may have changed substantially relative to true non-smokers, so even if the correct biological age is determined in other ways, it may be very beneficial to point out to the individual that certain CpG methylation levels that are indicative of smoking behavior indicate that this person has been severely affected by (passive) smoking.

[0315] This suggests that in some cases the preselected set may include additional CpG loci that, while not representative of biological age in the larger reference group, may still be relevant to a particular individual.

[0316] It should be noted that, given the association between aging behavior and methylation levels, the behavior of modifying methylation levels may be helpful. This can be done using appropriate means; among other things, it is understandable to assume that drugs may be part of such means. Therefore, when the methylation level of a certain CpG locus is changed relative to a control group, and such changes are found to be associated with adverse effects, drugs may help to avoid the biochemical adverse effects that caused the changes in methylation levels or prevent these changes.

[0317] Understanding this, a method for drug screening is also proposed, wherein a plurality of molecules that affect aging are screened, the method comprising determining, for a particular CpG locus, whether one molecule from the plurality of screened molecules has a positive effect on the methylation level of the CpG locus. In particular, this can be accomplished by a determination made at least in part in silico.

[0318] Therefore, in the age determination method according to the present invention, after a first set of genomic DNA sequences has been selected and the ages of a series of individuals have been determined, and the methylation levels of genomic DNA sequences other than those in the set have been determined for at least some individuals in the series, the group of individuals can be modified to include at least some individuals in the series, and in view of the methylation levels obtained for other genomic DNA sequences determined for at least some individuals in the series, it can be determined whether the set of genomic DNA sequences should be modified.

[0319] Thus, even if each single determination in the series yields an acceptable result, i.e., a result with a reasonably small and easily acceptable confidence interval, the determination of biological age is repeatedly revised over the course of the measurement using more and more data obtained in the series. As indicated above, the repeated iterations that are repeatedly performed may simply involve using the methylation levels obtained for an individual to modify the regression parameters obtained from the statistical analysis and used to calculate the age of the individual; or the repeated iterations may determine that the overall set should be revised, i.e., additional DNA sequences should be added and / or DNA sequences currently under consideration should be discarded.

[0320] It should be understood that this will improve the overall quality even when the result itself is acceptable. However, in the case of changing the set itself by adding additional DNA sequences, and in the case of the number of available genomic DNA sequences that can be selected into the set being very few, the pool or concentrated DNA sequences that the selection is made for should be carefully limited, so that the additional sequences added are really helpful. Therefore, at least in some cases, it is considered to be beneficial to start with a very large number of genomic DNA sequences with methylation levels that can be associated with age, and then reduce this large number of genomic DNA sequences to be considered, so as to select, especially repeatedly and often during standard measurement (such as, each time can be the 8th, 10th or 100th individual), or when there are x% (such as x=10%, 20%, 25%, 33%, 50%, 66%, 75%, 100%) other individuals that can be added to the reference group. Therefore, the set should be carefully selected, and it is usually recommended to use the step of multiple selections to determine useful preselection.

[0321] For example, first, the methylation levels of genomic DNA sequences of hundreds of individuals can be measured for approximately 800,000 (800000) different genomic DNA sequences. From the obtained data set, thousands of genomic DNA sequences can be selected, and the main components of the obtained methylation level data set can be determined relative to the actual age of the patient, for example, based on principal component analysis. Then, for the selected thousands of genomic DNA sequences, additional measurements can be performed on hundreds or more (e.g., thousands) of individuals, and from the resulting data set, hundreds of genomic DNA sequences, such as 384 DNA sequences, can be selected, and the methylation levels of these sequences can be detected by a DNA chip with 384 different or oligonucleotide spots.

[0322] Likewise, in view of further principal component analysis, in view of the values ​​of the respective methylation levels, in view of several methylation levels of different genomic DNA sequences that are highly correlated, and so on, the number of genomic DNA sequences can be reduced from thousands to 384 genomic DNA sequences.

[0323] After the final selection has been made and the set of genomic DNA sequences is small enough to economically determine all methylation levels (this is the case for 384 different genomic DNA sequences or 96 genomic DNA sequences), the set can be determined from the remaining genomic DNA sequences, but methylation levels can also be determined for all remaining methylated DNA sequences without incurring excessive cost.

[0324] In some embodiments of the invention, when determining whether the set or optimal age determination method should be modified based on the methylation levels obtained, the determination is made based on as large a set of individuals as possible. Therefore, additional data other than the methylation levels of the set can also be provided for at least some individuals other than the individuals in the reference group currently used. Then, in view of the methylation levels obtained for the additional individuals, it is decided whether the set or optimal age determination method should be modified.

[0325] It should be noted that the information related to other individuals is usually used in the following determination, i.e. the determination made about the best age calculation method, or the determination made about selecting genomic DNA sequences into or out of the set by simply expanding a given individual group. However, in some cases, it may be useful to delete individuals from the reference group or to divide the reference group into several groups simultaneously, wherein each group has individuals of specific attributes. One reason for excluding individuals from the reference group previously used may be because a large number of other individuals are added to the reference group, so that when analyzing the entire group before and the entire group after adding the individual, it may be found that the methylation level of the previously existing individual now constitutes a statistical outlier.

[0326] In addition, it may happen that a first detection method for detecting methylation levels (such as a detection method that measures the methylation levels of approximately 850,000 CpG loci) has been used for preselection, while the actual measurement is performed using a method that can only determine the methylation levels of much fewer CpG loci; and the methylation levels of these CpG loci show different methylation levels with the same CpG in cross-comparison. In this case, although it may be useful initially to rely on the initial measurements obtained by the first means, once a sufficiently large database of the second method that is actually used to provide the methylation levels of the set is available, those data obtained using the more complex first method can be deleted. Another reason why deletion is beneficial is that if a sufficiently large number of individuals with a common attribute have finally been sampled, individuals who do not have this attribute are to be deleted from the reference group. For example, if it is determined after a period of time that the set and the optimal age determination method should be determined to obtain the best results for non-obese perfectly trained athletes, obese individuals can be deleted from the initial reference group.

[0327] Although the set and / or optimal age determination method based on methylation levels may be modified only after a large amount of individual data is obtained, the following determination may be made additionally and / or alternatively: the set and / or optimal method should be re-evaluated for the methylation level of a particular individual if at least one of the following conditions is met: some or all of the methylation levels detected in the genomic DNA sequence are considered to be too low; the expected age of a single individual deviates too far from the individual's known chronological age; the predicted ages of multiple individuals show systematic deviations from the known chronological ages of multiple individuals; the predicted ages of multiple individuals are scattered around the known chronological ages of the individuals, and the variance is considered to be too large; the predicted ages of multiple individuals show systematic deviations from the known chronological ages of the individuals; the number of individuals whose ages have been determined based on a given set has reached a predetermined number; a specified time has passed since the last reselection.

[0328] In the case where the above conditions are met, it can be immediately determined that the set and / or the best method needs to be iterated or re-evaluated, and / or it can be determined that such iterative iteration is delayed until data can be obtained from a sufficient number of such individuals. Another reason for delaying iterative iteration is that such iterative iteration can only be performed within a specific time interval; basically, in all these cases, information related to the individual is stored before iterative iteration, in particular the methylation level detected in the genomic DNA sequence and the known chronological age of the individual; then, iterative iteration is performed using all the stored information.

[0329] In some embodiments, the quality of the determined age is determined by comparing it with the known chronological age. In most cases, a known confidence interval can be used as a measure of quality. A very large confidence interval can indicate that the determined age is unreliable. Moreover, once a larger group of individuals has been examined, the determined age is likely not to differ too far from the ages of other individuals previously determined. In other words, once a larger reference group has been examined and the determined biological age of a new individual shows that its aging behavior is much faster or slower than other individuals whose data analysis has been performed before, it is unlikely to be wrong, especially when it is not known that there are other factors that affect aging. In this case, it can be determined that although the age may be correct, its quality cannot be evaluated in a satisfactory manner. Nevertheless, in this case, the determined age will be pointed out to the individual, even if it cannot be guaranteed that the determined age is reliable, it can be assumed that the determined age is reliable, and the individual can be advised to take action. For example, if a particularly fast aging behavior that has never been observed in a large group of individuals is observed before, it is impossible to evaluate the quality of advanced age relative to the true chronological age, and the individual may need to consult a doctor.

[0330] Therefore, the present invention relates to the following items:

[0331] 1. A method for determining an age indicator, comprising the following steps:

[0332] (a) Provide multiple individual training data sets, each individual training data set includes:

[0333] (i) DNA methylation levels of a set of genomic DNA sequences, and

[0334] (ii) chronological age, and

[0335] (b) applying a regression method including a least absolute shrinkage and selection operator (LASSO) on the training data set to thereby determine the age indicator and a reduced training data set,

[0336] wherein the independent variable is the methylation level of the genomic DNA sequence, and preferably, wherein the dependent variable is age,

[0337] The age indicators include:

[0338] (i) a subset of said set of genomic DNA sequences, and

[0339] (ii) at least one coefficient for each genomic DNA sequence contained in the set, and

[0340] The reduced training data set includes all data in the training data set except the DNA methylation level of the genomic DNA sequence eliminated by the LASSO.

[0341] 2. A method for determining the age of an individual, comprising the steps of:

[0342] (a) Provide multiple individual training data sets, each individual training data set includes:

[0343] (i) DNA methylation levels of a set of genomic DNA sequences, and

[0344] (ii) chronological age, and

[0345] (b) applying a regression method including a least absolute shrinkage and selection operator (LASSO) on the training data set to thereby determine the age indicator and a reduced training data set,

[0346] wherein the independent variable is the methylation level of the genomic DNA sequence, and preferably, wherein the dependent variable is age,

[0347] The age indicators include:

[0348] (i) a subset of said set of genomic DNA sequences as a set, and

[0349] (ii) at least one coefficient for each genomic DNA sequence contained in the set, and

[0350] wherein the reduced training data set includes all data in the training data set except the DNA methylation level of the genomic DNA sequence eliminated by the LASSO, and

[0351] (c) providing the DNA methylation level of the individual whose age is to be determined, said DNA methylation level being the DNA methylation level of at least 80%, preferably 100%, of said genomic DNA sequences comprised in said age indicator, and

[0352] (d) determining the age of the individual based on the DNA methylation level of the individual and the age indicator, preferably, wherein the determined age may be different from the chronological age of the individual.

[0353] 3. The method according to item 1 or 2, wherein the regression method further comprises applying stepwise regression after the LASSO.

[0354] 4. A method according to item 3, wherein the stepwise regression is applied to the reduced training data set.

[0355] 5. A method according to any one of items 1 to 4, wherein the set included in the age indicator is smaller than the set of genomic DNA sequences.

[0356] 6. A method according to any one of items 1 to 5, wherein the set included in the age indicator is smaller than the set of genomic DNA sequences included in the reduced training data set.

[0357] 7. A method according to any one of items 3 to 6, wherein the stepwise regression is a two-way elimination, wherein statistically insignificant independent variables are removed, preferably wherein the significance level is 0.05.

[0358] 8. The method according to any one of items 1 to 7, wherein the LASSO is performed using the biglasso R software package, preferably by applying the "cv.biglasso" command, preferably wherein "nfold" is 20.

[0359] 9. A method according to any one of items 1 to 8, wherein the regression method does not include ridge regression (L2 regularization) or the L2 regularization parameter / λ parameter is 0.

[0360] 10. The method according to any one of items 1 to 9, wherein the LASSO L1 regularization parameter / α parameter is 1.

[0361] 11. A method according to any one of items 1 to 10, wherein the age indicator is iteratively updated, the iterative update comprising adding data of at least one other individual to the training data in each iteration, thereby iteratively expanding the training data set.

