Application of DNA methylation site marker in improvement of age prediction performance of machine learning model

Abstract

The invention discloses an application of a DNA methylation site marker in improvement of age prediction performance of a machine learning model, and provides a richer feature basis for learning of various methylation-age association modes by machine learning in combination with a newly discovered marker on the basis of known age-related methylation sites (ageCpGs). Therefore, the methylation prediction capability baseline of methods such as machine learning is improved, and the prediction accuracy can be effectively enhanced.

Description

Technical Field

[0001] The present invention relates to the field of model building, and specifically to the application of DNA methylation site markers in improving the age prediction performance of machine learning models. Background Art

[0002] The epigenetic clock is a method for estimating biological age by measuring the methylation status of specific CpG sites based on the correlation between DNA methylation levels and age. This technology has garnered widespread attention in recent years in fields such as aging research, disease diagnosis, and forensics. DNA methylation is an epigenetic modification that typically occurs on the cytosine residues of CpG dinucleotides. Methylation levels vary with age, and the methylation status of certain CpG sites is highly correlated with age.

[0003] Age prediction models are constructed by analyzing methylation levels at multiple CpG sites and combining them with statistical or machine learning models (such as elastic net regression and deep learning). These models map methylation data to age, generating an "epigenetic age." Representative models include the Horvath epigenetic clock (2013, using 353 CpG sites, based on elastic net regression, and applicable to a variety of tissues and cell types); the Hannum clock (2013, based on 71 CpG sites, primarily targeting blood samples); and PhenoAge and GrimAge (which focus more on epigenetic age associated with health and mortality risk, incorporating disease and lifestyle factors).

[0004] Previous research and technology in this field have primarily focused on improving prediction performance across different species, tissues (or pan-tissues), and populations, or on enhancing data and algorithms. However, there are still deficiencies in the diverse identification of methylation-age association patterns and the in-depth exploration of age-related methylation sites. Identifying more methylation-age association patterns will help uncover more potential sites, expanding upon the currently known sites in this area and further improving the age prediction performance of machine learning models. Summary of the Invention

[0005] In view of this, one of the objects of the present invention is to provide an application of a DNA methylation site marker in improving the age prediction performance of a machine learning model; a second object of the present invention is to provide a method for establishing a machine learning model for predicting age performance.

[0006] In order to achieve the above object, the present invention provides the following technical solutions: The invention relates to an application of DNA methylation site markers in improving the age prediction performance of a machine learning model, wherein the DNA methylation site markers are 283 sites: cg05437692, cg00743094, cg20676788, cg09294739,cg16072688, cg01331772, cg12912690, cg03468072, cg04845871, cg13935689,cg22254983, cg19715771, cg18450254, cg00842595, cg12526298, cg07850604,cg18582010, cg22158769, cg19401340, cg22622440, cg13860281, cg27427318,cg14911690, cg09066326, cg22990198, cg16997203, cg04945753, cg21304158,cg04685228, cg11622162, cg21927946, cg23193759, cg26049998, cg00034101,cg03930153, cg12951726, cg07303143, cg16513459, cg10397932, cg16750801,cg23669043, cg06688910, cg05005256, cg09978533, cg02156296, cg06832008,cg11039241, cg26513192, cg08943494, cg12892303, cg06998507, cg19730691,cg00541544, cg06285575, cg08854834, cg06493386, cg07227769, cg14699734,cg25590826, cg27198696, cg03476007, cg27274542, cg18736141, cg22719623,cg11932158, cg08271909, cg03767531, cg25940946, cg25753024, cg11617964,cg03760191, cg06037631, cg14901671, cg02571874, cg10002915, cg12943155,cg06870470, cg24611351, cg07164639, cg04865726,cg25637829, cg13100137,cg12253200, cg03065175, cg08972190, cg01899130, cg24020215, cg01055562,cg00689213, cg08511485, cg03301331, cg23461714, cg03131366, cg23602058,cg25826226, cg22900415, cg10247798, cg12024906, cg16154155, cg20068209,cg02481642, cg14284824, cg12659626, cg13488078, cg00522276, cg19996355,cg13362112, cg22807378, cg10203610, cg11495544, cg24580071, cg11604426,cg00570618, cg06872257, cg00343092, cg21328810, cg07971188, cg26052635,cg05221455, cg16069986, cg20592391, cg00412842, cg05331060, cg02447229,cg25090514, cg26700215, cg19514721, cg19065177, cg06196801, cg16300030,cg18623216, cg06073351, cg19136673, cg01278041, cg01514490, cg08398371,cg21223094, cg19449917, cg00491255, cg10012394, cg19022697, cg18393958,cg12577010, cg21089930, cg19150852, cg21205139, cg13026729, cg21915678,cg19789466, cg19343530, cg20076442, cg00685614, cg09565014, cg18899220,cg02804091, cg12036735, cg16888547, cg03087372, cg01344797, cg12713583,cg01894038, cg06580318, cg08822715, cg02375320,cg05599106, cg02691192,cg26530319, cg07518532, cg23907053, cg24338780, cg01580888, cg04742719,cg06231995, cg25909396, cg09285095, cg02396676, cg23008153, cg24675344,cg08939418, cg05917988, cg00655552, cg13251842, cg15519474, cg12111714,cg03433048, cg13029847, cg01447660, cg05694021, cg01890712, cg18038207,cg11176990, cg01585860, cg09595912, cg15967278, cg10905180, cg04880546,cg17945323, cg01511232, cg02047547, cg15205547, cg24217948, cg10860819,cg16330755, cg09432234, cg09320690, cg25974380, cg02017450, cg07810039,cg07372034, cg23595222, cg25150953, cg20600850, cg09310092, cg18448426,cg05302531, cg00804078, cg06445343, cg24170040, cg24536146, cg07813031,cg22128645, cg00254470, cg10625705, cg03390636, cg13221767, cg26229990,cg04340651, cg26158023, cg10977734, cg21767759, cg20012607, cg12096762,cg17696044, cg22243000, cg20404336, cg17400113, cg21523251, cg15016701,cg18008766, cg00593462, cg03607951, cg02966332, cg06458239, cg14956327,cg02560388, cg27665659, cg07430472, cg18778196,cg01411823, cg18853490,cg07685728, cg02699829, cg13719901, cg18565342, cg25402787, cg05478824,cg19837174, cg22236626, cg14627089, cg12546181, cg26833936, cg17671280,cg01207684, cg06260709, cg23062810, cg06413398, cg22598841, cg13165009,cg09432590, cg10006442, cg02903680, cg06648759, cg05714552, cg11787522,cg13485366, cg18434712, cg26350754, cg22162281, cg27295118, cg04596060, cg12967723, cg07810190, cg15743533. ,

[0007] Preferably, the machine learning model is an elastic net regression model or a neural network model.

