Pet phenotypic age and phenotypic age acceleration / deceleration prediction tool based on machine learning

Abstract

The present disclosure provides a system for generating a multi-component senescence index for an individual companion animal, the system including digital biomarkers, biological biomarkers, and subjective assessment methods to predict the phenotypic age and phenotypic age acceleration / deceleration (phenotypic age higher or lower than actual age) of dogs and cats. The disclosure also provides a method for slowing phenotypic aging in a companion animal in need thereof. The method includes determining an index according to one aspect of the present disclosure and providing personalized health, diet, and nutritional measures for a companion animal according to the index of the companion animal. The method can also be considered to be a method for improving accelerated aging of animals. In some embodiments, personalized health measures include diets that decelerate phenotypic aging and / or improve accelerated phenotypic aging.

Description

[0001] Cross-references to related applications

[0002] This application claims the benefit of U.S. Provisional Application No. 63 / 532,360, filed August 12, 2023, entitled “Machine Learning-Based PhenotypicAge and Phenotypic Age Acceleration / deceleration Prediction Tool for Pets,” which is incorporated herein by reference as fully set forth. Background Technology

[0003] This disclosure relates to tools for determining the phenotypic age of companion animals (also referred to as "pets," "dogs," or "cats," and which are used interchangeably herein) and for identifying whether such phenotypic age in an individual companion animal is accelerating or decelerating, as well as for predicting lifespan. This disclosure also relates to a method for slowing phenotypic aging in companion animals in need of this function.

[0004] Pets are important in human lives. They provide companionship, entertainment, and can even improve people's health. However, one drawback of owning a pet is that pets don't live as long as people, and people experience sadness and grief when they lose their pets. Not all pets, even those of the same initial age and breed, have the same predicted lifespan. Furthermore, not all pets age at the same rate; some show signs of aging faster than expected based on their actual age, while others show signs of aging slower. Two pets of the same breed and age can have very different aging trajectories, with one pet in poor health and developing age-related diseases relatively early, while the other is in excellent health and develops age-related diseases relatively late or not at all. Poor health can lead to a shorter lifespan, but pets can also live for many years with age-related diseases, which can limit a pet's physical abilities, cognitive abilities, care requirements, and quality of life during its lifetime. The effects of a pet's aging trajectory also affect pet parents, for example, through increased veterinary and home care requirements, limited ability to perform daily activities, medication costs, and the inability to enjoy activities previously shared. Because of the potential impact of their pets' aging process, pet parents want to provide the best care for their pets and ensure they remain healthy for as long as possible. One obstacle to choosing what is potentially the best care for a pet is a lack of understanding about the aging process. Pet lifespans vary considerably. This is especially true for dogs, some breeds of which may only live six years, while others can live twenty years or more. Lifespans within a breed are also quite variable and are influenced by genetics, healthcare practices, feeding and nutrition, exercise, environment, and other factors.

[0005] The guidelines for senior pets, vaccinations, diagnostic and screening tests, and age-related care differ between pets in their early life stages and those in their later stages. However, defining the senior life stage in dogs requires identifying dogs in the last 25% of their estimated lifespan, a threshold that is difficult for both pet parents and veterinarians to determine. See Bartges, J. et al., “AAHA Canine Life Stage Guidelines” (2012), JAAHA 48:1. Furthermore, dogs are among the most phenotypically diverse mammal species, with lifespans varying considerably by breed in addition to genetic, nutritional, and environmental factors (Ruple, A. et al., “Dog Models of Aging” (2022), Annu. Rev. Anim. Biosci. 10:419-439).

[0006] In cats, the definition of senior life stage is more clearly defined (cats over 10 years of age are considered senior). See Quimby, J. et al., “2021 AAHA / AAFP Feline Life Stage Guidelines” (2021) JAAHA 57:2. However, cat lifespans can vary considerably; the typical lifespan of a cat is reported to be between 13 and 17 years, some cats are known to live past 20 years, and the longest-lived cat on record lived past 38 years. Even within the context of a “typical” lifespan, pets may spend 25–40% of their lives as “senior,” suggesting that chronological age alone may not be sufficient to reflect the aging process. A better understanding of the aging process in specific pets will allow pet parents and veterinarians to provide more targeted and personalized care.

[0007] Some aspects of the aging process may be easier to identify, such as graying fur, but pets don't always exhibit easily identifiable changes in appearance or behavior as they age, thus failing to provide pet parents or veterinarians with clues about their pet's aging process. Furthermore, even when signs appear, pet parents and veterinarians may not recognize the signs of aging in their pets (including dogs and cats) until aging becomes more pronounced. Other aspects of the aging process may be invisible without health metrics. Such health metrics would allow pet parents to understand their pet's current phenotypic age (which may differ from the pet's chronological age) and whether their pet is experiencing an accelerated or decelerated rate of aging compared to expectations based on chronological age. Based on this information, pet parents could make modifications to modifiable factors such as care and nutrition to support their pet's longest and healthiest possible life. Many common parameters, including measures of body composition (such as body condition score or body fat index) and / or weight, disease counts, clinical chemistry blood measures, nutritional status, markers of inflammation and organ function, and blood cell counts and characteristics, can be used alone or in combination with or without chronological age to derive indices that estimate a pet's phenotypic age (i.e., an aging index based on a range of observed characteristics, rather than a measure calculated using only a known date of birth), accelerating / decelerating phenotypic age, aging rate patterns over time (patterns of the difference between phenotypic age and chronological age over time), lifespan, healthy lifespan, life expectancy, health expectations, and longevity. In addition to these parameters, measures such as indicators of physical activity and mobility, pain, sleep, behavior, cognition, location / proximity information, temperature, and subjective pet parent information, as well as less common parameters (including health markers from blood, urine, feces, saliva, skin and mucous membranes, sweat, tears, and tissues, including genetic markers, epigenetic markers, chromosomal changes, gene expression markers, endocrine markers, brain health markers, stress markers, immune function or inflammatory markers, metabolic markers, organ function markers, neurological parameters, body composition parameters, and / or microbiome parameters (such as composition, diversity, function), when used alone or in combination with or without more common parameters in various combinations, can indicate age-related changes well before they can be identified by observers such as pet parents and veterinarians.

[0008] Relatively simple frameworks have been developed for assessing aging beyond chronological age. One example is the DISHA framework (Landsberg, Nichol, and Araujo, Vet Clin Small Anim 42 (2012) 749-768), which pet parents and veterinarians can use to assess signs of cognitive aging in pets. The DISHA framework refers to signs such as disorientation, reduced social interaction, changes in sleep-wake cycles, loss of previous toilet training, increased anxiety, and changes in activity levels. However, even with such frameworks, it remains difficult for some pet parents or veterinarians to identify, observe, and / or quantify the assessed features. For example, pet parents and veterinarians may have difficulty identifying, observing, and / or quantifying changes in sleep-wake cycles, disorientation, or changes in social interaction. Furthermore, while the DISHA framework provides guidance on behavioral signs of cognitive aging, it lacks clear and standardized guidelines for diagnosing cognitive dysfunction in pets. Moreover, cognitive health is only one component of the aging process in companion animals.

[0009] Existing techniques describe the use of a single biomarker or a limited number of biological inputs primarily in age-related studies of pets. None of these techniques utilize larger biological inputs or a combination of clinical, biological, numerical, behavioral, and phenotypic inputs along with machine learning algorithms to predict the phenotypic age of pets. Summary of the Invention

[0010] While others have evaluated the use of physical activity in predicting age / longevity indices or mortality in human populations, incorporating indices, alone or in combination with chronological age, into objective measures of physical and cognitive aging for the purpose of quantifying the aging process in dogs and cats is novel and useful. Furthermore, the development of one or more panels specific to aging dogs and / or cats (e.g., panels that may include chronological age and parameters measured in blood, urine, fecal saliva, skin and mucous membranes, sweat, tears, and / or tissues; panels consisting of health parameters that may include physical activity, mobility, sleep, and other behavioral measures; panels that may include behavioral and location information; panels that may include pain measures, cognitive measures, and other information reported by pet parents, etc.) and the incorporation of one or more of these panels into aging indices will add information that pet parents and veterinarians cannot observe and complement indices based solely on chronological age. The disclosed system can replace or complement currently used, relatively simple frameworks for assessing age-related changes, such as DISHA for cognitive changes.

[0011] This disclosure addresses the lack of knowledge about the accelerated or decelerated phenotypic age of individual pets, patterns of the rate of aging over time (the pattern of difference between phenotypic age and chronological age over time), lifespan, healthy lifespan, life expectancy, health prospect, and longevity by providing companion animals with a tool for predicting phenotypic age. Data from wearable devices (e.g., collar-type accelerometers) in cats and dogs reveal behavioral changes between young and older animals, including differences in the frequency and duration of physical activity, health changes such as mobility challenges, anxiety-related behaviors and patterns of sleep and wakefulness timing, and variability between animals of similar chronological age. As used herein, healthy lifespan refers to the amount of time a subject is healthy without chronic and debilitating diseases, typically measured in years.

[0012] Compared to older participants, younger participants were observed to rest more, sleep less, and spend more time running and walking. See, for example, [link to relevant documentation]. Figure 1 Based on this, it is possible to include these measures in the measurement of aging. By comparing patterns of behavioral data such as running, walking, sleep, and resting using wearable devices across age groups, it has been observed that infants (18 months or younger) differ from adults (18 months–7 years) in sleep, walking, and running, exhibiting longer median durations of walking and running, and shorter median durations of sleep compared to adults; adults differ from older adults (>7 years) in walking, running, and sleep, exhibiting longer median durations of walking and running, and shorter median durations of sleep compared to older adults; and infants differ from older adults on all assessed measures, exhibiting longer median durations of resting, walking, and running, and shorter median durations of sleep compared to older adults. This suggests that different life stage categories are characterized by different behavioral patterns.

[0013] Subjects categorized as having mobility issues differed from healthy adult and healthy older dogs. Dogs with mobility issues differed from those without mobility issues in running and walking behaviors, exhibiting a shorter time spent running, a lower median duration of running behavior, and a lower median duration of walking behavior compared to dogs without mobility issues. See also Figure 2Dogs with mobility issues differ from healthy adult dogs in their walking and running behaviors, exhibiting a lower median duration of walking, a lower median duration of running, and a smaller range of running compared to healthy adult dogs. The median duration of walking and running behaviors in dogs with mobility issues is similar to that observed in healthy older dogs; however, healthy older dogs show more variability in the duration of running behaviors compared to dogs with mobility issues. Many older dogs may suffer from unidentified and undiagnosed mobility problems, such as osteoarthritis, which may instead be attributed to the natural aging process. See also Figure 3 Therefore, comparisons between healthy adult dogs, “healthy” senior dogs, and dogs identified as having mobility problems allow for the establishment of measures that can improve the identification of mobility problems in dogs that have not yet been diagnosed. For both walking and running, median duration of these behaviors was very similar between healthy senior dogs and dogs with mobility problems, but healthy senior dogs showed a much larger range of time spent running compared to dogs with mobility problems, and dogs with mobility problems showed a larger range of time spent walking compared to healthy senior dogs. Healthy adult dogs showed higher median walking and running durations compared to healthy senior dogs, and a larger range of walking durations compared to either healthy senior dogs or dogs with mobility problems. Since mobility problems become increasingly common with age, health measures that can highlight the development of mobility problems early, before they become clinically significant or begin to interfere with daily life, provide the best opportunity for intervention care and nutritional changes to support the healthy mobility of aging pets for as long as possible. In this process, measures that can differentiate groups based on health-related behavioral changes that have not yet been identified by pet parents or diagnosed by a veterinarian are an essential tool.

[0014] Anxiety is a broad marker that may be associated with characteristics of aging processes such as disorientation, reduced social interaction, changes in sleep-wake cycles, regression of previous toilet training, increased anxiety, and changes in activity levels as measured by the DISHA (Landsberg, Nichol, and Araujo, Vet Clin Small Anim 42 (2012) 749-768). Behavioral patterns in adult dogs classified as anxious but otherwise healthy were compared to those in healthy adult dogs classified as not anxious. Anxious dogs appeared to sleep slightly more and rest slightly less throughout the day. Figure 4 A and Figure 4 b). Further analysis of the timing of sleep behavior revealed that this difference was primarily observed during the night, particularly during resting behavior ( Figure 5 A and Figure 5B), because daytime sleep and rest rates are very close ( Figure 6 A and Figure 6 B).

[0015] In one aspect, this disclosure relates to a system for generating a multi-component aging index for an individual companion animal to predict phenotypic age and phenotypic age acceleration / deceleration in dogs and cats based on at least one of measuring digital biomarkers, conventional biomarkers, and subjective assessment methods. The system may optionally also include determining at least one of sex, sterilization status, and life stage.

[0016] In some embodiments, the biomarker panel includes two or more conventional biomarkers selected from: CBC / chemical parameters, fecal microbiome, fecal metabolites, urinary microbiome, urinary metabolites, blood metabolites and blood biomarkers (including albumin, creatinine, glucose, lymphocyte percentage, mean cell volume, erythrocyte distribution width, alkaline phosphatase, white blood cell count, SDMA, circulating peptides including Aβ42, post-circulating biotin, immunoglobulins, immunoglobulin M, growth hormone (GH) / insulin-1 growth factor-1 (IGF-1)) as well as DNA biomarkers, SNPs and genetic variations.

[0017] In some embodiments, the system further includes a system comprising at least one wearable device. In some embodiments, the wearable device measures physical activity. In some embodiments, the measured physical activity includes at least one of walking, running, resting, jumping, sleep duration, sleep quality, and sleep patterns.

[0018] In some embodiments, the system also includes a subjective assessment. In some embodiments, the subjective assessment is conducted via at least one of a pet parent questionnaire and a veterinary questionnaire.

[0019] In some embodiments, the system further includes clinical characteristics. In some embodiments, clinical characteristics include at least one of actual age, weight, BCS, BFI, temperature, respiratory rate, and heart rate.

