A process for identifying health outcomes through epigenetic prediction

Abstract

A method and system for determining a time-to-event of a health outcome in a patient. Embodiments herein are used to predict scores and ratios that are useful in determining the time-to-event. In some aspects, a treatment response is generated based on the methods and systems.

Description

A PROCESS FOR IDENTIFYING HEALTH OUTCOMES THROUGH EPIGENETIC PREDICTION

[0001] The present application claims priority to U.S. provisional application serial number 63 / 656,986 filed June 6, 2024, the entire contents of which are incorporated by reference and relied upon. STATEMENT OF GOVERNMENT SUPPORT

[0002] This invention was made with government support under MD013312 awarded by the National Institutes of Health. The government has certain rights in the invention. FIELD

[0003] The present disclosure generally relates to methods and systems for predicting patient health outcomes using machine learning models. BACKGROUND

[0004] Previous work from Thompson et al (npj Genomic Med., vol.7, no.50, pp.1–11, Aug. 2022, which is incorporated herein in its entirety by reference) has extensively demonstrated the utility of methylation risk scores (MRS), which are linear combinations of CpG methylation states, for imputation of clinically interesting phenotypes, including labs, medications and diagnosis codes, prior to the methylation sample collection time. It shows significant improvements over the current state-of-the-art polygenetic risk scores (PRS) and the baseline model, which only includes demographic information, in a broad spectrum of phenotypes investigated. MRS also achieves better reproducibility when applied to out-of- sample validation datasets and is less susceptible to severe confounding effects caused by population structure, a common issue that hinders the application and versatility of PRS.

[0005] With community health initiatives routinely collecting patient EHR data years after their original blood collection time, there has been an accumulation of a trove of data that offers a unique opportunity to check whether methylation patterns can be used for future risk prediction on prospective clinically interesting phenotypes. Specifically, there remain two important questions: (1) does the methylation value tell us the relative risk of certain individuals getting a specific diagnosis in the future compared to the background general population, and (2) what is the probability of certain individuals getting a specific diagnosis within a certain time period (e.g., one year) after the methylation blood collection?298325604.1 - 1 -

[0006] Various embodiments of the present disclosure address one or more of the above questions and / or one or more of the above described shortcomings. BRIEF SUMMARY

[0007] In one or more embodiments, a method is provided for obtaining information useful to determine a time-to-event to a health outcome for a patient. In one or more embodiments, the method comprises receiving, at a processor, epigenetic data of the patient. In one or more embodiments, the method comprises analyzing the received first set of epigenetic data via the processor, using a trained machine learning model, to generate a first relative risk score based on the methylation fraction of specific epigenetic markers. In one or more embodiments, the method comprises analyzing the received first set of epigenetic data via the processor, using a trained machine learning model, to generate a first hazard curve based on the methylation fraction of specific epigenetic markers. In one or more embodiments, the method comprises analyzing the received epigenetic data via the processor, using a trained machine learning model, to predict a relative risk score and / or a hazard curve of the time-to-event to the health outcome based on the methylation fraction of specific epigenetic markers.

[0008] In one or more embodiments, a method is provided for treating a patient with a therapeutic intervention. In one or more embodiments, the method comprises receiving, at a processor, a first set epigenetic of data from the patient collected at a time before the patient has been given an administration of the therapeutic intervention. In one or more embodiments, the method comprises analyzing the received first set of epigenetic data via the processor, using a trained machine learning model, to generate a first relative risk score based on the methylation fraction of specific epigenetic markers. In one or more embodiments, the method comprises analyzing the received first set of epigenetic data via the processor, using a trained machine learning model, to generate a first hazard curve based on the methylation fraction of specific epigenetic markers. In one or more embodiments, the method comprises analyzing the received first set of epigenetic data via the processor, using a trained machine learning model, to generate a first relative risk score and / or a first hazard curve based on the methylation fraction of specific epigenetic markers. In one or more embodiments, the method comprises receiving, at a processor, a second set of epigenetic data from the patient collected at a time after the patient has been given the administration of the therapeutic intervention. In one or more embodiments, the method comprises analyzing the received second set of epigenetic data via the processor, using the trained machine learning model, to generate a second relative risk score and / or a second hazard curve based on the methylation fraction of specific epigenetic markers.298325604.1 - 2 -In one or more embodiments, the method comprises comparing the first and second relative risk scores and / or first and second hazard curves to generate a comparison. In one or more embodiments, the method comprises administering an effective amount of the therapeutic intervention. The effective amount may be determined based on the comparison.

[0009] In one or more embodiments, the trained machine learning model is trained by tuning training data including a time before obtaining the epigenetic data from a training set patient in which the training set patient did not have the health outcome. In one or more embodiments, the trained machine learning model is trained by tuning training data including a time after obtaining the epigenetic data from the training set patient in which the training set patient was diagnosed with the health outcome. In one or more embodiments, the trained machine learning model is trained by tuning training data including a time before obtaining the epigenetic data from a training set patient in which the training set patient did not have the health outcome and / or a time after obtaining the epigenetic data from the training set patient in which the training set patient was diagnosed with the health outcome.

[0010] In one or more embodiments, the health outcome is developing osteoporosis. In one or more embodiments, the health outcome is developing sepsis. In one or more embodiments, the health outcome is a fluid disorder, which may be an electrolyte disorder. In one or more embodiments, the health outcome is an iron deficiency anemia.

[0011] In one or more embodiments, including when the health outcome is developing osteoporosis, the specific epigenetic markers comprise or consist of any or the top 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110, 111, 112, 113, 114, 115, 116, 117, 118, 119, 120, 121, 122, 123, 124, 125, 126, 127, 128, 129, 130, 131, 132, 133, 134, 135, 136, 137, 138, 139, 140, 141, 142, 143, 144, 145, 146, 147, 148, 149, 150, 151, 152, 153, 154, 155, 156, 157, 158, 159, 160, 161, 162, 163, 164, 165, 166, 167, 168, 169, 170, 171, 172, 173, 174, 175, 176, 177, 178, 179, 180, 181, 182, 183, 184, 185, 186, 187, 188, 189, 190, 191, 192, 193, 194, 195, 196, 197, 198, 199, 200, 201, 202, 203, 204, 205, 206, 207, 208, 209, 210, 211, 212, 213, 214, 215, 216, 217, 218, 219, 220, 221, 222, 223, 224, 225, 226, 227, 228, 229, 230, 231, 232, 233, 234, 235, 236, 237, 238, 239, 240, 241, 242, 243, 244, 245, 246, 247, 248, 249, 250, or any range derivable therein, CpG location designations identified in Table 1. In one or more embodiments, the epigenetic markers listed in Table 1 are in ranked order for the impact on the relative risk score and / or298325604.1 - 3 -hazard curve for osteoporosis. In one or more embodiments, the specific epigenetic markers comprise or consist of the CpG location designations, cg11047325 and cg18022864. In one or more embodiments, wherein the health outcome is developing osteoporosis, the specific epigenetic markers comprise or consist of the CpG location designations, cg11047325 and cg18022864.

[0012] In one or more embodiments, including when the health outcome is developing sepsis, the specific epigenetic markers comprise or consist of the CpG location designations, cg20494891 and cg23608915. In one or more embodiments, including when the health outcome is developing sepsis, the specific epigenetic markers comprise or consist of the CpG location designations, cg20494891. In one or more embodiments, including when the health outcome is developing osteoporosis, the specific epigenetic markers comprise or consist of any or the top 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, or any range derivable therein, CpG location designations identified in Table 2. In one or more embodiments, including when the health outcome is developing osteoporosis, the specific epigenetic markers comprise or consist of any or the top 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, or any range derivable therein, CpG location designations identified in Table 3. In one or more embodiments, including when the health outcome is developing osteoporosis, the specific epigenetic markers comprise or consist of any or the top 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, or any range derivable therein, CpG location designations identified in Table 4. In one or more embodiments, including when the health outcome is developing osteoporosis, the specific epigenetic markers comprise or consist of any or the top 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or any range derivable therein, CpG location designations identified in Table 5. In one or more embodiments, the epigenetic markers listed in any of Tables 2-5 are in ranked order for the impact on the relative risk score and / or hazard curve for sepsis.

[0013] In one or more embodiments, including when the health outcome is a fluid disorder, the specific epigenetic markers comprise or consist of any or the top 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, or any range derivable therein, CpG location designations identified in Table 6. In one or more embodiments, including when the health outcome is a fluid disorder, the specific epigenetic markers comprise or consist of any or the top 1, 2, 3, 4, 5, 6, 7, or any range derivable therein, CpG location designations identified in Table 7. In one or more embodiments, the epigenetic markers listed in either Table 6 or Table 7 are in ranked order for the impact on the relative risk score and / or hazard curve for fluid disorder.298325604.1 - 4 -

[0014] In one or more embodiments, including when the health outcome is a fluid disorder, the specific epigenetic markers comprise or consist of any or the top 1, 2, 3, 4, or any range derivable therein, CpG location designations identified in Table 8. In one or more embodiments, including when the health outcome is a fluid disorder, the specific epigenetic markers comprise or consist of any or the top 1, 2, 3, or any range derivable therein, CpG location designations identified in Table 9. In one or more embodiments, the epigenetic markers listed in either Table 8 or Table 9 are in ranked order for the impact on the relative risk score and / or hazard curve for iron deficiency anemia.

[0015] In one or more embodiments, the time before obtaining the epigenetic data from a training set patient in which the training set patient did not have the health outcome is at least, at most, or exactly 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110, 111, 112, 113, 114, 115, 116, 117, 118, 119, 120, 121, 122, 123, 124, 125, 126, 127, 128, 129, 130, 131, 132, 133, 134, 135, 136, 137, 138, 139, 140, 141, 142, 143, 144, 145, 146, 147, 148, 149, 150, 151, 152, 153, 154, 155, 156, 157, 158, 159, 160, 161, 162, 163, 164, 165, 166, 167, 168, 169, 170, 171, 172, 173, 174, 175, 176, 177, 178, 179, 180, 181, 182, 183, 184, 185, 186, 187, 188, 189, 190, 191, 192, 193, 194, 195, 196, 197, 198, 199, 200, 201, 202, 203, 204, 205, 206, 207, 208, 209, 210, 211, 212, 213, 214, 215, 216, 217, 218, 219, 220, 221, 222, 223, 224, 225, 226, 227, 228, 229, 230, 231, 232, 233, 234, 235, 236, 237, 238, 239, 240, 241, 242, 243, 244, 245, 246, 247, 248, 249, 250, 251, 252, 253, 254, 255, 256, 257, 258, 259, 260, 261, 262, 263, 264, 265, 266, 267, 268, 269, 270, 271, 272, 273, 274, 275, 276, 277, 278, 279, 280, 281, 282, 283, 284, 285, 286, 287, 288, 289, 290, 291, 292, 293, 294, 295, 296, 297, 298, 299, 300, 301, 302, 303, 304, 305, 306, 307, 308, 309, 310, 311, 312, 313, 314, 315, 316, 317, 318, 319, 320, 321, 322, 323, 324, 325, 326, 327, 328, 329, 330, 331, 332, 333, 334, 335, 336, 337, 338, 339, 340, 341, 342, 343, 344, 345, 346, 347, 348, 349, 350, 351, 352, 353, 354, 355, 356, 357, 358, 359, 360, 361, 362, 363, 364, 365 days, or any range derivable therein.

[0016] In one or more embodiments, the method comprises an analyzing step that comprises using a cox survival model. The cox survival model may be a penalized cox survival model. In one or more embodiments, the analyzing step comprises using a logistic regression, which may be a penalized logistic regression. In one or more embodiments, the analyzing step298325604.1 - 5 -comprises using a trained neural network, a trained non-linear survival analysis model, a trained Bayesian time-to-event model, or a trained linear or non-linear time series model.

[0017] In one or more embodiments, the therapeutic intervention comprises one or more of a small molecule, a biologic, a supplement, radiation, phototherapy, heat therapy, cold therapy, surgery, a lifestyle change, psychotherapy, or a combination thereof. The biologic may comprise one or more proteins, carbohydrates, nucleic acids, lipids, cells, vaccines, hormones, and / or metabolites. A lifestyle change may, in some embodiments, include a change in exercise frequency and / or intensity, a diet change, stopping or reducing smoking, stopping or reducing alcohol consumption, or a combination thereof.

[0018] In one or more embodiments, where the health outcome or disease is osteoporosis, the therapeutic intervention may comprise a bisphosphonate, denosumab, teriparatide, abaloparatide, romosozumab, and / or a lifestyle change.

[0019] In one or more embodiments, where the health outcome or disease is sepsis, the therapeutic intervention may comprise an antibiotic and / or intravenous fluids.

[0020] In one or more embodiments, where the health outcome or disease is a fluid disorder, the therapeutic intervention may comprise fluids, electrolyte supplements, and / or a lifestyle change.

[0021] In one or more embodiments, where the health outcome or disease is an iron deficiency anemia, the therapeutic intervention may comprise iron supplements.

[0022] In one or more embodiments, the patient has, or is suspected of having, a disease. The disease may be a cancer, an infection, an immune disease, a lung disease, a brain disease, a heart disease, a skin disease, a kidney disease, a liver disease, a gastrointestinal disease, a bone disease, a blood disease, a ligament disease, a nerve disease, a bladder disease, a pancreas disease, a gallbladder disease, a reproductive-organ disease, and / or a muscular disease.

[0023] In one or more embodiments, a system comprises one or more data processors; and a non-transitory computer readable storage medium containing instructions which, when executed on the one or more data processors, cause the one or more data processors to perform part or all of any one or more of the methods described herein.

[0024] In one or more embodiments, a computer-program product tangibly embodied in a non-transitory machine-readable storage medium is provided that includes instructions configured to cause one or more data processors to perform part or all of any one or more of the methods described herein.