[0362] 12. The method according to item 11, wherein in one update round, the added data for each other individual includes the following DNA methylation level of the individual:

[0363] (i) at least 5%, preferably 50%, and more preferably 100% of the genomic DNA sequence set included in the initial training data set or any extended training data set;

[0364] Methylation levels, and / or

[0365] (ii) DNA methylation levels of the genomic DNA sequences contained in the reduced training dataset.

[0366] 13. A method according to item 11 or 12, wherein all genomic DNA sequences (independent variables) that are not present in all individuals who contributed data to the extended training data set are removed from the extended training data set.

[0367] 14. A method according to any one of items 11 to 13, wherein in one updating round, for each of the other individuals, the set of genomic DNA sequences to which the added methylation levels belong is the same.

[0368] 15. A method according to any one of items 11 to 14, wherein an updating round comprises applying the LASSO on the expanded training data set to thereby determine an updated age indicator and / or an updated reduced training data set.

[0369] 16. A method according to any one of items 11 to 15, wherein the training data set to which data of at least one other individual is added is the reduced training data set, and the reduced training data set can be either the initial training data set or the updated reduced training data set.

[0370] 17. A method according to item 16, wherein the reduced training data set is a previous reduced training data set in an iteration.

[0371] 18. A method according to any one of items 11 to 17, wherein one updating round comprises applying stepwise regression to the reduced training data set to determine an updated age indicator.

[0372] 19. A method according to any one of items 1 to 18, wherein in one update round, data of at least one individual is removed from the training data set and / or the reduced training data set.

[0373] 20. A method according to any one of items 11 to 19, wherein the addition and / or removal of individual data depends on at least one characteristic of the individual, wherein the characteristic is ethnicity, gender, actual age, residence, place of birth, at least one disease and / or at least one lifestyle factor, wherein the lifestyle factor is selected from medication consumption, exposure to environmental pollutants, shift work or stress.

[0374] 21. The method according to any one of items 1 to 20, wherein the mass of the age indicator is determined, wherein the determining of the mass comprises the following steps:

[0375] (a) providing a plurality of test data sets of individuals who have not contributed data to the training data set, wherein the test data set of each individual comprises:

[0376] (i) said DNA methylation level of said set of genomic DNA sequences included in said age indicator, and

[0377] (ii) chronological age; and

[0378] (b) determining said quality of said age indicator by statistical evaluation and / or domain boundary evaluation,

[0379] The statistical evaluation includes:

[0380] (i) determining the age of the individuals included in the test data set,

[0381] (ii) correlating the determined age with the chronological age of the individual and determining at least one statistical parameter describing the correlation, and

[0382] (iii) determining whether the statistical parameter indicates that the age indicator has acceptable quality, preferably, wherein the statistical parameter is selected from the coefficient of determination (R 2 ) and mean absolute error (MAE), where R 2 greater than 0.50, preferably greater than 0.70, preferably greater than 0.90, preferably greater than 0.98 and / or MAE less than 6 years, preferably less than 4 years,

[0383] Preferably at most 1 year indicates acceptable quality, and

[0384] The domain boundary evaluation includes

[0385] (iv) determining said domain boundaries of said age indicator,

[0386] wherein the domain boundaries are minimum and maximum DNA methylation levels for each genomic DNA sequence included in the age indicator, and wherein the minimum and maximum DNA methylation levels are present in the training data set that has been used to determine the age indicator, and

[0387] (v) determining whether the test data set exceeds the domain boundary, wherein not exceeding the domain boundary indicates acceptable quality.

[0388] 22. A method according to any one of items 1 to 21, wherein the training data set and / or the test data set comprises at least 10, preferably at least 30 individuals, preferably at least 200 individuals, preferably, wherein the training data set comprises at least 200 individuals and the test data set comprises at least 30 individuals.

[0389] 23. A method according to item 21 or 22, wherein the age indicator is updated when the quality of the age indicator is unacceptable.

[0390] 24. A method according to any one of items 11 to 23, wherein the age of the individual is determined based on the DNA methylation level of the individual and the updated age indicator.

[0391] 25. A method according to any one of items 2 to 24, wherein the age of the individual is determined solely using the age indicator when the individual has not yet contributed data to the training data set used to generate the age indicator.

[0392] 26. A method according to any one of items 1 to 25, wherein the age indicator is no longer updated further when the number of individuals included in the data has reached a predetermined value and / or a predetermined time has passed since the last update.

[0393] 27. A method according to any one of items 1 to 26, wherein the set of genomic DNA sequences included in the training data set is preselected from genomic DNA sequences whose methylation levels can be associated with the chronological age.

[0394] 28. The method according to item 27, wherein the preselected set comprises at least 400,000, preferably at least 800,000 genomic DNA sequences.

[0395] 29. A method according to any one of items 1 to 28, wherein the genomic DNA sequences included in the training data set do not overlap with each other and / or each allele appears only once.

[0396] 30. A method according to any one of items 1 to 29, wherein the reduced training data set comprises at least 90, preferably at least 100, preferably at least 140 genomic DNA sequences.

[0397] 31. A method according to any one of items 1 to 30, wherein the reduced training data set comprises less than 5000, preferably less than 2000, preferably less than 500, preferably less than 350, preferably less than 300 genomic DNA sequences.

[0398] 32. A method according to any one of items 1 to 31, wherein the age indicator comprises at least 30, preferably at least 50, preferably at least 60, preferably at least 80 genomic DNA sequences.

[0399] 33. A method according to any one of items 1 to 32, wherein the age indicator comprises less than 300, preferably less than 150, preferably less than 110, preferably less than 100, preferably less than 90 genomic DNA sequences.

[0400] 34. A method according to any one of items 1 to 33, wherein the DNA methylation level of the genomic DNA sequence of an individual is measured in a sample of biological material of the individual containing the genomic DNA sequence.

[0401] 35. The method according to item 34, wherein the sample comprises buccal mucosal cells.

[0402] 36. The method according to any one of items 34 or 35, further comprising the step of obtaining the sample, wherein the sample is obtained non-invasively.

[0403] 37. A method according to any one of items 34 to 36, wherein the DNA methylation level is measured by: methylation sequencing, bisulfate sequencing, PCR method, high-resolution melting curve analysis (HRM), methylation-sensitive single nucleotide primer extension (MS-SnuPE), methylation-sensitive single-stranded conformation analysis, methylation-sensitive enzyme digestion counting (MSCC), base-specific cleavage / MALDI-TOF, combined bisulfate restriction enzyme digestion analysis (COBRA), methylated DNA co-immunoprecipitation (MeDIP), microarray-based method, bead array-based method, pyrosequencing and / or direct sequencing without bisulfate treatment (nanopore technology).

[0404] 38. A method according to any one of items 34 to 37, wherein the DNA methylation level of the individual's genomic DNA sequence is measured by base-specific cleavage / MALDI-TOF and / or PCR method, preferably, wherein the base-specific cleavage / MALDI-TOF is Agena technology and the PCR method is methylation-specific PCR.

[0405] 39. A method according to any one of items 34 to 38, wherein the DNA methylation level of the genomic DNA sequence included in the age indicator is determined in a sample of biological material containing the genomic DNA sequence of the individual whose age is to be determined.

[0406] 40. A collection of genomic DNA sequences, comprising at least 10, preferably at least 50, preferably at least 70, preferably all of the following: cg11330075, cg25845463, cg22519947, cg21807065, cg09001642, cg18815943, cg06335143, cg01636910, cg10501210, cg03324695, cg19432688, cg22540792, cg11176990, cg00097800, cg27320127, cg09805798, cg03526652, cg0946048 9. cg18737844, cg07802350, cg10522765, cg12548216, cg00876345, cg157 61531, cg05990274, cg05972734, cg03680898, cg16593468, cg19301963, cg 12732998, cg02536625, cg24088134, cg24319133, cg03388189, cg0510677 0. cg08686931, cg25606723, cg07782620, cg16781885, cg14231565, cg1833 9380, cg25642673, cg10240079, cg19851481, cg17665505, cg13333913, cg 07291317, cg12238343, cg08478427, cg07625177, cg03230469, cg13154327 , cg16456442, cg26430984, cg16867657, cg24724428, cg08194377, cg1054 3136, cg12650870, cg00087368, cg17760405, cg21628619, cg01820962, cg1 6999154, cg22444338, cg00831672, cg08044253, cg08960065, cg07529089 , cg11607603, cg08097417, cg07955995, cg03473532, cg06186727, cg04733 826, cg20425444, cg07513002, cg14305139, cg13759931, cg14756158, cg0 8662753, cg13206721, cg04287203, cg18768299, cg05812299, cg04028695,cg07120630, cg17343879, cg07766948, cg08856941, cg16950671, cg01520297, cg27540719, cg24954665, c g05211227, cg06831571, cg19112204, cg12804730, cg08224787, cg13973351, cg21165089, cg05087008, cg 05396610, cg23677767, cg21962791, cg04320377, cg16245716, cg21460868, cg09275691, cg19215678, cg0 8118942, cg16322747, cg12333719, cg23128025, cg27173374, cg02032962, cg18506897, cg05292016, cg16 673857, cg04875128, cg22101188, cg07381960, cg06279276, cg22077936, cg08457029, cg20576243, cg099 65557, cg03741619, cg04525002, cg15008041, cg16465695, cg16677512, cg12658720, cg27394136, cg1468 1176, cg07494888, cg14911690, cg06161948, cg15609017, cg10321869, cg15743533, cg19702785, cg16267121, cg13460409, cg19810954, cg06945504, cg06153788 and cg20088545, or a fragment thereof comprising at least 70%, preferably at least 90% of the contiguous nucleotide sequence.

[0407] 41. The collection of genomic DNA sequences according to item 39, comprising at least 4, preferably at least 10, preferably at least 30, preferably at least 70, preferably all of the following: cg11330075, cg00831672, cg27320127, cg27173374, cg14681176, cg06161948, cg08224787, cg05396610, cg15609017, cg09805798, cg19215678, cg12333719, cg03741619, cg16677512, cg03230469, cg19851481 543136, cg07291317, cg26430984, cg16950671, cg16867657, cg22077936, cg08044253, cg12548216, cg05211227, cg13759931, cg08686931, cg07955 995, cg07529089, cg01520297, cg00087368, cg05087008, cg24724428, cg1 9112204, cg04525002, cg08856941, cg16465695, cg08097417, cg21628619 , cg09460489, cg13460409, cg25642673, cg19702785, cg18506897, cg2116 5089, cg27540719, cg21807065, cg18815943, cg23677767, cg07802350, cg 11176990, cg10321869, cg17343879, cg08662753, cg14911690, cg1280473 0. cg16322747, cg14231565, cg10501210, cg09275691, cg15008041, cg058 12299, cg24319133, cg12658720, cg20576243, cg03473532, cg07381960, c g05106770, cg04320377, cg19432688, cg22519947, cg06831571, cg081943 77. cg01636910, cg14305139, cg04028695, cg15743533, cg03680898, cg20 088545, cg13333913, cg19301963, cg13973351, cg16781885, cg04287203,cg27394136, cg10240079, cg02536625 and cg23128025, or a fragment thereof comprising at least 70%, preferably at least 90% of the contiguous nucleotide sequence.

[0408] 42. The collection of genomic DNA sequences according to item 41, comprising at least 4, preferably at least 10, preferably all of the following: cg11330075, cg00831672, cg27320127, cg27173374, cg14681176, cg06161948, cg08224787, cg05396610, cg15609017, cg09805798, cg19215678, cg 12333719, cg03741619, cg03230469, cg19851481, cg10543136, cg07291317, cg26430984, cg16950671, cg16867657, cg13973351, cg16781885, cg04287203, cg27394136, cg10240079, cg02536625, and cg23128025.

[0409] 43. The collection of genomic DNA sequences according to item 41 or 42, comprising at least 4, preferably all, of the following: cg11330075, cg00831672, cg27320127, cg10240079, cg02536625 and cg23128025.