[0008] 2. A method for establishing a machine learning model for predicting age performance uses a set of 708 human age-related DNA methylation site markers. Using DNA methylation sequencing data, the dataset is split into a training set and a test set at a ratio of 9:1. The elastic net regression model is trained with hyperparameters max_iter=100000, alpha=0.001, and l1_ratio=0.01. The 708 sites are: cg24466241, cg11873854, cg20591472, cg06942701,cg12765028, cg08234504, cg12878812, cg10475970, cg13697378, cg10743390,cg18902090, cg13378934, cg16290275, cg13665684, cg26875805, cg10635122,cg04427498, cg01542719, cg08534628, cg04455430, cg25410668, cg07416237,cg23950157, cg12189835, cg10874644, cg03607117, cg19784428, cg06819923,cg03545227, cg09232937, cg27067781, cg00748589, cg11220950, cg09068158,cg03440125, cg03915012, cg19505546, cg13848598, cg12462816, cg05575054,cg24667575, cg15557036, cg24136205, cg01557798, cg01820962, cg11126134,cg20761322, cg15480367, cg11076306, cg07219494, cg15962547, cg22736354,cg19722847, cg01097384, cg19663246, cg17243289, cg09809672, cg25047092,cg04999352, cg07553761, cg21801378, cg05991454, cg21117668, cg05266781,cg02281167, cg12899747, cg00930873, cg00481951, cg00503840, cg22108374, cg24871743,cg18964582, cg07057579, cg05463027, cg23973429, cg15711521,cg00745389, cg04751276, cg15243034, cg19711579, cg06335143, cg17593472,cg21572722, cg05213896, cg02022733, cg02318971, cg24698655, cg16541931,cg14010720, cg21683284, cg23718736, cg18445088, cg14934413, cg20461912,cg16181396, cg00484358, cg20249566, cg16477091, cg08957484, cg09553242,cg24774812, cg06933824, cg17804348, cg12848065, cg22809047, cg08622677,cg15931839, cg10189695, cg23077820, cg00394316, cg13390332, cg00327072,cg14556683, cg22901840, cg07485775, cg06226630, cg18898125, cg07502389,cg21184711, cg08541518, cg03555227, cg25349154, cg15894389, cg10805721,cg12373771, cg16054275, cg23605843, cg15393490, cg14069287, cg03643998,cg20665157, cg02018902, cg08955995, cg17885226, cg23606718, cg10665321,cg05371578, cg02650266, cg03020208, cg05156137, cg23124451, cg04861640,cg18335796, cg06821999, cg06874016, cg25256723, cg18465694, cg10778288,cg19761273, cg13210467, cg11649376, cg08885800, cg05308819, cg20426994,cg18064714,cg00419702, cg05207048, cg03350900, cg22156456, cg20773033,cg06448705, cg09221867, cg06639320, cg22943590, cg01282174, cg26578682,cg04528819, cg27367526, cg22282410, cg14614094, cg02711397, cg10242160,cg01844642, cg15485728, cg03054277, cg07803108, cg24853724, cg23186333,cg12052661, cg09099868, cg21992250, cg09578568, cg14692377, cg00602811,cg01278873, cg18343474, cg26543112, cg20119148, cg08993878, cg21186299,cg26921969, cg12623930, cg06707978, cg01606885, cg02352240, cg01763090,cg00303541, cg13347071, cg14676592, cg00145253, cg02741882, cg00753885,cg17861230, cg16219603, cg12079303, cg09082287, cg03957108, cg01770755,cg02924487, cg18338984, cg10151367, cg22947000, cg07894352, cg23510764,cg18667659, cg00573770, cg21159778, cg23174607, cg14176752, cg03473532,cg22796704, cg18240400, cg16932827, cg09644065, cg22930275, cg01141812,cg24430580, cg13385220, cg18826637, cg07547549, cg07920503, cg23078123,cg21567504, cg19671120, cg00834536, cg20515136, cg19381811, cg04875128,cg02383785,cg05404236, cg12451153, cg11970349, cg26161329, cg00590036,cg17783401, cg15957394, cg01153166, cg06419432, cg12948621, cg09278098,cg05700079, cg05675373, cg06493994, cg22793142, cg22682811, cg22303418,cg01302656, cg00439658, cg07850154, cg20222376, cg12939283, cg11847992,cg11826475, cg05106770, cg16008966, cg24212517, cg03671597, cg17729667,cg09118625, cg19702785, cg13001142, cg08097417, cg05175318, cg16686396,cg19392831, cg16969368, cg06526721, cg13649056, cg13806741, cg04105923,cg23538901, cg20153322, cg24891133, cg19991948, cg04581938, cg07082267,cg04453050, cg15936446, cg26343883, cg21393163, cg03384000, cg26005082,cg11071401, cg25439632, cg17326555, cg07963234, cg24433287, cg17592231,cg25533247, cg02154074, cg20359994, cg18603228, cg15726426, cg02900766,cg25453381, cg17110586, cg18633600, cg06737494, cg27401724, cg13929658,cg24150153, cg16419235, cg02286081, cg20329210, cg21878650, cg11741201,cg17621438, cg23488804, cg21632975, cg07211259, cg05740045, cg18933331,cg01196788,cg10906284, cg03530962, cg21911021, cg00448560, cg25584930,cg17168836, cg13654588, cg04503319, cg23500537, cg11807280, cg00741624,cg04104132, cg13672300, cg21899500, cg09143195, cg04090392, cg18987410,cg18267374, cg07101980, cg01692968, cg19230755, cg17241310, cg04084157,cg14674720, cg11705975, cg00059225, cg25478614, cg21709871, cg04819580,cg03431918, cg02721902, cg09692396, cg14134497, cg05024939, cg04836038,cg20219381, cg17955729, cg21570988, cg00006081, cg04398950, cg00329615,cg07080372, cg06372850, cg16193278, cg14829814, cg03032497, cg18236477,cg11693709, cg16705273, cg10523019, cg01243823, cg08904363, cg27015931,cg16867657, cg21296230, cg23480341, cg01528542, cg07408456, cg15736994,cg20751926, cg25148589, cg16313343, cg03725309, cg22197050, cg23534245,cg00048759, cg02928699, cg20149168, cg19848940, cg16717122, cg04684267,cg19344626, cg03467555, cg03975694, cg04419610, cg12757011, cg00740914,cg16688437, cg01908954, cg19451698, cg25693132, cg15481583, cg25427880,cg21657577,cg24436906, cg12534424, cg21867345, cg01560871, cg17183905,cg11084334, cg22358580, cg01995468, cg03399905, cg12934382, cg17321954,cg06570224, cg12145907, cg16762684, cg27320127, cg15804973, cg04604946,cg05301866,cg05437692, cg00743094, cg20676788, cg09294739, cg16072688,cg01331772, cg12912690, cg03468072, cg04845871, cg13935689, cg22254983,cg19715771, cg18450254, cg00842595, cg12526298, cg07850604, cg18582010,cg22158769, cg19401340, cg22622440, cg13860281, cg27427318, cg14911690,cg09066326, cg22990198, cg16997203, cg04945753, cg21304158, cg04685228,cg11622162, cg21927946, cg23193759, cg26049998, cg00034101, cg03930153,cg12951726, cg07303143, cg16513459, cg10397932, cg16750801, cg23669043,cg06688910, cg05005256, cg09978533, cg02156296, cg06832008, cg11039241,cg26513192, cg08943494, cg12892303, cg06998507, cg19730691, cg00541544,cg06285575, cg08854834, cg06493386, cg07227769, cg14699734, cg25590826,cg27198696, cg03476007, cg27274542, cg18736141, cg22719623, cg11932158,cg08271909,cg03767531, cg25940946, cg25753024, cg11617964, cg03760191,cg06037631, cg14901671, cg02571874, cg10002915, cg12943155, cg06870470,cg24611351, cg07164639, cg04865726, cg25637829, cg13100137, cg12253200,cg03065175, cg08972190, cg01899130, cg24020215, cg01055562, cg00689213,cg08511485, cg03301331, cg23461714, cg03131366, cg23602058, cg25826226,cg22900415, cg10247798, cg12024906, cg16154155, cg20068209, cg02481642,cg14284824, cg12659626, cg13488078, cg00522276, cg19996355, cg13362112,cg22807378, cg10203610, cg11495544, cg24580071, cg11604426, cg00570618,cg06872257, cg00343092, cg21328810, cg07971188, cg26052635, cg05221455,cg16069986, cg20592391, cg00412842, cg05331060, cg02447229, cg25090514,cg26700215, cg19514721, cg19065177, cg06196801, cg16300030, cg18623216,cg06073351, cg19136673, cg01278041, cg01514490, cg08398371, cg21223094,cg19449917, cg00491255, cg10012394, cg19022697, cg18393958, cg12577010,cg21089930, cg19150852, cg21205139, cg13026729, cg21915678, cg19789466,cg19343530,cg20076442, cg00685614, cg09565014, cg18899220, cg02804091,cg12036735, cg16888547, cg03087372, cg01344797, cg12713583, cg01894038,cg06580318, cg08822715, cg02375320, cg05599106, cg02691192, cg26530319,cg07518532, cg23907053, cg24338780, cg01580888, cg04742719, cg06231995,cg25909396, cg09285095, cg02396676, cg23008153, cg24675344, cg08939418,cg05917988, cg00655552, cg13251842, cg15519474, cg12111714, cg03433048,cg13029847, cg01447660, cg05694021, cg01890712, cg18038207, cg11176990,cg01585860, cg09595912, cg15967278, cg10905180, cg04880546, cg17945323,cg01511232, cg02047547, cg15205547, cg24217948, cg10860819, cg16330755,cg09432234, cg09320690, cg25974380, cg02017450, cg07810039, cg07372034,cg23595222, cg25150953, cg20600850, cg09310092, cg18448426, cg05302531,cg00804078, cg06445343, cg24170040, cg24536146, cg07813031, cg22128645,cg00254470, cg10625705, cg03390636, cg13221767, cg26229990, cg04340651,cg26158023, cg10977734, cg21767759, cg20012607, cg12096762, cg17696044,cg22243000,cg20404336, cg17400113, cg21523251, cg15016701, cg18008766,cg00593462, cg03607951, cg02966332, cg06458239, cg14956327, cg02560388,cg27665659, cg07430472, cg18778196, cg01411823, cg18853490, cg07685728,cg02699829, cg13719901, cg18565342, cg25402787, cg05478824, cg19837174,cg22236626, cg14627089, cg12546181, cg26833936, cg17671280, cg01207684,cg06260709, cg23062810, cg06413398, cg22598841, cg13165009, cg09432590,cg10006442, cg02903680, cg06648759, cg05714552, cg11787522, cg13485366,cg18434712, cg26350754, cg22162281, cg27295118, cg04596060, cg12967723, cg07810190, cg15743533. ,