[0025] Also disclosed are one or more of the following enumerated aspects:298325604.1 - 6 -1. A computer-implemented method for determining a time-to-event of a health outcome for a patient comprising: receiving, by one or more processors, an epigenetic dataset obtained from a patient, wherein the epigenetic dataset includes methylation fractions of specific epigenetic markers; and applying, by the one or more processors, a trained machine learning model to the epigenetic dataset to predict a relative risk score and / or a hazard curve of the time-to-event of the health outcome. 2. The method of Aspect 1, further including: training the trained machine learning model with epigenetic training datasets obtained from a plurality of study patients, wherein the epigenetic training datasets include a time value before obtaining epigenetic data from a study patient without the health outcome and / or a time value after obtaining epigenetic data from a study patient with the health outcome. 3. The method of Aspect 1 or 2, wherein the health outcome is developing osteoporosis. 4. The method of any one of Aspects 1-3, wherein the specific epigenetic markers comprise or consist of at least three CpG location designations identified in Table 1. 5. The method of Aspect 4, wherein the specific epigenetic markers comprise or consist of at least 3 of the top CpG location designations identified in Table 1. 6. The method of any one of Aspects 1-5, wherein the epigenetic markers listed in Table 1 are in ranked order for the impact on the relative risk score and / or hazard curve for osteoporosis. 7. The method of any one of Aspects 1-3, wherein the specific epigenetic markers comprise or consist of the CpG location designations, cg11047325 and cg18022864. 8. The method of Aspect 1, wherein the health outcome is developing sepsis. 9. The method of Aspect 1 or 8, wherein the specific epigenetic markers comprise or consist of the CpG location designations, cg20494891 and cg23608915. 10. The method of Aspect 1 or 8, wherein the specific epigenetic markers comprise or consist of the CpG location designation, cg20494891. 11. The method of Aspect 1 or 8, wherein the specific epigenetic markers comprise or consist of at least three CpG location designations identified in Table 2. 12. The method of Aspect 11, wherein the specific epigenetic markers comprise or consist of at least 3 of the top CpG location designations identified in Table 2.298325604.1 - 7 -13. The method of any one of Aspects 11 or 12, wherein the epigenetic markers listed in Table 2 are in ranked order for the impact on the relative risk score and / or hazard curve for sepsis. 14. The method of Aspect 1 or 8, wherein the specific epigenetic markers comprise or consist of at least three CpG location designations identified in Table 3. 15. The method of Aspect 14, wherein the specific epigenetic markers comprise or consist of at least 3 of the top CpG location designations identified in Table 3. 16. The method of any one of Aspects 14 or 15, wherein the epigenetic markers listed in Table 3 are in ranked order for the impact on the relative risk score and / or hazard curve for sepsis. 17. The method of Aspect 1 or 8, wherein the specific epigenetic markers comprise or consist of at least three CpG location designations identified in Table 4. 18. The method of Aspect 17, wherein the specific epigenetic markers comprise or consist of at least 3 of the top CpG location designations identified in Table 4. 19. The method of any one of Aspects 17 or 18, wherein the epigenetic markers listed in Table 4 are in ranked order for the impact on the relative risk score and / or hazard curve for sepsis. 20. The method of Aspect 1 or 8, wherein the specific epigenetic markers comprise or consist of at least three CpG location designations identified in Table 5. 21. The method of Aspect 20, wherein the specific epigenetic markers comprise or consist of at least 3 of the top CpG location designations identified in Table 5. 22. The method of any one of Aspects 20 or 21, wherein the epigenetic markers listed in Table 5 are in ranked order for the impact on the relative risk score and / or hazard curve for sepsis. 23. The method of Aspect 1, wherein the health outcome is a fluid disorder. 24. The method of Aspect 23, wherein the fluid disorder is an electrolyte disorder. 25. The method of Aspect 23 or 24, wherein the specific epigenetic markers comprise or consist of at least three CpG location designations identified in Table 6. 26. The method of Aspect 25, wherein the specific epigenetic markers comprise or consist of at least 3 of the top CpG location designations identified in Table 6. 27. The method of any one of Aspects 25 or 26, wherein the epigenetic markers listed in Table 6 are in ranked order for the impact on the relative risk score and / or hazard curve for the fluid disorder.298325604.1 - 8 -28. The method of Aspect 23 or 24, wherein the specific epigenetic markers comprise or consist of at least three CpG location designations identified in Table 7. 29. The method of Aspect 28, wherein the specific epigenetic markers comprise or consist of at least 3 of the top CpG location designations identified in Table 7. 30. The method of any one of Aspects 28 or 29, wherein the epigenetic markers listed in Table 7 are in ranked order for the impact on the relative risk score and / or hazard curve for the fluid disorder. 31. The method of Aspect 1, wherein the health outcome is an iron deficiency anemia. 32. The method of Aspect 1 or 31, wherein the specific epigenetic markers comprise or consist of at least three CpG location designations identified in Table 8. 33. The method of Aspect 32, wherein the specific epigenetic markers comprise or consist of at least 3 of the top CpG location designations identified in Table 8. 34. The method of any one of Aspects 32 or 33, wherein the epigenetic markers listed in Table 8 are in ranked order for the impact on the relative risk score and / or hazard curve for the iron deficiency anemia. 35. The method of Aspect 1 or 31, wherein the specific epigenetic markers comprise or consist of at least three CpG location designations identified in Table 9. 36. The method of Aspect 35, wherein the specific epigenetic markers comprise or consist of at least 3 of the top CpG location designations identified in Table 9. 37. The method of any one of Aspects 35 or 36, wherein the epigenetic markers listed in Table 9 are in ranked order for the impact on the relative risk score and / or hazard curve for the iron deficiency anemia. 38. The method of any one of Aspects 2-37, wherein the time before obtaining the epigenetic data from a training set patient in which the training set patient did not have the health outcome is at least, at most, or exactly 90 days, 180 days, 365 days, or any range derivable therein. 39. The method of any one of Aspects 2-38, wherein the time after obtaining the epigenetic data from the training set patient in which the training set patient was diagnosed with the health outcome is at least, at most, or exactly 10 days, 14 days, 30 days, 60 days, or any range derivable therein. 40. The method of any one of Aspects 1-39, wherein the analyzing step comprises using a cox survival model. 41. The method of Aspect 40, wherein the cox survival model is a penalized cox survival model.298325604.1 - 9 -42. The method of any one of Aspects 1-41, wherein the analyzing step comprises using a logistic regression. 43. The method of Aspect 42, wherein the logistic regression is a penalized logistic regression. 44. The method of any one of Aspects 1-43, wherein the analyzing step comprises using a trained neural network, a trained non-linear survival analysis model, a trained Bayesian time- to-event model, or a trained linear or non-linear time series model. 45. A method for treating a patient with a therapeutic intervention comprising receiving, at a processor, a first set epigenetic of data from the patient collected at a time before the patient has been given an administration of the therapeutic intervention; analyzing the received first set of epigenetic data via the processor, using a trained machine learning model, to generate a first relative risk score and / or a first hazard curve based on the methylation fraction of specific epigenetic markers; receiving, at a processor, a second set of epigenetic data from the patient collected at a time after the patient has been given the administration of the therapeutic intervention; analyzing the received second set of epigenetic data via the processor, using the trained machine learning model, to generate a second relative risk score and / or a second hazard curve based on the methylation fraction of specific epigenetic markers; comparing the first and second relative risk scores and / or first and second hazard curves to generate a comparison; and administering an effective amount of the therapeutic intervention, wherein the effective amount is determined based on the comparison. 46. The method of Aspect 45, wherein the trained machine learning model is trained by tuning training data including a time before obtaining the epigenetic data from a training set patient in which the training set patient did not have the health outcome and / or a time after obtaining the epigenetic data from the training set patient in which the training set patient was diagnosed with the health outcome. 47. The method of Aspect 45, wherein the therapeutic intervention comprises one or more of a small molecule, a biologic, radiation, phototherapy, heat therapy, cold therapy, surgery, exercise, a diet change, psychotherapy, or a combination thereof.298325604.1 - 10 -48. The method of Aspect 47, wherein the biologic comprises one or more proteins, carbohydrates, nucleic acids, lipids, cells, vaccines, hormones, and / or metabolites. 49. The method of any one of Aspects 45-48, wherein the patient has, or is suspected of having, a disease. 50. The method of Aspect 49 wherein the disease is a cancer, an infection, an immune disease, a lung disease, a brain disease, a heart disease, a skin disease, a kidney disease, a liver disease, a gastrointestinal disease, a bone disease, a blood disease, a ligament disease, a nerve disease, a bladder disease, a pancreas disease, a gallbladder disease, a reproductive- organ disease, and / or a muscular disease. 51. The method of any one of Aspects Aspect 45-50, wherein the health outcome is developing osteoporosis. 52. The method of any one of Aspects 45-51, wherein the specific epigenetic markers comprise or consist of at least three CpG location designations identified in Table 1. 53. The method of Aspect 52, wherein the specific epigenetic markers comprise or consist of at least 3 of the top CpG location designations identified in Table 1. 54. The method of any one of Aspects 52-53, wherein the epigenetic markers listed in Table 1 are in ranked order for the impact on the relative risk score and / or hazard curve for osteoporosis. 55. The method of any one of Aspects 45-51, wherein the specific epigenetic markers comprise or consist of the CpG location designations, cg11047325 and cg18022864. 56. The method of any one of Aspects 51-55, wherein the therapeutic intervention comprises a bisphosphonate, denosumab, teriparatide, abaloparatide, romosozumab, and / or a lifestyle change. 57. The method of any one of Aspects 45-50, wherein the health outcome is developing sepsis. 58. The method of Aspect 45 or 57, wherein the specific epigenetic markers comprise or consist of the CpG location designations, cg20494891 and cg23608915. 59. The method of Aspect 45 or 57, wherein the specific epigenetic markers comprise or consist of the CpG location designation, cg20494891. 60. The method of Aspect 45 or 57, wherein the specific epigenetic markers comprise or consist of at least three CpG location designations identified in Table 2. 61. The method of Aspect 60, wherein the specific epigenetic markers comprise or consist of at least 3 of the top CpG location designations identified in Table 2.298325604.1 - 11 -62. The method of any one of Aspects 60 or 61, wherein the epigenetic markers listed in Table 2 are in ranked order for the impact on the relative risk score and / or hazard curve for sepsis. 63. The method of Aspect 45 or 57, wherein the specific epigenetic markers comprise or consist of at least three CpG location designations identified in Table 3. 64. The method of Aspect 63, wherein the specific epigenetic markers comprise or consist of at least 3 of the top CpG location designations identified in Table 3. 65. The method of any one of Aspects 63 or 64, wherein the epigenetic markers listed in Table 3 are in ranked order for the impact on the relative risk score and / or hazard curve for sepsis. 66. The method of Aspect 45 or 57, wherein the specific epigenetic markers comprise or consist of at least three CpG location designations identified in Table 4. 67. The method of Aspect 66, wherein the specific epigenetic markers comprise or consist of at least 3 of the top CpG location designations identified in Table 4. 68. The method of any one of Aspects 66 or 67, wherein the epigenetic markers listed in Table 4 are in ranked order for the impact on the relative risk score and / or hazard curve for sepsis. 69. The method of Aspect 45 or 57, wherein the specific epigenetic markers comprise or consist of at least three CpG location designations identified in Table 5. 70. The method of Aspect 69, wherein the specific epigenetic markers comprise or consist of at least 3 of the top CpG location designations identified in Table 5. 71. The method of any one of Aspects 69 or 70, wherein the epigenetic markers listed in Table 5 are in ranked order for the impact on the relative risk score and / or hazard curve for sepsis. 72. The method of any one of Aspects 57-71, wherein the therapeutic intervention comprises an antibiotic and / or intravenous fluids. 73. The method of Aspect 72, wherein the antibiotic comprises a broad-spectrum antibiotic. 74. The method of Aspect 45, wherein the health outcome is a fluid disorder. 75. The method of Aspect 74, wherein the fluid disorder is an electrolyte disorder. 76. The method of Aspect 74 or 75, wherein the specific epigenetic markers comprise or consist of at least three CpG location designations identified in Table 6. 77. The method of Aspect 76, wherein the specific epigenetic markers comprise or consist of at least 3 of the top CpG location designations identified in Table 6.298325604.1 - 12 -78. The method of any one of Aspects 76 or 77, wherein the epigenetic markers listed in Table 6 are in ranked order for the impact on the relative risk score and / or hazard curve for the fluid disorder. 79. The method of Aspect 74 or 75, wherein the specific epigenetic markers comprise or consist of at least three CpG location designations identified in Table 7. 80. The method of Aspect 79, wherein the specific epigenetic markers comprise or consist of at least 3 of the top CpG location designations identified in Table 7. 81. The method of any one of Aspects 79 or 80, wherein the epigenetic markers listed in Table 7 are in ranked order for the impact on the relative risk score and / or hazard curve for the fluid disorder. 82. The method of any one of Aspects 74-81, wherein the therapeutic intervention comprises fluids, electrolyte supplements, and / or a lifestyle change. 83. The method of Aspect 45, wherein the health outcome is an iron deficiency anemia. 84. The method of Aspect 45 or 83, wherein the specific epigenetic markers comprise or consist of at least three CpG location designations identified in Table 8. 85. The method of Aspect 84, wherein the specific epigenetic markers comprise or consist of at least 3 of the top CpG location designations identified in Table 8. 86. The method of any one of Aspects 84 or 85, wherein the epigenetic markers listed in Table 8 are in ranked order for the impact on the relative risk score and / or hazard curve for the iron deficiency anemia. 87. The method of Aspect 45 or 83, wherein the specific epigenetic markers comprise or consist of at least three CpG location designations identified in Table 9. 88. The method of Aspect 87, wherein the specific epigenetic markers comprise or consist of at least 3 of the top CpG location designations identified in Table 9. 89. The method of any one of Aspects 87 or 88, wherein the epigenetic markers listed in Table 9 are in ranked order for the impact on the relative risk score and / or hazard curve for the iron deficiency anemia. 90. The method of any one of Aspects 83-89, wherein the therapeutic intervention comprises iron supplements. 91. The method of any one of Aspects 46-90, wherein the time before obtaining the epigenetic data from a training set patient in which the training set patient did not have the health outcome is at least, at most, or exactly 90 days, 180 days, 365 days, or any range derivable therein.298325604.1 - 13 -92. The method of any one of Aspects 46-91, wherein the time after obtaining the epigenetic data from the training set patient in which the training set patient was diagnosed with the health outcome is at least, at most, or exactly 10 days, 14 days, 30 days, 60 days, or any range derivable therein. 93. The method of any one of Aspects 45-92, wherein the analyzing step comprises using a cox survival model. 94. The method of Aspect 93, wherein the cox survival model is a penalized cox survival model. 95. The method of any one of Aspects 45-94, wherein the analyzing step comprises using a logistic regression. 96. The method of Aspect 95, wherein the logistic regression is a penalized logistic regression. 97. The method of any one of Aspects 45-96, wherein the analyzing step comprises using a trained neural network, a trained non-linear survival analysis model, a trained Bayesian time-to-event model, or a trained linear or non-linear time series model. 98. A method for treating a subject predicted to develop osteoporosis within a set time-to- event, the method comprising: administering a therapeutically effective dose of an osteoporosis therapy to the subject, wherein the subject has been determined to have a relative risk score for developing osteoporosis within a time-to-event that exceeds a pre-determined threshold relative risk score based on a trained machine learning model that determines time-to-event for developing osteoporosis of subjects using methylation fractions of specific epigenetic markers associated with osteoporosis. 99. The method of Aspect 98, wherein the specific epigenetic markers associated with osteoporosis comprise or consist of at least three CpG location designations identified in Table 1.298325604.1 - 14 -100. The method of Aspect 99, wherein the specific epigenetic markers associated with osteoporosis comprise or consist of at least of the top CpG location designations identified in Table 1. 101. The method of any one of Aspects 98-100, wherein the epigenetic markers listed in Table 1 are in ranked order for the impact on the relative risk score and / or hazard curve for osteoporosis. 102. The method of Aspect 98, wherein the specific epigenetic markers comprise or consist of the CpG location designations, cg11047325 and cg18022864. 103. A method for treating a subject predicted to develop sepsis within a set time-to-event, the method comprising: administering a therapeutically effective dose of antibacterial therapy to the subject, wherein the subject has been determined to have a relative risk score for developing sepsis within a time-to-event that exceeds a pre-determined threshold relative risk score based on a trained machine learning model that determines time-to-event for developing sepsis of subjects using methylation fractions of specific epigenetic markers associated with sepsis. 104. The method of Aspect 103, wherein the specific epigenetic markers associated with sepsis comprise or consist of cg20494891 and cg23608915. 105. The method of Aspect 103, wherein the specific epigenetic markers associated with sepsis comprise or consist of cg20494891. 106. The method of Aspect 103, wherein the specific epigenetic markers comprise or consist of at least three CpG location designations identified in Table 2. 107. The method of Aspect 106, wherein the specific epigenetic markers comprise or consist of at least 3 of the top CpG location designations identified in Table 2.298325604.1 - 15 -108. The method of any one of Aspects 106 or 107, wherein the epigenetic markers listed in Table 2 are in ranked order for the impact on the relative risk score and / or hazard curve for sepsis. 109. The method of Aspect 103, wherein the specific epigenetic markers comprise or consist of at least three CpG location designations identified in Table 3. 110. The method of Aspect 109, wherein the specific epigenetic markers comprise or consist of at least 3 of the top CpG location designations identified in Table 3. 111. The method of any one of Aspects 109 or 110, wherein the epigenetic markers listed in Table 3 are in ranked order for the impact on the relative risk score and / or hazard curve for sepsis. 112. The method of Aspect 103, wherein the specific epigenetic markers comprise or consist of at least three CpG location designations identified in Table 4. 113. The method of Aspect 112, wherein the specific epigenetic markers comprise or consist of at least 3 of the top CpG location designations identified in Table 4. 114. The method of any one of Aspects 112 or 113, wherein the epigenetic markers listed in Table 4 are in ranked order for the impact on the relative risk score and / or hazard curve for sepsis. 115. The method of Aspect 103, wherein the specific epigenetic markers comprise or consist of at least three CpG location designations identified in Table 5. 116. The method of Aspect 112, wherein the specific epigenetic markers comprise or consist of at least 3 of the top CpG location designations identified in Table 5. 117. The method of any one of Aspects 112 or 113, wherein the epigenetic markers listed in Table 5 are in ranked order for the impact on the relative risk score and / or hazard curve for sepsis.298325604.1 - 16 -118. A method for treating a subject predicted to develop an electrolyte disorder within a set time-to-event, the method comprising: administering a therapeutically effective dose of an electrolyte disorder therapy to the subject, wherein the subject has been determined to have a relative risk score for developing an electrolyte disorder within a time-to-event that exceeds a pre-determined threshold relative risk score based on a trained machine learning model that determines time-to-event for developing an electrolyte disorder of subjects using methylation fractions of specific epigenetic markers associated with an electrolyte disorder. 119. The method of Aspect 118, wherein the specific epigenetic markers comprise or consist of at least three CpG location designations identified in Table 6. 120. The method of Aspect 119, wherein the specific epigenetic markers comprise or consist of at least 3 of the top CpG location designations identified in Table 6. 121. The method of any one of Aspects 119 or 120, wherein the epigenetic markers listed in Table 6 are in ranked order for the impact on the relative risk score and / or hazard curve for an electrolyte disorder. 122. The method of Aspect 118, wherein the specific epigenetic markers comprise or consist of at least three CpG location designations identified in Table 7. 123. The method of Aspect 122, wherein the specific epigenetic markers comprise or consist of at least 3 of the top CpG location designations identified in Table 7. 124. The method of any one of Aspects 122 or 123, wherein the epigenetic markers listed in Table 7 are in ranked order for the impact on the relative risk score and / or hazard curve for an electrolyte disorder. 125. A method for treating a subject predicted to develop anemia within a set time-to- event, the method comprising:298325604.1 - 17 -administering a therapeutically effective dose of an anemia therapy to the subject, wherein the subject has been determined to have a relative risk score for developing an electrolyte disorder within a time-to-event that exceeds a pre-determined threshold relative risk score based on a trained machine learning model that determines time-to-event for developing anemia of subjects using methylation fractions of specific epigenetic markers associated with anemia. 126. The method of Aspect 125, wherein the specific epigenetic markers comprise or consist of at least three CpG location designations identified in Table 8. 127. The method of Aspect 126, wherein the specific epigenetic markers comprise or consist of at least 3 of the top CpG location designations identified in Table 8. 128. The method of any one of Aspects 126 or 127, wherein the epigenetic markers listed in Table 8 are in ranked order for the impact on the relative risk score and / or hazard curve for anemia. 129. The method of Aspect 125, wherein the specific epigenetic markers comprise or consist of at least three CpG location designations identified in Table 9. 130. The method of Aspect 129, wherein the specific epigenetic markers comprise or consist of at least 3 of the top CpG location designations identified in Table 9. 131. The method of any one of Aspects 129 or 130, wherein the epigenetic markers listed in Table 9 are in ranked order for the impact on the relative risk score and / or hazard curve for anemia. 132. A method for treating a patient with a therapeutic intervention, the method comprising: administering a therapeutically effective dose of the therapeutic intervention to the subject, wherein the subject has been determined to have a relative risk score for developing an health outcome within a time-to-event that exceeds a pre-determined threshold relative298325604.1 - 18 -risk score based on a trained machine learning model that determines time-to-event for developing the health outcome of subjects using methylation fractions of specific epigenetic markers associated with anemia. 133. The method of Aspect 132, wherein the trained machine learning model is trained by tuning training data including a time before obtaining the epigenetic data from a training set patient in which the training set patient did not have the health outcome and / or a time after obtaining the epigenetic data from the training set patient in which the training set patient was diagnosed with the health outcome. 134. The method of Aspect 132 or 133, wherein the therapeutic intervention comprises one or more of a small molecule, a biologic, radiation, phototherapy, heat therapy, cold therapy, surgery, exercise, a diet change, psychotherapy, or a combination thereof. 135. The method of Aspect 134, wherein the biologic comprises one or more proteins, carbohydrates, nucleic acids, lipids, cells, vaccines, hormones, and / or metabolites. 136. The method of any one of Aspects 132-135, wherein the patient has, or is suspected of having, a disease. 137. The method of Aspect 136 wherein the disease is a cancer, an infection, an immune disease, a lung disease, a brain disease, a heart disease, a skin disease, a kidney disease, a liver disease, a gastrointestinal disease, a bone disease, a blood disease, a ligament disease, a nerve disease, a bladder disease, a pancreas disease, a gallbladder disease, a reproductive- organ disease, and / or a muscular disease. 138. A system for determining a time-to-event of a health outcome, the system comprising: a processor; and memory storing instructions that, when executed by the processor, cause the processor to: receive an epigenetic dataset obtained from a patient, wherein the epigenetic dataset includes methylation fractions of specific epigenetic markers; and298325604.1 - 19 -apply, by the one or more processors, a trained machine learning model to the epigenetic dataset to predict a relative risk score and / or a hazard curve of the time-to-event of the health outcome. 139. A system for determining a time-to-event of osteoporosis, the system comprising: a processor; and memory storing instructions that, when executed by the processor, cause the processor to: receive an epigenetic dataset obtained from a patient, wherein the epigenetic dataset includes methylation fractions of specific epigenetic markers; and apply, by the one or more processors, a trained machine learning model to the epigenetic dataset to predict a relative risk score and / or a hazard curve of the time-to-event of osteoporosis. 140. A system for determining a time-to-event of sepsis, the system comprising: a processor; and memory storing instructions that, when executed by the processor, cause the processor to: receive an epigenetic dataset obtained from a patient, wherein the epigenetic dataset includes methylation fractions of specific epigenetic markers; and apply, by the one or more processors, a trained machine learning model to the epigenetic dataset to predict a relative risk score and / or a hazard curve of the time-to-event of sepsis. 141. A system for determining a time-to-event of an electrolyte disorder, the system comprising: a processor; and memory storing instructions that, when executed by the processor, cause the processor to: receive an epigenetic dataset obtained from a patient, wherein the epigenetic dataset includes methylation fractions of specific epigenetic markers; and298325604.1 - 20 -apply, by the one or more processors, a trained machine learning model to the epigenetic dataset to predict a relative risk score and / or a hazard curve of the time-to-event of the electrolyte disorder. 142. A system for determining a time-to-event of anemia, the system comprising: a processor; and memory storing instructions that, when executed by the processor, cause the processor to: receive an epigenetic dataset obtained from a patient, wherein the epigenetic dataset includes methylation fractions of specific epigenetic markers; and apply, by the one or more processors, a trained machine learning model to the epigenetic dataset to predict a relative risk score and / or a hazard curve of the time-to-event of anemia. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] The present disclosure is described in conjunction with the appended figures:

[0027] FIG. 1 shows histograms of time-to-event distribution. Samples are filtered as described in the Long term risk model. Patients who present a certain health outcome, i.e. the phecode, after a set time from when the sample was taken, i.e. the post-draw clearance time, are shown in orange. Time-to-event is defined as the number of days after blood draw when the first occurrence of a diagnosis associated with the phecode is diagnosed. Patients who never had the phecode after the post-draw clearance time to present day are shown in white, and time- to-event is defined as number of days after blood draw to the present day, which is the time of that patient’s latest phecode registered in the EHR system. (a) Osteoporosis with 90 days pre- and 10 days post-draw clearance. (b) Sepsis with 90 days pre- and 10 days post-draw clearance. (c) Disorders of fluid, electrolyte with 180 days pre- and 30 days post-draw clearance. (d) Iron deficiency anemias with 365 days pre- and 60 days post-draw clearance.

[0028] FIG. 2 shows performance comparison between 4 models using covar.base: baseline covars, covars.full: baseline covars and cell type composition derived from different sets of methylation data including, meth.100k (the top 100k most variable CpGs) and meth.full (a full set of methylation CpGs) under 90 days pre- and 10 days post-draw clearance time. In general, cell type composition shows a strong signal over the base line covars. The last panel298325604.1 - 21 -indicates that there are a few phecodes that clearly benefit from using the full set of methylation CpG: the middle top dots achieving close to 0.7 concordance with methylation while the base line covars only have 0.55.

[0029] FIG.3 shows similar performance comparison between 4 models but for 180 days pre and 30 days pre and post draw clearance time

[0030] FIG.4 shows similar performance comparison between 4 models but for 365 days pre and 60 days pre and post draw clearance time

[0031] FIG.5. Fixed window model precision recall AUC and ROC AUC scatter plots for 3 different time windows. As the post draw clearance and end times are increased, more samples are added and the models using methylation CpGs gradually start to show better signals in certain phecodes. Each dot indicates a phecode. Number of phecodes investigated under different time windows are different due to the fact that the inventors filter on the minimal sample size and prevalence.

[0032] FIG. 6 shows binarized prediction results. Sepsis with 90 days of pre-draw clearance, 14 days post draw clearance time and 365 days of post-draw end time. Out of sample ROC-AUC and PR-AUC. Upper left: model using only baseline covariates (age, sex, BMI, Self reported ancestry). Upper right: model using full set of covariates: baseline covariates and estimation of cell type compositions. Lower left: model using all methylation CpGs only. Lower right: model using all methylation CpGs and full set of baseline covariates. Lower right seems slightly better than lower left. However, manually checking the features selected by the model reveals that there is no covariates selected. Thus this discrepancy can be viewed as the inherent variation of the model.