[0020] In some embodiments, the system further includes one or more environmental sensors. In some embodiments, the one or more environmental sensors include one or more sensors that detect at least one of location, location-based behavior, and activity. In some embodiments, activity includes at least one of proximity to the pet's parent, play, timing and frequency of feeding, location of eating, drinking, urination and defecation, body posture, posture estimation, tail position, body position, and motion tracking over time.

[0021] In some embodiments, the system also includes repeatability measures of at least one of clinical, digital, and biological data.

[0022] In some embodiments, the system further includes at least one of the following modes: eating, drinking, urinating, and defecating.

[0023] In some embodiments, the system also includes signs of at least one of emotional health, cognitive health, fear, anxiety, stress, dementia, social interaction with humans, and interaction with other animals.

[0024] In some embodiments, the system further includes a veterinary assessment. In some embodiments, the veterinary assessment includes at least one of the following: gastrointestinal diseases, genitourinary diseases, kidney diseases, skin diseases, respiratory diseases, neurological diseases, muscle diseases, ophthalmic diseases, hearing diseases, cardiovascular diseases, cancer, oral health, endocrine diseases, infectious diseases, immune function, inflammation, orthopedic diseases, mobility impairment, and pain.

[0025] In one aspect, this disclosure relates to a method for mitigating phenotypic aging in companion animals in need, comprising determining an index using the system or method described herein and, on that basis, providing the companion animal with at least one of a customized measure of health, diet, or nutrition.

[0026] In one aspect, this disclosure provides a system for generating a multi-component aging index for individual companion animals, the system comprising digital biomarkers, conventional (biological) biomarkers, a range of health parameters, and subjective assessment methods to predict phenotypic age (i.e., an aging index based on a range of observed characteristics, rather than a measure calculated using only known birth dates), accelerated / decelerated phenotypic age, patterns of aging rates over time (patterns of the difference between phenotypic age and chronological age over time), lifespan, healthy lifespan, life expectancy, health expectations, and longevity in dogs and cats. Multi-component aging indices include, but are not limited to, measurements from wearable devices of specific forms of physical activity and / or mobility and / or behavior, measured individually or in combination, including but not limited to walking, running, resting, jumping, scratching, shaking, licking, chewing, sniffing or ground tracking, exploring, tail wagging, playing bowing, panting, circling, climbing (e.g., climbing stairs), digging, vocalization, abnormal gait, pacing, eating, drinking, urinating, defecating, vomiting, pain, lameness, reproductive and / or mating behaviors, seizures, sleep health (including but not limited to sleep duration, sleep timing, variability in sleep duration, variability in sleep quality, and regularity in sleep timing), subjective assessment methods (such as pet parent questionnaires and veterinary questionnaires), and clinical characteristics, including but not limited to chronological age, body composition (such as body condition score, body fat index, or objectively measured body composition parameters), weight, temperature, respiration, heart rate, pulse, sex, neutering status, life stage, and disease count.

[0027] In some embodiments, the index also utilizes environmental sensors to assess location and proximity, thereby assessing location / proximity-based behaviors and activities, including but not limited to health signs such as intake and appropriate or inappropriate output behaviors (such as eating, drinking, urinating, defecating, vomiting, and hairballs), illness signs such as gastrointestinal diseases, urogenital and / or kidney diseases, skin diseases, respiratory diseases, nervous system diseases, muscle diseases, eye diseases, hearing diseases, cardiovascular diseases, cancer, oral health, endocrine diseases, infectious diseases, immune function and / or inflammation, and orthopedic diseases and / or mobility / pain, and signs of emotional / cognitive health, including but not limited to fear, anxiety, stress, dementia, cognitive function, attention, and social interactions with humans and other animals.

[0028] In some embodiments, the index also utilizes, alone or in combination, blood, urine, feces, saliva, skin and mucous membranes, sweat, tears, tissue, or other health parameters, such as, but not limited to, clinical blood chemical measurements, nutritional status, markers of inflammation and organ function, and panels of blood cell counts, genetic markers, epigenetic markers, chromosomal changes, gene expression markers, endocrine markers, brain health markers, stress markers, immune function and / or inflammatory markers, metabolic markers, organ function markers, neurological parameters, and / or microbial composition / diversity / functional measures. These parameters may include, but are not limited to, albumin, creatinine, glucose, lymphocyte percentage, mean cell volume, erythrocyte distribution width, alkaline phosphatase, white blood cell count, SDMA, circulating peptides such as Aβ42, post-circulating biotin, immunoglobulins such as immunoglobulin M, growth hormone (GH) / insulin-1 (IGF-1), and DNA biomarkers (including SNPs). In some embodiments, the index further utilizes DNA methylome.

[0029] In some embodiments, the index also incorporates data from several months or years, rather than just a single sample. This data includes, but is not limited to, observations of proximity to pet parents, playtime, feeding timing and frequency, location of eating, drinking, urination and defecation, body posture (e.g., via posture estimation), tail position, body position, motion tracking over time, and the pooling of repeated measures of selected clinical, digital, and biological data using an expanded set of environmental sensors. This provides a phenotypic age index that is continuously updated as the pet provides a continuous stream of data through connected health devices as part of a connected health IoT ecosystem. Such devices include, but are not limited to, cameras, microphones, eye-tracking devices, touchscreens, smart trash cans, smart pet beds, smart scales, smart feeders, smart collars or harnesses, smart clothing, smart thermometers, smart toys, smart water fountains, and smart pet gates, in addition to one or more wearable sensors with accelerometer / gyroscope / magnetometer / thermometer / light sensor / Wi-Fi and Bluetooth capabilities and near-field beacons. Machine learning algorithms are implemented in assistive technologies or in the cloud using data recorded by smart devices, including wearable sensors or other smart devices, via a distributed computing topology (“edge computing”).

[0030] Therefore, this disclosure relates to a method for identifying the interaction between measures and likelihood of a pet or subject being healthy or unhealthy, the method comprising: (a) quantifying one or more measures of the pet over a time period; (b) calculating a score corresponding to the measure relative to a control subject; and (c) associating the score with the likelihood of the subject being unhealthy or healthy based on the measure or magnitude of the measure or magnitude.

[0031] This disclosure also relates to a method for determining the age of a subject, the method comprising: (a) quantifying one or more measures; (b) calculating a score corresponding to one or more measures from the subject; and (c) associating the score with the likelihood that the subject has a particular age (e.g., phenotypic age).

[0032] This disclosure also relates to a method for identifying subjects who may respond to treatment for age-related diseases, the method comprising: (a) calculating a score corresponding to one or more measures; and (b) associating the score with the subject’s probability of having a statistically relevant disease, wherein if the score is above a first threshold, the subject may respond to treatment for the age-related disease, and wherein if the score is below the first threshold, the subject is unlikely to respond to treatment for the disease.

[0033] This disclosure also relates to a method for identifying subjects who may respond to a nutritional intervention for mitigating the effects of aging, the method comprising: (a) calculating a score corresponding to one or more measures; and (b) associating the score with the occurrence of age-related changes, wherein if the score is above a first threshold, the subject may respond to the nutritional intervention, and wherein if the score is below the first threshold, the subject is unlikely to respond to the nutritional intervention.

[0034] This disclosure also relates to a method for predicting the likelihood of a subject responding to or not responding to treatment for an age-related disease, the method comprising: (a) compiling a number of metrics from a population that has presented one or more metrics, wherein the population includes the subject; (b) calculating the quantity or frequency of a metric by sensing one or more metrics associated with the age-related disease to predict the age or behavior associated with the age-related disease; (c) calculating a score; (d) associating the score with the likelihood of the subject having the disease based on the metric; and (e) selecting a treatment or intervention for the subject based on the metric or score.

[0035] This disclosure also relates to a method for predicting the likelihood of a subject responding or not responding to a nutritional intervention for mitigating the effects of aging, the method comprising: (a) compiling a number of measures from a population that has presented one or more measures, wherein the population includes the subject; (b) calculating the amount or frequency of a measure by sensing one or more measures associated with the disease to predict disease-associated age or behavior; (c) calculating a score; (d) associating the score with the occurrence of age-related changes based on the measure; and (e) selecting treatment for the subject based on the measure or score.

[0036] This disclosure also relates to a computer software / program product encoded on a computer-readable storage medium, wherein the computer program product includes instructions for: (a) identifying a measure associated with the subject's age; and (b) calculating a score that quantifies the measure.

[0037] This disclosure also relates to a system for predicting the age of a subject (e.g., phenotypic age, lifespan, healthy lifespan, or acceleration / deceleration of aging), the system comprising: (a) a processor operable to perform a program; (b) a memory associated with the processor; (c) a database associated with said processor and said memory; and (d) a program stored in the memory and executable by the processor, the program being operable to: (i) perform an analysis on the subject with a measure of disease-associated impairment; (ii) identify a measure of disease-associated functional impairment; and (iii) calculate a score corresponding to said measure.

[0038] In one aspect, this disclosure relates to a system including a biosensor, at least one computer storage memory, and a controller. In some embodiments, the biosensor includes at least one or all of the following: a solid support comprising an inner cavity and an outer surface; a strip operatively linked to the outer surface of the solid support; and circuitry positioned within the inner cavity. In some embodiments, the circuitry includes at least one of a first position sensor and a first motion sensor. In some embodiments, each of the sensors is in electrical communication with the controller.

[0039] In one aspect, this disclosure relates to a method for determining whether a subject's age is accelerating or decelerating. The method includes measuring one or a combination of activity metrics of the subject over a time period; determining a mobility score of the subject relative to a control subject of the same age; classifying the subject as active if the mobility score is equal to or higher than the control mobility score for the subject's age; or classifying the subject as inactive if the mobility score is lower than the control mobility score for the subject's age.

[0040] In one aspect, this disclosure relates to a method for determining the phenotypic age of a subject, the method comprising measuring one or a combination of activity measures of the subject over a period of time; determining a mobility score of the subject relative to a control subject of the same age; classifying the subject as healthy if the mobility score is equal to or higher than a control mobility score relative to the subject's age; or classifying the subject as unhealthy if the mobility score is lower than a control mobility score relative to the subject's age; and / or (d) determining the subject's age based on the mobility score.

[0041] In one aspect, this disclosure relates to a computer program product encoded on a computer-readable storage medium. The computer program product includes instructions for: receiving data from one or more biosensors on a subject; calculating a mobility score based on the data; and determining the subject's activity level or age based on the mobility score.

[0042] In one aspect, this disclosure relates to a biosensor. The biosensor includes a top and a bottom outer surface separated by a height. The outer surface and the height define an inner cavity including at least one sensor. In some embodiments, the at least one sensor is at least one of: a gyroscope, at least a first pressure sensor, at least a first temperature sensor, and at least a first accelerometer. In some embodiments, the biosensor further includes a controller. In some embodiments, each of the at least one sensor is in electrical communication with the controller via circuitry.

[0043] In one aspect, this disclosure relates to a smart bed. In some embodiments, the smart bed includes a frame, a base, and sensing devices.

[0044] In one aspect, this disclosure relates to a smart room. In some embodiments, the smart room includes at least one of: (1) one or more biosensors as described herein, and (2) one or more smart beds as described herein. Attached Figure Description

[0045] Figure 1 A box plot is shown, comparing activity-related behaviors among healthy dogs in each age group. The box plot displays the median, IQR, minimum, maximum, and outliers, indicated by red crosses.

[0046] Figure 2 A box plot is shown, comparing forward movement behavior between dogs diagnosed with and without mobility difficulties. The box plot displays the median, IQR, minimum, maximum, and outliers.

[0047] Figure 3 Box plots are shown, comparing forward movement behavior among dogs diagnosed with mobility difficulties, older dogs not diagnosed with mobility difficulties, and healthy adult dogs.

[0048] Figure 4 Box plots were shown, comparing sleep and rest rates between anxious and non-anxious subjects over 24 hours.

[0049] Figure 5 Box plots were shown, comparing sleep and rest rates between anxious and non-anxious subjects throughout the night (00:00–06:00).

[0050] Figure 6 Box plots were shown, comparing sleep and rest rates between anxious and non-anxious subjects throughout the day (06:00–00:00).

[0051] Figure 7 An example of a wearable sensor in the form of a lightweight digital tracker worn on a collar is shown.

[0052] Figure 8 It demonstrates how wearable devices can capture three-dimensional motion.

[0053] Figure 9 It shows how 3D motion is translated into behavior.

[0054] Figures 10A-10E The diagram shows the components and activity flow of a collar-mounted motion sensor (CMAS). Figure 10A The external view of CMAS is displayed. Figure 10B The internal view of CMAS is displayed. Figure 10CThis shows a view of the printed circuit board in CMAS. Figure 10D This shows a second view of the printed circuit board in CMAS. Figure 10E A flowchart illustrating the information within CMAS is shown.

[0055] Figure 11A -D shows various views of the updated CMAS utilizing HPN1 technology. HPN1 technology includes a data collection rate for the accelerometer at 104 Hz or optionally 50 Hz or 100 Hz. Additionally, HPN1 technology optionally includes a configurable data collection rate. Sensors with HPN1 technology may also include a gyroscope and / or one or more other specific sensing devices; such as the others listed herein.

[0056] Figure 12A and Figure 12B This demonstrates a new metal collar design that overcomes the problems of breakage and assembly associated with nylon collars.

[0057] Figure 13A and Figure 13B The data displayed is the signal amplitude data from the wearable sensor. Figure 13A The pattern for scratching is displayed. Figure 13B The running mode is displayed.

[0058] Figure 14 This demonstrates how health status can be characterized by multivariate behavioral changes.

[0059] Figure 15 This shows an example of using computer vision to quantify behavior in a group of pets.

[0060] Figure 16A and Figure 16B It demonstrates how computer vision can simplify the complex geometry of detected individuals into simpler shapes, such as ellipses, thereby allowing dogs to orient themselves relative to their environment, objects, and / or each other.

[0061] Figures 17A-17D The representation of audio data is shown, which can characterize vocalizations to quantify the type, timing, duration, or intensity of animal behavior, to quantify the number and type of animals involved in the behavior, and provides context for other measurements such as posture, behavior, and / or interaction captured by wearable devices and / or computer vision.

[0062] Figure 18A and Figure 18B It shows how computer vision can focus on a target area on a pet and look for changes.