[0009] Alternatively, a set of 708 human age-related DNA methylation site markers was used, and then DNA methylation sequencing data was used to divide the dataset into a training set and a test set at a ratio of 9:1. The training-validation set was used for model training using a ten-fold cross-validation.

[0010] The beneficial effects of the present invention are as follows: the present invention discloses a set of 708 human age-related DNA methylation site markers, all of which are within the site coverage of the Illumina HumanMethylation450, IlluminaHumanMethylation850, and IlluminaHumanMethylationEPIC sequencing platforms, including 283 newly discovered DNA methylation sites and 425 reported sites. The new DNA methylation sites are compatible with the reported sites and include a variety of age-methylation association patterns, which helps the model learn the age-methylation association, thereby improving the predictive ability and effectively enhancing the model's predictive performance. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] In order to make the purpose, technical solutions and beneficial effects of the present invention more clear, the present invention provides the following drawings for illustration: Figure 1 The figure shows the age-methylation trajectory of the representative sites among the 708 human age-related DNA methylation site markers published in the present invention; Figure 2 The neural network architecture used for age prediction in the present invention; Figure 3 Figure 3 is the model performance of age prediction using a neural network based on the complete 708 ageCpGs proposed in the present invention.

[0012] Figure 4 This is the performance of the model for age prediction using a neural network based on the ageCpGs proposed in this invention. FULL indicates the complete 708 ageCpGs, new only indicates 283 newly discovered ageCpGs, and reported only indicates 425 reported ageCpGs. DETAILED DESCRIPTION

[0013] The present invention will be further described below with reference to the accompanying drawings and specific embodiments so that those skilled in the art can better understand the present invention and implement it. However, the embodiments are not intended to limit the present invention.

[0014] Using DNA methylation sequencing data from normal human samples (filter terms: Healthy, Normal, Control, Wild-type, Adjacent Normal Tissue form Cancer Patients) from public databases (GEO and ArrayExpress), a specially modified epigenomic association study (EWAS) based on a linear mixed model (LMM) was used to identify five groups of human age-associated DNA methylation markers (age-CpGs) with distinct methylation-age association patterns. The union of these five age-CpG groups yielded 708 sites on the Illumina HumanMethylation450 platform. These included 425 previously reported age-CpGs (Gregory Hannum et al. 2013; Horvath S. 2013; Weidner, CI, et al. 2024; Jana Naue et al. 2017; O Vershinina et al. 2021) and 283 newly discovered sites.