[0033] FIG. 7 shows relative risk prediction results under the default model using the original CpG set. For each panel, the inventors show the estimated hazard curve for 2 patients, one with the highest estimated risk and one with the lowest estimated risk: (a) for osteoporosis with 90 days pre clearance and 10 days post clearance, (b) for sepsis with 90 days pre clearance and 10 days post clearance, (c) for sepsis with 180 days pre clearance and 30 days post clearance, (d) for disorders of fluid, electrolyte with 180 days pre clearance and 30 days post clearance, (e) for iron deficiency anemias with 365 days pre clearance and 60 days post clearance.298325604.1 - 22 -DETAILED DESCRIPTION I. Overview

[0034] Disclosed herein are methods, systems, workflows, and / or techniques that are useful in determining health outcomes by analyzing DNA methylation, in particular CpG methylation. The disclosure provides certain, specific CpG sites that are used in the methods, systems, workflows, and / or techniques described herein. While specific CpG sites are disclosed, it is also contemplated that analogous CpG sites may also be used. Similarly, it is also contemplated that CpGs within a linkage disequilibrium block may also be used. In some aspects, the CpGs used in the methods, systems, workflows and / or techniques are adjacent to specific CpGs disclosed herein. In some aspects, the CpGs used in the methods, systems, workflows and / or techniques are in the same genetic locus to specific CpGs disclosed herein. In some aspects, the CpGs used in the methods, systems, workflows and / or techniques are within the same gene as specific CpGs disclosed herein. In some aspects, the CpGs used in the methods, systems, workflows and / or techniques are within 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110, 111, 112, 113, 114, 115, 116, 117, 118, 119, 120, 121, 122, 123, 124, 125, 126, 127, 128, 129, 130, 131, 132, 133, 134, 135, 136, 137, 138, 139, 140, 141, 142, 143, 144, 145, 146, 147, 148, 149, 150, 151, 152, 153, 154, 155, 156, 157, 158, 159, 160, 161, 162, 163, 164, 165, 166, 167, 168, 169, 170, 171, 172, 173, 174, 175, 176, 177, 178, 179, 180, 181, 182, 183, 184, 185, 186, 187, 188, 189, 190, 191, 192, 193, 194, 195, 196, 197, 198, 199, 200, 201, 202, 203, 204, 205, 206, 207, 208, 209, 210, 211, 212, 213, 214, 215, 216, 217, 218, 219, 220, 221, 222, 223, 224, 225, 226, 227, 228, 229, 230, 231, 232, 233, 234, 235, 236, 237, 238, 239, 240, 241, 242, 243, 244, 245, 246, 247, 248, 249, 250, 251, 252, 253, 254, 255, 256, 257, 258, 259, 260, 261, 262, 263, 264, 265, 266, 267, 268, 269, 270, 271, 272, 273, 274, 275, 276, 277, 278, 279, 280, 281, 282, 283, 284, 285, 286, 287, 288, 289, 290, 291, 292, 293, 294, 295, 296, 297, 298, 299, 300 base pairs, or any range derivable therein, of specific CpGs disclosed herein.

[0035] The description below provides exemplary implementations of the methods and systems described herein. Tables 1-9 provide specific epigenetic markers for health outcomes described herein, including osteoporosis, sepsis, fluid disorders, and iron deficiency anemia. In one or more embodiments, the specific epigenetic marker lists are shown in ranked order for the effect on predicting the time-to-event. In one or more embodiments, the specific epigenetic298325604.1 - 23 -marker lists are a reduced set. Sparsity in encouraged by presenting a leaner, more compact version of the bigger model list. The decision of which model to use (i.e., default or reduced) is context specific. For example, the reduced set may generalize better for populations that are more distinct from the population of patients served by a specific institution, based on which population was used to train the models.

[0036] Table 1: CpG Location Designations in ranked order for effect on predicting time to osteoporosis trained with a 90 day pre-draw clearance time, 10 day post-draw clearance time.298325604.1 - 24 -298325604.1 - 25 -298325604.1 - 26 -298325604.1 - 27 -298325604.1 - 28 -298325604.1 - 29 -298325604.1 - 30 -298325604.1 - 31 -

[0037] Table 2: CpG Location Designations in ranked order for effect on predicting time to sepsis trained with a 180 day pre-draw clearance time and 30 day post-draw clearance time.298325604.1 - 32 -

[0038] Table 3: CpG Location Designations in ranked order for effect on predicting time to sepsis trained with a 180 day pre-draw clearance time, 30 day post-draw clearance time, and encouraging sparsity.298325604.1 - 33 -

[0039] Table 4: CpG Location Designations in ranked order for effect on predicting time to sepsis trained with a 90 day pre-draw clearance time, 14 day post-draw clearance time, and a 365 post-draw end time.298325604.1 - 34 -298325604.1 - 35 -298325604.1 - 36 -

[0040] Table 5: CpG Location Designations in ranked order for effect on predicting time to sepsis trained with a 90 day pre-draw clearance time, 14 day post-draw clearance time, a 365 post-draw end time, and encouraging sparsity.

[0041] Table 6: CpG Location Designations in ranked order for effect on predicting time to a fluid disorder trained with a 180 day pre-draw clearance time, and 30 day post-draw clearance time.298325604.1 - 37 -

[0042] Table 7: CpG Location Designations in ranked order for effect on predicting time to a fluid disorder trained with a 180 day pre-draw clearance time, 30 day post-draw clearance time and encouraging sparsity.

[0043] Table 8: CpG Location Designations in ranked order for effect on predicting time to an iron deficiency anemia trained with a 365 day pre-draw clearance time, and 60 day post- draw clearance time.

[0044] Table 9: CpG Location Designations in ranked order for effect on predicting time to an iron deficiency anemia trained with a 180 day pre-draw clearance time, and 60 day post- draw clearance time and encouraging sparsity.298325604.1 - 38 -II. Exemplary Descriptions of Terms

[0045] As used herein, the term “set of” means one or more. For example, a set of items includes one or more items.

[0046] As used herein, the phrase “at least one of,” when used with a list of items, means different combinations of one or more of the listed items may be used and only one of the items in the list may be needed. The item may be a particular object, thing, step, operation, process, or category. In other words, “at least one of” means any combination of items or number of items may be used from the list, but not all of the items in the list may be required. For example, without limitation, “at least one of item A, item B, or item C” means item A; item A and item B; item B; item A, item B, and item C; item C; item B and item C; or item A and C. In some cases, “at least one of item A, item B, or item C” means, but is not limited to, two of item A, one of item B, and ten of item C; four of item B and seven of item C; or some other suitable combination.

[0047] As used herein, “substantially” means sufficient to work for the intended purpose. The term “substantially” thus allows for minor, insignificant variations from an absolute or perfect state, dimension, measurement, result, or the like such as would be expected by a person of ordinary skill in the field but that do not appreciably affect overall performance. When used with respect to numerical values or parameters or characteristics that can be expressed as numerical values, “substantially” means within ten percent.

[0048] The terms “biological sample,” “biological specimen,” or “biospecimen” as used herein, generally refers to a specimen taken by sampling so as to be representative of the source of the specimen, typically, from a subject. A biological sample can be representative of an organism as a whole, specific tissue, cell type, or category or sub-category of interest. The biological sample can include a macromolecule. The biological sample can include a small molecule. The biological sample can include a virus. The biological sample can include a cell or derivative of a cell. The biological sample can include an organelle. The biological sample can include a cell nucleus. The biological sample can include a rare cell from a population of cells. The biological sample can include any type of cell, including without limitation prokaryotic cells, eukaryotic cells, bacterial, fungal, plant, mammalian, or other animal cell type, mycoplasmas, normal tissue cells, tumor cells, or any other cell type, whether derived from single cell or multicellular organisms. The biological sample can include a constituent of a cell. The biological sample can include nucleotides (e.g., ssDNA, dsDNA, RNA), organelles, amino acids, peptides, proteins, carbohydrates, glycoproteins, or any combination thereof. The298325604.1 - 39 -biological sample can include a matrix (e.g., a gel or polymer matrix) comprising a cell or one or more constituents from a cell (e.g., cell bead), such as DNA, RNA, organelles, proteins, or any combination thereof, from the cell. The biological sample may be obtained from a tissue of a subject. The biological sample can include a hardened cell. Such hardened cells may or may not include a cell wall or cell membrane. The biological sample can include one or more constituents of a cell but may not include other constituents of the cell. An example of such constituents may include a nucleus or an organelle. The biological sample may include a live cell. The live cell can be capable of being cultured. The sample may include at least one of a plasma, blood, or serum sample.

[0049] The term “biomarker,” as used herein, generally refers to any measurable substance taken as a sample from a subject whose presence is indicative of some phenomenon. Non- limiting examples of such phenomenon can include a disease state, a condition, or exposure to a compound or environmental condition. In various embodiments described herein, biomarkers may be used for diagnostic purposes (e.g., to diagnose a health state, a disease state). The term “biomarker” can be used interchangeably with the term “marker.”

[0050] The term “disease state,” which may be used interchangeably with “health outcome,” as used herein, generally refers to a condition that affects the structure or function of an organism. Non-limiting examples of causes of disease states may include pathogens, immune system dysfunctions, cell damage caused by aging, cell damage caused by other factors (e.g., trauma and cancer). Disease states can include any state of a disease whether symptomatic or asymptomatic. Disease states can include disease stages of a disease progression. Disease states can cause minor, moderate, or severe disruptions in structure or function of an organism (e.g., a subject).

[0051] As used herein, a “model” may include one or more algorithms, one or more mathematical techniques, one or more machine learning algorithms, statistical algorithms, statistical learning algorithms, statistical models, or a combination thereof.

[0052] As used herein, “machine learning” may be the practice of using algorithms to parse data, learn from it, and then make a determination or prediction about something in the world. Machine learning uses algorithms that can learn from data without relying on rules-based programming. A machine learning algorithm may include a parametric model, a nonparametric model, a deep learning model, a neural network, a linear discriminant analysis model, a quadratic discriminant analysis model, a support vector machine, a random forest algorithm, a nearest neighbor algorithm, a combined discriminant analysis model, a k-means clustering algorithm, a supervised model, an unsupervised model, logistic regression model, a298325604.1 - 40 -multivariable regression model, a penalized multivariable regression model, or a combination thereof, or another type of model.

[0053] As used herein, “relative fraction” may refer to a quantitative value generated from a methylation assay and / or analysis. In various embodiments, the quantitative value may relate to the amount of methylation present at a particular location on a nucleic acid.

[0054] As used herein, “methylation fraction” may have its plain and ordinary meaning and may refer to a comparison of two or more abundances. In various embodiments, the comparison may comprise comparing the abundance of methylated nucleotides, such as cytosines, at a particular location to the total number of nucleotides, such as cytosines, at the particular location in the sample. In various embodiments, the comparison may comprise comparing the abundance of methylated nucleotides, such as cytosines, at a particular location to the total number of amount of nucleic acid in the sample. In various embodiments, a methylation fraction can be expressed as a ratio. In other embodiments, a methylation fraction can be expressed as a percentage.

[0055] As used herein, the term “subject” may be used interchangeably with “patient” and includes an individual that has been diagnosed with, determined to have, suspected of having, or has symptoms of one or more health outcomes, including osteoporosis, sepsis, an electrolyte disorder, and anemia. The patient may be a human patient. III. Overview of Exemplary Workflow

[0056] In one or more embodiments, epigenetic data is collected from a patient, using any method described herein. In one or more embodiments, the epigenetic data is analyzed using a trained model, including any model described herein, to predict one or more scores. The score(s) may be indicative of a time-to-event of a health outcome, including any health outcome described herein. In one or more embodiments, the patient is treated for the health outcome by performing one or more of the methods described herein before and after administering a therapeutic intervention. In one or more embodiments, a change of the score from before to after administering the therapeutic intervention is indicative of whether or not to change the therapeutic intervention administration, including changing the amount and / or type of intervention.298325604.1 - 41 -III.A. Data Acquisition III.A.1. Detection of methylated DNA

[0057] Aspects of the methods include assaying nucleic acids to determine expression levels and / or methylation levels of nucleic acids. Aspects of the disclosure include the detection of one or more CpG islands, such as at least, at most, or exactly 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 CpG islands (or any range derivable therein). Each biomarker may comprise or consist of at least or at most or exactly 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 CpG islands (or any range derivable therein). Assays for the detection of methylated DNA are known in the art. Exemplary methods are described herein. III.A.1.a. Bisulfite Sequencing

[0058] The bisulfite treatment of DNA mediates the deamination of cytosine into uracil, and these converted residues will be read as thymine, as determined by PCR-amplification and subsequent Sanger sequencing analysis. However, 5 mC residues are resistant to this conversion and, so, will remain read as cytosine. Thus, comparing the Sanger sequencing read from an untreated DNA sample to the same sample following bisulfite treatment enables the detection of the methylated cytosines. With the advent of next-generation sequencing (NGS) technology, this approach can be extended to DNA methylation analysis across an entire genome. To ensure complete conversion of non-methylated cytosines, controls may be incorporated for bisulfite reactions.

[0059] Whole genome bisulfite sequencing (WGBS) is similar to whole genome sequencing, except for the additional step of bisulfite conversion. Sequencing of the 5 mC- enriched fraction of the genome is not only a less expensive approach, but it also allows one to increase the sequencing coverage and, therefore, precision in revealing differentially- methylated regions. Sequencing could be done using any existing NGS platform; Illumina and Life Technologies both offer kits for such analysis.