[0410] 44. A collection of genomic DNA sequences according to any one of items 40 to 43, comprising complementary sequences in addition to and / or instead of said collection of genomic DNA sequences.

[0411] 45. A gene set comprising at least 10, preferably at least 30, preferably at least 50, preferably at least 70, preferably all of the following:

[0412] SIM bHLH transcription factor 1 (SIM1); microtubule-associated protein 4 (MAP4); protein kinase Cζ (PRKCZ); glutamate ionotropic receptor AMPA-type subunit 4 (GRIA4); BCL10, immune signaling engenderer (BCL10); 5' nucleotidase domain-containing protein 1 (NT5DC1); suppressor of tumorigenicity 7 (ST7); protein kinase Cη (PRKCH); glial cell line-derived neurotrophic factor (GDNF); muskelin 1 protein (MKLN1); exocyst complex component 6B

[0413] (EXOC6B); protein S (PROS1); calcium voltage-gated channel subunit alpha 1D (CACNA1D); kelch-like family member 42 (KLHL42); OTU deubiquitinase 7A (OTUD7A); death-associated protein (DAP); coiled-coil domain-containing protein 179 (CCDC179); iodothyronine deiodinase 2 (DIO2); transient receptor potential cation channel subfamily V member 3 (TRPV3); MT-RNR2-like protein 5 (MTRNR2L5); filamin B (FLNB); furin, paired alkaline amino acid lyase (FURIN); solute carrier family 25 member 17 (SLC25A17); G-patch binding domain-containing protein 1 (GPATCH1); UDP-GlcNAc:βGalβ-1,3-N-acetylglucosamine transferase 9 (B3GNT9); zyg-11 family member A, cell cycle regulator (ZYG11A); epilepsy-associated protein 6 (SEZ6L); myosin X (MYO10); acetyl-CoA carboxylase alpha (ACACA); G protein subunit alphai1 (GNAI1); CUE domain-containing protein 2 (CUEDC2); homeobox D13 (HOXD13); Kruppel-like factor 14 (KLF14); solute carrier family 1 member 2 (SLC1A2); acetoacetyl-CoA synthetase (AACS); ankyrin repeat-containing and sterile alpha motif-containing protein 1A (ANKS1A); microRNA 7641-2 (MIR7641-2); collagen V alpha 1 chain (COL5A1); arsenite methyltransferase (AS3MT); solute carrier family 26 member 5 (SLC26A5); nucleoporin 107 (NUP107); intergenic nonprotein long coding RNA 1797 (LINC01797); myosin IC (MYO1C); ankyrin repeat domain 37 (ANKRD37); phosphodiesterase 4C (PDE4C); EF-hand domain-containing protein 1 (EFHC1); uncharacterized LOC375196 (LOC375196); ELOVL lipid Fatty acid elongase 2 (ELOVL2); WAS protein family member 3 (WASF3); chromosome 17 open reading frame 82 (C17orf82); G protein-coupled receptor 158 (GPR158); F-box and leucine-rich repeat protein 7 (FBXL7); rippled transcription repressor 3 (RIPPLY3); VPS37C subunit of ESCRT-1 (VPS37C); polypeptide N-acetylgalactosaminyltransferase-like protein 6 (GALNTL6); DENN domain-containing protein 3 (DENND3); nuclear receptor co-repressor 2 (NCOR2); endothelial PAS domain-containing protein 1 (EPAS1); PBX homeobox protein 4 (PBX4);Long intergenic nonprotein-coding RNA 1531 (LINC01531); family with sequence similarity 110 member A (FAM110A); glycosyltransferase 8 domain-containing protein 1 (GLT8D1); G protein subunit gamma 2 (GNG2); MT-RNR2-like protein 3 (MTRNR2L3); zinc finger protein 140 (ZNF140); kinase inhibitor 1 (KSR1); protein disulfide isomerase family A member 5 (PDIA5); spermatogenesis-associated protein 7 (SPATA7); pantothenate kinase 1 (PANK1); ubiquitin-specific protease 4 (USP4); G protein subunit alpha q (GNG2). AQ); potassium voltage-gated channel modifier subfamily S member 1 (KCNS1); DNA polymerase gamma 2, auxiliary subunit (POLG2); stork head box 2 (STOX2); neuronalin 3 (NRXN3); BMS1, ribosome biogenesis factor (BMS1); forkhead box E3 (FOXE3); NADH:ubiquinone oxidoreductase subunit A10 (NDUFA10); relaxin family peptide receptor 3 (RXFP3); GATA binding protein 2 (GATA2); isoprenoid synthase domain-containing protein (ISPD); adenosine deaminase, RNA-specific B1 (AD ARB1); Wnt family member 7B (WNT7B); pleckstrin and Sec7 domain-containing protein 3 (PSD3); membrane-anchored junction protein (MAJIN); pyridine nucleotide-disulfide oxidoreductase domain 1 (PYROXD1); cingulin-like protein 1 (CGNL1); chromosome 7 open reading frame 50 (C7orf50); MORN repeat-containing protein 1 (MORN1); atlastin GTPase protein 2 (ATL2); WD repeat and FYVE domain-containing protein 2 (WDFY2); transmembrane protein 13 6 (TMEM136); inositol polyphosphate-5-phosphatase A (INPP5A); TBC1 domain family member 9 (TBC1D9); interferon regulatory factor 2 (IRF2); sirtuin 7 (SIRT7) type XXIII collagen alpha 1 chain (COL23A1); guanine monophosphate synthase (GMPS); potassium double pore domain channel subfamily K member 12 (KCNK12); SIN3-HDAC complex-associated factor (SINHCAF); hemoglobin subunit ε1 (HBE1); and tudor domain-containing protein 1 (TDRD1).

[0414] 46. ​​The gene set according to item 45, comprising at least 5, preferably at least 10, preferably at least 30, preferably all of the following: ISPD, KCNK12, GNG2, SIRT7, GPATCH1, GRIA4, LINC01531, LOC101927577, NCOR2, WASF3, TRPV3, ACACA, GDNF, EFHC1, MYO10, COL23A1, TDRD1, ELOVL2, GNAI1, MAP4, CCDC179, KLF14, ST7, INPP5A, SIM1, SLC1A2, AS3MT, KSR1, DSCR6, IRF2, KCN S1, NRXN3, C11orf85, HBE1, FOXE3, TMEM136, HOXD13, LOC375196, PANK1, MIR107, COL5A1, PBX4, ZNF140, GALNTL6, NUP107, LOC100507250, MTRNR2L5, C17orf82, MKLN1, FURIN, KLHL42, MORN1, ANKS1A, BCL10, DENND3, FAM110A, PROS1, WNT7B, FBXL7, GATA2, VPS37C, NRP1, POLG2, ANKRD37, GMPS, and WDFY2.

[0415] 47. The gene set according to item 45, comprising at least 5, preferably at least 10, preferably at least 20, preferably all of the following:

[0416] Microtubule-associated protein 4 (MAP4); protein kinase Cζ (PRKCZ); glutamate ionotropic receptor AMPA-type subunit 4 (GRIA4); suppressor of tumorigenicity 7 (ST7); protein kinase Cη (PRKCH); calcium voltage-gated channel subunit alpha 1D (CACNA1D); death-associated protein (DAP); transient receptor potential cation channel subfamily V member 3 (TRPV3); furin, paired alkaline amino acid lyase (FURIN); acetyl-CoA carboxylase alpha (ACACA); G protein subunit alphai1 (GNAI1); solute carrier family 1 member 2 (SLC1A2); phosphodiesterase 4C (PDE4C); ELOVL fatty acid elongase 2 (ELOVL2); nuclear receptor corepressor 2 (NCOR2); endothelial PAS-containing structure domain protein 1 (EPAS1); G protein subunit gamma 2 (GNG2); pantothenate kinase 1 (PANK1); ubiquitin-specific protease 4 (USP4); G protein subunit alpha q (GNAQ); potassium voltage-gated channel modifier subfamily S member 1 (KCNS1); DNA polymerase gamma 2, auxiliary subunit (POLG2); NADH:ubiquinone oxidoreductase subunit A10 (NDUFA10); relaxin family peptide receptor 3 (RXFP3); isoprenoid synthase domain-containing protein (ISPD); inositol polyphosphate-5-phosphatase A (INPP5A); sirtuin 7 (SIRT7); guanine monophosphate synthase (GMPS); SIN3-HDAC complex-associated factor (SINHCAF); tudor domain-containing protein 1 (TDRD1).

[0417] 48. A collection of genomic DNA sequences according to any one of items 40 to 44 or a gene set according to any one of items 45 to 47, obtained by the method according to items 2 to 39,

[0418] wherein according to the method, the set of genomic DNA sequences is included in the reduced training data set and / or in the age indicator, and

[0419] The gene set is obtained by selecting genes encoding proteins, microRNAs or long non-coding RNAs from the collection of genomic DNA sequences.

[0420] 49. A collection of genomic DNA sequences according to any one of items 40 to 44 or 48, or a gene set according to any one of items 45 to 48, for use in diagnosing the health status of an individual.

[0421] 50. The set of genomic DNA sequences or gene set used according to item 49, wherein the health state includes at least one aging-related disease, at least one phenotype associated with at least one aging-related disease, and / or the state of cancer,

[0422] The state indicates whether the disease or the phenotype associated with the disease exists, or the stage at which it is located.

[0423] 51. The collection of genomic DNA sequences or gene sets used according to item 50, wherein the aging-related diseases are Alzheimer's disease, Parkinson's disease, atherosclerosis, cardiovascular disease, cancer, arthritis, cataracts, osteoporosis, type 2 diabetes, hypertension, age-related macular degeneration and / or benign prostatic hyperplasia.

[0424] 52. Use of a collection of genomic DNA sequences according to any one of items 40 to 44 or 48, or a gene set according to any one of items 45 to 48, in determining the health status of an individual.

[0425] 53. The use according to item 52, wherein the health status includes blood pressure, body weight, immune cell level, inflammation level and / or cognitive function of the individual.

[0426] 54. A method for diagnosing the health status and / or health care status of an individual, comprising the step of providing a collection of genomic DNA sequences according to any one of items 40 to 44 or 48 or a gene set according to any one of items 45 to 48.

[0427] 55. The method according to item 54 further comprises determining the methylation level of the genomic DNA sequence in a biological sample of the individual containing the genomic DNA sequence.

[0428] 56. The method according to any one of items 54 or 55, wherein the health state comprises the state of at least one aging-associated disease, at least one phenotype associated with at least one aging-associated disease, and / or cancer,

[0429] Preferably, the aging-related disease is Alzheimer's disease, Parkinson's disease, atherosclerosis, cardiovascular disease, cancer, arthritis, cataracts, osteoporosis, type 2 diabetes, hypertension, age-related macular degeneration and / or benign prostatic hyperplasia, and / or

[0430] The health status includes blood pressure, weight, immune cell levels, inflammation levels and / or cognitive function of the individual.

[0431] 57. The method according to any one of items 55 or 56, wherein the biological sample is obtained non-invasively, preferably by buccal swab.

[0432] 58. An in silico and / or in vitro screening method for identifying molecules that influence aging, comprising the step of providing a collection of genomic DNA sequences according to any one of items 40 to 44 or 48 or a gene set according to any one of items 45 to 48,

[0433] Wherein the molecule, when administered to a subject, ameliorates, prevents and / or reverses at least one senescence-associated disease, at least one phenotype associated with at least one senescence-associated disease, and / or cancer.

[0434] 59. The method according to item 58 further comprises the step of determining the DNA methylation level of at least one of the genomic DNA sequences.

[0435] 60. A method according to item 58 or 59, wherein when administered to an individual, the identified molecule increases and / or decreases the DNA methylation level of at least one of the genomic DNA sequences in the individual.

[0436] 61. A method according to item 60, wherein the DNA methylation level is altered so that it is associated with a younger chronological age than before the alteration.