[0015] The 425 reported loci are as follows: cg24466241, cg11873854, cg20591472, cg06942701,cg12765028, cg08234504, cg12878812, cg10475970, cg13697378, cg10743390,cg18902090, cg13378934, cg16290275, cg13665684, cg26875805, cg10635122,cg04427498, cg01542719, cg08534628, cg04455430, cg25410668, cg07416237,cg23950157, cg12189835, cg10874644, cg03607117, cg19784428, cg06819923,cg03545227, cg09232937, cg27067781, cg00748589, cg11220950, cg09068158,cg03440125, cg03915012, cg19505546, cg13848598, cg12462816, cg05575054,cg24667575, cg15557036, cg24136205, cg01557798, cg01820962, cg11126134,cg20761322, cg15480367, cg11076306, cg07219494, cg15962547, cg22736354,cg19722847, cg01097384, cg19663246, cg17243289, cg09809672, cg25047092,cg04999352, cg07553761, cg21801378, cg05991454, cg21117668, cg05266781,cg02281167, cg12899747, cg00930873, cg00481951, cg00503840, cg22108374,cg24871743, cg18964582, cg07057579, cg05463027, cg23973429, cg15711521,cg00745389, cg04751276, cg15243034, cg19711579, cg06335143, cg17593472,cg21572722,cg05213896, cg02022733, cg02318971, cg24698655, cg16541931,cg14010720, cg21683284, cg23718736, cg18445088, cg14934413, cg20461912,cg16181396, cg00484358, cg20249566, cg16477091, cg08957484, cg09553242,cg24774812, cg06933824, cg17804348, cg12848065, cg22809047, cg08622677,cg15931839, cg10189695, cg23077820, cg00394316, cg13390332, cg00327072,cg14556683, cg22901840, cg07485775, cg06226630, cg18898125, cg07502389,cg21184711, cg08541518, cg03555227, cg25349154, cg15894389, cg10805721,cg12373771, cg16054275, cg23605843, cg15393490, cg14069287, cg03643998,cg20665157, cg02018902, cg08955995, cg17885226, cg23606718, cg10665321,cg05371578, cg02650266, cg03020208, cg05156137, cg23124451, cg04861640,cg18335796, cg06821999, cg06874016, cg25256723, cg18465694, cg10778288,cg19761273, cg13210467, cg11649376, cg08885800, cg05308819, cg20426994,cg18064714, cg00419702, cg05207048, cg03350900, cg22156456, cg20773033,cg06448705, cg09221867, cg06639320, cg22943590, cg01282174, cg26578682,cg04528819,cg27367526, cg22282410, cg14614094, cg02711397, cg10242160,cg01844642, cg15485728, cg03054277, cg07803108, cg24853724, cg23186333,cg12052661, cg09099868, cg21992250, cg09578568, cg14692377, cg00602811,cg01278873, cg18343474, cg26543112, cg20119148, cg08993878, cg21186299,cg26921969, cg12623930, cg06707978, cg01606885, cg02352240, cg01763090,cg00303541, cg13347071, cg14676592, cg00145253, cg02741882, cg00753885,cg17861230, cg16219603, cg12079303, cg09082287, cg03957108, cg01770755,cg02924487, cg18338984, cg10151367, cg22947000, cg07894352, cg23510764,cg18667659, cg00573770, cg21159778, cg23174607, cg14176752, cg03473532,cg22796704, cg18240400, cg16932827, cg09644065, cg22930275, cg01141812,cg24430580, cg13385220, cg18826637, cg07547549, cg07920503, cg23078123,cg21567504, cg19671120, cg00834536, cg20515136, cg19381811, cg04875128,cg02383785, cg05404236, cg12451153, cg11970349, cg26161329, cg00590036,cg17783401, cg15957394, cg01153166, cg06419432, cg12948621, cg09278098,cg05700079,cg05675373, cg06493994, cg22793142, cg22682811, cg22303418,cg01302656, cg00439658, cg07850154, cg20222376, cg12939283, cg11847992,cg11826475, cg05106770, cg16008966, cg24212517, cg03671597, cg17729667,cg09118625, cg19702785, cg13001142, cg08097417, cg05175318, cg16686396,cg19392831, cg16969368, cg06526721, cg13649056, cg13806741, cg04105923,cg23538901, cg20153322, cg24891133, cg19991948, cg04581938, cg07082267,cg04453050, cg15936446, cg26343883, cg21393163, cg03384000, cg26005082,cg11071401, cg25439632, cg17326555, cg07963234, cg24433287, cg17592231,cg25533247, cg02154074, cg20359994, cg18603228, cg15726426, cg02900766,cg25453381, cg17110586, cg18633600, cg06737494, cg27401724, cg13929658,cg24150153, cg16419235, cg02286081, cg20329210, cg21878650, cg11741201,cg17621438, cg23488804, cg21632975, cg07211259, cg05740045, cg18933331,cg01196788, cg10906284, cg03530962, cg21911021, cg00448560, cg25584930,cg17168836, cg13654588, cg04503319, cg23500537, cg11807280, cg00741624,cg04104132,cg13672300, cg21899500, cg09143195, cg04090392, cg18987410,cg18267374, cg07101980, cg01692968, cg19230755, cg17241310, cg04084157,cg14674720, cg11705975, cg00059225, cg25478614, cg21709871, cg04819580,cg03431918, cg02721902, cg09692396, cg14134497, cg05024939, cg04836038,cg20219381, cg17955729, cg21570988, cg00006081, cg04398950, cg00329615,cg07080372, cg06372850, cg16193278, cg14829814, cg03032497, cg18236477,cg11693709, cg16705273, cg10523019, cg01243823, cg08904363, cg27015931,cg16867657, cg21296230, cg23480341, cg01528542, cg07408456, cg15736994,cg20751926, cg25148589, cg16313343, cg03725309, cg22197050, cg23534245,cg00048759, cg02928699, cg20149168, cg19848940, cg16717122, cg04684267,cg19344626, cg03467555, cg03975694, cg04419610, cg12757011, cg00740914,cg16688437, cg01908954, cg19451698, cg25693132, cg15481583, cg25427880,cg21657577, cg24436906, cg12534424, cg21867345, cg01560871, cg17183905,cg11084334, cg22358580, cg01995468, cg03399905, cg12934382, cg17321954,cg06570224,cg12145907, cg16762684, cg27320127, cg15804973, cg04604946,cg05301866, 283 newly discovered loci: cg05437692, cg00743094, cg20676788,cg09294739, cg16072688, cg01331772, cg12912690, cg03468072, cg04845871,cg13935689, cg22254983, cg19715771, cg18450254, cg00842595, cg12526298,cg07850604, cg18582010, cg22158769, cg19401340, cg22622440, cg13860281,cg27427318, cg14911690, cg09066326, cg22990198, cg16997203, cg04945753,cg21304158, cg04685228, cg11622162, cg21927946, cg23193759, cg26049998,cg00034101, cg03930153, cg12951726, cg07303143, cg16513459, cg10397932,cg16750801, cg23669043, cg06688910, cg05005256, cg09978533, cg02156296,cg06832008, cg11039241, cg26513192, cg08943494, cg12892303, cg06998507,cg19730691, cg00541544, cg06285575, cg08854834, cg06493386, cg07227769,cg14699734, cg25590826, cg27198696, cg03476007, cg27274542, cg18736141,cg22719623, cg11932158, cg08271909, cg03767531, cg25940946, cg25753024,cg11617964, cg03760191, cg06037631, cg14901671, cg02571874, cg10002915,cg12943155, cg06870470, cg24611351, cg07164639, cg04865726, cg25637829,cg13100137, cg12253200,cg03065175, cg08972190, cg01899130, cg24020215,cg01055562, cg00689213, cg08511485, cg03301331, cg23461714, cg03131366,cg23602058, cg25826226, cg22900415, cg10247798, cg12024906, cg16154155,cg20068209, cg02481642, cg14284824, cg12659626, cg13488078, cg00522276,cg19996355, cg13362112, cg22807378, cg10203610, cg11495544, cg24580071,cg11604426, cg00570618, cg06872257, cg00343092, cg21328810, cg07971188,cg26052635, cg05221455, cg16069986, cg20592391, cg00412842, cg05331060,cg02447229, cg25090514, cg26700215, cg19514721, cg19065177, cg06196801,cg16300030, cg18623216, cg06073351, cg19136673, cg01278041, cg01514490,cg08398371, cg21223094, cg19449917, cg00491255, cg10012394, cg19022697,cg18393958, cg12577010, cg21089930, cg19150852, cg21205139, cg13026729,cg21915678, cg19789466, cg19343530, cg20076442, cg00685614, cg09565014,cg18899220, cg02804091, cg12036735, cg16888547, cg03087372, cg01344797,cg12713583, cg01894038, cg06580318, cg08822715, cg02375320, cg05599106,cg02691192, cg26530319,cg07518532, cg23907053, cg24338780, cg01580888,cg04742719, cg06231995, cg25909396, cg09285095, cg02396676, cg23008153,cg24675344, cg08939418, cg05917988, cg00655552, cg13251842, cg15519474,cg12111714, cg03433048, cg13029847, cg01447660, cg05694021, cg01890712,cg18038207, cg11176990, cg01585860, cg09595912, cg15967278, cg10905180,cg04880546, cg17945323, cg01511232, cg02047547, cg15205547, cg24217948,cg10860819, cg16330755, cg09432234, cg09320690, cg25974380, cg02017450,cg07810039, cg07372034, cg23595222, cg25150953, cg20600850, cg09310092,cg18448426, cg05302531, cg00804078, cg06445343, cg24170040, cg24536146,cg07813031, cg22128645, cg00254470, cg10625705, cg03390636, cg13221767,cg26229990, cg04340651, cg26158023, cg10977734, cg21767759, cg20012607,cg12096762, cg17696044, cg22243000, cg20404336, cg17400113, cg21523251,cg15016701, cg18008766, cg00593462, cg03607951, cg02966332, cg06458239,cg14956327, cg02560388, cg27665659, cg07430472, cg18778196, cg01411823,cg18853490, cg07685728,cg02699829, cg13719901, cg18565342, cg25402787,cg05478824, cg19837174, cg22236626, cg14627089, cg12546181, cg26833936,cg17671280, cg01207684, cg06260709, cg23062810, cg06413398, cg22598841,cg13165009, cg09432590, cg10006442, cg02903680, cg06648759, cg05714552,cg11787522, cg13485366, cg18434712, cg26350754, cg22162281, cg27295118,cg04596060, cg12967723, cg07810190, cg15743533, The p values ​​of all age-CpGs are shown in Table 1 , and the results (Experimental Data - Table 1 ) show that all reached the significance level.