[0060] Bisulfite sequencing methods include reduced representation bisulfite sequencing (RRBS), where only a fraction of the genome is sequenced. In RRBS, enrichment of CpG-rich regions is achieved by isolation of short fragments after MspI digestion that recognizes CCGG sites (and it cut both methylated and unmethylated sites). It ensures isolation of ~85% of CpG islands in the human genome. Then, the same bisulfite conversion and library preparation is performed as for WGBS. The RRBS procedure normally requires ~100 ng - 1 µg of DNA. III.A.1.b. Hybridization

[0061] Certain aspects concern nucleic acids that hybridize to other nucleic acids under particular hybridization conditions. Methods for hybridizing nucleic acids are well known in298325604.1 - 42 -the art. See, e.g., Current Protocols in Molecular Biology, John Wiley and Sons, N.Y. (1989), 6.3.1-6.3.6. Certain aspects concern a moderately stringent hybridization condition using a prewashing solution containing 5× sodium chloride / sodium citrate (SSC), 0.5% SDS, 1.0 mM EDTA (pH 8.0), hybridization buffer of about 50% formamide, 6×SSC, and a hybridization temperature of 55° C. (or other similar hybridization solutions, such as one containing about 50% formamide, with a hybridization temperature of 42° C), and washing conditions of 60° C. in 0.5×SSC, 0.1% SDS. A stringent hybridization condition hybridizes in 6×SSC at 45° C., followed by one or more washes in 0.1×SSC, 0.2% SDS at 68° C. Furthermore, one of skill in the art can manipulate the hybridization and / or washing conditions to increase or decrease the stringency of hybridization such that nucleic acids comprising nucleotide sequence that are at least 65%, at least 70%, at least 75%, at least 80%, at least 85%, at least 90%, at least 95%, at least 96%, at least 97%, at least 98% or at least 99% identical to each other typically remain hybridized to each other.

[0062] The parameters affecting the choice of hybridization conditions and guidance for devising suitable conditions are set forth by, for example, Sambrook, Fritsch, and Maniatis (Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y., chapters 9 and 11 (1989); Current Protocols in Molecular Biology, Ausubel et al., eds., John Wiley and Sons, Inc., sections 2.10 and 6.3-6.4 (1995), both of which are herein incorporated by reference in their entirety for all purposes) and can be readily determined by those having ordinary skill in the art based on, for example, the length and / or base composition of the DNA. III.A.1.c. Probes

[0063] In another aspect, nucleic acid molecules are suitable for use as primers or hybridization probes for the detection or purification of nucleic acid sequences.

[0064] Probes based on the desired sequence of a nucleic acid can be used to detect the nucleic acid or similar nucleic acids, for example, transcripts encoding a polypeptide of interest. The probe can comprise a label group, e.g., a radioisotope, a fluorescent compound, an enzyme, or an enzyme co-factor. Such probes can be used to isolate or purify certain nucleic acids. III.A.2. Sequencing

[0065] DNA, including bisulfite-converted DN,A could be used for the amplification of the region of interest followed by sequencing. Primers can be designed around the CpG island and used for PCR amplification of bisulfite-converted DNA. The resulting PCR products could be cloned and sequenced. Accordingly, aspects of the disclosure may include sequencing nucleic acids to detect methylation of nucleic acids and / or biomarkers. In some aspects, the298325604.1 - 43 -methods of the disclosure include a sequencing method. Sequencing methods useful for certain aspects may include those described below. Other sequencing methods known in the art may also be used, in some aspects. III.A.2.a. Massively parallel signature sequencing (MPSS).

[0066] The first of the next-generation sequencing technologies, massively parallel signature sequencing (or MPSS), was developed in the 1990s at Lynx Therapeutics. MPSS was a bead-based method that used a complex approach of adapter ligation followed by adapter decoding, reading the sequence in increments of four nucleotides. This method made it susceptible to sequence-specific bias or loss of specific sequences. Because the technology was so complex, MPSS was only performed 'in-house' by Lynx Therapeutics and no DNA sequencing machines were sold to independent laboratories. Lynx Therapeutics merged with Solexa (later acquired by Illumina) in 2004, leading to the development of sequencing-by- synthesis, a simpler approach acquired from Manteia Predictive Medicine, which rendered MPSS obsolete. However, the essential properties of the MPSS output were typical of later "next-generation" data types, including hundreds of thousands of short DNA sequences. In the case of MPSS, these were typically used for sequencing cDNA for measurements of gene expression levels. Indeed, the powerful Illumina HiSeq2000, HiSeq2500 and MiSeq systems are based on MPSS. III.A.2.b. Polony sequencing.

[0067] The Polony sequencing method, developed in the laboratory of George M. Church at Harvard, was among the first next-generation sequencing systems and was used to sequence a full genome in 2005. It combined an in vitro paired-tag library with emulsion PCR, an automated microscope, and ligation-based sequencing chemistry to sequence an E. coli genome at an accuracy of >99.9999% and a cost approximately 1 / 9 that of Sanger sequencing. The technology was licensed to Agencourt Biosciences, subsequently spun out into Agencourt Personal Genomics, and eventually incorporated into the Applied Biosystems SOLiD platform, which is now owned by Life Technologies. III.A.2.c. 454 pyrosequencing.

[0068] A parallelized version of pyrosequencing was developed by 454 Life Sciences, which has since been acquired by Roche Diagnostics. The method amplifies DNA inside water droplets in an oil solution (emulsion PCR), with each droplet containing a single DNA template attached to a single primer-coated bead that then forms a clonal colony. The sequencing machine contains many picoliter-volume wells each containing a single bead and sequencing enzymes. Pyrosequencing uses luciferase to generate light for detection of the individual298325604.1 - 44 -nucleotides added to the nascent DNA, and the combined data are used to generate sequence read-outs. This technology provides intermediate read length and price per base compared to Sanger sequencing on one end and Solexa and SOLiD on the other. III.A.2.d. Illumina (Solexa) sequencing.

[0069] Solexa, now part of Illumina, developed a sequencing method based on reversible dye-terminators technology, and engineered polymerases, that it developed internally. The terminated chemistry was developed internally at Solexa and the concept of the Solexa system was invented by Balasubramanian and Klennerman from Cambridge University's chemistry department. In 2004, Solexa acquired the company Manteia Predictive Medicine in order to gain a massivelly parallel sequencing technology based on "DNA Clusters", which involves the clonal amplification of DNA on a surface. The cluster technology was co-acquired with Lynx Therapeutics of California. Solexa Ltd. later merged with Lynx to form Solexa Inc. In this method, DNA molecules and primers are first attached on a slide and amplified with polymerase so that local clonal DNA colonies, later coined "DNA clusters", are formed. To determine the sequence, four types of reversible terminator bases (RT-bases) are added and non-incorporated nucleotides are washed away. A camera takes images of the fluorescently labeled nucleotides, then the dye, along with the terminal 3' blocker, is chemically removed from the DNA, allowing for the next cycle to begin. Unlike pyrosequencing, the DNA chains are extended one nucleotide at a time and image acquisition can be performed at a delayed moment, allowing for very large arrays of DNA colonies to be captured by sequential images taken from a single camera.

[0070] Decoupling the enzymatic reaction and the image capture allows for optimal throughput and theoretically unlimited sequencing capacity. With an optimal configuration, the ultimately reachable instrument throughput is thus dictated solely by the analog-to-digital conversion rate of the camera, multiplied by the number of cameras and divided by the number of pixels per DNA colony required for visualizing them optimally (approximately 10 pixels / colony). In 2012, with cameras operating at more than 10 MHz A / D conversion rates and available optics, fluidics and enzymatics, throughput can be multiples of 1 million nucleotides / second, corresponding roughly to one human genome equivalent at 1x coverage per hour per instrument, and one human genome re-sequenced (at approx. 30x) per day per instrument (equipped with a single camera). III.A.2.e. SOLiD sequencing.

[0071] Applied Biosystems' (now a Thermo Fisher Scientific brand) SOLiD technology employs sequencing by ligation. Here, a pool of all possible oligonucleotides of a fixed length298325604.1 - 45 -are labeled according to the sequenced position. Oligonucleotides are annealed and ligated; the preferential ligation by DNA ligase for matching sequences results in a signal informative of the nucleotide at that position. Before sequencing, the DNA is amplified by emulsion PCR. The resulting beads, each containing single copies of the same DNA molecule, are deposited on a glass slide. The result is sequences of quantities and lengths comparable to Illumina sequencing. This sequencing by ligation method has been reported to have some issue sequencing palindromic sequences. III.A.2.f. Ion Torrent semiconductor sequencing.

[0072] Ion Torrent Systems Inc. (now owned by Thermo Fisher Scientific) developed a system based on using standard sequencing chemistry, but with a novel, semiconductor based detection system. This method of sequencing is based on the detection of hydrogen ions that are released during the polymerization of DNA, as opposed to the optical methods used in other sequencing systems. A microwell containing a template DNA strand to be sequenced is flooded with a single type of nucleotide. If the introduced nucleotide is complementary to the leading template nucleotide it is incorporated into the growing complementary strand. This causes the release of a hydrogen ion that triggers a hypersensitive ion sensor, which indicates that a reaction has occurred. If homopolymer repeats are present in the template sequence multiple nucleotides will be incorporated in a single cycle. This leads to a corresponding number of released hydrogens and a proportionally higher electronic signal. III.A.2.g. DNA nanoball sequencing.

[0073] DNA nanoball sequencing is a type of high throughput sequencing technology used to determine the entire genomic sequence of an organism. The company Complete Genomics uses this technology to sequence samples submitted by independent researchers. The method uses rolling circle replication to amplify small fragments of genomic DNA into DNA nanoballs. Unchained sequencing by ligation is then used to determine the nucleotide sequence. This method of DNA sequencing allows large numbers of DNA nanoballs to be sequenced per run and at low reagent costs compared to other next generation sequencing platforms. However, only short sequences of DNA are determined from each DNA nanoball which makes mapping the short reads to a reference genome difficult. This technology has been used for multiple genome sequencing projects. III.A.2.h. Heliscope single molecule sequencing.

[0074] Heliscope sequencing is a method of single-molecule sequencing developed by Helicos Biosciences. It uses DNA fragments with added poly-A tail adapters which are attached to the flow cell surface. The next steps involve extension-based sequencing with298325604.1 - 46 -cyclic washes of the flow cell with fluorescently labeled nucleotides (one nucleotide type at a time, as with the Sanger method). The reads are performed by the Heliscope sequencer. The reads are short, up to 55 bases per run, but recent improvements allow for more accurate reads of stretches of one type of nucleotides. This sequencing method and equipment were used to sequence the genome of the M13 bacteriophage. III.A.2.i. Single molecule real time (SMRT) sequencing.

[0075] SMRT sequencing is based on the sequencing by synthesis approach. The DNA is synthesized in zero-mode wave-guides (ZMWs) – small well-like containers with the capturing tools located at the bottom of the well. The sequencing is performed with use of unmodified polymerase (attached to the ZMW bottom) and fluorescently labelled nucleotides flowing freely in the solution. The wells are constructed in a way that only the fluorescence occurring by the bottom of the well is detected. The fluorescent label is detached from the nucleotide at its incorporation into the DNA strand, leaving an unmodified DNA strand. According to Pacific Biosciences, the SMRT technology developer, this methodology allows detection of nucleotide modifications (such as cytosine methylation). This happens through the observation of polymerase kinetics. This approach allows reads of 20,000 nucleotides or more, with average read lengths of 5 kilobases. IV. Therapeutic Methods

[0076] Disclosed herein are methods of treating a patient that has, is suspected of having, determined to have, diagnosed with, and / or has symptoms of one or more health outcomes including sepsis, osteoporosis, an electrolyte disorder, and / or anemia. The methods can include administering a therapeutic to the patient. The therapeutic may be known to treat or prevent the health outcome. IV.A. Sepsis

[0077] One or more embodiments relate to a method comprising the step of administering to a patient, who has been determined to have sepsis, a therapeutically effective amount of one or more antibiotic agents. The determination of sepsis may be based on clinical criteria, laboratory findings, or a combination thereof, in addition to the method described herein.

[0078] The antibiotic agents may be selected from one or more classes known to be effective in the treatment of sepsis. These include beta-lactam antibiotics such as penicillins (e.g., piperacillin / tazobactam), cephalosporins (e.g., ceftriaxone, cefotaxime, ceftazidime), and carbapenems (e.g., meropenem, imipenem, ertapenem). Glycopeptides such as vancomycin298325604.1 - 47 -may be used, particularly for coverage of methicillin-resistant Staphylococcus aureus (MRSA). Aminoglycosides, including gentamicin and tobramycin, may be administered in combination with beta-lactams to achieve synergistic bactericidal activity. Monobactams such as aztreonam may be used, especially in patients with beta-lactam allergies. Lincosamides, such as clindamycin, may be employed adjunctively in cases of suspected toxin-mediated syndromes. Nitroimidazoles, such as metronidazole, may be included for anaerobic coverage in intra- abdominal or polymicrobial infections.

[0079] In certain embodiments, the method may further comprise the administration of antifungal agents, such as echinocandins (e.g., micafungin), in patients at risk for fungal sepsis.

[0080] The antibiotics may be administered intravenously, with dosing regimens adjusted based on patient-specific factors including renal function, severity of illness, and pharmacokinetic / pharmacodynamic parameters. The therapy is initiated as soon as possible following the diagnosis of sepsis and may be modified based on microbiological culture results and clinical response.

[0081] This method may be implemented in conjunction with other supportive measures, including fluid resuscitation, vasopressor therapy, and organ support, as part of a comprehensive sepsis management protocol. IV.B. Osteoporosis

[0082] One or more embodiments relate to a method comprising the step of administering to a patient, who has been determined to have sepsis, a therapeutically effective amount of one or more therapeutic agents.