[0437] 62. A method according to any one of items 58 to 61, wherein a gene set according to items 45 to 48 is provided, and wherein the method further comprises a step of determining the activity of at least one protein encoded by the gene set.

[0438] 63. A method according to item 62, wherein the identified molecule inhibits and / or enhances the activity of at least one protein encoded by the gene set.

[0439] 64. A method according to item 63, wherein the activity of the protein is altered so that it is associated with a younger chronological age than before the alteration.

[0440] 65. A chip comprising as spots a collection of genomic DNA sequences according to any one of items 40 to 44 or 48 or a gene set according to any one of items 45 to 48, wherein each sequence is contained in a separate spot.

[0441] 66. A kit comprising at least one unique primer pair,

[0442] wherein in each primer pair, one primer is a forward primer that binds to the reverse strand of one genomic DNA sequence included in the set of genomic DNA sequences according to any one of items 40 to 44 or 48, or to the reverse strand of one gene included in the gene set according to any one of items 45 to 48, and the other primer is a reverse primer that binds to the forward strand of the one genomic DNA sequence, or to the forward strand of the one gene,

[0443] Furthermore, the two nucleotides complementary to the 3' ends of the forward and reverse primers are more than 30 and less than 3000, preferably less than 1000 nucleotides apart.

[0444] 67. A kit comprising at least one probe complementary to one of the genomic DNA sequences included in the set of genomic DNA sequences according to any one of items 40 to 44 or 48 or one of the genes included in the gene set according to any one of items 45 to 48.

[0445] 68. The kit according to item 65 or 66, wherein the primer or probe binds specifically to methylated or unmethylated DNA, wherein unmethylated cytosine has been converted to uracil.

[0446] 69. A kit comprising the chip according to item 65.

[0447] 70. A kit according to any one of items 51 to 57, further comprising a container for biological material and / or buccal swab material.

[0448] 71. A kit according to any one of items 66 to 70, further comprising a material for extracting, purifying or amplifying genomic DNA from a biological sample, wherein the material is a spin column and / or an enzyme.

[0449] 72. A kit according to any one of items 66 to 71, further comprising bisulfite.

[0450] 73. A data carrier comprising the age indicator obtained by the method according to any one of items 2 to 39. A collection of genomic DNA sequences according to any one of items 40 to 44 or 48, and / or a gene set according to any one of items 45 to 48.

[0451] 74. A kit according to any one of items 66 to 72 or a data carrier according to item 73, further comprising a questionnaire for the individual whose age is to be determined, wherein the questionnaire can be blank or include information about the individual.

[0452] 75. A method according to any one of items 1 to 39, wherein the training data set, reduced training data set and / or added data also includes at least one factor related to a lifestyle or risk pattern associated with the individual.

[0453] 76. The method of claim 75, wherein the factor is selected from the group consisting of medication consumption, environmental pollutants, shift work and stress.

[0454] 77. A method according to any one of items 75 or 76, wherein the training dataset and / or the reduced training dataset is restricted to sequences whose DNA methylation level and / or activity / level of the encoded protein is associated with at least one of the lifestyle factors.

[0455] 78. The method of any one of 75 to 77, further comprising the step of determining at least one lifestyle factor associated with the difference between the determined age and the chronological age of the individual.

[0456] In other aspects, the present invention relates to the following items:

[0457] Item 79 concerns a method for determining the age of an individual:

[0458] It is based on the methylation levels of genomic DNA sequences found in an individual.

[0459] The following steps are involved:

[0460] Preselection:

[0461] preselecting a set of genomic DNA sequences from genomic DNA sequences having methylation levels that can be correlated with the age of the individual;

[0462] Determining the methylation level of the preselected genomic DNA sequence for a plurality of individuals;

[0463] selecting a collection of genomic DNA sequences from said preselected set,

[0464] So that:

[0465] the number of genomic DNA sequences in the collection is less than the number of genomic DNA sequences in the preselected set,

[0466] The ages of the plurality of individuals may be calculated based on the methylation levels of the sequences in the collection,

[0467] as well as

[0468] Statistical evaluation of the calculated ages indicated that the quality of the calculated ages was acceptable;

[0469] determining the methylation levels of sequences in the collection in a sample of biological material from the individual;

[0470] calculating the age of the individual based on the methylation levels of the sequences in the collection;

[0471] Calculate a statistical measure of the quality of the calculated ages;

[0472] determining whether the quality is acceptable based on the statistical measure;

[0473] If the quality is judged to be acceptable, output the age of the individual;

[0474] If the quality is judged to be unacceptable, determining whether it is necessary to reselect the genomic DNA sequence and modify the individual group to include the individual;

[0475] Based on the methylation levels of individuals in the determined modified group, the collection of genomic DNA sequences is reselected from the pre-selected subset.

[0476] Furthermore, the present invention has disclosed Item 80 related to the method for determining age according to Item No. 79 listed above, wherein the individual is a human.

[0477] Furthermore, the present invention has disclosed item 81 relating to a method for determining age according to the previously numbered items listed above, wherein at least one step is a computer-implemented step,

[0478] In particular, at least one of the following steps:

[0479] And preferably all the following steps:

[0480] selecting a collection of genomic DNA sequences from said preselected set,

[0481] So that:

[0482] the number of genomic DNA sequences in the collection is less than the number of genomic DNA sequences in the preselected set,

[0483] The ages of the plurality of individuals may be calculated based on the methylation levels of the sequences in the collection,

[0484] as well as

[0485] Statistical evaluation of the calculated ages indicated that the quality of the calculated ages was acceptable;

[0486] determining the methylation levels of sequences in the collection in a sample of biological material from the individual;

[0487] calculating the age of the individual based on the methylation levels of the sequences in the collection;

[0488] Calculate a statistical measure of the quality of the calculated ages;

[0489] determining whether the quality is acceptable based on the statistical measure;

[0490] If the quality is judged to be acceptable, output the age of the individual;

[0491] If the quality is judged to be unacceptable, it is determined that the genomic DNA sequence needs to be reselected,

[0492] Modify the individual group to include the individual;

[0493] Based on the methylation levels of individuals in the determined modified group, the collection of genomic DNA sequences is reselected from the pre-selected subset.

[0494] In addition, the present invention has disclosed item 82 relating to the method for determining age according to one of the previously numbered items listed above, wherein the methylation level of a genomic DNA sequence found in an individual is measured by at least one of the following:

[0495] Methylation sequencing / bisulfite sequencing; a PCR method, in particular at least one of methylation-specific PCR (MSP), real-time methylation-specific PCR, quantitative methylation-specific PCR (QMSP), COLD-PCR, PCR using a methylated DNA-specific binding protein, targeted multiplex PCR, real-time PCR and microarray-based PCR; high-resolution melting curve analysis (HRM); methylation-sensitive single nucleotide primer extension (MS-SnuPE); methylation-sensitive single-strand conformation analysis; methylation-sensitive restriction counting (MSCC); base-specific cleavage / MALDI-TOF, such as Agena; combined bisulfate restriction analysis (COBRA); methylated DNA co-immunoprecipitation (MeDIP); microarray-based methods; magnetic bead array-based methods; pyrosequencing; direct sequencing without bisulfate treatment (nanopore technology).

[0496] In addition, the present invention has disclosed an 83rd item related to the method for determining age according to one of the previously numbered items listed above, wherein the methylation level of the genomic DNA sequence is measured by at least one of the following:

[0497] Methylation sequencing / bisulfite sequencing, methylation-specific PCR (MSP), real-time methylation-specific PCR, quantitative methylation-specific PCR (QMSP), COLD-PCR, base-specific cleavage / MALDI-TOF (e.g., Agena), microarray-based methods, bead array-based methods, pyrosequencing.

[0498] In addition, the present invention has disclosed item 84, which proposes an age determination method according to one of the previously numbered items listed above, wherein:

[0499] The plurality of individuals for which the methylation levels of the preselected genomic DNA sequences are determined comprises at least 50, preferably at least 100, in particular at least 200 individuals.

[0500] Furthermore, the present invention has disclosed item 85 relating to the method for determining age according to one of the previously numbered items listed above, wherein the group of individuals is modified by adding individuals to the group.

[0501] In addition, the present invention has disclosed item 86 relating to a method for determining age according to one of the previously numbered items listed above, wherein modifying the group of individuals to include the individual comprises: eliminating at least one other individual from the group, in particular eliminating at least one other individual from the group due to factors unrelated to their age and / or the methylation levels of some or all of their genomic DNA sequences.

[0502] Furthermore, the present invention has disclosed item 87 relating to a method for determining age according to one of the previously numbered items listed above,

[0503] wherein after a first set of genomic DNA sequences has been selected, the ages of a series of individuals are determined,

[0504] and wherein for at least some individuals in the series, methylation levels of genomic DNA sequences other than those in the collection are determined,

[0505] The group of individuals is modified to include at least some of the individuals in the series, and it is determined whether the set of genomic DNA sequences should be modified in view of the methylation levels obtained for other genomic DNA sequences determined for at least some of the individuals in the series.

[0506] In addition, the present invention has disclosed an 88th item related to the age determination method according to the previous numbered items listed above, wherein:

[0507] For at least some of the individuals, determining the methylation levels of all genomic DNA sequences in a preselected set,

[0508] and wherein given the methylation levels among all of those methylation levels obtained for at least some of the individuals, it is determined whether the set of genomic DNA sequences should be modified.

[0509] Furthermore, the present invention has disclosed item 89 relating to the method for determining age according to the previously numbered items listed above, wherein it is determined to modify the set based on the methylation levels obtained for another individual if at least one or preferably several of the following conditions have been met:

[0510] Some or all of the methylation levels detected in the genomic DNA sequence are considered too low,

[0511] The predicted age of a single individual deviates too far from the individual's known chronological age.

[0512] The predicted ages of multiple individuals show systematic deviations from the known chronological ages of multiple individuals.

[0513] The predicted ages for multiple individuals are scattered around the known chronological ages of the individuals with variance that is considered too large, and the predicted ages for multiple individuals show systematic deviations from the known chronological ages of the individuals,

[0514] The number of individuals whose age is determined based on a given set has reached a predetermined number,

[0515] The specified amount of time has passed since the last reselection.

[0516] In addition, the present invention has disclosed item 90 related to the age determination method described in any of the previously numbered items listed above, wherein judging whether the quality is acceptable based on statistical measures includes: performing a statistical evaluation of the age taking into account the known actual age of at least some individuals, in particular taking into account whether the predicted age of a single individual deviates too much from the known actual age of the individual, in particular relative to known outlier behavior.

[0517] In addition, the present invention has disclosed item 91 related to the method for determining age according to one of the previously numbered items listed above, wherein the preselected set comprises at least 90 genomic DNA sequences, preferably at least 100 genomic DNA sequences, particularly preferably at least 140 genomic DNA sequences and / or the preselected set comprises less than 2000 genomic DNA sequences, in particular less than 500 genomic DNA sequences, in particular less than 350 genomic DNA sequences, in particular less than 170 genomic DNA sequences, in particular less than 150 genomic DNA sequences.

[0518] In addition, the present invention has disclosed item 92 relating to the method for determining age according to one of the previously numbered items listed above, wherein the selected collection comprises at least 30 genomic DNA sequences, preferably at least 50 genomic DNA sequences, particularly preferably at least 60 genomic DNA sequences and / or the selected collection comprises less than 150 genomic DNA sequences, in particular less than 110 genomic DNA sequences, in particular less than 100 genomic DNA sequences, in particular less than 90 genomic DNA sequences, in particular less than 80 genomic DNA sequences, in particular less than 70 genomic DNA sequences.