[0016] The age-methylation trajectory is shown in the figure Figure 1 The results show that it can be intuitively displayed that it is related to age.

[0017] Age prediction was validated using real-world healthy human DNA methylation sequencing data (3,625 samples, derived from public datasets from GEO and ArrayExpress). The results showed that 708 age-CpGs can serve as DNA methylation site markers for human age prediction and are suitable for machine learning-based age prediction. Both the elastic net regression model and the neural network model constructed based on the 708 sites achieved excellent performance.

[0018] In the training and prediction of the elastic net regression model, the 3625 samples were divided into a training set and a test set at a ratio of 9:1. The hyperparameters max_iter=100000, alpha=0.001, and l1_ratio=0.01 were set for model training. The mean square error (MSE) of the elastic net regression model based on the complete 708 age-CpGs in the test set was 0.0343, and R 2 The predicted and target values ​​for the model performance index calculations are based on Horvath-transformed age, not actual age, the same below. Compared with the elastic net regression model based on the reported 425 loci (MSE 0.0516, R2 =0.9338), expanding the 283 loci newly proposed in this invention to reduce the error of the model and enhance the performance of the model.

[0019] In the training and prediction of the neural network model, the 3625 samples were divided into a training set and a test set at a ratio of 9:1. The training set was used for model training using a 10-fold cross validation. The neural network architecture is as follows: Figure 2 The MAE of the neural network model based on the complete 708 age-CpGs in the test set is 2.187, R 2 Compared with the neural network model based on the reported 425 loci (MAE 2.495, R 2 0.956), expanding the 283 new sites proposed in this invention to reduce the error of the model and enhance the performance of the model. Figure 3 and Figure 4 shown.

[0020] In summary, this paper discloses a set of 708 human age-related DNA methylation site markers, compatible with sequencing data from both new and existing Illumina platforms. This set enables rapid and accurate prediction of human age. 283 of these sites are newly discovered markers, effectively enhancing prediction accuracy.

[0021] The above embodiments are merely preferred embodiments for the purpose of fully illustrating the present invention, and the scope of protection of the present invention is not limited thereto. Equivalent substitutions or modifications made by those skilled in the art based on the present invention are within the scope of protection of the present invention. The scope of protection of the present invention shall be subject to the claims.

Claims

1. Application of DNA methylation site markers to improve the age prediction performance of machine learning models, characterized by: The DNA methylation site markers are 283 sites: cg05437692, cg00743094, cg20676788,cg09294739, cg16072688, cg01331772, cg12912690, cg03468072, cg04845871,cg13935689, cg22254983, cg19715771, cg18450254, cg00842595, cg12526298,cg07850604, cg18582010, cg22158769, cg19401340, cg22622440, cg13860281,cg27427318, cg14911690, cg09066326, cg22990198, cg16997203, cg04945753,cg21304158, cg04685228, cg11622162, cg21927946, cg23193759, cg26049998,cg00034101, cg03930153, cg12951726, cg07303143, cg16513459, cg10397932,cg16750801, cg23669043, cg06688910, cg05005256, cg09978533, cg02156296,cg06832008, cg11039241, cg26513192, cg08943494, cg12892303, cg06998507,cg19730691, cg00541544, cg06285575, cg08854834, cg06493386, cg07227769,cg14699734, cg25590826, cg27198696, cg03476007, cg27274542, cg18736141,cg22719623, cg11932158, cg08271909, cg03767531, cg25940946, cg25753024,cg11617964, cg03760191, cg06037631, cg14901671, cg02571874, cg10002915,cg12943155, cg06870470, cg24611351, cg07164639, cg04865726, cg25637829,cg13100137,cg12253200, cg03065175, cg08972190, cg01899130, cg24020215,cg01055562, cg00689213, cg08511485, cg03301331, cg23461714, cg03131366,cg23602058, cg25826226, cg22900415, cg10247798, cg12024906, cg16154155,cg20068209, cg02481642, cg14284824, cg12659626, cg13488078, cg00522276,cg19996355, cg13362112, cg22807378, cg10203610, cg11495544, cg24580071,cg11604426, cg00570618, cg06872257, cg00343092, cg21328810, cg07971188,cg26052635, cg05221455, cg16069986, cg20592391, cg00412842, cg05331060,cg02447229, cg25090514, cg26700215, cg19514721, cg19065177, cg06196801,cg16300030, cg18623216, cg06073351, cg19136673, cg01278041, cg01514490,cg08398371, cg21223094, cg19449917, cg00491255, cg10012394, cg19022697,cg18393958, cg12577010, cg21089930, cg19150852, cg21205139, cg13026729,cg21915678, cg19789466, cg19343530, cg20076442, cg00685614, cg09565014,cg18899220, cg02804091, cg12036735, cg16888547, cg03087372, cg01344797,cg12713583, cg01894038, cg06580318, cg08822715, cg02375320, cg05599106,cg02691192,cg26530319, cg07518532, cg23907053, cg24338780, cg01580888,cg04742719, cg06231995, cg25909396, cg09285095, cg02396676, cg23008153,cg24675344, cg08939418, cg05917988, cg00655552, cg13251842, cg15519474,cg12111714, cg03433048, cg13029847, cg01447660, cg05694021, cg01890712,cg18038207, cg11176990, cg01585860, cg09595912, cg15967278, cg10905180,cg04880546, cg17945323, cg01511232, cg02047547, cg15205547, cg24217948,cg10860819, cg16330755, cg09432234, cg09320690, cg25974380, cg02017450,cg07810039, cg07372034, cg23595222, cg25150953, cg20600850, cg09310092,cg18448426, cg05302531, cg00804078, cg06445343, cg24170040, cg24536146,cg07813031, cg22128645, cg00254470, cg10625705, cg03390636, cg13221767,cg26229990, cg04340651, cg26158023, cg10977734, cg21767759, cg20012607,cg12096762, cg17696044, cg22243000, cg20404336, cg17400113, cg21523251,cg15016701, cg18008766, cg00593462, cg03607951, cg02966332, cg06458239,cg14956327, cg02560388, cg27665659, cg07430472, cg18778196, cg01411823,cg18853490,cg07685728, cg02699829, cg13719901, cg18565342, cg25402787,cg05478824, cg19837174, cg22236626, cg14627089, cg12546181, cg26833936,cg17671280, cg01207684, cg06260709, cg23062810, cg06413398, cg22598841,cg13165009, cg09432590, cg10006442, cg02903680, cg06648759, cg05714552,cg11787522, cg13485366, cg18434712, cg26350754, cg22162281, cg27295118,cg04596060, cg12967723, cg07810190, cg15743533。, 2. The use of the DNA methylation site marker according to claim 1 in improving the age prediction performance of a machine learning model, characterized in that: The machine learning model is an elastic net regression model or a neural network model.