[0083] In one or more embodiments, the method includes the administration of therapeutic agents that promote bone density and reduce fracture risk. The therapeutic agent(s) may comprise bisphosphonates, such as alendronate, risedronate, ibandronate, and zoledronic acid. The therapeutic agents may be administered orally or intravenously.

[0084] The therapeutic agents may comprise denosumab. In some embodiments, denosumab is administered via subcutaneous injection. In some embodiments denosumab is administered every six months.

[0085] The therapeutic agents may comprise selective estrogen receptor modulators (SERMs) such as raloxifene, parathyroid hormone analogs such as teriparatide and abaloparatide, and sclerostin inhibitors such as romosozumab.

[0086] The therapeutic agents may comprise antibiotic agents in patients who concurrently present with or are at risk for systemic infections, including sepsis.298325604.1 - 48 -IV.C. Electrolyte disorders

[0087] Certain embodiments relate to methods for treating patients diagnosed with electrolyte disorders, which are characterized by abnormal concentrations of essential minerals in the body’s fluids. Electrolytes, including sodium, potassium, calcium, magnesium, chloride, phosphate, and bicarbonate, are critical for maintaining fluid balance, nerve conduction, muscle function, and acid-base homeostasis. Electrolyte disorders can include the dysregulation of one or more of such electrolytes (or others). Electrolyte disorders may result from a variety of causes, including dehydration, renal dysfunction, endocrine disorders, medication effects, or acute illness.

[0088] One or more embodiments relate to methods comprising the administration of one or more therapeutic agents to a patient who has been determined to have an electrolyte disorder.

[0089] The therapeutic agents administered may vary depending on the specific electrolyte abnormality. For example, in cases of hyponatremia (low sodium), treatment may include intravenous administration of isotonic or hypertonic saline solutions to restore sodium levels. In hypernatremia (high sodium), controlled rehydration with hypotonic fluids may be employed. Hypokalemia (low potassium) may be treated with oral or intravenous potassium chloride, while hyperkalemia (high potassium) may require administration of calcium gluconate to stabilize cardiac membranes, insulin with glucose to shift potassium intracellularly, and agents such as sodium polystyrene sulfonate or patiromer to promote potassium excretion.

[0090] For hypocalcemia, calcium gluconate or calcium chloride may be administered intravenously, and vitamin D analogs may be used to enhance calcium absorption. Hypercalcemia may be managed with intravenous hydration, bisphosphonates, or calcitonin. Magnesium sulfate is commonly used to treat hypomagnesemia, while hypermagnesemia may require intravenous calcium and supportive measures. Phosphate imbalances may be corrected with oral or intravenous phosphate preparations or phosphate binders, depending on whether the condition is hypo- or hyperphosphatemia. Bicarbonate therapy may be used to correct metabolic acidosis, while acidifying agents may be used in cases of metabolic alkalosis. The therapeutic regimen may be administered orally or intravenously. Dosing and duration of therapy are determined based on the degree of electrolyte abnormality, underlying etiology, renal function, and response to treatment.298325604.1 - 49 -IV.D. Anemia

[0091] One or more embodiments relate to methods comprising the administration of one or more therapeutic agents to a patient who has been determined to have anemia.

[0092] The therapeutic agents administered may vary depending on the underlying etiology of the anemia. The therapeutic agent may comprise iron supplementation, which can be administered orally or intravenously. Oral iron preparations may include ferrous sulfate, ferrous gluconate, or ferrous fumarate, while intravenous formulations may include iron dextran, iron sucrose, or ferric carboxymaltose. The therapeutic agent can comprise cyanocobalamin, hydroxocobalamin, and / or folic acid, which can be administered orally or intramuscularly.

[0093] In some embodiments, the therapeutic agent comprises erythropoiesis-stimulating agents (ESAs) such as epoetin alfa or darbepoetin alfa, which may be administered to stimulate red blood cell production. In some embodiments, the therapeutic agent comprises luspatercept and / or lenalidomide. In some embodiments, red blood cell transfusions are administered to the patient.

[0094] The therapeutic regimen may be administered orally, intravenously, or subcutaneously. Dosing and duration of therapy are determined based on the type and cause of anemia, the patient’s clinical status, and response to treatment. V. Representative Experimental Results V.A. Electronic health record data

[0095] De-identified electronic health record data for this study was extracted from the perioperative data warehouse (PDW), a custom-built, robust data warehouse containing all patients who have undergone surgery at UCLA Health since the implementation of UCLA’s EMR (EPIC Systems, Madison, WI, USA) in March 2013. The PDW, which has been described previously, has a two-stage design. First, data are extracted from EPIC’s Clarity database into 29 tables organized around three distinct concepts: patients, surgical procedures, and health system encounters. Then, these data are used to populate a series of 4000 distinct measures and metrics such as procedure duration, admission ICD codes, lab results, and medication orders.

[0096] Methylation and genotype samples were collected using blood from 831 patients as part of the UCLA ATLAS precision health initiative between October 26, 2016 and December 10, 2018. Retrospective data collection and analysis was approved by the UCLA IRB. Patient298325604.1 - 50 -Recruitment and Sample Collection for Precision Health Activities at UCLA is an approved study by the UCLA Institutional Review Board (UCLA IRB). IRB17-001013. All necessary patient / participant consent has been obtained and the appropriate institutional forms have been archived. V.B. Patient group

[0097] The samples were collected from patients before undergoing surgery with general anesthesia at UCLA Health, and the patients had not undergone surgery in the 30 days prior to blood sample collection. Of these patients, 302 were selected for inclusion based on the presence of acute kidney injury (AKI), defined as an Acute Kidney Injury Network (AKIN) classification of one or greater, after undergoing surgery. An additional 348 patients were risk- matched controls, with either glomerular filtration rate (GFR) less than or equal to 38 (210 patients), or GFR >38 and a propensity risk score that matched case patients (348 patients). The propensity score was created using available EHR features such as age, weight, BMI, and other preoperative features that were measured in the hospital. Within the control group, the inventors also performed a similar procedure ascertained on whether individuals were a heart attack case. Controls for heart attack patients were also selected using propensity scoring. V.C. Long term risk model

[0098] For the Long term risk model described herein, the inventors define the following two key concepts.

[0099] Pre-draw Clearance time (days): This quantity is denoted as time_inc in all the figures. The inventors tune the model to predict which patients will present the phenotype among patients who did not present the phenotype for at least this number of days before the blood draw. The inventors have models that use different number of days – the idea is that some phenotypes may occur only once in a lifetime, whereas other phenotypes may occur multiple times. This component allows one to tune models differently, for example, in cases where one wants to consider the possibility of recurrence of a condition.

[0100] Post-draw Clearance time (days): This quantity is denoted as time_lb in all the figures. The inventors only aim to predict which patients will present the phenotype at least this number of days after the blood draw. The reason why different models may use a different number of days is because different phenotypes have different dynamics and therefore setting the post-draw clearance time differently can lead to a different interpretation of the predictive model.298325604.1 - 51 -For example, for a condition that develops very slowly one may want to consider a relatively long number of days since otherwise being able to predict the condition right after the blood draw would imply that the individual already had the condition at the time of draw (and one may care about future prediction rather than a diagnostic test of existing conditions).

[0101] A pre-defined pair of Pre-draw Clearance time and Post-draw Clearance time sets the tone for the long term risk model. One can then carve out a clean subset of the patients present in the UCLA ATLAS for training and evaluation. Long term risk model outputs 2 things: (1) a relative risk score, which is essentially a linear combination of (a subset of) features, indicates the relative risk of a given sample compared to the background population; and (2) a hazard curve (Fig 7), which is an estimated chance of not getting the phenotype of interest across time (days) after methylation blood collection for a given patient.

[0102] To evaluate the model, the inventors use the concordance metric, which measures the probability of predicted event times of two randomly selected individuals have the same relative order as their true event times. In other words, if the model performs well, as indicated by a high concordance score, patients with higher relative risk scores should have the phenotype sooner than those with lower relative risk scores. A naive model that randomly assigns the relative risk score will achieve concordance score 0.5, indicating that it gets the correct order 50% of the time.

[0103] Note that both outputs are relative to the baseline hazard from the portion of data used for the training. Thus, the inventors report the out-of-sample mean and standard deviation (std) of the concordance scores across models learned and evaluated under different folds of the data.

[0104] In some aspects, the model needs an extra constant term, added to all the samples, after computing the linear combination with the provided coefficient and each sample’s CpG methylation ratio. This number can have more to do with model calibration than with the actual discriminate capacity, and can correct shifts that cause the population to have artificially higher risk or lower risk than expected. V.D. Fixed window risk model

[0105] In addition, the inventors also report a model that predicts a slightly different and simpler question: whether a given patient is likely to get certain phenotype within a fixed window. In addition to the Pre-draw Clearance time (days) and the Post-draw Clearance time (days) defined above. The inventors introduce one more concept: Post-draw End time (days): This quantity is denoted as time_ub in all the figures. The inventors aimed to predict298325604.1 - 52 -which patients will present the phenotype at most this number of days after the blood draw. It is not unreasonable to assume the predictive power of methylation patterns declines as the inventors focus on diagnoses that will onset in the more and more distant future. For example, for certain acute condition that develops within a short period of time, one might lose all the predictive power a year after the initial blood collection.

[0106] A pre-defined combination of Pre-draw Clearance time, Post-Draw Clearance time and Post-Draw End time sets the tone for the Fixed window risk model. The model only outputs one number for each patient: the estimated probability of getting the phenotype within the specified fixed window, which is also just a linear combination of (a subset of) features.

[0107] To evaluate the model, the inventors use the ROC auc, which provides an aggregate measure of how well is the discriminative capacity of the model across all possible classification thresholds. Similar to the concordance metrics used in the Long term risk model, if the model performs well, as indicated by a high ROC auc (Fig 6), the estimated probability of getting the phenotype should be higher for a randomly selected true positive patient than a randomly selected negative patient. A naive model achieve the 0.5 ROC auc.

[0108] Note that the Fixed window risk model output is not relative to the population background. Thus, the inventors report the final out-of-sample performance evaluated between the cross-validated predicted outcome (i.e., the concatenated predictor across all folds) and the true outcome. V.E. Diagnosis codes

[0109] Similar to what is described in the original Thompson et al paper. International Classification of Diseases, Ninth Revision (ICD-9) and International Classification of Diseases, Tenth Revision (ICD- 10) codes are a standard set of diagnosis codes, primarily used for billing purposes. While these codes provide a standardized methodology for describing a diagnosis, they are very specific. To map these specific diagnosis codes into meaningful, distinct diseases / traits, Denny et al. aggregated the ICD codes into phenotype codes (phecodes)42,43. Specifically, for each patient, the inventors queried all diagnoses prior to the methylation sample collection date, and used the phecode (version 1.2) mapping to aggregate ICD-9 and ICD- 10 codes to unique, meaningful phenotypes. If a patient’s diagnosis record had both ICD-9 and ICD-10 labels, the ICD-10 to phecode mapping was used instead of the ICD-9 to phecode mapping.298325604.1 - 53 -

[0110] Each phecode was treated as a binary variable, indicating the presence or absence of a relevant diagnosis code between the Post-draw Clearance time and current date for the Long term risk model or between the Post-draw Clearance time and Post-draw End time for the Fixed window risk model 1. V.F. Preprocessing of methylation array data

[0111] The inventors measured methylation data for 831 individuals based on their DNA sampled from whole blood using the EPIC Illumina array. To generate beta-normalized methylation levels at each CpG, the inventors ran the default pipeline of ENmix (1.22.0) on the raw probe data (IDAT files), which performs background correction, RELIC dye bias correction, and RCP probe-type bias adjustment. The inventors removed from the analysis CpGs that coincided with SNP loci as well as CpGs on the sex chromosomes. The inventors also filtered out outlier samples, defined as having a PC score more than 4 standard deviations away from the average PC score in the first two principal components of the CpG data. In the imputation tasks, the inventors removed sites with low variability (standard deviation ≤ 0.02) leading to a total of 269,471 sites used to train the models. V.G. Cell type proportion estimate

[0112] Cell type proportion is estimated by the default function in ENmix. It performs the algorithm of Houseman et al. with purified celltype panels for EPIC array. Only 50 probes (by default) are used for computing the quadratic optimization. V.H. Penalized Cox

[0113] The inventors trained a penalized cox-proportional hazard model with Lasso regularization. The ideal strength of the penalization is determined by cross validation. When training the model, the inventors first standardized the raw input CpG methylation fraction at each site before applying and tuning the strength of the regularization. The standardization procedure forces each CpGs to have constant variance. Thus, the learned effect size can be directly interpreted as the magnitude of the contribution in the final model. The list of CpGs reported are ranked by the absolute value of learned effect size.

[0114] In practice the inventors report two ranked lists of CpGs: a default version which is learned under the optimal regularization strength determined by the cross validation procedure298325604.1 - 54 -in the training dataset, and a leaner version, learned under a slightly more stringent regularization strength set to one standard deviation larger than the optimal one.

[0115] The final evaluation is not done on the concatenation of all folds’ out-of-sample prediction. Instead, due to the nature of a Cox’s model output, in other words, hazards are described relative to the training set, the inventors collect each out-of-sample fold’s prediction and C-index separately and look at the average C-index. V.I. Penalized Logistic regression

[0116] The inventors trained a penalized logistic regression model with Lasso regularization. The ideal strength of the penalization is determined by cross validation. When training the model, the inventors first standardized the raw input CpG methylation fraction at each site before applying and tuning the strength of the regularization. The standardization procedure forces each CpGs to have constant variance. Thus, the learned effect size can be directly interpreted as the magnitude of the contribution in the final model. The list of CpGs reported are ranked by the absolute value of learned effect size.