[0063] Figures 19A-19D This is an example of a smart bed. Figure 19A and Figure 19B An example of a smart pet bed is shown. Figure 19C A more detailed schematic diagram of an exemplary smart bed is shown. Figure 19D An example of a smart room is shown.

[0064] Figure 20A and Figure 20B A schematic diagram of an exemplary smart toy used to generate exponential data is shown.

[0065] Figure 21 This shows how a cat's activity changes with age.

[0066] Figure 22 It shows the daily activity curve of a "normal" cat.

[0067] Figure 23 Daily activity curves were compared among all cats (regardless of age or health status), healthy adult cats, and adult cats with arthritis.

[0068] Figure 24 It shows the daily activities of a healthy adult cat.

[0069] Figure 25 The daily activities of an older cat with a history of health problems are shown.

[0070] Figure 26 This shows the daily activities of a five-month-old kitten.

[0071] Figure 27 The cat accelerometer in use is shown.

[0072] Figure 28 The intelligent system set up for the cat is shown.

[0073] Figure 29A and Figure 29B It describes conversational metrics for both aged and non-aged individuals.

[0074] Figures 30A to 30J The conversational data of the pet group is described.

[0075] Figure 31 The study design was described in relation to dogs that underwent dietary changes.

[0076] Figure 32 A set of research design endpoints is described.

[0077] Figure 33 A set of metrics is described for measurement on wearable sensors.

[0078] Figure 34 This paper outlines the methodology for research designs that include wearable sensors on subjects.

[0079] Figures 35A to 35F A set of data related to pet activity measurements was described.

[0080] Figures 36A to 36X An embodiment of a wearable sensor incorporated into a collar for a domestic animal is depicted.

[0081] Figure 37 A modified embodiment of a wearable sensor incorporated into a pet collar with HPN1 is depicted.

[0082] Figure 38 The design of a sensor incorporated into a pet's bedding is described.

[0083] Figures 39A-39E An embodiment of a sensor incorporated into pet bedding is described.

[0084] Figure 40A and Figure 40B A set of measurements to be measured is described using an embodiment of sensors incorporated into pet bedding.

[0085] Figure 41A and Figure 41B The circuitry of a sensor embedded in a collar device or an embodiment of a sensor incorporated into pet bedding is described.

[0086] Figure 42 The metric detection system components associated with the devices supporting the network are described.

[0087] Figure 43 The image depicts a web-based application and computer program product that displays one or more metrics.

[0088] Figure 44 The overall survival rate among felines is shown. As shown in Table 11a, 109 out of 721 felines (15.1%) died during the 3-year follow-up period; Kaplan-Meier curves for the feline population as a whole are presented.

[0089] Figure 45A , Figure 45B and Figure 45C Kaplan-Meier curves for the highest and lowest 20% of felines in PhenoAgeAccel (Model 1) are shown.

[0090] Figure 46A , Figure 46B and Figure 46C Kaplan-Meier curves for the highest and lowest 20% of felines in PhenoAgeAccel (Model 2) are shown.

[0091] Figure 47A , Figure 47B and Figure 47C Kaplan-Meier curves for the highest and lowest 20% of felines in PhenoAgeAccel (Model 3) are shown.

[0092] Figure 48A and Figure 48B The relationship between phenotypic age, actual age, and PhenoAgeAccel (Model 1) of felines is shown. Figure 48A The correlation between phenotypic age and chronological age was shown. Figure 48B The PhenoAgeAccel distribution is shown.

[0093] Figure 49A and Figure 49B The relationship between phenotypic age, actual age, and PhenoAgeAccel (Model 2) of felines is shown. Figure 49A The correlation between phenotypic age and chronological age was shown. Figure 49B The PhenoAgeAccel distribution is shown.

[0094] Figure 50A and Figure 50B The relationship between phenotypic age, actual age, and PhenoAgeAccel (Model 3) of felines is shown. Figure 50A The correlation between phenotypic age and chronological age was shown. Figure 50B The PhenoAgeAccel distribution is shown.

[0095] Figure 51 The receiver operating characteristic curves for 2-year mortality are shown using a feline model.

[0096] Figure 52 The overall survival rate among the dogs is shown. As shown in Table 11b, 89 out of 709 dogs (12.6%) died during the 3-year follow-up period; Kaplan-Meier curves for the dogs as a whole.

[0097] Figure 53A , Figure 53B and Figure 53C The Kaplan-Meier curves for the top 20% and bottom 20% of dogs in PhenoAgeAccel (Model 1) are shown.

[0098] Figure 54A , Figure 54B and Figure 54C The Kaplan-Meier curves for the top 20% and bottom 20% of dogs in PhenoAgeAccel (Model 2) are shown.

[0099] Figure 55A , Figure 55B and Figure 55C The Kaplan-Meier curves for the top 20% and bottom 20% of dogs in PhenoAgeAccel (Model 3) are shown.

[0100] Figure 56A and Figure 56B The relationship between phenotypic age, actual age, and PhenoAgeAccel (Model 1) of dogs is shown. Figure 56A The correlation between phenotypic age and chronological age was shown. Figure 56B The PhenoAgeAccel distribution is shown.

[0101] Figure 57A and Figure 57B The relationship between phenotypic age, actual age, and PhenoAgeAccel (Model 2) of dogs is shown. Figure 57A The correlation between phenotypic age and chronological age was shown. Figure 57B The PhenoAgeAccel distribution is shown.

[0102] Figure 58A and Figure 58B The relationship between phenotypic age, actual age, and PhenoAgeAccel (Model 3) of dogs is shown. Figure 58A The correlation between phenotypic age and chronological age was shown. Figure 58B The PhenoAgeAccel distribution is shown.

[0103] Figure 59 The receiver operational characteristic curves for 2-year mortality using a canine model are shown. Detailed Implementation

[0104] For illustrative purposes, the principles of this disclosure are described with reference to various exemplary embodiments thereof. Although certain embodiments of this disclosure have been specifically described herein, those skilled in the art will readily recognize that the same principles are equally applicable to and can be used in other compositions and methods. Before explaining the disclosed embodiments in detail, it should be understood that this disclosure is not necessarily limited in its application to the details of any particular embodiment disclosed. The terminology used herein is for descriptive purposes only and not for limiting purposes.

[0105] As used herein and in the appended claims, the singular forms “a,” “an,” and “the” include plural references unless the context otherwise requires. The singular form of any class of ingredients refers not only to one ingredient in that class but also to mixtures of those ingredients. The terms “a” (or “an”), “one or more,” and “at least one” are used interchangeably herein. The terms “comprising,” “including,” and “having” are used interchangeably. The term “comprising” should be construed as “including, but not limited to.” The term “comprising” should be construed as “including but not limited to.”

[0106] As used throughout, "range" is used as a shorthand to describe each value within the range. Any value within the range can be chosen as the endpoint of the range. Therefore, the range 1-5 specifically includes 1, 2, 3, 4, and 5, as well as subranges such as 2-5, 3-5, 2-3, 2-4, 1-4, etc. When referring to a number, "approximately" means any number within 15% of that number.

[0107] Unless otherwise stated, abbreviations and symbols used in this document have their ordinary meanings. The abbreviation "wt.%" indicates a weight percentage relative to the pet food ingredient. The symbol "°" refers to degrees, such as temperature or angle. The symbols "h", "min", "mL", "nm", and "µm" represent hours, minutes, milliliters, nanometers, and micrometers, respectively. The abbreviation "UV-VIS" refers to a spectrometer or spectroscopy, indicating ultraviolet-visible light. The abbreviation "rpm" indicates revolutions per minute.

[0108] As used herein, in some embodiments, "fundamentally equal" means within a known range associated with an abnormal or normal range for a given measurement, and "fundamentally equal" means that the range of the relevant term is approximately + / - 10% of the equal value, where the equal value will be associated with any metric modified by that term. For example, if the control sample is from a patient with the disease, in some embodiments, fundamentally equal is within + / - 10% of the abnormal range of the abnormal value. If the control sample is from a patient known not to have the measured disease, fundamentally equal is within the normal range of the given metric. In some embodiments, if the subject is a control subject, the probability of having an age-related disease by any method disclosed herein is approximately 1.01 to approximately 2.00 times the probability of having an age-related disease.

[0109] As used herein, the terms “subject,” “individual,” or “patient” are used interchangeably to refer to any animal, including mammals such as mice, rats, other rodents, rabbits, dogs, cats, pigs, cattle, sheep, horses, or primates. In some embodiments, the term “subject” is used throughout the specification to describe the animal from which samples were collected. In some embodiments, the subject is a domestic pet. The term “patient” is used interchangeably for the purpose of analyzing those diseases specific to a particular subject, such as a dog. In some instances in the description of this disclosure, the term “patient” will refer to a human pet that is inactive or active, or a healthy or unhealthy pet. In some embodiments, the subject may be a dog suspected of having an age-related disease or identified as being at risk of developing an age-related disease. In some embodiments, the subject may be diagnosed as resistant to one or more treatments used to treat a symptom or disease troubling the subject. In some embodiments, the subject may be a non-human animal from which measurements were obtained.

[0110] As used herein, the terms “comprising” (and any form of inclusion, such as “comprise”, “comprises”, and “comprised”), “having” (and any form of having, such as “have” and “has”), “including” (and any form of including, such as “includes” and “include)”), or “containing” (and any form of containing, such as “contains” and “contain)”) are inclusive or open-ended and do not exclude additional, unlisted elements or method steps.

[0111] As used herein, the terms “treatment,” “curative,” or “medical” can refer to therapeutic procedures and / or preventative or preventative measures aimed at preventing or mitigating (alleviating) an undesirable physical condition, disease, or symptom, or achieving a beneficial or desired clinical outcome. For the purposes of the embodiments described herein, a beneficial or desired clinical outcome includes, but is not limited to, relief of symptoms; reduction in the severity of the condition, disease, or symptom; a stable (i.e., non-worsening) state of the condition, disease, or symptom; a delayed onset or slowing of the progression of the condition, disease, or symptom; an improvement or relief of the condition, disease, or symptom state (whether partial or complete), whether detectable or undetectable; an improvement in at least one measurable physical parameter that may not necessarily be discernible to the patient; or an enhancement or improvement of the condition, disease, or symptom. Treatment may also include eliciting a clinically significant response without excessive levels of side effects. Treatment also includes prolonging the survival of the pet compared to the expected survival of an untreated pet.

[0112] As used herein, the terms “diagnosis,” “diagnostic stage,” or variations thereof refer to identifying the nature of a physiological condition, disease, or symptom. In some embodiments, diagnosing a subject means identifying whether a subject has an age-related disease. In some embodiments, diagnosis means distinguishing between an unhealthy and a healthy pet or subject.

[0113] As used in this article, a “control sample” or “reference sample” refers to a sample of quantities that are known to exist, not exist, or be measured, and is used to compare with an experimental sample.

[0114] A “score” is a numerical value that can be assigned or generated based on the presence, absence, or quantity of the substrate or enzyme disclosed herein, after normalization of the values. In some embodiments, the score is normalized relative to the original control data values.

[0115] All references cited in this document are incorporated herein by reference. In the event of any conflict between the definitions in this disclosure and the definitions in the cited references, this disclosure shall prevail.

[0116] Detecting signs of aging before they become clinically significant provides opportunities for intervention, such as nutritional or lifestyle changes, which have the potential to alter the aging trajectory. Therefore, early detection through non-invasive means, such as monitoring physical body measures, such as weight, indicators of body composition, such as body condition scores or body fat index, alone or in combination, and other health parameters, including physical activity and mobility, sleep parameters, behavioral parameters such as frequency or duration of behaviors, pain measures, cognitive measures, location and / or proximity information such as frequency and duration of occurrence in specific areas, temperature, and subjective information reported by pet parents and / or veterinarians, provides opportunities to identify age-related changes. These signs can be enhanced by using health biomarkers derived from blood, urine, feces, saliva, skin and mucous membranes, sweat, tears, and tissues, as well as other health parameters (e.g., clinical chemistry parameters, nutritional status markers, blood cell counts, genetic markers, chromosomal changes, gene expression markers, endocrine markers, brain health markers, stress markers, immune function and / or inflammation markers, metabolic markers, organ function markers (e.g., liver, heart, lungs, etc.), neurological parameters, body composition parameters, and / or microbiome parameters (such as composition, diversity, function), etc.), which provide deeper information about the type of age-related changes and the affected systems, and can characterize the degree of age-related changes (acceleration or deceleration). Machine learning methods can be used to develop models capable of identifying behavioral and activity patterns, as well as digital and traditional biomarkers, clinical parameters, and subjective information indicative of aging in dogs and cats reported by pet parents and / or veterinarians. These models can also be applied to predict phenotypic age, accelerated / decelerated phenotypic age, patterns of aging rates over time (patterns of difference between phenotypic age and chronological age over time), lifespan, healthy lifespan, life expectancy, health prediction, and longevity in new dog and cat groups. Digital biomarkers include data collected by the sensors described herein. Non-limiting examples of digital biomarkers include wearable device data, such as total activity, walking time and / or days, running time and / or days, sleep time and / or days, sleep quality, scratching time and / or days, shaking time and / or days, duration of activity events (walking, running, scratching, shaking, sleeping), and data on eating, drinking, urination, defecation, and location and proximity. Other non-limiting examples of digital biomarkers include measures of memory and learning, such as those from digital toys and / or video games for pets (e.g., PupPod). These may include measures of at least one of associative learning, learning speed, task switching, or attention. Further non-limiting examples of digital biomarkers include weight, heart rate, respiratory rate, and temperature from wearable devices or smart pet beds.In some embodiments, digital biomarkers include at least one of running time and / or days, longest duration of uninterrupted running, walking time and / or days, longest duration of uninterrupted walking, stride length, walking speed, sleep duration, sleep duration stability, sleep timing, sleep timing stability, sleep quality, and sleep quality stability. In some embodiments, digital biomarkers include all of running time and / or days, longest duration of uninterrupted running, walking time and / or days, longest duration of uninterrupted walking, stride length, walking speed, sleep duration, sleep duration stability, sleep timing, sleep timing stability, sleep quality, and sleep quality stability.