3. A method for establishing a machine learning model for predicting age performance, characterized by: A set of 708 human age-related DNA methylation site markers was used, and then DNA methylation sequencing data was used. The dataset was divided into training and test sets at a ratio of 9:

1. The elastic net regression model was selected and set with hyperparameters max_iter=100000, alpha=0.001, and l1_ratio=0.01 for model training. The 708 sites were: cg24466241, cg11873854, cg20591472,cg06942701, cg12765028, cg08234504, cg12878812, cg10475970, cg13697378,cg10743390, cg18902090, cg13378934, cg16290275, cg13665684, cg26875805,cg10635122, cg04427498, cg01542719, cg08534628, cg04455430, cg25410668,cg07416237, cg23950157, cg12189835, cg10874644, cg03607117, cg19784428,cg06819923, cg03545227, cg09232937, cg27067781, cg00748589, cg11220950,cg09068158, cg03440125, cg03915012, cg19505546, cg13848598, cg12462816,cg05575054, cg24667575, cg15557036, cg24136205, cg01557798, cg01820962,cg11126134, cg20761322, cg15480367, cg11076306, cg07219494, cg15962547,cg22736354, cg19722847, cg01097384, cg19663246, cg17243289, cg09809672,cg25047092, cg04999352, cg07553761, cg21801378, cg05991454, cg21117668,cg05266781, cg02281167, cg12899747, cg00930873, cg00481951, cg00503840,cg22108374, cg24871743, cg18964582, cg07057579,cg05463027, cg23973429,cg15711521, cg00745389, cg04751276, cg15243034, cg19711579, cg06335143,cg17593472, cg21572722, cg05213896, cg02022733, cg02318971, cg24698655,cg16541931, cg14010720, cg21683284, cg23718736, cg18445088, cg14934413,cg20461912, cg16181396, cg00484358, cg20249566, cg16477091, cg08957484,cg09553242, cg24774812, cg06933824, cg17804348, cg12848065, cg22809047,cg08622677, cg15931839, cg10189695, cg23077820, cg00394316, cg13390332,cg00327072, cg14556683, cg22901840, cg07485775, cg06226630, cg18898125,cg07502389, cg21184711, cg08541518, cg03555227, cg25349154, cg15894389,cg10805721, cg12373771, cg16054275, cg23605843, cg15393490, cg14069287,cg03643998, cg20665157, cg02018902, cg08955995, cg17885226, cg23606718,cg10665321, cg05371578, cg02650266, cg03020208, cg05156137, cg23124451,cg04861640, cg18335796, cg06821999, cg06874016, cg25256723, cg18465694,cg10778288, cg19761273, cg13210467, cg11649376, cg08885800, cg05308819,cg20426994, cg18064714, cg00419702, cg05207048,cg03350900, cg22156456,cg20773033, cg06448705, cg09221867, cg06639320, cg22943590, cg01282174,cg26578682, cg04528819, cg27367526, cg22282410, cg14614094, cg02711397,cg10242160, cg01844642, cg15485728, cg03054277, cg07803108, cg24853724,cg23186333, cg12052661, cg09099868, cg21992250, cg09578568, cg14692377,cg00602811, cg01278873, cg18343474, cg26543112, cg20119148, cg08993878,cg21186299, cg26921969, cg12623930, cg06707978, cg01606885, cg02352240,cg01763090, cg00303541, cg13347071, cg14676592, cg00145253, cg02741882,cg00753885, cg17861230, cg16219603, cg12079303, cg09082287, cg03957108,cg01770755, cg02924487, cg18338984, cg10151367, cg22947000, cg07894352,cg23510764, cg18667659, cg00573770, cg21159778, cg23174607, cg14176752,cg03473532, cg22796704, cg18240400, cg16932827, cg09644065, cg22930275,cg01141812, cg24430580, cg13385220, cg18826637, cg07547549, cg07920503,cg23078123, cg21567504, cg19671120, cg00834536, cg20515136, cg19381811,cg04875128, cg02383785, cg05404236, cg12451153,cg11970349, cg26161329,cg00590036, cg17783401, cg15957394, cg01153166, cg06419432, cg12948621,cg09278098, cg05700079, cg05675373, cg06493994, cg22793142, cg22682811,cg22303418, cg01302656, cg00439658, cg07850154, cg20222376, cg12939283,cg11847992, cg11826475, cg05106770, cg16008966, cg24212517, cg03671597,cg17729667, cg09118625, cg19702785, cg13001142, cg08097417, cg05175318,cg16686396, cg19392831, cg16969368, cg06526721, cg13649056, cg13806741,cg04105923, cg23538901, cg20153322, cg24891133, cg19991948, cg04581938,cg07082267, cg04453050, cg15936446, cg26343883, cg21393163, cg03384000,cg26005082, cg11071401, cg25439632, cg17326555, cg07963234, cg24433287,cg17592231, cg25533247, cg02154074, cg20359994, cg18603228, cg15726426,cg02900766, cg25453381, cg17110586, cg18633600, cg06737494, cg27401724,cg13929658, cg24150153, cg16419235, cg02286081, cg20329210, cg21878650,cg11741201, cg17621438, cg23488804, cg21632975, cg07211259, cg05740045,cg18933331, cg01196788, cg10906284, cg03530962,cg21911021, cg00448560,cg25584930, cg17168836, cg13654588, cg04503319, cg23500537, cg11807280,cg00741624, cg04104132, cg13672300, cg21899500, cg09143195, cg04090392,cg18987410, cg18267374, cg07101980, cg01692968, cg19230755, cg17241310,cg04084157, cg14674720, cg11705975, cg00059225, cg25478614, cg21709871