[0117] In practice the inventors report two ranked lists of CpGs: a default version which is learned under the optimal regularization strength determined by the cross validation procedure in the training dataset, and a leaner version, learned under a slightly more stringent regularization strength set to one standard deviation larger than the optimal one.

[0118] Finally, the inventors collect each out-of-sample fold’s prediction and evaluate the model performance in terms of ROC-AUC on the concatenation of all folds’ out-of-sample prediction. V.J. Methylation Prediction

[0119] In Table 10, shown below, each row represents a model for a health outcome and each column indicates the details of this model and the associated coefficients. Phecodes are codes developed to group electronic health record billing codes into clinically meaningful phenotypes.

[0120] Table 10: Model Performance298325604.1 - 55 -*Includes a post-draw end time of 365 days **ROC-AUC ***PR-AUC VI. Additional Considerations

[0121] Any headers and / or sub-headers between sections and subsections of this document are included solely for the purpose of improving readability and do not imply that features cannot be combined across sections and subsection. Accordingly, sections and subsections do not describe separate embodiments.

[0122] While the present teachings are described in conjunction with various embodiments, it is not intended that the present teachings be limited to such embodiments. On the contrary, the present teachings encompass various alternatives, modifications, and equivalents, as will be appreciated by those of skill in the art. The present description provides preferred exemplary embodiments, and is not intended to limit the scope, applicability or configuration of the disclosure. Rather, the present description of the preferred exemplary embodiments will provide those skilled in the art with an enabling description for implementing various embodiments.

[0123] It is understood that various changes may be made in the function and arrangement of elements without departing from the spirit and scope as set forth in the appended claims. Thus, such modifications and variations are considered to be within the scope set forth in the appended claims. Further, the terms and expressions which have been employed are used as terms of description and not of limitation, and there is no intention in the use of such terms and expressions of excluding any equivalents of the features shown and described or portions thereof, but it is recognized that various modifications are possible within the scope of the invention claimed.

[0124] In describing the various embodiments, the specification may have presented a method and / or process as a particular sequence of steps. However, to the extent that the method298325604.1 - 56 -or process does not rely on the particular order of steps set forth herein, the method or process should not be limited to the particular sequence of steps described, and one skilled in the art can readily appreciate that the sequences may be varied and still remain within the spirit and scope of the various embodiments.

[0125] Some embodiments of the present disclosure include a system including one or more data processors. In some embodiments, the system includes a non-transitory computer readable storage medium containing instructions which, when executed on the one or more data processors, cause the one or more data processors to perform part or all of one or more methods and / or part or all of one or more processes disclosed herein. Some embodiments of the present disclosure include a computer-program product tangibly embodied in a non-transitory machine-readable storage medium, including instructions configured to cause one or more data processors to perform part or all of one or more methods and / or part or all of one or more processes disclosed herein.

[0126] Specific details are given in the present description to provide an understanding of the embodiments. However, it is understood that the embodiments may be practiced without these specific details. For example, circuits, systems, networks, processes, and other components may be shown as components in block diagram form in order not to obscure the embodiments in unnecessary detail. In other instances, well-known circuits, processes, algorithms, structures, and techniques may be shown without unnecessary detail in order to avoid obscuring the embodiments.298325604.1 - 57 -

Claims

CLAIMS What is claimed is:

1. A computer-implemented method for determining a time-to-event of a health outcome for a patient comprising: receiving, by one or more processors, an epigenetic dataset obtained from a patient, wherein the epigenetic dataset includes methylation fractions of specific epigenetic markers; and applying, by the one or more processors, a trained machine learning model to the epigenetic dataset to predict a relative risk score and / or a hazard curve of the time-to-event of the health outcome.

2. The method of claim 1, further including: training the trained machine learning model with epigenetic training datasets obtained from a plurality of study patients, wherein the epigenetic training datasets include a time value before obtaining epigenetic data from a study patient without the health outcome and / or a time value after obtaining epigenetic data from a study patient with the health outcome.

3. The method of claim 1 or 2, wherein the health outcome is developing osteoporosis.

4. The method of any one of claims 1-3, wherein the specific epigenetic markers comprise or consist of at least three CpG location designations identified in Table 1.

5. The method of claim 4, wherein the specific epigenetic markers comprise or consist of at least 3 of the top CpG location designations identified in Table 1.

6. The method of any one of claims 1-5, wherein the epigenetic markers listed in Table 1 are in ranked order for the impact on the relative risk score and / or hazard curve for osteoporosis.

7. The method of any one of claims 1-3, wherein the specific epigenetic markers comprise or consist of the CpG location designations, cg11047325 and cg18022864.298325604.1 - 58 -8. The method of claim 1, wherein the health outcome is developing sepsis.

9. The method of claim 1 or 8, wherein the specific epigenetic markers comprise or consist of the CpG location designations, cg20494891 and cg23608915.

10. The method of claim 1 or 8, wherein the specific epigenetic markers comprise or consist of the CpG location designation, cg20494891.

11. The method of claim 1 or 8, wherein the specific epigenetic markers comprise or consist of at least three CpG location designations identified in Table 2.

12. The method of claim 11, wherein the specific epigenetic markers comprise or consist of at least 3 of the top CpG location designations identified in Table 2.

13. The method of any one of claims 11 or 12, wherein the epigenetic markers listed in Table 2 are in ranked order for the impact on the relative risk score and / or hazard curve for sepsis.

14. The method of claim 1 or 8, wherein the specific epigenetic markers comprise or consist of at least three CpG location designations identified in Table 3.

15. The method of claim 14, wherein the specific epigenetic markers comprise or consist of at least 3 of the top CpG location designations identified in Table 3.

16. The method of any one of claims 14 or 15, wherein the epigenetic markers listed in Table 3 are in ranked order for the impact on the relative risk score and / or hazard curve for sepsis.

17. The method of claim 1 or 8, wherein the specific epigenetic markers comprise or consist of at least three CpG location designations identified in Table 4.

18. The method of claim 17, wherein the specific epigenetic markers comprise or consist of at least 3 of the top CpG location designations identified in Table 4.298325604.1 - 59 -19. The method of any one of claims 17 or 18, wherein the epigenetic markers listed in Table 4 are in ranked order for the impact on the relative risk score and / or hazard curve for sepsis.

20. The method of claim 1 or 8, wherein the specific epigenetic markers comprise or consist of at least three CpG location designations identified in Table 5.

21. The method of claim 20, wherein the specific epigenetic markers comprise or consist of at least 3 of the top CpG location designations identified in Table 5.

22. The method of any one of claims 20 or 21, wherein the epigenetic markers listed in Table 5 are in ranked order for the impact on the relative risk score and / or hazard curve for sepsis.

23. The method of claim 1, wherein the health outcome is a fluid disorder.

24. The method of claim 23, wherein the fluid disorder is an electrolyte disorder.

25. The method of claim 23 or 24, wherein the specific epigenetic markers comprise or consist of at least three CpG location designations identified in Table 6.

26. The method of claim 25, wherein the specific epigenetic markers comprise or consist of at least 3 of the top CpG location designations identified in Table 6.

27. The method of any one of claims 25 or 26, wherein the epigenetic markers listed in Table 6 are in ranked order for the impact on the relative risk score and / or hazard curve for the fluid disorder.

28. The method of claim 23 or 24, wherein the specific epigenetic markers comprise or consist of at least three CpG location designations identified in Table 7.

29. The method of claim 28, wherein the specific epigenetic markers comprise or consist of at least 3 of the top CpG location designations identified in Table 7.298325604.1 - 60 -30. The method of any one of claims 28 or 29, wherein the epigenetic markers listed in Table 7 are in ranked order for the impact on the relative risk score and / or hazard curve for the fluid disorder.

31. The method of claim 1, wherein the health outcome is an iron deficiency anemia.

32. The method of claim 1 or 31, wherein the specific epigenetic markers comprise or consist of at least three CpG location designations identified in Table 8.

33. The method of claim 32, wherein the specific epigenetic markers comprise or consist of at least 3 of the top CpG location designations identified in Table 8.

34. The method of any one of claims 32 or 33, wherein the epigenetic markers listed in Table 8 are in ranked order for the impact on the relative risk score and / or hazard curve for the iron deficiency anemia.

35. The method of claim 1 or 31, wherein the specific epigenetic markers comprise or consist of at least three CpG location designations identified in Table 9.

36. The method of claim 35, wherein the specific epigenetic markers comprise or consist of at least 3 of the top CpG location designations identified in Table 9.

37. The method of any one of claims 35 or 36, wherein the epigenetic markers listed in Table 9 are in ranked order for the impact on the relative risk score and / or hazard curve for the iron deficiency anemia.

38. The method of any one of claims 2-37, wherein the time before obtaining the epigenetic data from a training set patient in which the training set patient did not have the health outcome is at least, at most, or exactly 90 days, 180 days, 365 days, or any range derivable therein.

39. The method of any one of claims 2-38, wherein the time after obtaining the epigenetic data from the training set patient in which the training set patient was diagnosed with the298325604.1 - 61 -health outcome is at least, at most, or exactly 10 days, 14 days, 30 days, 60 days, or any range derivable therein.

40. The method of any one of claims 1-39, wherein the analyzing step comprises using a cox survival model.

41. The method of claim 40, wherein the cox survival model is a penalized cox survival model.

42. The method of any one of claims 1-41, wherein the analyzing step comprises using a logistic regression.

43. The method of claim 42, wherein the logistic regression is a penalized logistic regression.

44. The method of any one of claims 1-43, wherein the analyzing step comprises using a trained neural network, a trained non-linear survival analysis model, a trained Bayesian time- to-event model, or a trained linear or non-linear time series model.

45. A method for treating a patient with a therapeutic intervention comprising receiving, at a processor, a first set epigenetic of data from the patient collected at a time before the patient has been given an administration of the therapeutic intervention; analyzing the received first set of epigenetic data via the processor, using a trained machine learning model, to generate a first relative risk score and / or a first hazard curve based on the methylation fraction of specific epigenetic markers; receiving, at a processor, a second set of epigenetic data from the patient collected at a time after the patient has been given the administration of the therapeutic intervention; analyzing the received second set of epigenetic data via the processor, using the trained machine learning model, to generate a second relative risk score and / or a second hazard curve based on the methylation fraction of specific epigenetic markers;298325604.1 - 62 -comparing the first and second relative risk scores and / or first and second hazard curves to generate a comparison; and administering an effective amount of the therapeutic intervention, wherein the effective amount is determined based on the comparison.

46. The method of claim 45, wherein the trained machine learning model is trained by tuning training data including a time before obtaining the epigenetic data from a training set patient in which the training set patient did not have the health outcome and / or a time after obtaining the epigenetic data from the training set patient in which the training set patient was diagnosed with the health outcome.

47. The method of claim 45, wherein the therapeutic intervention comprises one or more of a small molecule, a biologic, radiation, phototherapy, heat therapy, cold therapy, surgery, exercise, a diet change, psychotherapy, or a combination thereof.

48. The method of claim 47, wherein the biologic comprises one or more proteins, carbohydrates, nucleic acids, lipids, cells, vaccines, hormones, and / or metabolites.

49. The method of any one of claims 45-48, wherein the patient has, or is suspected of having, a disease.

50. The method of claim 49 wherein the disease is a cancer, an infection, an immune disease, a lung disease, a brain disease, a heart disease, a skin disease, a kidney disease, a liver disease, a gastrointestinal disease, a bone disease, a blood disease, a ligament disease, a nerve disease, a bladder disease, a pancreas disease, a gallbladder disease, a reproductive- organ disease, and / or a muscular disease.

51. The method of any one of claims claim 45-50, wherein the health outcome is developing osteoporosis.

52. The method of any one of claims 45-51, wherein the specific epigenetic markers comprise or consist of at least three CpG location designations identified in Table 1.298325604.1 - 63 -53. The method of claim 52, wherein the specific epigenetic markers comprise or consist of at least 3 of the top CpG location designations identified in Table 1.

54. The method of any one of claims 52-53, wherein the epigenetic markers listed in Table 1 are in ranked order for the impact on the relative risk score and / or hazard curve for osteoporosis.

55. The method of any one of claims 45-51, wherein the specific epigenetic markers comprise or consist of the CpG location designations, cg11047325 and cg18022864.

56. The method of any one of claims 51-55, wherein the therapeutic intervention comprises a bisphosphonate, denosumab, teriparatide, abaloparatide, romosozumab, and / or a lifestyle change.

57. The method of any one of claims 45-50, wherein the health outcome is developing sepsis.

58. The method of claim 45 or 57, wherein the specific epigenetic markers comprise or consist of the CpG location designations, cg20494891 and cg23608915.

59. The method of claim 45 or 57, wherein the specific epigenetic markers comprise or consist of the CpG location designation, cg20494891.

60. The method of claim 45 or 57, wherein the specific epigenetic markers comprise or consist of at least three CpG location designations identified in Table 2.

61. The method of claim 60, wherein the specific epigenetic markers comprise or consist of at least 3 of the top CpG location designations identified in Table 2.

62. The method of any one of claims 60 or 61, wherein the epigenetic markers listed in Table 2 are in ranked order for the impact on the relative risk score and / or hazard curve for sepsis.298325604.1 - 64 -63. The method of claim 45 or 57, wherein the specific epigenetic markers comprise or consist of at least three CpG location designations identified in Table 3.

64. The method of claim 63, wherein the specific epigenetic markers comprise or consist of at least 3 of the top CpG location designations identified in Table 3.

65. The method of any one of claims 63 or 64, wherein the epigenetic markers listed in Table 3 are in ranked order for the impact on the relative risk score and / or hazard curve for sepsis.

66. The method of claim 45 or 57, wherein the specific epigenetic markers comprise or consist of at least three CpG location designations identified in Table 4.

67. The method of claim 66, wherein the specific epigenetic markers comprise or consist of at least 3 of the top CpG location designations identified in Table 4.