[0519] In addition, the present invention has disclosed item 93 related to the age determination method according to one of the previously numbered items listed above, wherein the reselected collection comprises at least 30 genomic DNA sequences, preferably at least 50 genomic DNA sequences, particularly preferably at least 60 genomic DNA sequences and / or the selected collection comprises less than 150 genomic DNA sequences, in particular less than 110 genomic DNA sequences, in particular less than 100 genomic DNA sequences, in particular less than 90 genomic DNA sequences, in particular less than 80 genomic DNA sequences, in particular less than 70 genomic DNA sequences.

[0520] In addition, the present invention has disclosed item 94 related to the age determination method described in one of the previously numbered items listed above, wherein the number of genomic DNA sequences in the reselected collection is different from the number of genomic DNA sequences in the initially selected collection.

[0521] In addition, the present invention has disclosed item 95 relating to the age determination method according to one of the previously numbered items listed above, wherein at least one genomic DNA sequence included in the selected set is not included in the genomic DNA sequences of the reselected set.

[0522] In addition, the present invention has disclosed item 96 related to the age determination method described in one of the previously numbered items listed above, wherein, in view of the known ages of the individuals in the group, the age is determined by statistical analysis of the methylation levels of the genomic DNA sequences in the collection, in particular by using the coefficients obtained for each genomic DNA sequence in the collection in a multivariate linear regression of the methylation level values ​​based on the known ages of the individuals in the group.

[0523] In addition, the present invention has disclosed item 97 related to the age determination method described in one of the previously numbered items listed above, wherein the methylation level value is determined based on the methylation level after centering and / or normalizing the obtained level, and wherein the methylation level value is statistically analyzed.

[0524] Furthermore, the present invention has disclosed an item 98 relating to an age determination method according to one of the previously numbered items listed above, wherein independently of a determination of the quality of the measurements, the calculated individual age is output before reselecting the set.

[0525] In addition, the present invention has disclosed item 99 relating to a method for determining age according to one of the previously numbered items listed above, wherein the calculated age of an individual is judged to be unacceptable and the age is output only after a reselection has been performed on a set of genomic DNA sequences and the age has been recalculated for the reselected set.

[0526] In addition, the present invention has disclosed item 100, which proposes an age determination method according to one of the previously numbered items listed above, wherein the selection of genomic DNA sequences is based on a statistical analysis of values ​​associated with the methylation levels of genomic DNA sequences of multiple individuals, in particular a statistical analysis using at least one regression method to identify relevant CpG loci, in particular at least one of the principal component analysis method, LASSO / elastic network regression method and / or XPG Boost method for identifying relevant CpGs.

[0527] In addition, the present invention has disclosed item 101, which relates to a kit, which comprises at least a container for biological material for an individual, wherein the biological material is obtained and / or prepared in a manner that allows the age determination method described in one of the previous numbered items listed above; the kit also comprises an information carrier, which carries information related to patient identification; the kit also comprises instructions for performing or how to perform the following items according to one of the previous methods of the numbered items listed above, and / or provides data for the generation of a data carrier, wherein the data carrier contains age-related data determined by the method described in the previous method of the numbered items listed above.

[0528] and / or providing a data carrier comprising age-related data determined by a method according to the aforementioned method of the previously numbered item listed above.

[0529] In addition, the present invention has disclosed item 102, which relates to a method for assessing the difference between chronological age and biological age, the method comprising determining age based on methylation levels according to a method described in one of the previously numbered items listed above, and comparing the determined biological age with the known chronological age.

[0530] In addition, the present invention has disclosed item 103, which relates to a method for evaluating the difference between chronological age and biological age according to the previously numbered items listed above, wherein the differences between the multiple individuals are determined, the values ​​of factors that may or may not affect the differences for the multiple individuals are determined, and the factors that have a greater impact on the difference between chronological age and biological age in a large number of individuals are determined.

[0531] In addition, the present invention has disclosed Item 104, which relates to a method for screening multiple molecules that affect aging, the method comprising the following steps: determining multiple genomic DNA sequences that are closely related to biological age, in particular by referring to the genomic DNA sequences selected for the collection in the method of Item 79 listed above, and determining whether one of the multiple molecules has a positive effect on the methylation level of the genomic DNA sequence, in particular by computer simulation.

[0532] In addition, the present invention has disclosed Item 105, which relates to a method for determining the age of an individual based on the evaluation of the methylation level of a genomic DNA sequence selected from a plurality of individuals, wherein the plurality of individuals includes the individual.

[0533] In addition, the present invention has disclosed item 106, which relates to a chip, the chip comprising a plurality of spots suitable for determining methylation levels, particularly less than 500 spots, preferably less than 385 spots, particularly less than 193 spots, particularly less than 160 spots, the spots comprising at least one and preferably several spots specifically suitable for determining the methylation level of at least one of the following: cg11330075, cg25845463, cg22519947, cg21807065, cg09001642, cg18815943, cg06335143, cg01636910, cg10501210, cg03324695, cg19 432688, cg22540792, cg11176990, cg00097800, cg09805798, cg03526652, cg09460489, cg18737844, cg07802350, cg10522765, cg12548216, cg00876 345, cg15761531, cg05990274, cg05972734, cg03680898, cg16593468, cg1 9301963, cg12732998, cg02536625, cg24088134, cg24319133, cg03388189 , cg05106770, cg08686931, cg25606723, cg07782620, cg16781885, cg1423 1565, cg18339380, cg25642673, cg10240079, cg19851481, cg17665505, cg 13333913, cg07291317, cg12238343, cg08478427, cg07625177, cg0323046 9. cg13154327, cg16456442, cg26430984, cg16867657, cg24724428, cg081 94377, cg10543136, cg12650870, cg00087368, cg17760405, cg21628619, c g01820962, cg16999154, cg22444338, cg00831672, cg08044253, cg089600 65. cg07529089, cg11607603, cg08097417, cg07955995, cg03473532, cg06 186727, cg04733826, cg20425444, cg07513002, cg14305139, cg13759931,cg14756158, cg08662753, cg13206721, cg04287203, cg18768299, cg05812299, cg04028695, cg07120630, cg17343 879, cg07766948, cg08856941, cg16950671, cg01520297, cg27540719, cg24954665, cg05211227, cg06831571, cg1 9112204, cg12804730, cg08224787, cg13973351, cg21165089, cg05087008, cg05396610, cg23677767, cg21962791 , cg04320377, cg16245716, cg21460868, cg09275691, cg19215678, cg08118942, cg16322747, cg12333719, cg2312 8025, cg27173374, cg02032962, cg18506897, cg05292016, cg16673857, cg04875128, cg22101188, cg07381960, cg 06279276, cg22077936, cg08457029, cg20576243, cg09965557, cg03741619, cg04525002, cg15008041, cg1646569 5. cg16677512, cg12658720, cg27394136, cg14681176, cg07494888, cg14911690, cg06161948, cg15609017, cg10321869, cg15743533, cg19702785, cg16267121, cg13460409, cg19810954, cg06945504, cg06153788 and cg20088545.

[0534] In addition, the present invention has disclosed item 107, which relates to a method for determining an age indicator of an individual in a series of individuals, wherein the determination is based on the methylation level of a genomic DNA sequence found in the individual, wherein the age indicator of the individual is provided in the following manner: a method that relies on a statistical evaluation of the methylation level of the genomic DNA sequences of the multiple individuals, and is characterized in that the age indicator of the individual is provided in the following manner: a method that relies on a statistical evaluation of the methylation level of the genomic DNA sequences of the multiple individuals. The method of statistically evaluating the age level of a plurality of individuals is different from the plurality of individuals involved in the previous statistical evaluation and also used to determine the age indicator of the previous individuals in the series, the plurality of individuals differ in that the plurality of individuals used for the first statistical evaluation has been modified by at least introducing at least one additional previous individual in the series, and wherein, preferably, the age indicator of the individuals is provided in the following way: at least two different statistical evaluations of the two different plurality of individuals result in a change in at least one coefficient used when calculating the age indicator from the methylation level of the collection, and / or result in methylation levels of different genomic DNA sequences or discovered CgP loci being taken into account.

[0535] In some aspects, a method is disclosed for determining an age indicator for an individual in a series of individuals based on methylation levels of genomic DNA sequences, wherein a collection of genomic DNA sequences is selected and the age indicator for the individual is provided in a manner that continuously improves a statistical evaluation of previous measurements to obtain a better model. BRIEF DESCRIPTION OF THE DRAWINGS

[0536] Figure 1 : Performance of LASSO. A set of 148 cg sites were identified as optimal. The four plots show the Lasso regression and its performance. In all four plots, the vertical dashed line represents the automatic threshold selected for the number of selected variables. All plots report the mean as well as the range intervals resulting from 20 cross-validation runs. Different axes show different model metrics according to the biglasso package. The top two plots report the cross-validation error and the coefficient of determination (R 2 ), the following two figures report two specific parameters in the R implementation of LASSO regression: signal-to-noise ratio and <bs>. Details are in https: / / cran.rstudio.com / web / packages / biglasso / biglasso.pdf.

[0537] Figure 2 : Performance of age indicators obtained by LASSO and subsequent stepwise regression. Shown are the chronological age (actual age) and the determined age (predicted age) for the 259 individuals in the training dataset and the 30 individuals in the test dataset. No correlation or significant difference was observed between the training and test datasets. The coefficient of variation R is shown. 2 Based on combined training and testing data.

[0538] Figure 3 : Correlation between representative CpG sites and chronological age. Individuals of the combined training and test data were grouped based on their full age (>48 years, 25-48 years, and <25 years; "Old", "Middle-aged", and "Young", respectively). The distribution of DNA methylation levels ("values") of eight representative CpG sites for each age group is shown. Genes contained in the CpG sites are annotated.

[0539] Figure 4 : CG sites and the CG site set as described by Horvath in "Genome Biology, 2013, 14: R115" overlap. The Venn diagram reports the amount of overlap between the 148 genomic DNA sequences (CpGs) determined here by applying LASSO (IME-Cerascreen) and the list of 353 CpGs reported by Horvath in "Genome Biology, 2013, 14: R115". See also Figure 5 .

[0540] Figure 5 :This paper applies LASSO (IME-Cerascreen) and subsequent stepwise regression (IME_Cerascreen_8) Overlap of identified CG sites. The CG site set described by Horvath in "Genome Biology, 2013, 14: R115" Overlap . See also Figure 4 .

[0541] Example

[0542] Example 1: Measuring CpG methylation levels in DNA of biological samples

[0543] For a very large number of CpGs (approximately 850,000 (850,000)), their respective methylation levels have been measured by:

[0544] Cheek swabs were used to collect cheek cells from multiple test persons, and genomic DNA was purified from cheek cells using the QIAamp 96 DNA Swab BioRobot kit (Qiagen, Hilden, Germany). The purified genomic DNA was treated with sodium bisulfite using the Zymo EZ DNA Methylation Kit (Zymo, Irvine, CA, USA). This treatment converts unmethylated cytosine to uracil, while methylated cytosine remains unchanged.

[0545] The Infinium MethylationEPIC kit (Illumina TM , San Diego, California, USA) to perform all subsequent steps. Briefly, the bisulfite-treated samples were denatured and neutralized in preparation for amplification. The amplified DNA was then isothermally amplified overnight and enzymatically fragmented. The fragmented DNA was precipitated with isopropanol, collected by centrifugation at 4°C, and resuspended in hybridization buffer. The fragmented, resuspended DNA samples were then distributed onto the Infinium MethylationEPIC microbead chip (Illumina TM ) and placed the microbead chip on an Illumina TM The samples were hybridized to the bead chip by incubating overnight in a hybridization oven, by annealing the fragments to the locus-specific 50mers covalently linked to the beads.

[0546] Unhybridized and nonspecifically hybridized DNA was washed away, and the bead chip was prepared for staining and extension in the capillary flow chamber. Using the captured DNA as a template, single-base extension of oligonucleotides was performed on the bead chip, so that the fluorescent label was incorporated into the bead chip to determine the methylation level of the queried CpG site. The bead chip was scanned with the iScan system, and the fluorophore of the single-base extension product on the bead was excited with a laser, and a high-resolution image of the light emitted from the fluorophore was recorded. The GenomeStudio Methylation Module (Illumina TM ) analysis data, this module allows the calculation of the β value for each analyzed CpG.