,cg04819580, cg03431918, cg02721902, cg09692396, cg14134497, cg05024939,cg04836038, cg20219381, cg17955729, cg21570988, cg00006081, cg04398950,cg00329615, cg07080372, cg06372850, cg16193278, cg14829814, cg03032497,cg18236477, cg11693709, cg16705273, cg10523019, cg01243823, cg08904363,cg27015931, cg16867657, cg21296230, cg23480341, cg01528542, cg07408456,cg15736994, cg20751926, cg25148589, cg16313343, cg03725309, cg22197050,cg23534245, cg00048759, cg02928699, cg20149168, cg19848940, cg16717122,cg04684267, cg19344626, cg03467555, cg03975694, cg04419610, cg12757011,cg00740914, cg16688437, cg01908954, cg19451698, cg25693132, cg15481583,cg25427880, cg21657577, cg24436906, cg12534424,cg21867345, cg01560871,cg17183905, cg11084334, cg22358580, cg01995468, cg03399905, cg12934382,cg17321954, cg06570224, cg12145907, cg16762684, cg27320127, cg15804973,cg04604946, cg05301866,cg05437692, cg00743094, cg20676788, cg09294739,cg16072688, cg01331772, cg12912690, cg03468072, cg04845871, cg13935689,cg22254983, cg19715771, cg18450254, cg00842595, cg12526298, cg07850604,cg18582010, cg22158769, cg19401340, cg22622440, cg13860281, cg27427318,cg14911690, cg09066326, cg22990198, cg16997203, cg04945753, cg21304158,cg04685228, cg11622162, cg21927946, cg23193759, cg26049998, cg00034101,cg03930153, cg12951726, cg07303143, cg16513459, cg10397932, cg16750801,cg23669043, cg06688910, cg05005256, cg09978533, cg02156296, cg06832008,cg11039241, cg26513192, cg08943494, cg12892303, cg06998507, cg19730691,cg00541544, cg06285575, cg08854834, cg06493386, cg07227769, cg14699734,cg25590826, cg27198696, cg03476007, cg27274542, cg18736141, cg22719623,cg11932158, cg08271909, cg03767531, cg25940946,cg25753024, cg11617964,cg03760191, cg06037631, cg14901671, cg02571874, cg10002915, cg12943155,cg06870470, cg24611351, cg07164639, cg04865726, cg25637829, cg13100137,cg12253200, cg03065175, cg08972190, cg01899130, cg24020215, cg01055562,cg00689213, cg08511485, cg03301331, cg23461714, cg03131366, cg23602058,cg25826226, cg22900415, cg10247798, cg12024906, cg16154155, cg20068209,cg02481642, cg14284824, cg12659626, cg13488078, cg00522276, cg19996355,cg13362112, cg22807378, cg10203610, cg11495544, cg24580071, cg11604426,cg00570618, cg06872257, cg00343092, cg21328810, cg07971188, cg26052635,cg05221455, cg16069986, cg20592391, cg00412842, cg05331060, cg02447229,cg25090514, cg26700215, cg19514721, cg19065177, cg06196801, cg16300030,cg18623216, cg06073351, cg19136673, cg01278041, cg01514490, cg08398371,cg21223094, cg19449917, cg00491255, cg10012394, cg19022697, cg18393958,cg12577010, cg21089930, cg19150852, cg21205139, cg13026729, cg21915678,cg19789466, cg19343530, cg20076442, cg00685614,cg09565014, cg18899220,cg02804091, cg12036735, cg16888547, cg03087372, cg01344797, cg12713583,cg01894038, cg06580318, cg08822715, cg02375320, cg05599106, cg02691192,cg26530319, cg07518532, cg23907053, cg24338780, cg01580888, cg04742719,cg06231995, cg25909396, cg09285095, cg02396676, cg23008153, cg24675344,cg08939418, cg05917988, cg00655552, cg13251842, cg15519474, cg12111714,cg03433048, cg13029847, cg01447660, cg05694021, cg01890712, cg18038207,cg11176990, cg01585860, cg09595912, cg15967278, cg10905180, cg04880546,cg17945323, cg01511232, cg02047547, cg15205547, cg24217948, cg10860819,cg16330755, cg09432234, cg09320690, cg25974380, cg02017450, cg07810039,cg07372034, cg23595222, cg25150953, cg20600850, cg09310092, cg18448426,cg05302531, cg00804078, cg06445343, cg24170040, cg24536146, cg07813031,cg22128645, cg00254470, cg10625705, cg03390636, cg13221767, cg26229990,cg04340651, cg26158023, cg10977734, cg21767759, cg20012607, cg12096762,cg17696044, cg22243000, cg20404336, cg17400113,cg21523251, cg15016701,cg18008766, cg00593462, cg03607951, cg02966332, cg06458239, cg14956327,cg02560388, cg27665659, cg07430472, cg18778196, cg01411823, cg18853490,cg07685728, cg02699829, cg13719901, cg18565342, cg25402787, cg05478824,cg19837174, cg22236626, cg14627089, cg12546181, cg26833936, cg17671280,cg01207684, cg06260709, cg23062810, cg06413398, cg22598841, cg13165009,cg09432590, cg10006442, cg02903680, cg06648759, cg05714552, cg11787522,cg13485366, cg18434712, cg26350754, cg22162281, cg27295118, cg04596060,cg12967723, cg07810190, cg15743533。, 4. A method for establishing a machine learning model for predicting age performance, characterized by: A set of 708 human age-related DNA methylation site markers was used, and then DNA methylation sequencing data was used to divide the dataset into training and test sets in a 9:1 ratio. The training-validation set was used for model training using ten-fold cross-validation.

Images