[0117] Once pet parents and veterinarians have information about where their individual pet falls on the lifespan timeline, how their pet's phenotypic age compares to its actual age, and whether the current rate of aging is faster or slower than expected, and can observe patterns in the rate of aging over time, pet parents and veterinarians can then optimize the care they give their pets and make decisions that may affect the pet's overall lifespan, aging trajectory, and / or quality of life.

[0118] This disclosure relates to tools and systems for determining the phenotypic age of companion animals and identifying whether the phenotypic age of a given pet differs from its chronological age, thereby determining whether the pet's aging is accelerating or decelerating. This disclosure also relates to tools and systems for determining whether the lifespan or healthy lifespan of a given pet differs from what might be expected. This disclosure further relates to a method for mitigating phenotypic aging in companion animals where this is necessary. This disclosure addresses the lack of knowledge about an individual pet's lifespan, life expectancy, healthy lifespan, health expectation, longevity, and rate of aging by providing companion pets with phenotypic age prediction tools. Wearable device data for cats and dogs show behavioral changes between young and older animals, including differences in the amount and timing of physical activity, changes in pathological states such as mobility problems, behaviors such as anxiety, and circadian rhythms or patterns of sleep and wakefulness. In some embodiments, the disclosed metrics relate to the lifespan, life expectancy, healthy lifespan, health expectation, longevity, and rate of aging of an individual subject or group of subjects.

[0119] This disclosure relates to a system, composition, and a series of methods for analyzing samples from subjects, to accurately diagnose, predict, or classify subjects as aging or unhealthy due to age. This disclosure also relates to a system, composition, and a series of methods for analyzing pet activity or behavior from subjects, to accurately diagnose, predict, or classify subjects as aging or having a certain phenotype, age, or life expectancy compared to control subjects or a control group. In some embodiments, the system of the present invention includes means for detecting and / or quantifying or observing data, including measurements of the subject or pet, frequency of behavior or activity; and correlating this data with the subject's medical history to predict clinical outcomes, treatment plans, preventative medical plans, or effective therapies.

[0120] like Figure 1 As shown, younger subjects tended to rest more and sleep less, and spent more time running and walking. Based on this, and given the evidence for this trend in observational studies of canine behavior, these measures can be included in the measurement of aging. Table A shows significant differences between the juvenile group and the adult group in sleep, walking, and rest; significant differences between the adult group and the older group in walking, running, sleep, and rest; and significant differences between the juvenile and older groups across all measured measures.

[0121] Table A: Index data for sleep, resting, and walking.

[0122]

[0123] Given that these behaviors appear to change with age, it is expected that observations at a higher temporal resolution will reflect these changes. By visualizing the daily median values, it is anticipated that these changes will become apparent over a sufficiently long period.

[0124] like Figure 2 As shown, for dogs with mobility issues, the median time spent walking and running was slightly reduced; however, it is possible that some dogs classified as healthy senior dogs may have undiagnosed mobility issues, which are instead attributed to old age, potentially making the differences between groups appear artificially smaller. Many senior dogs suffer from mobility difficulties such as arthritis without being diagnosed. Therefore, comparisons between healthy adult dogs (n=34), “healthy” senior dogs (n=30), and dogs diagnosed with mobility difficulties (n=10) allow for the development of measures to identify undiagnosed dogs. Figure 3As shown, for both walking and running, the median display rates were very similar between healthy older dogs and dogs with mobility issues, although the behavioral expression patterns differed between the two groups. Particularly for running, dogs with mobility issues showed a significantly smaller range of running duration compared to healthy older dogs. Compared to healthy adult dogs, dogs with mobility issues showed a lower median running duration and a significantly smaller range of running duration. Dogs with mobility issues also showed a lower median duration of daily walking behavior and a slightly smaller range of daily walking duration compared to healthy adult dogs. These findings suggest that variations in the median daily duration and / or range of daily running for activity and / or forward movement behaviors such as walking and / or running can be indicators of mobility issues in dogs, helping to differentiate them from healthy dogs. Running behavior may be a particularly sensitive indicator of mobility issues in dogs, and the median daily running duration and / or range of running duration in a dog group may be able to distinguish dogs with mobility issues from healthy older dogs and / or healthy adult dogs.

[0125] Anxiety is a broad marker that may be related to the DISHA measure, such as disorientation, changes in social interaction, alterations in sleep-wake cycles, indoor fecal impaction, and activity levels. Adult dogs were categorized as either anxiety-prone but otherwise healthy (n=6) or healthy but not anxiety-prone (n=26). Figure 4 As shown, anxious dogs appear to sleep more and rest less. Further analysis of sleep behavior timing revealed that this difference was more pronounced at night. Figure 5 ), because daytime sleep and resting periods are more similar in duration ( Figure 6 ).

[0126] In one aspect, this disclosure provides a system for generating a multi-component aging index for individual companion animals, comprising digital biomarkers, conventional (biological) biomarkers, and subjective assessment methods to predict phenotypic age and phenotypic age acceleration / deceleration (phenotypic age higher or lower than actual age), patterns of aging rate over time, and longevity in dogs and cats. Systems used to generate multi-component indices include, but are not limited to, wearable devices for measuring the body and specific forms of physical activity, such as walking, running, resting, jumping, scratching, shaking, licking, chewing, sniffing, or ground tracking, exploring, tail wagging, playing, bowing, panting, circling, climbing (e.g., climbing stairs), digging, vocalizing, abnormal gait, pacing, eating, drinking, urinating, defecating, vomiting, pain, lameness, reproductive and / or mating behaviors, seizures, sleep health (including but not limited to sleep duration, sleep timing, variability in sleep duration, sleep quality, variability in sleep quality, and regularity of sleep timing), subjective assessment methods (such as pet parent questionnaires and veterinary questionnaires), and clinical characteristics, including but not limited to actual age, body composition (such as body condition score, body fat index, or objectively measured body composition parameters), weight, temperature, respiration, heart rate, pulse, and symptom counts.

[0127] In some embodiments, the index also utilizes environmental sensors to assess location, thereby assessing location-based behaviors and activities, including but not limited to health signs such as intake and appropriate or inappropriate output behaviors (such as the occurrence of eating, drinking, urination, defecation, and vomiting), illness signs such as gastrointestinal diseases, urinary and / or kidney diseases, skin diseases, respiratory diseases, nervous system diseases, cardiovascular diseases, oral health, endocrine diseases, infectious diseases, immune function and / or inflammation, and orthopedic diseases and / or mobility / pain, and signs of emotional / cognitive health, including but not limited to fear, anxiety, stress, dementia, cognitive function, attention, and social interactions with humans and other animals.

[0128] In some embodiments, the index also utilizes two, three, four, five, six, seven, eight, nine, or more biomarker panels. Biomarker panels include, but are not limited to, at least one conventional biomarker selected from: clinical chemistry parameters, nutritional status markers, blood cell counts, genetic markers, chromosomal changes, gene expression markers, endocrine markers, brain health markers, stress markers, immune function and / or inflammation markers, metabolic markers, organ function markers (e.g., liver, heart, lung, etc.), neurological parameters, body composition parameters, and / or microbiome parameters (such as composition, diversity, function, etc.). Biomarker panels may include, but are not limited to, at least one conventional biomarker selected from: albumin, creatinine, glucose, lymphocyte percentage, mean corpuscular volume, erythrocyte distribution width, alkaline phosphatase, white blood cell count, SDMA, circulating peptides such as Aβ42, post-circulating biotin, immunoglobulins such as immunoglobulin M, growth hormone (GH) / insulin-1 (IGF-1), and DNA biomarkers (including SNPs). This index may also include physical body measures such as weight, indicators of body composition such as body condition scores or body fat percentages, and other health parameters including physical activity and mobility, sleep parameters, behavioral parameters (such as the frequency or duration of behaviors), pain measures, cognitive measures, location and / or proximity information (such as the frequency and duration of occurrence in specific areas), temperature, and subjective information reported by pet parents. See the following non-limiting examples of the use of traditional biomarkers. The following examples include non-limiting examples of analyses using multiple traditional biomarkers to predict phenotypic age and acceleration / deceleration of aging. They show individual metric differences between surviving and non-surviving cats and dogs over a 3-year observation period. This example uses data from hundreds of dogs and cats from a PNC population observed for many years. It also shows how individual parameters of dogs at least 5 years old varied across different populations. This example shows which blood markers had a high proportion of abnormally high or abnormally low metric outcomes in a population of older dogs. The analysis shows that elevated or decreased values ​​of certain biomarkers may be common in older dogs and indicates an association between the parameter and age when considering one parameter at a time and studying it cross-sectionally (at a certain moment, rather than over several years).

[0129] Traditional biomarkers include common biochemical parameters.

[0130] Common biochemical parameters measured in older dogs show deviations from normal values ​​and may be useful markers of aging. For example, in a population of approximately 485 healthy, non-fasting dogs aged 5 years and older registered at veterinary clinics across the United States, the following distribution of normal and abnormal blood values ​​was observed:

[0131] albumin

[0132] a. High: 2.9%

[0133] b. Low: 0.6%

[0134] c. Normal: 96.5%

[0135] Glucose (non-fasting)

[0136] a. High: 0.0%

[0137] b. Low: 4.3%

[0138] c. Normal: 95.7%

[0139] magnesium

[0140] a. High: 11.1%

[0141] b. Low: 0.4%

[0142] c. Normal: 88.5%

[0143] phosphorus

[0144] a. High: 0.0%

[0145] b. Low: 5.4%

[0146] c. Normal: 94.6%

[0147] AST (aspartate aminotransferase)

[0148] a. High: 0.0%

[0149] b. Low: 5.4%

[0150] c. Normal: 94.6%

[0151] GGT (gamma-glutamyl transferase)

[0152] a. High: 0.8%

[0153] b. Low: 19.8%

[0154] c. Normal: 79.4%

[0155] BUN (blood urea nitrogen)

[0156] a. High: 2.5%

[0157] b. Low: 1.0%

[0158] c. Normal: 96.5%

[0159] TRIG (triglycerides)

[0160] a. High: 18.6%

[0161] b. Low: 0.6%

[0162] c. Normal: 80.8%

[0163] CHOL (cholesterol)

[0164] a. High: 15.3%

[0165] b. Low: 0.0%

[0166] c. Normal: 84.7%

[0167] GLOBU (globulin)

[0168] a. High: 3.7%

[0169] b. Low: 0.0%

[0170] c. Normal: 96.3%

[0171] CO2 (bicarbonate)

[0172] a. High: 0.0%

[0173] b. Low: 3.1%

[0174] c. Normal: 96.9%

[0175] ANGAP (Anion Gap)

[0176] a. High: 0.2%

[0177] b. Low: 16.5%

[0178] c. Normal: 83.3%

[0179] RBCs (red blood cells)

[0180] a. High: 5.6%

[0181] b. Low: 0.0%

[0182] c. Normal: 94.4%

[0183] HCT (hematocrit)

[0184] a. High: 10.3%

[0185] b. Low: 0.0%

[0186] c. Normal: 89.7%

[0187] HGB (hemoglobin)

[0188] a. High: 8.4%

[0189] b. Low: 0.0%

[0190] c. Normal: 91.6%

[0191] MCV (mean corpuscular volume)

[0192] a. High: 10.1%

[0193] b. Low: 4.7%

[0194] c. Normal: 85.2%

[0195] PLT (platelet-rich plasma)

[0196] a. High: 4.3%

[0197] b. Low: 0.4%

[0198] c. Normal: 95.3%

[0199] ALKP (alkaline phosphatase)

[0200] a. High: 6.4%

[0201] b. Low: 6.4%

[0202] c. Normal: 87.2%

[0203] In the embodiments described herein, at least one of the above may be included as a conventional biomarker. For example, in the embodiments described herein, by examining these conventional biomarkers for a given dog, the dog's likely age category can be determined. For example, the normal number of healthy dogs with high magnesium is 0. If a healthy dog ​​is found to have high magnesium, the data from healthy dogs shows that those dogs with high magnesium are disproportionately older dogs. Some older dogs still have normal magnesium, but if a healthy dog ​​is found to have high magnesium, then this dog should be considered to be more likely in the "older" dog category, even if the dog's true age is unknown. Similar reasoning applies to HGB (hemoglobin). All the conventional biomarkers listed above are shown because the distribution of these biomarkers in older healthy dogs differs from what we would expect in younger healthy dogs. Conventional biomarkers also include those in Tables 11a and 11b below. Some embodiments include one or more conventional biomarkers in Tables 11a and / or 11b. Some embodiments include methods or systems utilizing one or more conventional biomarkers in Tables 11a and / or 11b, as in the following examples. In some embodiments, conventional biomarkers for feline subjects include HCT, MCH, WBC, RDW, Bun / Creat, BUN, chloride, sodium, and animal age. In some embodiments, conventional biomarkers for feline subjects include RBC, MCHC, MCH, MCV, RDW, chloride, and triglycerides. In some embodiments, conventional biomarkers for feline subjects include creatinine, glucose, MCV, albumin, RDW, lymphocytes, WBC, ALP, and animal age. In some embodiments, conventional biomarkers for canine subjects include MCHC, MCH, RDW, ALT, albumin, chloride, sodium, eosinophils, lymphocytes, and animal age. In some embodiments, conventional biomarkers for canine subjects include RBC, MCH, RDW, ALT, and Na / K ratio. In some embodiments, conventional biomarkers for canine subjects include creatinine, glucose, MCV, albumin, RDW, lymphocytes, WBC, ALP, and animal age.

[0204] In some embodiments, a conventional biomarker is a biomarker whose distribution in older healthy animals differs from the expected distribution in younger healthy animals.

[0205] In some embodiments, as a supplement to or alternative to traditional biomarkers, the index further utilizes epigenetic markers, including DNA methylome data.