[0547] Through this process, more than 850,000 (850,000) different Illumina TM The methylation levels of CpGs defined in the present invention are shown in Figure 2, and numerical values ​​for the methylation levels of each of the more than 850,000 (850,000) different CpGs are provided. This was done for a large number of samples, each from a different individual. The values ​​have been normalized so that 0 corresponds to the minimum possible methylation value for a CpG and 1 corresponds to the maximum methylation value for a CpG. It is worth noting that 1 also corresponds to 100% or complete methylation.

[0548] Example 2: Measurement of CpG methylation levels by base-specific cleavage / MALDI-TOF (Agena)

[0549] To determine the methylation levels of a pre-selected set of several hundred different CpGs, the EpiTYPER DNA Methylation Analysis Kit from Agena Bioscience (San Diego, CA, USA) was used. In this example, 384 methylation levels of 384 different CpGs have been determined.

[0550] Similarly, cheek cells were collected from multiple individuals using cheek swabs, and genomic DNA was purified from cheek cells using the QIAamp 96 DNA Swab BioRobot kit (Qiagen, Hilden, Germany). The purified genomic DNA was treated with sodium bisulfite using the Zymo EZ DNA Methylation Kit (Zymo, Irvine, CA, USA). This treatment converts unmethylated cytosine to uracil, while methylated cytosine remains unchanged.

[0551] Subsequently, the target region containing the CpG of interest was amplified by PCR using one specific primer pair per target region, each containing a reverse primer for a T7 promoter tag.

[0552] The PCR products were then treated with shrimp alkaline phosphatase to remove unreacted nucleotides from the sample and in vitro transcribed using T7 RNA polymerase. The resulting RNA transcripts were specifically cleaved at uracil residues and distributed onto the SpectroCHIP array. The chip was placed into a MALDI-TOF mass spectrometer for data acquisition, and the resulting data were analyzed using EpiTYPER software.

[0553] From the results, numerical values ​​for each methylation level of the 384 different CpGs are provided.The numerical values ​​are again normalized so that 0 corresponds to the minimum possible methylation value for a CpG and 1 (100%) corresponds to the maximum methylation value for a CpG.

[0554] The methylation levels of 384 different genomic DNA sequences were determined by the method of Example 2, compared to approximately 850,000 (850000) different genomic DNA sequences. It was noted that the analysis cost according to Example 2 was significantly lower, less than 1 / 5 of the cost when applied.

[0555] Example 3: Measurement of CpG methylation levels by methylation-specific PCR (msPCR)

[0556] To determine the methylation levels of a preselected set of 192 different CpGs, real-time quantitative methylation-specific PCR (msPCR) was performed according to the following protocol:

[0557] For each of the 192 CpG-containing target regions to be analyzed, a specific set of three oligonucleotides was designed, containing one forward primer and two reverse primers. The two reverse primers were designed so that one reverse primer had a G at the 3' end that was complementary to the methylated, unaltered C, while the second forward primer had an A at the 3' end that was complementary to the converted uracil.

[0558] Then, cheek cells were collected from multiple individuals using cheek swabs, and genomic DNA was purified from cheek cells using the QIAamp 96 DNA SwabBioRobot kit (Qiagen, Hilden, Germany). The purified genomic DNA was treated with sodium bisulfite using the Zymo EZ DNA Methylation Kit (Zymo, Irvine, CA, USA). This treatment converts unmethylated cytosine to uracil, while methylated cytosine remains unchanged.

[0559] To determine the methylation level of CpGs contained in the sample, two PCR reactions were initiated for each set of three oligonucleotides, the first PCR reaction using the forward primer and the first of the two reverse primers, and the second PCR reaction using the forward primer and the second of the two reverse primers. Using real-time quantitative msPCR and TaqMan probes specific for each amplified target region, the methylation level of each CpG was determined.

[0560] From the results, numerical values ​​for the methylation levels of each of the 192 different CpGs are provided.The values ​​are again normalized so that 0 corresponds to the minimum possible methylation value for a CpG and 1 (100%) corresponds to the maximum methylation value for a CpG.

[0561] Since the number of different genomic DNA sequences is less than that of the method in Example 2, this method is extremely competitive in terms of cost.

[0562] Example 4: Generating an age predictor using LASSO

[0563] Unless otherwise stated, the DNA methylation levels of 289 individuals (259 for the training dataset and 30 for the test dataset) have been determined as described in Example 1. Briefly, the Infinium MethylationEPIC microbead chip (Illumina TM ) determined the DNA methylation levels of 850,000 different genomic DNA sequences from cheek swab samples. The methylation levels were normalized to beta values ​​using the program R v3.4.2, so that their values ​​can be between 0 and 1. The dataset, i.e., the training dataset, is a data matrix with the structure of Table 1.

[0564] Table 1

[0565] ID Chronological age CG1 CG2 … CG850000 Individual 1 28 0.2 1.0 … 0.1 Individual 2 8 … … … … … … … … … … Individual 259 65 … … … …

[0566] Using the statistical software R v3.4.1 and the biglasso package, LASSO regression was performed using the following command:

[0567] cvfit<-cv.biglasso(Vars800bm,Age,seed=2401,nfolds=20),

[0568] Where Vars800bm is the training dataset associated with the example matrix shown in Table 1, with cg loci as the independent variables and age as the dependent variable to be modeled; seed is the number used by the random generator; nfolds is the number of cross-validation repetitions that must be used to build the model. A value of 20 was used for cross-validation. The Biglasso package is as follows: "The biglasso Package: A Memory-and Computation-Effic Solver for LASSO ModelFitting with Big Data in R" by Yaohui Zeng and Patrick Breheny in arXiv:1701.05936v2[statCO] published on March 11, 2018.

[0569] The formula of the model (age indicator) obtained by LASSO regression is:

[0570] Age = +53.9126 * cg27320127 + 43.1588 * cg16267121 + 31.5464 * cg00831672 + 30.4384 * cg27173374 + 26.5197 * cg16867657 + 20.9302 * cg14681176 + 19.0975 * cg25606723 + 16.8674 * cg11607603 + 16.6092 * cg08097417 + 15.0595 * cg11330075 + 14.5786 * cg12333719 + 14.1955 * cg10543136 + 13.6743 * cg21807065 + 12.4988 * cg19851481 + 12.1954 * cg08224787 + 11.7822 * cg19702785 + 11.7706 * cg13759931 + 11.6845 * cg19112204 + 11.4521 * cg07955995 + 10.869 * cg18815943 + 10.829 * cg24724428 + 10.7537 * cg22101188 + 10.4571 * cg19215678 + 9.551 * cg22519947 + 9.5225 * cg06161948 + 9.3932 * cg16677512 + 9.2647 * cg05396610 + 8.9059 * cg21628619 + 8.7864 * cg15609017 + 8.6846 * cg24954665 + 8.5015 * cg25642673 + 8.284 * cg07802350 + 7.9408 * cg05087008 + 7.8335 * cg12548216 + 7.7144 * cg09965557 + 7.6203 * cg16999154 + 7.6057 * cg12238343 + 7.5126 * cg08044253 + 7.0673 * cg16465695 + 6.939 * cg13206721 + 6.6733 * cg09001642 + 6.1215 * cg11176990 + 6.0675 * cg07625177 + 6.0657 * cg05292016 + 5.9961 * cg16593468 + 5.9511 * cg07291317 + 5.5409 * cg18506897 + 5.4739 * cg07120630 + 5.2279 * cg08662753 + 5.1938 * cg24088134 + 5.1655 * cg00097800 + 4.8623 * cg16950671 + 4.6431 * cg16245716 + 4.6364 * cg06279276 + 4.6224*cg08686931+4.1089*cg27540719+4.0082*cg07529089+3.9294*cg06945504+3.8147*cg23677767+3.7304*cg07766948+3.7296*cg00876345+3.541*cg05972734+3.5305*cg22540792+3.4169*cg08118942+3.1845*cg02032962+3.1329*cg09460489+3.0723*cg22444338+3.0498*cg08856941+2.8317*cg03741619+2.7707*cg03230469+2.6979*cg06153788+2.6678*cg10522765+2.6533*cg14911690+2.5934*cg06186727+2.5488*cg03526652+2.5152*cg01520297+2.4409*cg09805798+2.3836*cg07513002+2.3539*cg08960065+2.3285*cg06335143+2.3044*cg16673857+2.2379*cg05990274+2.0254*cg04525002+1.9303*cg13154327+1.8016*cg07494888+1.7889*cg03388189+1.7543*cg08478427+1.7476*cg18768299+1.6312*cg21165089+1.6196*cg17665505+1.613*cg13460409+1.5347*cg14305139+1.4346*cg12804730+1.2032*cg04875128+1.2025*cg05211227+1.1767*cg18737844+1.1712*cg21460868+1.15*cg26430984+1.135*cg10321869+1.0067*cg14756158+1.0021*cg16322747+0.9948*cg17343879+0.9605*cg22077936+0.7994*cg18339380+0.5436*cg00087368+0.3003*cg05812299+0.281*cg12732998+0.0507*cg16456442+0.0277*cg17760405+0.0165*cg12658720–0.2038*cg08457029–0.4098*cg21962791–0.4232*cg15761531–0.4506*cg19810954–0 .4626*cg20425444–0.5866*cg23128025–0.6731*cg25845463–0.6945*cg03324695–1.0445*cg01636910– 1.4555*cg12650870–1.8012*cg01820962–2.2813*cg07782620–2.4468*cg04320377–2.6024*cg09275691 –2.6286*cg15008041–2.7124*cg20576243–3.4046*cg13973351–3.5199*cg08194377–3.5713*cg0738196 0–4.0608*cg10240079–4.2758*cg14231565–4.8117*cg24319133–4.8449*cg03680898–5.694*cg1930196 3–6.83*cg03473532–7.515*cg13333913–8.0702*cg05106770–8.3397*cg04287203–9.4713*cg27394136– 9.4931*cg10501210–10.8424*cg19432688–12.9786*cg02536625–13.2229*cg04028695–14.2271*cg1678 1885–14.728*cg15743533–14.9252*cg04733826–15.7917*cg20088545–16.5954*cg06831571–367.4866. .

[0571] The age indicator includes 148 words, such as +16.6092*cg08097417, where the positive sign indicates that the methylation level is positively correlated with age, and the negative sign indicates that the methylation level is negatively correlated with age. The numbered cg refers to the genomic DNA sequence according to the nfiniumMethylationEPIC microbead chip; and the absolute value of the coefficient multiplied by the cg indicates the importance of the cg.

[0572] Multiple model performance checks confirmed that the selection of 148 cg sites was optimal ( Figure 1 ).

[0573] The age indicator has the following properties: 2 = 0.72, selected variables = 148 (non-zero coefficients), where R 2 is the coefficient of determination. The statistic was determined from an independent test data set consisting of data from 30 individuals (about 10%) that were different from the 259 (289-30) individuals used in the training data set but were taken from the same population of 289 individuals.

[0574] In addition, LASSO has been applied to the data of 64 or 150 individuals out of the 289 individuals (Table 2).

[0575] Table 2

[0576]

[0577] This suggests that the performance of the LASSO improves when additional individuals' data are iteratively added to the dataset and the age indicator is iteratively updated.

[0578] Example 5: Generating an age predictor using LASSO and subsequent stepwise regression

[0579] The reduced training data set obtained after performing LASSO was subjected to stepwise regression (Example 4) to extract the most significant set of cg sites / CpGs to optimize the model. The reduced training data set (IME_blasso[,-1]) is the same as the training data set used in Example 4, except that it only retains 148 columns related to the 148 cg sites selected by LASSO.