68. The method of any one of claims 66 or 67, wherein the epigenetic markers listed in Table 4 are in ranked order for the impact on the relative risk score and / or hazard curve for sepsis.

69. The method of claim 45 or 57, wherein the specific epigenetic markers comprise or consist of at least three CpG location designations identified in Table 5.

70. The method of claim 69, wherein the specific epigenetic markers comprise or consist of at least 3 of the top CpG location designations identified in Table 5.

71. The method of any one of claims 69 or 70, wherein the epigenetic markers listed in Table 5 are in ranked order for the impact on the relative risk score and / or hazard curve for sepsis.

72. The method of any one of claims 57-71, wherein the therapeutic intervention comprises an antibiotic and / or intravenous fluids.298325604.1 - 65 -73. The method of claim 72, wherein the antibiotic comprises a broad-spectrum antibiotic.

74. The method of claim 45, wherein the health outcome is a fluid disorder.

75. The method of claim 74, wherein the fluid disorder is an electrolyte disorder.

76. The method of claim 74 or 75, wherein the specific epigenetic markers comprise or consist of at least three CpG location designations identified in Table 6.

77. The method of claim 76, wherein the specific epigenetic markers comprise or consist of at least 3 of the top CpG location designations identified in Table 6.

78. The method of any one of claims 76 or 77, wherein the epigenetic markers listed in Table 6 are in ranked order for the impact on the relative risk score and / or hazard curve for the fluid disorder.

79. The method of claim 74 or 75, wherein the specific epigenetic markers comprise or consist of at least three CpG location designations identified in Table 7.

80. The method of claim 79, wherein the specific epigenetic markers comprise or consist of at least 3 of the top CpG location designations identified in Table 7.

81. The method of any one of claims 79 or 80, wherein the epigenetic markers listed in Table 7 are in ranked order for the impact on the relative risk score and / or hazard curve for the fluid disorder.

82. The method of any one of claims 74-81, wherein the therapeutic intervention comprises fluids, electrolyte supplements, and / or a lifestyle change.

83. The method of claim 45, wherein the health outcome is an iron deficiency anemia.

84. The method of claim 45 or 83, wherein the specific epigenetic markers comprise or consist of at least three CpG location designations identified in Table 8.298325604.1 - 66 -85. The method of claim 84, wherein the specific epigenetic markers comprise or consist of at least 3 of the top CpG location designations identified in Table 8.

86. The method of any one of claims 84 or 85, wherein the epigenetic markers listed in Table 8 are in ranked order for the impact on the relative risk score and / or hazard curve for the iron deficiency anemia.

87. The method of claim 45 or 83, wherein the specific epigenetic markers comprise or consist of at least three CpG location designations identified in Table 9.

88. The method of claim 87, wherein the specific epigenetic markers comprise or consist of at least 3 of the top CpG location designations identified in Table 9.

89. The method of any one of claims 87 or 88, wherein the epigenetic markers listed in Table 9 are in ranked order for the impact on the relative risk score and / or hazard curve for the iron deficiency anemia.

90. The method of any one of claims 83-89, wherein the therapeutic intervention comprises iron supplements.

91. The method of any one of claims 46-90, wherein the time before obtaining the epigenetic data from a training set patient in which the training set patient did not have the health outcome is at least, at most, or exactly 90 days, 180 days, 365 days, or any range derivable therein.

92. The method of any one of claims 46-91, wherein the time after obtaining the epigenetic data from the training set patient in which the training set patient was diagnosed with the health outcome is at least, at most, or exactly 10 days, 14 days, 30 days, 60 days, or any range derivable therein.

93. The method of any one of claims 45-92, wherein the analyzing step comprises using a cox survival model.298325604.1 - 67 -94. The method of claim 93, wherein the cox survival model is a penalized cox survival model.

95. The method of any one of claims 45-94, wherein the analyzing step comprises using a logistic regression.

96. The method of claim 95, wherein the logistic regression is a penalized logistic regression.

97. The method of any one of claims 45-96, wherein the analyzing step comprises using a trained neural network, a trained non-linear survival analysis model, a trained Bayesian time- to-event model, or a trained linear or non-linear time series model.

98. A method for treating a subject predicted to develop osteoporosis within a set time-to- event, the method comprising: administering a therapeutically effective dose of an osteoporosis therapy to the subject, wherein the subject has been determined to have a relative risk score for developing osteoporosis within a time-to-event that exceeds a pre-determined threshold relative risk score based on a trained machine learning model that determines time-to-event for developing osteoporosis of subjects using methylation fractions of specific epigenetic markers associated with osteoporosis.

99. The method of claim 98, wherein the specific epigenetic markers associated with osteoporosis comprise or consist of at least three CpG location designations identified in Table 1.298325604.1 - 68 -100. The method of claim 99, wherein the specific epigenetic markers associated with osteoporosis comprise or consist of at least of the top CpG location designations identified in Table 1.

101. The method of any one of claims 98-100, wherein the epigenetic markers listed in Table 1 are in ranked order for the impact on the relative risk score and / or hazard curve for osteoporosis.

102. The method of claim 98, wherein the specific epigenetic markers comprise or consist of the CpG location designations, cg11047325 and cg18022864.

103. A method for treating a subject predicted to develop sepsis within a set time-to-event, the method comprising: administering a therapeutically effective dose of antibacterial therapy to the subject, wherein the subject has been determined to have a relative risk score for developing sepsis within a time-to-event that exceeds a pre-determined threshold relative risk score based on a trained machine learning model that determines time-to-event for developing sepsis of subjects using methylation fractions of specific epigenetic markers associated with sepsis.

104. The method of claim 103, wherein the specific epigenetic markers associated with sepsis comprise or consist of cg20494891 and cg23608915.

105. The method of claim 103, wherein the specific epigenetic markers associated with sepsis comprise or consist of cg20494891.298325604.1 - 69 -106. The method of claim 103, wherein the specific epigenetic markers comprise or consist of at least three CpG location designations identified in Table 2.

107. The method of claim 106, wherein the specific epigenetic markers comprise or consist of at least 3 of the top CpG location designations identified in Table 2.

108. The method of any one of claims 106 or 107, wherein the epigenetic markers listed in Table 2 are in ranked order for the impact on the relative risk score and / or hazard curve for sepsis.

109. The method of claim 103, wherein the specific epigenetic markers comprise or consist of at least three CpG location designations identified in Table 3.

110. The method of claim 109, wherein the specific epigenetic markers comprise or consist of at least 3 of the top CpG location designations identified in Table 3.

111. The method of any one of claims 109 or 110, wherein the epigenetic markers listed in Table 3 are in ranked order for the impact on the relative risk score and / or hazard curve for sepsis.

112. The method of claim 103, wherein the specific epigenetic markers comprise or consist of at least three CpG location designations identified in Table 4.

113. The method of claim 112, wherein the specific epigenetic markers comprise or consist of at least 3 of the top CpG location designations identified in Table 4.298325604.1 - 70 -114. The method of any one of claims 112 or 113, wherein the epigenetic markers listed in Table 4 are in ranked order for the impact on the relative risk score and / or hazard curve for sepsis.

115. The method of claim 103, wherein the specific epigenetic markers comprise or consist of at least three CpG location designations identified in Table 5.

116. The method of claim 112, wherein the specific epigenetic markers comprise or consist of at least 3 of the top CpG location designations identified in Table 5.

117. The method of any one of claims 112 or 113, wherein the epigenetic markers listed in Table 5 are in ranked order for the impact on the relative risk score and / or hazard curve for sepsis.

118. A method for treating a subject predicted to develop an electrolyte disorder within a set time-to-event, the method comprising: administering a therapeutically effective dose of an electrolyte disorder therapy to the subject, wherein the subject has been determined to have a relative risk score for developing an electrolyte disorder within a time-to-event that exceeds a pre-determined threshold relative risk score based on a trained machine learning model that determines time-to-event for developing an electrolyte disorder of subjects using methylation fractions of specific epigenetic markers associated with an electrolyte disorder.298325604.1 - 71 -119. The method of claim 118, wherein the specific epigenetic markers comprise or consist of at least three CpG location designations identified in Table 6.

120. The method of claim 119, wherein the specific epigenetic markers comprise or consist of at least 3 of the top CpG location designations identified in Table 6.

121. The method of any one of claims 119 or 120, wherein the epigenetic markers listed in Table 6 are in ranked order for the impact on the relative risk score and / or hazard curve for an electrolyte disorder.

122. The method of claim 118, wherein the specific epigenetic markers comprise or consist of at least three CpG location designations identified in Table 7.

123. The method of claim 122, wherein the specific epigenetic markers comprise or consist of at least 3 of the top CpG location designations identified in Table 7.

124. The method of any one of claims 122 or 123, wherein the epigenetic markers listed in Table 7 are in ranked order for the impact on the relative risk score and / or hazard curve for an electrolyte disorder.

125. A method for treating a subject predicted to develop anemia within a set time-to- event, the method comprising: administering a therapeutically effective dose of an anemia therapy to the subject, wherein the subject has been determined to have a relative risk score for developing an electrolyte disorder within a time-to-event that exceeds a pre-determined threshold relative298325604.1 - 72 -risk score based on a trained machine learning model that determines time-to-event for developing anemia of subjects using methylation fractions of specific epigenetic markers associated with anemia.

126. The method of claim 125, wherein the specific epigenetic markers comprise or consist of at least three CpG location designations identified in Table 8.

127. The method of claim 126, wherein the specific epigenetic markers comprise or consist of at least 3 of the top CpG location designations identified in Table 8.

128. The method of any one of claims 126 or 127, wherein the epigenetic markers listed in Table 8 are in ranked order for the impact on the relative risk score and / or hazard curve for anemia.

129. The method of claim 125, wherein the specific epigenetic markers comprise or consist of at least three CpG location designations identified in Table 9.

130. The method of claim 129, wherein the specific epigenetic markers comprise or consist of at least 3 of the top CpG location designations identified in Table 9.

131. The method of any one of claims 129 or 130, wherein the epigenetic markers listed in Table 9 are in ranked order for the impact on the relative risk score and / or hazard curve for anemia.298325604.1 - 73 -132. A method for treating a patient with a therapeutic intervention, the method comprising: administering a therapeutically effective dose of the therapeutic intervention to the subject, wherein the subject has been determined to have a relative risk score for developing an health outcome within a time-to-event that exceeds a pre-determined threshold relative risk score based on a trained machine learning model that determines time-to-event for developing the health outcome of subjects using methylation fractions of specific epigenetic markers associated with anemia.

133. The method of claim 132, wherein the trained machine learning model is trained by tuning training data including a time before obtaining the epigenetic data from a training set patient in which the training set patient did not have the health outcome and / or a time after obtaining the epigenetic data from the training set patient in which the training set patient was diagnosed with the health outcome.

134. The method of claim 132 or 133, wherein the therapeutic intervention comprises one or more of a small molecule, a biologic, radiation, phototherapy, heat therapy, cold therapy, surgery, exercise, a diet change, psychotherapy, or a combination thereof.

135. The method of claim 134, wherein the biologic comprises one or more proteins, carbohydrates, nucleic acids, lipids, cells, vaccines, hormones, and / or metabolites.

136. The method of any one of claims 132-135, wherein the patient has, or is suspected of having, a disease.298325604.1 - 74 -137. The method of claim 136 wherein the disease is a cancer, an infection, an immune disease, a lung disease, a brain disease, a heart disease, a skin disease, a kidney disease, a liver disease, a gastrointestinal disease, a bone disease, a blood disease, a ligament disease, a nerve disease, a bladder disease, a pancreas disease, a gallbladder disease, a reproductive- organ disease, and / or a muscular disease.

138. A system for determining a time-to-event of a health outcome, the system comprising: a processor; and memory storing instructions that, when executed by the processor, cause the processor to: receive an epigenetic dataset obtained from a patient, wherein the epigenetic dataset includes methylation fractions of specific epigenetic markers; and apply, by the one or more processors, a trained machine learning model to the epigenetic dataset to predict a relative risk score and / or a hazard curve of the time-to-event of the health outcome.

139. A system for determining a time-to-event of osteoporosis, the system comprising: a processor; and memory storing instructions that, when executed by the processor, cause the processor to: receive an epigenetic dataset obtained from a patient, wherein the epigenetic dataset includes methylation fractions of specific epigenetic markers; and apply, by the one or more processors, a trained machine learning model to the epigenetic dataset to predict a relative risk score and / or a hazard curve of the time-to-event of osteoporosis.

140. A system for determining a time-to-event of sepsis, the system comprising: a processor; and298325604.1 - 75 -memory storing instructions that, when executed by the processor, cause the processor to: receive an epigenetic dataset obtained from a patient, wherein the epigenetic dataset includes methylation fractions of specific epigenetic markers; and apply, by the one or more processors, a trained machine learning model to the epigenetic dataset to predict a relative risk score and / or a hazard curve of the time-to-event of sepsis.

141. A system for determining a time-to-event of an electrolyte disorder, the system comprising: a processor; and memory storing instructions that, when executed by the processor, cause the processor to: receive an epigenetic dataset obtained from a patient, wherein the epigenetic dataset includes methylation fractions of specific epigenetic markers; and apply, by the one or more processors, a trained machine learning model to the epigenetic dataset to predict a relative risk score and / or a hazard curve of the time-to-event of the electrolyte disorder.

142. A system for determining a time-to-event of anemia, the system comprising: a processor; and memory storing instructions that, when executed by the processor, cause the processor to: receive an epigenetic dataset obtained from a patient, wherein the epigenetic dataset includes methylation fractions of specific epigenetic markers; and apply, by the one or more processors, a trained machine learning model to the epigenetic dataset to predict a relative risk score and / or a hazard curve of the time-to-event of anemia.298325604.1 - 76 -

Images