[0206] In some embodiments, the index also incorporates data from several months or years, rather than just a single sample. This data includes, but is not limited to, observations using an extended set of environmental sensors to observe proximity to pet parents, playtime, timing and frequency of feeding, timing and frequency of drinking, timing and frequency of urination, timing and frequency of defecation, location of eating, drinking, urination, and defecation, body posture (e.g., via posture estimation), tail position, body position, motion tracking over time, and the pooling of repeated measurements of selected clinical, digital, and biological data. In some embodiments, this provides a phenotypic age index that is continuously updated as the pet provides a continuous stream of data through connected health devices as part of a connected health IoT ecosystem. Such devices include, but are not limited to, cameras, microphones, eye-tracking devices, touchscreens, smart trash cans, smart pet beds, smart scales, smart feeders, smart collars or harnesses, smart clothing, smart thermometers, smart toys, smart water fountains, and smart pet gates. Machine learning algorithms are implemented in smart devices (including wearable sensors or other smart devices), via distributed computing topologies (“edge computing”), or in the cloud.

[0207] In some embodiments, the systems and / or methods described herein include a combination of at least two, three, four, five, six, seven, eight, or nine biomarkers to diagnose and / or recommend appropriate treatment plans for the respective pet. The biomarkers may be conventional biomarkers. Biomarkers may be selected from those listed in the examples below. For example, in some embodiments, at least two, three, four, five, six, seven, eight, or nine of the biomarkers are selected from the group consisting of creatinine, HCT, MCH, MCHC, RBC, WBC, RDW, Bun / Creat ratio, BUN, chloride, sodium, triglycerides, glucose, MCV, albumin, RDW, lymphocytes, eosinophils, ALT, and Na / K ratio.

[0208] In some embodiments, conventional biomarkers for canine subjects according to the present invention include any combination of at least two, three, four, five, six, seven, eight, and nine of the following: creatinine, glucose, MCV, albumin, RDW, lymphocytes, WBC, basophils, eosinophils, monocytes, neutrophils, ALT, ALP, chloride, cholesterol, globulin, albumin to globulin ratio, phosphorus, magnesium, sodium, potassium, total protein, and triglycerides.

[0209] In some embodiments, the method according to the invention may consider any combination of 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 or 29 to 30, 31, 32, 33, 34, 35, 36, 37, 38, 39 or 40 of any conventional biomarker. In some embodiments, the method according to the invention may consider any combination of 3 to 30, 31, 32, 33, 34, 35, 36, 37, 38, 39 or 40 of any conventional biomarker. Conventional biomarkers may be selected from those described herein. Conventional biomarkers may be selected from those listed above. In some embodiments, the index also utilizes epigenetic markers, including DNA methylome data, as a supplement or alternative to the individual or combination of conventional biomarkers.

[0210] In some embodiments, phenotypic indices may include any combination of two, three, four, five, six, seven, eight, or nine data points over several months or years, rather than a single sample. These data points may be obtained, but are not limited to, by using an extended set of environmental sensors to observe proximity to pet parents, play, timing and frequency of feeding, timing and frequency of drinking, timing and frequency of urination, timing and frequency of defecation, location of eating, drinking, urination, and defecation, body posture, tail position, body position, movement tracking over time, and the pooling of repeated measurements of selected clinical, numerical, and biological data.

[0211] wearable devices

[0212] Wearable sensors are lightweight digital trackers worn on a collar (see...) Figure 7 ). refer to Figure 7 The wearable sensor 701 shown includes a sensor 710 attached to a collar 720. The wearable sensor 710 provides information about the dog's movement, which helps in understanding the dog's health. Figure 8 The image shows a wearable sensor 701 for capturing three-dimensional motion. (Reference) Figure 9 The 3D motion is converted into motion axes, which can include forward / backward, up / down, and left / right. Motion axes can indicate behaviors; for example, sleeping, resting, walking, running, scratching, shaking, etc.

[0213] A collar-mounted motion sensor (CMAS) is a wearable sensor. In some embodiments, the CMAS includes an accelerometer. In some embodiments, the accelerometer is capable of capturing motion data at 100 Hz. (Reference) Figure 10ASensor 1010 is included in the CMAS configuration of some embodiments described herein. Despite its small size, some embodiments of the CMAS may contain 130 electronic components, six-axis motion detection provided by an accelerometer (100 Hz sampled acceleration) and a gyroscope (50 Hz sampled gyroscope), Wi-Fi and BTLE communication, 2GB of memory with 256KB of flash memory, FIFO data output, and approximately 10 days of battery life. See also Figure 10B-10E . refer to Figure 11A-11D Another embodiment of sensor 1110 is shown, which includes a charger and incorporates HPN1 technology. As shown, sensor 1110 includes a USB port 1120, charging contacts 1130, and a light guide 1140. Sensor 1110 can provide 6-axis motion detection using only 59 electronic components, where an accelerometer and gyroscope provide variable sampling in a single component, combined with Wi-Fi and BTLE communication, 8GB flash memory, FIFO data output, approximately 10 days of battery life, proximity detection, smaller footprint and weight, and a dual-neckband attachment option. In some embodiments, the wearable sensor includes... Figure 10A CMAS.

[0214] refer to Figure 12A-12B An embodiment including a durable collar 1220 is shown. The durable collar 1220 may be a detachable collar. The durable collar 1220 can be appropriately sized to suit the body proportions and weight of each pet as needed, providing a safe and practical attachment point for collar-mounted wearable sensors. The durable collar 1220 may include a ball 1240 and a latching mechanism 1230, as well as a sensor mounting base 1250. In one embodiment, the durable collar 1220 is a metal collar design, which addresses the issues of security, practicality, and durability, and allows for a custom fit for each pet.

[0215] The 3D motion data collected using such sensors provides training material for algorithms and the continuous measurements required to quantify behavior. High-resolution data is transmitted in great detail (see...). Figure 13A and Figure 13B Frequent reading can provide easier categorization of behaviors and models, or conclusions drawn from frequent reading may be more sensitive, specific, and accurate. In embodiments, models using multiple variables to predict changes in behavioral patterns can be used to characterize health status. For example, see... Figure 14The diagram illustrates a grid that cross-references a pet's health status, emotional state, and / or cognitive state (e.g., dermatitis, aging, arthritis, food allergies, happiness, anxiety, and / or others) with symptoms for each condition (e.g., scratching, jumping, eating, defecating, wagging tail, sleeping, and / or others). Robust algorithmic validation and performance metrics generate confidence in the health claims (e.g., see Griffies et al., BMC Veterinary Research (2018) 14:124), which is incorporated herein by reference as fully elaborated.

[0216] This disclosure relates to a sensor embedded, attached, or mounted in a collar, and a system including the sensor, wherein the device includes a sensor configured to measure heart rate, electrocardiogram, and pulse, such sensor being positioned within a housing. In some embodiments, this disclosure relates to a device including a sensor configured to measure heart rate, electrocardiogram, pulse, and respiratory rate, and such sensor being positioned within a housing, wherein the housing is physically attached to a collar. In some embodiments, the collar is wearable or designed to be worn by a pet. In some embodiments, the device is operatively connected to a controller, a display, and a computer program product having executable code to perform any of the method steps disclosed herein.

[0217] Computer Vision

[0218] Examples include computer vision, a field of artificial intelligence that enables computers to extract useful information from digital images, videos, or other visual inputs, thereby facilitating vision-based decision-making. One such use is identifying pet behavior from collections of images. Computer vision of pet groups is complex but important for understanding pet behavior. Figure 15 An example is shown. Detecting interactions is the first step in representation. Neural networks can segment regions in an image that contain individuals and further uniquely assign and track each individual. When individuals are close to each other, they can track and represent the interactions between them. For example... Figure 16A and Figure 16B As shown, embodiments include simplifying the complex geometry of detected individuals into ellipses or similar bounding boxes, which allows characterizing the dogs' orientation relative to each other. In some embodiments, pose estimation is included and allows for further refinement of the interaction representation. In some embodiments, tail, head, and ear positions all provide signals between the interacting parties. In some embodiments, the system may also monitor vocalizations to provide context to the interaction (see...). Figure 17A -D). In some embodiments, posture estimation can also be used to characterize body language, which aids in emotional representation and serves as input to phenotypic age models.

[0219] In some embodiments, computer vision is used to monitor parameters such as respiratory rate to aid in the assessment of physical and emotional health. Figure 18A and Figure 18B As shown, one embodiment includes computer vision, which can focus on a target area on a pet and identify changes (e.g., in the case of skin problems). In some embodiments, computer vision can also be used to measure physical characteristics such as length, width, height, perimeter, depth, volume, surface area, weight, and mass. In some embodiments, computer vision can also be used to estimate, score, or predict body composition parameters. Non-limiting examples of body composition parameters include 5-point or 9-point body condition rating scales, body fat index (e.g., a 20-70 scale), muscle condition rating, and body composition parameters. Non-limiting examples of body composition parameters include total mass (kg), fat mass (kg), lean mass (kg), fat mass (kg) / lean mass (kg), % fat mass, % lean mass, body fat distribution, lean mass distribution, water mass, water mass percentage, bone mineral density, bone mineral content, body compartment volume, frame size, and skeletal parameters.

[0220] Smart beds, smart rooms, and smart toys

[0221] Some embodiments include smart pet beds that can be used to provide non-invasive monitoring of a pet's weight, temperature, heart rate and / or pulse, respiratory rate, sleep parameters, audio parameters or vocalizations, and behavioral parameters such as social sleep. (See also...) Figure 19A and Figure 19B This illustrates an embodiment of a smart bed. Figure 19A The smart bed 1901 is shown, which includes legs 1910, a frame 1920, and a mattress 1930 extending within the frame 1920. Figure 19B An embodiment of a smart bed 1940 is shown, which includes a frame 1950 and legs 1960 supported on a base 1970. Detailed views of an exemplary smart bed 1971 are shown in... Figure 19C The smart bed 1971 may include a frame 1972, legs 1973, a base 1974, electronic components 1975, and a mounting bracket 1976 for the electronic components 1975. The smart bed 1971 may also include sensing equipment 1977. Sensing equipment 1977 may include the elements listed in Table B below.

[0222] Table B: Figure 19A and Figure 19B The items shown

[0223]

[0224] like Figure 19DAs shown, some embodiments include a smart room. The smart room may include cameras and an automated pet feeder system to monitor one or more of a pet's food intake and activity, either alone or in conjunction with wearable or other smart devices, and this data may be used to determine the pet's calorie requirements. Some embodiments include smart toys. In some embodiments, the smart toy is used as a data collection tool for detecting health-related endpoints. Exemplary, non-limiting health-related endpoints include cognitive function, signs of aging, physical strength, and oral health. In some embodiments, the smart toy is used for saliva collection. In some embodiments, the smart toy may also be used to measure bite strength / force, oral temperature, audio signals, video signals, and pH levels based on a separate smart toy IR measurement. In some embodiments, the smart room includes a microphone. In some embodiments, the smart room includes one or more smart beds.

[0225] Figure 20A An exemplary smart toy 2001 is shown. Smart toy 2001 includes a charging port 2010, a camera port 2020, a speaker / microphone port 2030, pressure and temperature sensors 2040, internal components 2040 (which may include at least one of an accelerometer, gyroscope, near-field beacon, clock, Wi-Fi, Bluetooth, or communication antenna), and a pH meter 2050. Smart toy 2001 may include a rubberized outer material. This material may be water-resistant or waterproof. The material may be suitable for chewing but safe for teeth. Figure 20B Modular smart toy components are shown. Any component of the smart toy can be housed in a material similar to or the same as the exterior material of the smart toy. Modular components can be releasably engaged with the surface of the smart toy, allowing them to be switched. Figure 20B The left panel shows modular components with a focus on separation anxiety, including a charging port with a water-resistant or waterproof lanyard, a camera port, a speaker / microphone port, and internal components (e.g., at least one of an accelerometer, gyroscope, near-field beacon, clock, Wi-Fi, Bluetooth, and communication antenna). Figure 20B The right panel shows the modular components with an oral health module. The oral health module includes pressure and temperature sensors, a pH meter, and a UV spectral sensor, as well as an optional miniaturized amperometric hydrogen sulfide sensor. In some embodiments, the smart toy can also be used to measure bite strength / force, oral temperature, audio signals, video signals, and pH values ​​based on a separate smart toy IR measurement.

[0226] algorithm

[0227] The following references provide examples of algorithms useful in this disclosure and are incorporated herein by reference in their entirety. These algorithms may be used in the embodiments herein. WO2022 / 066282 describes an algorithm that can successfully and reliably identify forward movement (gait) behavior in pets wearing collared devices. WO2022 / 031513 uses wearable devices to determine (estimate) biometric data of pets. In particular, this tool is an algorithm that uses measurements of animal activity to estimate energy expenditure (and thus estimate caloric requirements and recommended feeding amounts) based on an objective assessment of activity. WO2022 / 072049 demonstrates near real-time assessment / continuous observation of cats and dogs' responses to nutritional therapy, estimation of energy intake, and assessment of GI health and disease, responses to stress, systemic disease and / or aging, and abnormal excretion. These algorithms may also be paired with other tools that measure location or proximity to provide an assessment of the subject's location (indoor / outdoor).

[0228] Customized health, diet and nutrition

[0229] The measures and indices disclosed herein provide guidance for delivering personalized health, diet, and nutrition. For example, U.S. Provisional Application No. 63 / 241,173 and U.S. Pre-Publication Publication No. 20230073738, incorporated herein by reference, disclose a method, apparatus, and system comprising sensors configured to capture image data related to a pet for determining the pet’s body attributes, such as ideal body weight (IBW), animal body condition score (BCS), animal body fat index (BFI), weight of parts of the animal or pet, or other attributes including height, width, length, depth, diameter, radius, and circumference.

[0230] In some embodiments, activity data recorded using wearable sensors can be used to estimate energy expenditure during activity, and analysis of physical parameters and body composition can be used to refine these calculations. Body composition parameters and / or physical measurements and / or tools, such as computer vision or other machine learning models and algorithms, can be used to estimate weight goals and identify appropriate feeding algorithms based on weight gain, loss, or maintenance goals. Together with activity data, customized feeding recommendations can be made.