[0580] Stepwise regression was performed using the statistical software R v3.4.1 and the following command:

[0581] model_blasso<-step(lm(Age~.,data=IME_blasso[,-1]),direction="both”), where the direction of removing insignificant variables is "both", which means that both adding variables and removing variables are allowed.

[0582] After LASSO regression and subsequent stepwise regression, the formula of the model obtained (age indicator) is:

[0583] Age = +66.2822 * cg11330075 + 65.203 * cg00831672 + 55.7265 * cg27320127 + 44.4116 * cg27173374 + 38.3902 * cg14681176 + 37.8069 * cg06161948 + 36.6564 * cg08224787 + 31.9397 * cg05396610 + 30.1919 * cg15609017 + 28.089 * cg09805798 + 27.9392 * cg19215678 + 27.8502 * cg12333719 + 27.226 * cg03741619 + 27.0323 * cg16677512 + 25.9599 * cg03230469 + 25.3932 * cg19851481 + 24.5374 * cg10543136 + 22.5525 * cg07291317 + 21.8666 * cg26430984 + 20.3621 * cg16950671 + 20.3269 * cg16867657 + 19.7973 * cg22077936 + 18.7137 * cg08044253 + 18.2047 * cg12548216 + 18.1936 * cg05211227 + 18.0812 * cg13759931 + 17.6857 * cg08686931 + 17.5303 * cg07955995 + 16.1143 * cg07529089 + 14.8703 * cg01520297 + 14.6684 * cg00087368 + 14.4397 * cg05087008 + 14.4361 * cg24724428 + 14.3055 * cg19112204 + 14.2968 * cg04525002 + 14.2302 * cg08856941 + 13.3831 * cg16465695 + 11.8127 * cg08097417 + 11.7798 * cg21628619 + 11.3523 * cg09460489 + 11.2461 * cg13460409 + 10.6268 * cg25642673 + 10.4347 * cg19702785 + 9.7844 * cg18506897 + 9.5931 * cg21165089 + 9.093 * cg27540719 + 8.9361 * cg21807065 + 8.8577 * cg18815943 + 8.6138 * cg23677767 + 7.1699 * cg07802350 + 7.0528 * cg11176990 + 6.5416 * cg10321869 + 6.5049*cg17343879+5.8296*cg08662753+5.696*cg14911690+3.2983*cg12804730+3.1388*cg163227 47–4.8653*cg14231565–5.5608*cg10501210–6.047*cg09275691–6.35*cg15008041–9.1942*cg058 12299–9.3144*cg24319133–9.4566*cg12658720–9.8704*cg20576243–10.4082*cg03473532–10.64 29*cg07381960–11.1592*cg05106770–12.0021*cg04320377–12.3296*cg19432688–12.9858*cg2251 9947–13.7116*cg06831571–13.8029*cg08194377–13.8668*cg01636910–14.6975*cg14305139–15. 0408*cg04028695–16.3295*cg15743533–16.3314*cg03680898–18.6196*cg20088545–19.0952*cg1 3333913–19.3068*cg19301963–21.5752*cg13973351–23.0892*cg16781885–26.0415*cg04287203– 32.3606*cg27394136–48.0918*cg10240079–50.0227*cg02536625–63.4434*cg23128025–519.3495. .

[0584] The meanings of terms and statistics are as explained in Example 4. See Table 6 for more details on the cg sequences and coefficients.

[0585] Therefore, after applying stepwise regression, the number of selected variables will be further reduced. In fact, the age indicator contains only 88 genomic DNA sequences (cg sites / CpG).

[0586] Furthermore, the performance of the age indicator obtained by LASSO and subsequent stepwise regression is:

[0587] R 2 =0.9884 (training data); R 2 = 0.9929 (test data set containing data of 30 test individuals, as explained in Example 4). Thus, the performance is enhanced compared to the age indicator obtained by LASSO without stepwise regression.

[0588] The performance on the test data is as good as on the training dataset, indicating that the age indicator has excellent performance ( Figure 2 ). Moreover, this higher coefficient of determination value represents a significant improvement over prior art age indicators.

[0589] By grouping individuals based on their chronological age (combined training and testing datasets), it was confirmed that the methylation levels of representative CG sites selected by regression analysis correlated well with age groups ( Figure 3 ).

[0590] The age indicators and their determined values ​​are then compared in Table 3 with the age indicators in Horvath, Genome Biology, 2013, 14: R115:

[0591] Table 3

[0592]

[0593]

[0594] This demonstrates that, despite accounting for only approximately 25% of the number of genomic DNA sequences (independent variables), the age indicator obtained by LASSO+stepwise regression performs at least as well as, or even better than, the relevant prior art age indicators.

[0595] The small set of genomic DNA sequences included in the age indicator allows the use of alternative, ie simpler, methods (see Examples 2 and 3) to determine the DNA methylation level of the individual whose age is to be determined.

[0596] Furthermore, the set of cg sites identified by LASSO alone or by LASSO + subsequent stepwise regression has little overlap with the cg sites identified in Horvath, Genome Biology, 2013, 14: R115 ( Figure 4 and 5 ).

[0597] Example 6: Determining gene sets from cg sites / CpG sets

[0598] The list of cg sites determined by applying LASSO (Example 4) or LASSO + stepwise regression (Example 5) was filtered to filter out those cg sites that were completely contained in the gene. In the first list (Table 4), 106 (partially redundant) coding sequences and non-coding sequences, such as microRNAs or long non-coding RNAs, were selected based on the 148 CpGs determined by LASSO:

[0599] Table 4

[0600]

[0601]

[0602]

[0603]

[0604]

[0605]

[0606]

[0607] From the reduced gene set (Table 5), gene targets useful for drug therapy were selected from Table 4. In particular, the gene was selected if an in vitro assay for determining the activity or function of the encoded protein was known in the art.

[0608] Table 5

[0609]

[0610]

[0611]

[0612] Finally, from the 88 CpGs determined by LASSO+ stepwise regression, a list with 68 (partially redundant) coding sequences and non-coding sequences (such as microRNA or long non-coding RNA) was selected (Table 6). The table further shows the coefficients and standard errors of the various age indicators (see Example 5).

[0613] Table 6

[0614]

[0615]

[0616]

[0617]

[0618] Example 7: Iterative update of age indicator Based on the judgment of whether the domain boundary of the test data is outside the domain boundary of the age indicator training set, the age indicator is automatically updated according to the situation (proband; individual). The domain boundary is the minimum and maximum DNA methylation level of each genomic DNA sequence included in the age indicator. The minimum and maximum DNA methylation levels exist in the original training data set that has been used to determine the age indicator. If there are other individual values ​​input and replace the original minimum and maximum values ​​of each CpG, these values ​​will change at any time. Therefore, the minimum value of each CpG will decrease (if the minimum value is not yet 0), and the maximum value will increase (if it is not yet 1). In this way, the domain boundary of the age indicator will be extended to the optimal value, and the possibility of further updating the age indicator will be smaller and smaller.

[0619] The updating is done with the following R code:

[0620] ##%####################################################%##

[0621] ##

[0622] ####Predictions with a test data set####

[0623] ##

[0624] ##%####################################################%##

[0625] prdct<-data.frame(SampleID=newsamlesdf$SampleID,

[0626] pred_age=predict(model_blasso,newsamplesdf),stringsAsFactors=F)

[0627] plot(newsamplesdf$Age,prdct$pred_age,pch=16,col="red",xlab="RealAge",ylab=

[0628] "Predicted Age")

[0629] abline(0,1,col="red")

[0630] ##%######################################################%##

[0631] ##

[0632] ####If the predictions this way are####

[0633] ####not satisfactory need to run this####

[0634] ##

[0635] ##%######################################################%##

[0636] IME_blasso<-IME_blasso%>%dplyr::select(Age,everything())

[0637] domain<-data.frame(min=apply(as.matrix(IME_blasso[,-1]),2,min),

[0638] max=apply(as.matrix(IME_blasso[,-1]),2,max))

[0639] #calculate domain for new samples

[0640] domain_curr<-data.frame(min=apply(as.matrix(newsamplesdf),2,min),

[0641] max=apply(as.matrix(newsamplesdf),2,max))

[0642] ##%######################################################%##

[0643] ##

[0644] ####operative check for prediction####

[0645] ##

[0646] ####%######################################################%##

[0647] if(sum((domain$min-domain_curr$min)<0&(domain$max-domain_curr$max)>0)){

[0648] nnew<-NROW(newsamplesdf)

[0649] nn<-NROW(IME_blasso)

[0650] #add new probands to the training set

[0651] newIME_blasso<-rbind(IME_blasso,newsamplesdf)#concatenate the two set

[0652] #rerun the model model_blasso_new<-step(lm(Age~.,data=newIME_blasso),direction="both")

[0653] sstep<-summary(model_blasso_new)

[0654] sstep

[0655] ##check

[0656] par(mfrow=c(1,1))

[0657] plot(newIME_blasso$Age,model_blasso_new$fitted.values,

[0658] xlab="Real Age[red points=new points]",ylab="Predicted Age",

[0659] main=paste("Stepwise Regression with IME_newModel CpGs R 2 =",

[0660] round(sstep$r.squared,3),sep=""),pch=1)

[0661] abline(0,1,col="red")

[0662] errs<-newIME_blasso$Age-model_blasso_new$fitted.values

[0663] mae(errs)

[0664] postResample(newIME_blasso$Age,

[0665] as.vector(model_blasso_new$fitted.values))

[0666] points(newIME_blasso$Age[nn:(nn+nnew)],

[0667] as.vector(model_blasso_new$fitted.values[nn:(nn+nnew)]),col="red",

[0668] pch=16)

[0669] ##

[0670] predictions<-data.frame(Age=newIME_blasso$Age[nn:(nn+nnew)],

[0671] PredAge=model_blasso_new$fitted.values[nn:(nn+nnew)])

[0672] write.csv(predictions,"predictions.csv")

[0673] save(model_blasso_confy_new,file="model_blasso_new.lm")

[0674] #rm(newIME_blasso)

[0675] }else{

[0676] predicted<-predict.lm(model_blasso_confy,newsamplesdf)

[0677] plot(newsamplesdf$Age,predicted,pch=12,main="Predictions with IME_model")

[0678] abline(coef=c(model_blasso_new,1),col="red")

[0679] external_pred<-data.frame(PredAge=predicted,RAge=newsamplesdf$Age)

[0680] postResample(predicted,newsamplesdf$Age)

[0681] }

[0682] Example 8: Further statistical analysis and age prediction

[0683] DNA has been collected from approximately 200 individuals. These samples were all obtained in northern Germany, but in order to make the database broader, care has been taken not to exclude any individual due to factors such as chronological age, general health status, obesity, physical fitness level, drug consumption (including drugs such as nicotine and alcohol), etc. Therefore, this group is considered to be representative of the general population.

[0684] The CpG methylation levels of DNA in biological samples of about 100 individuals have been determined using the method of Example 1, and a relatively large number of CpGs, about 850,000 (850,000), have been obtained for each individual.

[0685] Given the volume of data and the computational cost of its analysis, the data were arbitrarily divided into smaller groups and the data of these smaller groups were then analyzed.

[0686] A principal component analysis was performed using the data of the first group of 16 individuals, and 10 main components were found that accounted for almost all the variance observed in the methylation levels of the CpGs of the samples of each group, of which the first two components were already responsible for 98% of the variance, which clearly showed that, despite the very large number of different CpG methylation levels considered, it is advisable to reduce the number. Based on the principal component analysis and using regression techniques, a predictor model was built for each group, which basically showed that the constructed model was still cluttered with some meaningless coefficients.

[0687] It was also determined that, apart from this, many of the coefficients determined were also found to be not statistically significant.