[0231] Some embodiments include observing the pet's interaction with food. This assessment can be used to measure ingestion behavior. This assessment can be used for such personalized, customized, and / or tailored health, diet, and nutrition recommendations, or subsequent personalized, customized, and / or tailored health, diet, and nutrition treatments. This includes eating speed, time in contact with the food bowl, number of bites, bite size, bite speed, time spent sniffing or exploring, time spent approaching or moving away from the bowl, the direction the pet faces toward the bowl versus toward other people or objects in the environment, body posture, time parameters (such as the duration between food presentation and the start of eating or the number of bites per minute), time spent chewing, time spent eating, time required to finish eating, time spent in different postures, tail direction and / or tail movement, interruption of eating, vomiting, regurgitation or expulsion of food, presence of chewing, chewing in different parts of the mouth (e.g., chewing only on one side of the mouth, as if biased toward one part of the mouth due to pain or sensitivity), the ratio of eating to other behaviors (such as drinking, sniffing, etc.), and facial expressions before, during, and after eating.

[0232] cat

[0233] Although the system according to this disclosure is described primarily based on its implementation in dogs, those skilled in the art will readily recognize that it is readily applicable to cats. For example, Figure 21 It shows how a cat's activity changes with age. Figure 22 It shows the daily activity curve of a "normal" cat. Figure 23 The daily activity curves of all cats, healthy cats, and cats with arthritis were compared. Figure 24 It shows the daily activities of a healthy adult cat. Figure 25 The daily activities of an older cat with a history of health problems are shown. Figure 26 This shows the daily activities of a five-month-old kitten. Figure 27 The cat accelerometer in use is shown. Figure 28 It shows the smart system and smart room set up for the cat.

[0234] Figure 29A and Figure 29B A graph depicting conversation metrics for both aged and non-aged individuals was used. Figure 29A Line graphs showing the change in the number of 15-minute periods of walking, running, or activity per day compared to baseline for subjects in the treatment and control groups. Figure 29B A violin plot is shown, which visualizes the distribution of the number of dogs exhibiting behavioral sessions of varying durations and behaviors.

[0235] Figures 30A to 30J The session data for the pet group is depicted. The bar chart shows the differences in session presentation between different subgroups.

[0236] Figure 31 Research designs associated with dogs undergoing dietary changes are described. Examples include methods along any research design described herein, as well as systems incorporating devices for data collection for method use. Figure 31 As shown, the first part of the study includes monitoring behavior while the dog is eating its regular food. The second part includes monitoring behavior while the dog is eating a new food. The study can implement at least one of the wearable sensors, smart toys, smart beds, or smart rooms described herein to monitor behavior and collect associated data. The study can also incorporate cameras for measuring intake behavior; for example, those described above. The study can also incorporate estimates such as BFI, BCS, weight, etc.; for example, those described above. Many different permutations are conceivable. The algorithms described above can be used, etc. The system can also implement the algorithms described herein to provide data display and / or data-based conclusions. The design can include periodic questionnaires; for example, monthly questionnaires completed by people observing the dog; for example, the dog's owner.

[0237] Figure 32 A set of research design endpoints is described. The primary endpoint can be quantified behavior. The primary endpoint can be data collected by the wearable device described in this paper. The primary endpoint can be data collected by at least one of the wearable sensors described in this paper, the smart toy described in this paper, the smart bed described in this paper, or the smart room described in this paper. The research design endpoints can be those described above for… Figure 31 The endpoints of the described research design.

[0238] Figure 33 A set of metrics measured on wearable sensors is described. These metrics may include behavior duration, episode count, behavior session, or compound behavior. Behavior duration may be a full day, a 6-hour period (or “quarter-day”), or a ratio of behavior duration. Episode counts may occur throughout a full day, within a 6-hour period (“quarter-day”), with rolling activity session counts—the number of times a subject exhibits a 5-minute interval of walking or running for more than 450 seconds within the next 15 minutes of the day, and fixed activity session counts—the number of times a subject exhibits a 15-minute interval of walking or running for more than 150 seconds of the day. Behavior sessions may include periods in which the target behavior occurs at least once and instances of the behavior are not separated by gaps exceeding 10% of the session duration. Duration may be 60 seconds (1 minute), 300 seconds (5 minutes), or 900 seconds (15 minutes). Compound behavior may involve “active” periods, which include all forward motor activities (walking + running), or “awake” behaviors, which include running, walking, scratching, shaking, and resting (everything except sleep).

[0239] Figure 34This paper outlines the methodology of the research design, which included wearable sensors on the subjects. In this study, there were two diets, multiple pre-feeding time points, and over 84 treatment time points. However, only a subset of the pre-feeding data and data from the first four weeks of the test period were used in the analysis. To estimate baseline (week 0) values, the baseline number of days up to day 0 was averaged for each animal. During the test period, time points were grouped into weekly intervals and averaged. Data were analyzed as factors of diet and week. Correlation between repeated measures (weeks) was interpreted. An appropriate covariance structure was selected using AICC fitting statistics. Trends over several weeks were analyzed using orthogonal polynomial contrastive analysis to test linear, quadratic, and cubic trends. These contrastive analyses were performed against the main effects each week and the trend of each diet over several weeks. Furthermore, the weekly means during the treatment period were compared to the baseline means for each diet using the SLICEDIFFTYPE=CONTROL option. The ADJUST=DUNNETT or ADJUST=SIMULATE options were used to calculate adjusted p-values ​​to adjust for inflation of the Type I error rate. Analysis was performed using all animals in the intention-to-treat and protocol-compliant populations, as well as subgroups of animals under 8 years of age and over 8 years of age within each population. All analyses were performed using PROC GLIMMIX in SAS® version 9.4. All p-values ​​were corrected for multiple comparisons.

[0240] Figures 35A to 3 5G depicts a set of data related to pet activity measurements.

[0241] Figures 36A to 36X An embodiment of a wearable sensor incorporated into a collar for a domestic animal is depicted. The sensor may include the basic elements of a sensor as described above. Figure 36A and Figure 36B It aligns various concepts for wearable sensors.

[0242] Figure 36C Wearable sensors designed for pets with different neck sizes are shown. The device on the left is labeled with a four-inch neck diameter, while the device on the right is labeled with a six-inch neck diameter. However, embodiments described herein include wearable sensors with collars sufficient for any neck size. Further embodiments include wearable sensors with adjustable collars that can adapt to the neck size of the animal to which they are applied.

[0243] Figure 36DA biosensor with trimmed corners is shown. The trimmed corner portions are the darkly shaded areas at the sharpest angles on the top surface of the sensor, as shown by the upper right and upper left corners of the sensor. One embodiment trims the corners to soften the overall profile of the sensor and remove the sharpest angles on the top surface. The portions that can be considered for trimming are those shown in the darkest gray shade.

[0244] Figure 36E and Figures 36M to 36O A wearable sensor based on Concept 2 is shown, such as Figure 36A and Figure 36B The markings include a first housing 3610 and a housing 3611. One or both of the first and second housings may be made of plastic. As shown, the first housing 3610 and the second housing 3611 can slide onto a main housing 3612. The main housing may also be made of plastic. Figure 36E The diagram also shows a base plate 3613 fitted to the bottom of the main housing 3612. The base plate may be aluminum. Fasteners may extend through the plate 3613 and the main housing 3612 and into the first housing 3610 and the second housing 3611. One or both of the first housing 3610 and the second housing 3611, also referred to as the first plug and the second plug, may be included to cover lugs, such as nylon lugs. The plugs may be used as locations for placing or encoding information about the animal. For example, medications may be indicated by name or different colors, and the animal name may be provided thereon.

[0245] Figure 36F , Figure 36P and Figure 36Q A wearable sensor based on Concept 2.5 (a variant of Concept 2) is shown, such as... Figure 36A and Figure 36B The device is marked as shown in the image. Concept 2.5 is similar to Concept 2, but features the housing of "Concept 3". The housing extends downwards at the ends. This may help prevent the base plate from being exposed and allows for a better fit around the animal's neck; for example, a cat or dog wearing it.

[0246] Figure 36G and Figures 36R to 36W A wearable sensor based on Concept 3 is shown, such as Figure 36A and Figure 36B As marked in the diagram. Concept 3 is similar to Concept 2.5, wherein the housing extends downward and also includes a locating pin 3620 and corresponding pin receivers 3621 and 3622 within the housing. The locating pin 3620 can be used to engage a collar 3623. When engaged with the collar 3623 and the pin receivers 3621, 3622. In one embodiment, the collar is a metal collar comprising links, wherein the last link 3624 engages the locating pin 3620. Figure 36VAs shown, the collar straps can rest against the back 3625.

[0247] Figure 36H , Figure 36I , Figure 36J , Figure 36K , Figure 36L A wearable sensor according to Concept 1 is shown. The general features are the same as described in Concept 2 above, with the following differences: A positioning pin 3631 is included in a first housing 3632 to engage with a recess on a main housing. The first housing includes a slot 3633 for engagement with a collar. Both the first and second housings can be metal to provide strength in the collar engagement. Figures 36J to 36L As shown, the collar can engage with the slot.

[0248] Figure 36X The HPN-103 wearable sensor is compared from top to bottom across Concept 1, Concept 2, Concept 2.5, and Concept 3 variants. The comparison highlights the center line of the first link on the collar of each concept.

[0249] Figure 37 A modified embodiment of a wearable sensor incorporated into a pet collar with an HPN1 accessory is depicted. A C-shaped clip is designed to secure the HPN1 via a snap-fit ​​frictional engagement.

[0250] Figure 38 A smart bed design is described, which includes sensors incorporated into the pet's bedding. The smart bed includes a padding 3810, which may be foam, battery holders 3820, 3821, a weighing sensor housing 3831, 3832, 3833 and 3834, and a weight frame 3840, which may be acrylic.

[0251] Figures 39A to 39E An embodiment of a sensor incorporated into pet bedding is described. Figure 39A A top view of the smart bed is shown, with the weight frame sliding underneath. Figure 39B A bottom view is shown. Figure 39C A side view is shown. This configuration of the smart bed may be smaller. It may have a footprint of approximately 24 inches by 18 inches by 3 inches (including two inches of foam). It can be fitted into the shell of a store-bought bed. It may weigh approximately 12 pounds. Figure 39D A weight frame with four half-bridge load cells is shown, serving as the legs of the device. The legs can support the weight of the animal. The load cells can be connected to an Hx711 analog-to-digital converter / multiplexer that can be connected to an MCU. Wiring can be routed under the frame, passing through holes under the foam, and can be fully insulated. The weight frame can be made of acrylic. Figure 39D The calibration of the weight frame with the calibration weight is shown. The calibration constant is 5000.

[0252] Figures 40A to 4 0C describes an embodiment of a set of metrics to be measured using sensors incorporated into pet bedding. Figure 40A The function of obtaining heart rate using BCG is shown. This is done via a weight unit, as the center of mass shifts due to the ejected blood. Other measurement methods may include light-based pulse readings, phonocardiography, and oscillometric methods. Figure 40B The features that can be added to the design described herein are shown. Respiratory rate can be monitored. The design may include a whealston bridge system for measuring changes in abdominal volume during inhalation and exhalation of the animal in bed. Temperature can also be monitored. An embodiment includes a thermistor device for measuring the contact temperature of the animal while it is lying in bed.

[0253] Figures 41A to 41B Circuitry for sensors embedded in collar devices or embodiments of sensors incorporated into pet bedding are depicted. The MCU may include a Seeduino or an Arduino nRF532840, preferably the latter, which offers lower power consumption, more GPIO pins, greater coding compatibility, and BLE support. The battery may be a 6V battery, which may be a K-state battery. A buck converter may be provided to reduce the power output to approximately 3.3V. Modularity may be a feature; for example, different batteries may be available. Batteries up to approximately 28V may be included in the embodiments described herein. These improvements can result in improved battery life. Improved battery life can reach up to approximately 460 consecutive hours of runtime. The illustrated embodiment is a fully integrated board containing all subsystems: (e.g., weight, respiratory rate, heart rate, and temperature).

[0254] Figure 42 The metric detection system components associated with the devices supporting the network are described.

[0255] Figure 43 The image depicts a web-based application and computer program product displaying one or more metrics. The nRF can be connected to the user's computer, or it can be battery-powered. The Arduino can perform all BLE operations and send data packets.

[0256] The results of the systems and methods described herein may lead to aging estimates that differ from the pet's actual age. For example, calculating a pet's phenotypic age may result in an age estimate that differs from the pet's actual age, such as a medium-sized pet that is 6 years old (adult) in chronological age, but whose phenotypic age is more consistent with a pet that is a few years older, such as 10 years old (senior). Therefore, in some embodiments, the methods described herein also include providing nutritional and care methods typically used for senior or mature pets, where the phenotypic age indicates that the pet is still senior or mature despite being younger in chronological age. In some embodiments, the method includes providing appropriate care based on the pet's phenotypic age, aging acceleration, aging deceleration, lifespan, and / or healthy lifespan. This may include modifying the schedule of veterinary examinations, diagnoses, and screening tests… The method may also include providing appropriate nutrition based on the pet's phenotypic age, aging acceleration, aging deceleration, lifespan, and / or healthy lifespan. This may include providing food appropriate for the pet's age and / or life stage, such as Hill's Scientific Diet Adult 7+ Senior Vitality Chicken & Rice Formula Dog Food and Hill's Scientific Diet Adult 7+ Senior Vitality Chicken & Rice Formula Cat Food. This may also include providing food with an appropriate nutritional profile for the pet's age and / or life stage. Examples of modifications to the nutritional profile of food to better suit older or mature dogs could include an energy level of 3.0 to 4.0 kcal / g dry matter (DM). Examples of modifications to the nutritional profile of food to better suit older or mature cats could include an energy level of 3.5 to 4.5 kcal / g dry matter (DM). For dogs and cats, appropriate nutrition may need to be personalized based on the results of phenotypic age, accelerated aging, lifespan, or healthy lifespan assessments.

[0257] These are merely illustrative examples, and when an animal's phenotypic age is determined to be less than its actual age, alternative nutritional approaches more suited to younger animals may be used. For instance, calculating a pet's phenotypic age can lead to an age estimate that differs from the pet's actual age in the opposite direction; for example, a medium-sized pet might have an actual age of 10 years (senior), but a phenotypic age more consistent with pets a few years younger, such as 6 years (adult). Nutritional recommendations for animals aging more slowly than expected will differ from those for older pets. For dogs and cats, appropriate nutrition may need to be personalized based on the results of phenotypic age, accelerated aging, lifespan, or healthy lifespan assessments.

[0258] In some embodiments, this disclosure provides a method for slowing phenotypic aging in companion animals in need. The method includes determining an index according to one aspect of this disclosure and providing the companion animal with personalized, customized, and / or tailored health, dietary, and nutritional measures based on the companion animal's index. This method can also be considered a method for improving accelerated aging or promoting / supporting decelerating aging or maintaining an animal's aging trajectory. In some embodiments, personalized health measures include diets that slow phenotypic aging and / or improve accelerated phenotypic aging.

[0259] Example

[0260] Understanding a pet's health and well-being presents unique challenges, especially given the communication gap between humans and animals. Relying solely on owner reports to understand a pet's health and well-being outside of a veterinary clinic has several limitations, including the subjective nature of such reports and the possibility that owners may overlook signs and behavioral changes that could indicate health problems.

[0261] Using wearable devices to assess animal behavior has emerged as a promising potential solution to address these challenges (Neethirajan, Suresh, “Recent advances in wearable sensors for animal health management,” Sensing and Biosensing Research 12 (2017): 15–29). Innovative tools such as wearable devices allow for continuous, objective monitoring of animal activity and behavior (Griffies, Joel D. et al., “Wearable sensorshown to specifically quantify pruritic behaviors in dogs,” BMC Veterinary Research 14 (2018): 1–10), enabling a more comprehensive and detailed understanding of their health.

[0262] Establishing behavioral norms is a crucial step in using wearable technology to monitor and identify early changes in animal health and well-being. Norms provide a baseline for comparing behavioral deviations, enabling early identification of potential health problems. However, raw behavioral data alone may not be sufficient to draw meaningful insights, thus necessitating the development of sensitive and specific algorithms for behavioral identification and quantification.

[0263] In traditional behavioral research, one of the key methods used to analyze animal behavior is the identification and measurement of behavioral blobs (Dawkins, Marian Stamp, Observing animal behavior: design and analysis of quantitative data, Oxford University Press, 2007). These blobs represent discrete episodes of a specific behavior, allowing researchers to investigate the frequency, duration, and temporal distribution of animal activity. However, this analytical approach may fail to capture the full complexity of animal behavioral patterns because it typically treats each blob as an isolated event without considering the broader context of the animal's overall activity patterns.

[0264] A range of derived measures can be used for animal behavior analysis. One such measure, conversationalization, involves classifying prolonged, continuous behaviors as “conversations,” treating these episodes not as isolated events but as part of a sustained sequence of behaviors. It provides a comprehensive view of animal behavioral patterns not only by examining individual behavioral units but also by examining the structure and continuity of these units over longer time intervals. Conversationalization goes beyond traditional measurements of isolated behavioral units to consider the broader temporal dimension of animal behavior. It acknowledges that individual behavioral units do not occur in isolation but are part of a continuous flow of activity. By considering these broader patterns, it becomes possible to uncover additional layers of information, such as the regularity of certain behaviors, the sequence of different behaviors, and the overall rhythm of the animal’s activity cycle.

[0265] While subjective assessments may accurately reflect animal behavior in some cases, they can be improved in terms of reliability and stability in others. Understanding the degree of consistency between these subjective assessments and objective measurements provided by sensor data will allow researchers to understand the relationship between observable dog behavior and their perceived health status, and to provide them with the option of collecting continuous, automated, and objective data in situations where sufficient information may be difficult to obtain through subjective assessments. Furthermore, in some cases, combining subjective assessments (such as owner observations and veterinary assessments) with automated assessments generated through wearable devices can provide a more comprehensive understanding of animal well-being and health. This interaction can help fine-tune the analysis and interpretation of data generated by wearable devices.

[0266] This example presents a preliminary examination of a large and growing longitudinal dataset collected from companion dogs. This dataset utilizes wearable technologies, derived metrics, and subjective assessments to provide a broad overview of animal health and well-being.

[0267] Methodology

[0268] Participants and Settings

[0269] Participants were recruited from existing employees of Hills Pet Nutrition. Employees who owned dogs were invited to participate in a cohort study involving wearable devices and questionnaires, with invitations issued between 2020 and 2023. Upon enrollment, participants were provided with a collar-style wearable device containing a three-axis accelerometer and instructions for setting up, charging, and using the device. To supplement this objective data collection, subjective assessments of the health and well-being of each dog were also collected. For this purpose, participants were intermittently invited via email to complete questionnaires, all of which were completed online.

[0270] Questionnaire data collection

[0271] The questionnaires distributed to study participants were carefully selected to cover various aspects of the dogs' health and quality of life. These included the Canine Transient Pain Inventory (CBPI) (Brown DC, Boston RC, Coyne JC, Farrar JT: Development and psychometric testing of an instrument designed to measure chronic pain in dogs with osteoarthritis, American Journal of Veterinary Research 68:631-637; 2007) to assess pain and its impact on dogs' lives; the Sleep and Nighttime Restlessness Assessment Scale (SNoRE) (Knazovicky, David et al., “Initial evaluation of nighttime restlessness in a naturally occurring canine model of osteoarthritis pain”, Peer J.3 (2015): e772.) sleep quality questionnaire; an internally developed aging questionnaire for monitoring signs of aging; and quality of life (QoL) (Lavan RP, Development and validation of a survey for quality of life assessment by owners of healthy dogs, Veterinary Research 68:631-637; 2007) to assess pain and its impact on dogs' lives; the Sleep and Nighttime Restlessness Assessment Scale (SNoRE) (Knazovicky, David et al., “Initial evaluation of nighttime restlessness in a naturally occurring canine model of osteoarthritis pain”, Peer J.3 (2015): e772.) to assess sleep quality; an internally developed aging questionnaire to monitor signs of aging; and quality of life (QoL) (Lavan RP, Development and validation of a survey for quality of life assessment by owners of healthy dogs, Veterinary Research 68:631-637; 2007). J. Sep 2013; 197(3): 578-82, doi:10.1016 / j.tvjl.2013.03.021, Epub April 29, 2013, PMID:23639368) questionnaires were used to assess overall well-being. In addition, veterinary assessment results reported by pet owners (“reported veterinary assessments”) were collected by distributing questionnaires to pet owners.

[0272] The CBPI questionnaire was distributed in 2023 and received 168 responses, the SNoRE questionnaire was distributed in 2021 and received 287 responses, the Aging Questionnaire was completed in 2022 and received 166 responses, and the QoL questionnaire[1] was distributed in 2022 and received 154 responses. Not all participants who completed the questionnaires had wearable device data.

[0273] Wearable data collection

[0274] Collar-type triaxial accelerometers worn by the dogs participating in the study allowed for 24 / 7 recording of motion data. These accelerometers were set to capture data at a rate of 100 Hz with a sensitivity of ±8g to ensure the precision and accuracy of the recorded motion data.

[0275] Raw accelerometer data was wirelessly transmitted to the cloud and processed through a classification pipeline designed to apply a range of behavioral recognition models (algorithms) to the accelerometer data. This tagged or simulated dog behaviors to categorize motion data into distinct classes, including running, walking, scratching, shaking, resting, and sleeping. The algorithms have demonstrated high sensitivity, specificity, and accuracy (Griffies, Joel D. et al., “Wearable sensor shown to specifically quantify pruritic behaviors in dogs,” BMC Veterinary Research 14 (2018): 1–10, which is incorporated herein by reference as fully elaborated), providing a solid foundation for analyzing animal behavior in this context. Examples of classification pipelines for various types of data can be found in US Prelicity Publications 2022 / 0087229 (System and Method for Monitoring Motion of an Animal), 2022 / 0044788 (System and Method for Determining Caloric Requirements of an Animal), US 2022 / 0104464 (System and Method for associating a signature of an animal movement and an animal activity), and 2022 / 0039358 (System and Method for Determining Caloric Requirements of an Animal Based on a Plurality of Durational Parameters). Data collected by any device or sensor described herein can be analyzed as set forth in any one or more of the foregoing.

[0276] During the study, data was continuously collected from collar-mounted triaxial accelerometers attached to the collars of each customer's dog. This methodology enabled the accumulation of a large dataset capturing the nuances of each dog's movement and behavior in real time over a period of more than three years, with recruitment taking place throughout the timeframe.

[0277] Wearable device data from each participant for the seven days prior to completing each questionnaire was used for association analysis with questionnaire results. All available data were used when comparing conversation metrics between subgroups.

[0278] Target subgroup analysis

[0279] This part of the work was conducted using different cohorts of 188 customer-owned dogs. Metric validation focused on a distinct group of dogs within this cohort. This group was designated “Aging Assessment” and comprised two subgroups: “Signs of Aging” and “No Reported Signs of Aging.” The “Signs of Aging” subgroup (n=53) included dogs that met our defined aging inclusion criteria: 5 years or older, 15 lbs or more, without multiple preselected diseases, on a therapeutic diet, and reportedly diagnosed with age-related diseases. The “No Reported Signs of Aging” subgroup (n=89) included dogs that met the aging inclusion criteria but had not been diagnosed with any age-related diseases.

[0280] Data Analysis

[0281] Conversational metrics

[0282] Our study explores the concept of a session, often referred to as a unit, within the context of animal behavior. Here, we use the term specifically to describe an extended display of target behavior that is robust to temporary deviations. We define a session by three parameters: minimum length, termination threshold, and minimum number of events. The study focuses primarily on the first two parameters, while the third parameter—the minimum number of events—is set as a single event within the context of our study.

[0283] The minimum length parameter plays a crucial role in shaping the interpretation of the results. Minimum length refers to the shortest duration (including termination) that the target behavior needs to be displayed to qualify the session. This is based on the selection of a reasonable duration for a single motor unit, considering three main durations—15 minutes, 5 minutes, and 1 minute.

[0284] The termination threshold refers to the maximum permissible gap between instances of a target behavior before an ongoing session is considered to have ended. Finding the right balance in determining the termination threshold is crucial; a threshold that is too short may result in too many sessions being dropped prematurely, while a threshold that is too long may lead to inappropriate merging of different sessions. Current research examined a range of values ​​and ultimately determined that a 10% threshold is appropriate because it positions the market at an inflection point, after a slow initial increase and before exponential growth.

[0285] A session is considered to be initiated by the first instance of the target behavior. Subsequent gaps in the target behavior are evaluated based on their length. If such a gap exceeds a termination threshold, the ongoing session is considered to have ended. Then, with the next occurrence of the target behavior, a new session is considered to have begun. Finally, all sessions that reach or exceed the minimum length are included in the final count. In the context of this study, three target behaviors were examined: walking, running, and walking or running, referred to as “activities”.

[0286] Validating conversationalization using aging assessment data

[0287] In this study, we propose a hypothesis centered on dogs identified as having age-related health conditions. We hypothesize that these dogs exhibit lower average daily conversation counts across different types of physical activity (i.e., walking, running, and general activity) against three defined minimum lengths: 15 minutes, 5 minutes, and 1 minute. To test this hypothesis, we generated conversation metrics from two distinct groups within the aging canine population: a control group consisting of dogs that did not report signs of age-related disease, and a target group consisting of dogs identified as having signs of age-related disease. The aim was to examine and compare the daily conversation counts between these two groups.

[0288] The nonparametric Mann-Whitney U test was used to identify differences between the control and target groups in the number of sessions and total duration of walking, running, and general activity behaviors. The Mann-Whitney U test was chosen because it can compare the non-normal distribution of independent variables between the two groups.

[0289] Subgroup comparison

[0290] Extensive metadata was collected for the cohort, enabling us to provide an overview of the various statistical groups within the population, as shown in Table 1 (below). These groups were defined based on several key characteristics, including age, living status (indoor, outdoor, or both), sex and reproductive status (sterile males, sterile females, intact males, intact females), and breed type (purebred vs. mixed). This metadata is crucial to the study as it provides rich contextual information about the different characteristics of our cohort. The Mann-Whitney U test was used to compare session variables among these subgroups. Visualizations of group means were created to allow for visual comparisons of mean measures between groups.

[0291] Table 1: Subgroup Participant Count

[0292]

[0293] Owner Assessment

[0294] The relationship between average weekly behavior rates and responses collected from questionnaires is another key aspect of this study. We sought to explore how the dogs' physical activity levels, measured by our defined conversation metric, correlated with assessments provided by owners.

[0295] To investigate this, the Spearman rank correlation coefficient, a nonparametric statistic that measures the strength and direction of the association between two rank variables, was used. This statistical tool is particularly well-suited for this study because it can identify linear and monotonic relationships, providing a robust measure of correlation independent of data distribution.

[0296] By applying Spearman's rank correlation coefficient, we were able to assess the correlation between activity measures derived from a conversation-based approach and subjective assessments provided by owners. This allowed us not only to validate our activity measures but also to understand the relationship between observable canine behavior and their perceived health status.

[0297] We performed the Mann-Whitney U test to determine activity differences between the control group (dogs without signs of age-related health conditions) and the target group (dogs reporting signs of age-related health conditions). Specifically, we tested differences in the counts of walking, running, and general activity sessions, with minimum durations of 1 minute, 5 minutes, and 15 minutes (60 seconds, 300 seconds, and 900 seconds, respectively). The U statistic and p-values ​​are listed in Table 2. A violin plot illustrates the distribution of values ​​for each measure, including a long tail of increasingly higher values. Visually, the differences between groups appear small, but are evident in the length of the tail, see [see table]. Figure 1 Furthermore, these differences were confirmed in the significance tests performed.

[0298] Table 2 shows the results of the Mann-Whitney U test between dogs showing signs of aging and those not showing signs of aging.

[0299] result

[0300] Validation of aging through conversationalization

[0301] We used the Mann-Whitney U test to determine activity differences between the control group (dogs without signs of age-related health conditions) and the target group (dogs reporting signs of age-related health conditions). Specifically, the test was applied to counts of walking, running, and general activity sessions, with minimum durations of 1 minute, 5 minutes, and 15 minutes (60 seconds, 300 seconds, and 900 seconds, respectively). The U statistic and p-values ​​are listed in Table 2. A violin plot shows the distribution of values ​​for each measure, including a long tail of increasingly higher values. Differences between groups were small but significant in the length of the tail, see [see table]. Figure 29A and Figure 29B .