[0688] With this in mind, data from the first large group (98 individuals) were analyzed with the aim of establishing a model in which the number of CpGs to be considered was significantly reduced while maintaining high statistical significance for all parameters. To this end, LASSO regression was first performed; note that LASSO regression is a well-known technique in the art and software packages for implementing Lasso regression are readily available. It should be noted that it is possible to distinguish whether the methylation level of a given CpG has a specific statistical relevance; this enables only CpGs with a certain relevance to be considered. In particular, in this regard, reference is made to "Thebiglasso Package:A Memory-and Computation-Effic Solver for LASSO ModelFitting with Big Data in R" published by Yaohui Zeng and Patrick Breheny in arXiv:1701.05936v2[statCO] on March 11, 2018. Using the best set of only 50 different determined CpGs selected by LASSO regression, attempts were made to further optimize the resulting model. This step was completed using the XgBoost algorithm. Please note that XgBoost is a well-known open source software library that provides a gradient boosting framework in multiple languages. It should be noted that XgBoost is used to modify the coefficients used in statistical models. For more details on the XgBoost algorithm and its implementation, please refer to "XGBoost: A ScalableTree Boosting System" published by T. Chen and C. Guestrin in arXiv:1603.02754v3 on June 10, 2016. For the purpose of disclosure, the contents of the cited documents are fully incorporated into this article in their entirety.

[0689] It was found that a high performance model that produced good regression coefficients could be obtained.

[0690] However, not content with the high regression coefficients obtained for the considered group, nor with keeping this high-performance model unchanged, the data from another 98 individuals were analyzed in the same manner as before. It was found that for the second group, about 78 CpGs should be considered in the model, and 8 of the 78 CpGs overlapped with the 50 CpGs selected for the first arbitrary group (98 individuals).

[0691] Then, another run was performed and it was determined that, in the combined set, 70 CpGs were selected very favorably out of the approximately 850,000 different CpGs originally considered. Of these 70 CpGs, only 10 overlapped with CpGs in the first set, only 12 overlapped with CpGs in the second set, and 8 overlapped with both sets.

[0692] Regression using XgBoost allowed the same high performance to be maintained after 20 rounds of cross validation.

[0693] This suggests that by statistical means (particularly LASSO regression, PCA or other means for distinguishing whether a particular CpG among a large number of CpGs is statistically relevant), the number of CpGs can be significantly reduced from an overall extremely large set to a fairly small set, thereby allowing cheap detection using the methods mentioned in Examples 2 and 3 above.

[0694] A usable model can then be built for only this smaller set, and despite the smaller number of CpGs considered, age can still be determined with high accuracy and small confidence intervals, especially by repeatedly iterating the parameters of the built statistical model.

[0695] In this way, despite the overall small number of CpGs considered, the ages initially determined will be very precise and their reliability will increase over time.< / bs>

Claims

1. A method for determining an age indicator, comprising the following steps: (a) Provide multiple individual training data sets, each individual training data set includes: (i) DNA methylation levels of a set of genomic DNA sequences, and (ii) chronological age, and (b) applying a regression method on the training data set comprising: (i) Least Absolute Shrinkage and Selection Operator (LASSO), thereby determining the reduced training dataset, and (ii) subsequent stepwise regression to determine said age indicator, The independent variable is the methylation level of the genomic DNA sequence, The age indicators include: (i) a subset of said set of genomic DNA sequences, and (ii) at least one coefficient for each genomic DNA sequence contained in the subset of the genomic DNA sequence set, the coefficient being a weight of the methylation level of the genomic DNA sequence, and The reduced training data set includes all data in the training data set except the DNA methylation levels of the genomic DNA sequences eliminated by the minimum absolute shrinkage and selection operator.

2. A method for determining the age of an individual, comprising the steps of: (a) Provide multiple individual training data sets, each individual training data set includes: (i) DNA methylation levels of a set of genomic DNA sequences, and (ii) chronological age, and (b) applying a regression method on the training data set comprising: (i) Least Absolute Shrinkage and Selection Operator (LASSO), thereby determining the reduced training dataset, and (ii) subsequent stepwise regression to determine the age indicator, The independent variable is the methylation level of the genomic DNA sequence, The age indicators include: (i) a subset of said set of genomic DNA sequences as a set, and (ii) at least one coefficient for each genomic DNA sequence contained in the subset of the genomic DNA sequence set, the coefficient being a weight of the methylation level of the genomic DNA sequence, and wherein the reduced training data set includes all data in the training data set except the DNA methylation levels of genomic DNA sequences eliminated by the minimum absolute shrinkage and selection operator, and (c) providing a DNA methylation level of the individual whose age is to be determined, said DNA methylation level being the DNA methylation level of at least 80% of said genomic DNA sequences included in said age indicator, and (d) determining the age of the individual based on the DNA methylation level of the individual and the age indicator; Wherein, each step of the method is executed by a computer.

3. The method according to claim 1 or 2, wherein the stepwise regression is applied to the reduced training data set.

4. The method according to claim 1 or 2, wherein the dependent variable is age.

5. The method of claim 1 or 2, wherein the stepwise regression is a two-way elimination, wherein statistically insignificant independent variables are removed.

6. The method according to claim 1 or 2, wherein the regression method does not include ridge regression (L2 regularization) or the L2 regularization parameter / λ parameter is 0, and / or, The LASSO L1 regularization parameter / alpha parameter is 1.

7. The method according to claim 1 or 2, wherein the age indicator is iteratively updated, the iterative updating comprising adding data of at least one other individual to the training data in each iteration, thereby iteratively expanding the training data set.

8. The method of claim 7, wherein one updating round comprises applying the LASSO on the expanded training data set to thereby determine an updated reduced training data set; and Stepwise regression is applied to the reduced training data set to determine updated age indicators.

9. The method according to claim 7, wherein the mass of the age indicator is determined, wherein the determining the mass comprises the following steps: (a) providing a plurality of test data sets of individuals who have not contributed data to the training data set, wherein the test data set of each individual comprises: (i) said DNA methylation level of said set of genomic DNA sequences included in said age indicator, and (ii) chronological age; and (b) determining said quality of said age indicator by statistical evaluation and / or domain boundary evaluation, The statistical evaluation includes: (i) determining the age of the individuals included in the test data set, (ii) correlating the determined age with the chronological age of the individual and determining a statistical parameter describing the correlation between the determined age and the chronological age of the individual, and (iii) determining whether the statistical parameter indicates that the age indicator has acceptable quality, wherein the statistical parameter is a coefficient of determination (R 2 ), where R 2 is greater than 0.90, indicating acceptable quality, and wherein the domain boundary evaluation includes (iv) determining said domain boundaries of said age indicator, wherein the domain boundaries are the minimum and maximum DNA methylation levels for each genomic DNA sequence included in the age indicator, and wherein said minimum and maximum DNA methylation levels are present in said training data set that has been used to determine said age indicator, and (v) determining whether the test data set exceeds the domain boundary, wherein not exceeding the domain boundary indicates acceptable quality; The age indicator is updated when the quality of the age indicator is unacceptable.

10. The method according to claim 1 or 2, wherein the age indicator comprises at least 80 genomic DNA sequences, and / or The age indicator comprises fewer than 300 genomic DNA sequences.

11. The method of claim 1 or 2, wherein the age indicator comprises less than 100 genomic DNA sequences.

12. A method for determining the age of an individual according to any one of claims 2, 8 and 9, wherein the training data set also includes at least one lifestyle factor, and the method also includes the step of determining at least one lifestyle factor associated with the difference between the determined age and the actual age of the individual.

13. A method for determining the age of an individual, comprising: (a) provide age indicators, including: (i) a collection of genomic DNA sequences, comprising: At least 70 DNA sequences selected from the group consisting of: cg11330075, cg00831672, cg27320127, cg27173374, cg14681176, cg06161948, cg08224787, cg05396610, cg15609017, cg09805798, cg19215678, cg12333719, cg03741619, cg16677512, cg03230469, cg19851481, cg10543136, cg07291317, cg26430984, cg16950671, cg168 67657, cg22077936, cg08044253, cg12548216, cg05211227, cg13759931, c g08686931, cg07955995, cg07529089, cg01520297, cg00087368, cg050870 08, cg24724428, cg19112204, cg04525002, cg08856941, cg16465695, cg08 097417, cg21628619, cg09460489, cg13460409, cg25642673, cg19702785, cg18506897, cg21165089, cg27540719, cg21807065, cg18815943, cg2367 7767, cg07802350, cg11176990, cg10321869, cg17343879, cg08662753, cg 14911690, cg12804730, cg16322747, cg14231565, cg10501210, cg0927569 1. cg15008041, cg05812299, cg24319133, cg12658720, cg20576243, cg034 73532, cg07381960, cg05106770, cg04320377, cg19432688, cg22519947, c g06831571, cg08194377, cg01636910, cg14305139, cg04028695, cg157435 33. cg03680898, cg20088545, cg13333913, cg19301963, cg13973351, cg16781885, cg04287203, cg27394136, cg10240079, cg02536625, and cg23128025; (ii) at least one coefficient for each genomic DNA sequence contained in the set, the coefficient being a weight of the methylation level of the genomic DNA sequence; wherein the age indicator comprises less than 300 genomic DNA sequences, (c) providing a DNA methylation level of the individual whose age is to be determined, said DNA methylation level being the DNA methylation level of at least 80% of said genomic DNA sequences included in said age indicator, and (d) determining the age of the individual based on the DNA methylation level of the individual and the age indicator; Wherein, each step of the method is executed by a computer.

14. The method according to claim 13, wherein the age indicator is obtained by the method for determining an age indicator according to any one of claims 1 or 3-11.

15. The method according to any one of claims 1, 8 and 9, wherein the method is a computer-implemented method.

16. The method according to any one of claims 1-2, 8, 9 and 13, wherein the DNA methylation level of the genomic DNA sequence of the individual is a methylation level measured in a biological material sample of the individual containing the genomic DNA sequence.

17. The method of claim 16, wherein the sample comprises buccal mucosal cells.

18. A chip comprising a collection of genomic DNA sequences, the collection comprising: At least 70 DNA sequences selected from the group consisting of: cg11330075, cg00831672, cg27320127, cg27173374, cg14681176, cg06161948, cg08224787, cg05396610, cg15609017, cg09805798, cg19215678, cg12333719, cg03741619, cg16677512, cg03230469, cg19851481, cg10543136, cg07291317, cg26430984, cg16950671, cg16867657, cg22077936, cg08044253, cg12548216, cg05211227, cg13759931, cg08686931, cg07955995, cg07529089, cg01520297, cg00087368, cg05087008, cg24724428, cg19112204, cg04525002, cg08856941, cg16465695, cg08097417, cg21628619, cg09460489, cg13460409, cg25642673, cg19702785, cg18506897, cg21165089, cg27540719, cg21807065, cg18815943, cg23677767, cg07802350, cg11176990, cg10321869, cg17343879, cg08662753, cg14911690, cg12804730, cg16322747, cg14231565, cg10501210, cg09275691, cg15008041, cg05812299, cg24319133, cg12658720, cg20576243, cg03473532, cg07381960, cg05106770, cg04320377, cg19432688, cg22519947, cg06831571, cg08194377, cg01636910, cg14305139, cg04028695, cg15743533, cg03680898, cg20088545, cg13333913, cg19301963, cg13973351, cg16781885, cg04287203, cg27394136, cg10240079, cg02536625 and cg23128025, wherein,The chip comprises less than 500 spots, each sequence is contained in a separate spot, and the chip is suitable for detecting the DNA methylation level of the collection of genomic DNA sequences.

19. Use of the chip according to claim 18 in preparing a diagnostic kit, wherein the diagnostic kit is used to detect the age, health status and / or health status of an individual.

20. A kit comprising: (i) a chip according to claim 18, and (ii) a container for biological material, cheek swab material, spin columns and / or enzymes for extracting, purifying or amplifying genomic DNA from a biological sample, and / or bisulfite.

Images