[0302] Table 2: Output of the Manny-Whitney U test for aging and non-aging

[0303]

[0304] Subgroup comparison

[0305] To compare conversational variables across groups, we performed the Mann-Whitney U test. We applied it in pairs to each group to identify significant differences in behavioral conversations within our defined subgroups, thereby elucidating the variability of behavior across different group subsets.

[0306] Our study incorporates a series of groups, each with its own unique characteristics and profile. In addition to tables, we have included key figures that visually represent the most important findings of our study.

[0307] Table 3: Most Important Measures in Subgroup Comparison Analysis

[0308]

[0309] We explored the correlation between session metrics and responses to the aging questionnaire, as shown in Table 4. A 15-minute run was positively correlated with owners' perceptions of their dog's endurance, indicating that more 15-minute run sessions were associated with increased endurance reported by pet parents (R=0.43). More 15-minute walk sessions were associated with decreased endurance reported by pet parents (R=-0.18). Overall, the number of 15-minute run sessions was not only the best distinguishing factor between dogs showing signs of aging and those without, but also the best predictor of endurance reported by pet parents. The second best distinguishing factor between dogs showing signs of aging and those without was the 5-minute run session, which was also the second best predictor of endurance reported by pet parents, linked to the total 5-minute activity session. Higher scores on the aging questionnaire indicated fewer signs of aging.

[0310] Table 4: Spearman correlation between the aging questionnaire and conversation metrics. Higher scores on the aging questionnaire indicate fewer signs of aging. Increased numbers mean increased calorie intake.

[0311]

[0312] Session counts correlated with SNoRE questionnaire responses (see Table 5). Dogs with more 15-minute walking sessions exhibited less activity during sleep (R = -0.35). 15-minute running sessions were positively correlated with reports of dreaming during sleep (R = 0.44), suggesting that longer running activities may affect the depth or nature of a dog's sleep. Furthermore, these prolonged running sessions were significantly associated with increased pacing during sleep, suggesting potential restlessness (R = 0.55). When considering the dog's overall sleep quality, it was clear that more frequent 1-minute total activity sessions corresponded to owners' perceived decline in their dog's sleep quality (R = 0.54). Overall, an increase in 15-minute sessions appeared to be associated with better sleep quality, while more 1-minute sessions were associated with poorer sleep quality. Higher scores on the SNoRE questionnaire indicated a negative impact on sleep quality.

[0313] Table 5: Spearman correlation matrix of snoring and conversation metrics. Higher scores on the SNoRE questionnaire indicate a negative impact on sleep quality. Increased numbers mean increased calorie intake.

[0314]

[0315] We correlated session metrics with the Quality of Life (QoL) questionnaire responses in Table 6. Notably, dogs’ responses to their owners were strongly correlated with their average 15-minute activity sessions per day (R=0.83). Furthermore, a higher frequency of 15-minute running sessions was observed in dogs whose owners perceived them as sleeping more (R=0.89). In contrast, dogs experiencing pain showed a strong negative correlation with 15-minute walking sessions (R=-0.93). Interestingly, dogs with frequent panting showed a significant positive correlation with their average 15-minute activity sessions (R=0.85), while dogs exhibiting general health changes had a significant negative correlation with 15-minute running sessions (R=-0.89). Scores from the QoL questionnaire were normalized, therefore higher scores indicate improved quality of life.

[0316] Table 6: Spearman correlation matrix for QOL and conversation metrics. Scores from the QoL questionnaire have been normalized, therefore higher scores indicate improved quality of life. Increased numbers mean increased calorie intake.

[0317]

[0318] For most pain-related measures, increased pain perception by the owner was consistent with decreased activity (see Table 7). High scores on the CBPI questionnaire indicated greater pain intensity or impact, and in addition to overall quality of life, higher scores indicated better quality of life. For example, over a 7-day period, perceived minimum pain was negatively correlated with many activity measures, including the 15-minute activity session (R=-0.23), suggesting that higher baseline pain perception corresponds to reduced activity. Current pain perception further reinforced this pattern, showing a significant negative correlation with the 60-second activity session (R=-0.38). In contrast, the dog's ability to stand up from a lying position showed a slight positive correlation with the 15-minute running session (R=0.13), indicating that this specific assessment of pain may not significantly impede the running unit. Notably, higher QoL (meaning better well-being and health) was significantly positively correlated with the 60-second activity session (R=0.38). This highlights that despite the challenges of pain, the dog's general well-being and health are associated with its activity level.

[0319] Table 7: Spearman correlation matrix for CBPI and conversation metrics. Higher scores on the CBPI questionnaire indicate greater pain intensity or impact of pain; in addition to overall quality of life, higher scores indicate better quality of life. Increased numbers imply increased calorie intake.

[0320]

[0321] Our findings reveal several noteworthy trends associated with session length measures, as shown in Table 2. For several longer session measures, particularly those involving walking, running, and activity behaviors within 15-minute and 5-minute intervals, dogs without signs of aging exhibited significantly higher values ​​than those showing signs of aging. This trend was particularly pronounced in the 15-minute session case. These findings support the hypothesis that session measures can distinguish between dogs exhibiting and not exhibiting signs of aging. Notably, the differences between groups in these 15-minute sessions suggest that conversationalization is a powerful tool for capturing the typical endurance decline in older dogs. This further underscores the utility of our methodology in elucidating subtle changes in animal behavior that might otherwise be overlooked. Further examination of the data revealed interesting trends at the 5-minute session level. In dogs without signs of aging, measures for running and activity behaviors were significantly higher, as shown in Table 2, suggesting that as dogs age, their ability to sustain longer periods of walking may be relatively preserved compared to their ability to maintain higher energy levels.

[0322] Mann-Whitney U tests conducted across different dog subgroups did not reveal significant differences across all behavioral comparisons, suggesting that not every comparison of “walking,” “running,” and “active” states over various durations was associated with signs of aging reported by the pet parents (see Table 3). Nevertheless, some interesting patterns did emerge. A significant difference was detected in the “walking_60” measure between dogs weighing less than 25 lbs and those weighing more than 55 lbs (p=0.008), suggesting that walking behavior patterns may vary based on weight. Furthermore, significant differences were found between male and female dogs in the “running_60” and “active_60” categories (p=0.005 and p=0.040, respectively), indicating potential sex-based differences in running and overall activity levels. These findings provide interesting directions for future research exploring the effects of weight and sex on canine behavior. Interestingly, we did not see significant differences between adult and older dogs, grouped solely based on chronological age. However, we observed behavioral differences when considering differences based on signs of aging. This suggests that actual age alone may not be sufficient to distinguish pets exhibiting age-related behavioral changes.

[0323] Our analysis of the correlations between conversational metrics and questionnaire responses indicates that while some correlations appear intuitive and compelling, others are less so, suggesting that not all owner assessments are correlated with behavioral data aggregated through conversationalization. Particularly noteworthy findings include the correlation between activity measures and CBPI responses. These correlations consistently align with the notion that conversational metrics are influenced by the pain experienced by the dog, as assessed by the owner, and that quality of life responses are positively correlated with conversational metrics, suggesting that dogs perceiving a higher quality of life exhibit more active behaviors. This suggests that, despite limitations, conversational metrics can still provide valuable insights into aspects of canine well-being closely related to observable behavior. Responses from the Quality of Life (QoL) questionnaire also revealed some promising trends. Generally, QoL responses and conversational metrics appear to be positively correlated, suggesting that dogs with higher QoL scores tend to exhibit longer or more frequent behavioral conversations. The 15-minute running session and the 15-minute total activity session were the best predictors of a dog's responsiveness to its owner, enjoyment of life, and having more good days than bad days. 15-minute running and total activity sessions were strongly negatively correlated with more sleep or healthy overall changes in dogs, but positively correlated with normal movement, frequent panting, self-grooming, and rocking. The ability to walk, run, or engage in 15-minute activity sessions was strongly positively correlated with quality of life. Overall, this suggests that the 15-minute session count has the potential to be a valuable component in assessing canine quality of life. In our analysis of responses to the Aging Questionnaire, we found that, overall, the number of 15-minute running sessions was not only the best distinguishing factor between dogs showing signs of aging and those not showing signs of aging, but also the best predictor of endurance reported by pet parents. The second best distinguishing factor between dogs showing signs of aging and those not showing signs of aging was the 5-minute running session, which was also the second best predictor of endurance reported by pet parents, linked to the 5-minute total activity session. This further suggests that longer session durations can capture energy expenditure beyond the capacity of dogs experiencing obvious signs of aging. The SNoRE questionnaire, designed to quantify the impact on sleep quality, generally indicated that an increase in 15-minute activity sessions was associated with better sleep quality, while more 1-minute sessions were associated with poorer sleep quality. This may be because the dog is not engaged in a sustained period of activity, but rather in shorter, scattered periods of activity, which is to be expected if sleep quality is poor.

[0324] Given these findings, we recognize the inherent complexity and challenges of using conversation metrics as a tool to interpret animal well-being and health. It is equally important to consider the directional nature of the inferences drawn. It is possible that when conversation metrics align with owner assessments, it may be because the metrics reflect the dog's underlying condition, or the dog's ability to maintain activity may be a factor that the owner perceives as an indicator of improvement in quality of life, sleep quality, pain experience, or signs of aging.

[0325] The exploration of correlations between variables can be further expanded. For example, modeling the interactions between variables can reveal complex relationships overlooked by simple correlations. More in-depth research can also examine the relationship between owner and veterinarian assessments, providing insights into the differences between professional and lay assessments of canine health and well-being. Furthermore, behavioral performance rates in dogs under different conditions can be investigated, allowing for an understanding of the impact of various diseases on canine behavior. Using different questionnaires to potentially explore different aspects of canine behavior and well-being can further enrich the dataset.

[0326] Integrating other measures of health and well-being can enhance our understanding of canine pain and behavior. These measures can include physiological indicators such as heart rate, body temperature, or cortisol levels, which can serve as objective measures of stress or discomfort. Furthermore, observational data on canine social behavior, sleep patterns, eating and elimination habits, and other daily activities can provide supplementary data sources, offering a holistic picture of canine well-being and health.

[0327] This study reveals preliminary findings that provide confidence in the information that wearable devices can provide, showing a significant correlation between certain wearable device metrics and canine quality of life and pain measures, opening the door to further investigation and validation.

[0328] Table 8: Supplementary Table for Standardized Data

[0329]

[0330] An improved understanding of the aging process in a given pet would allow for:

[0331] a. Pet parents and veterinarians provide more personalized care based on the pet's aging process. This may include modifications to veterinary examinations, diagnostic and screening tests, and vaccination schedules, as well as age- and / or life stage-related care or modifications (here and here), resulting in care better suited to the pet's rate of aging and life stage.

[0332] b. Provide appropriate nutrition based on the pet's phenotypic age, aging acceleration, or aging deceleration. This may include providing age- and / or life-stage-appropriate foods, such as Hill's Scientific Diet Adult 7+ Senior Vitality Chicken & Rice Formula Dog Food and Hill's Scientific Diet Adult 7+ Senior Vitality Chicken & Rice Formula Cat Food.

[0333] c. Provide the pet with food that has an appropriate nutritional balance according to its age and / or life stage.

[0334] When applied to dogs, this may include feeding them food containing at least one or all of the following:

[0335] a. 3.0 to 4.0 kcal / g dry matter (DM)

[0336] b. Fat content between 7% and 15% DM

[0337] c. At least 2% crude fiber of DM

[0338] d. Protein content of DM: 15% to 23%

[0339] e. Phosphorus content of DM: 0.3% to 0.7%

[0340] f. Sodium content of DM: 0.15% to 0.4%

[0341] g. At least 400 IU vitamin E / kg DM

[0342] h. At least 100 mg vitamin C / kg DM

[0343] i. 0.5 to 1.3 mg selenium / kg DM

[0344] The calcium content of 0.4% to 0.8% DM, and

[0345] k. A calcium:phosphorus ratio of not less than 1.1.

[0346] When applied to cats, this may include feeding them food containing at least one or all of the following:

[0347] a. 3.5 to 4.5 kcal / g dry matter (DM)

[0348] b. Fat content between 10% and 25% DM

[0349] c. Crude fiber up to 15% DM

[0350] d. Protein content of 30% to 45% DM

[0351] e. Phosphorus content of 0.5% to 0.7% DM

[0352] f. Sodium content of DM: 0.2% to 0.4%

[0353] g. Potassium content of at least 0.6% DM

[0354] magnesium content of h.05% to 0.1% DM

[0355] i. Lower urine acidification potential, with target urine pH values ​​of 6.4 and 6.6.

[0356] j. At least 500 IU of vitamin E / kg DM

[0357] 100 to 200 mg of Vitamin C per kg of DM

[0358] 0.5 to 1.3 mg selenium / kg DM

[0359] The calcium content of m.0.6 to 1.0% DM, and

[0360] n. The calcium:phosphorus ratio between 0.9:1 and 1.5:1.

[0361] Based on phenotypic age, the pet's parents and / or veterinarian may introduce or modify cognitive and / or behavioral enrichment activities in conjunction with or separately from the provision of the aforementioned foods. These activities may include, but are not limited to, modifying exercise routines (e.g., increasing / decreasing frequency or lengthening / shortening walking), play, interaction with smart toys or devices, introducing new toys or experiences, scent-related puzzles, jigsaw puzzle feeders, and increasing opportunities for interaction with people or other dogs.

[0362] Therefore, some embodiments of this document include feeding animals a diet based on their phenotypic age known or determined by any of the methods or systems described herein. Some embodiments of this document include modifying animal exercise routines based on their phenotypic age known or determined by any of the methods or systems described herein. Some embodiments of this document include modifying animal play routines based on their phenotypic age known or determined by any of the methods or systems described herein. Some embodiments of this document include modifying the availability of intelligent toys or devices for animals based on their phenotypic age known or determined by any of the methods or systems described herein. Some embodiments of this document include introducing animals with at least one of the following: new toys, experiences, puzzles, scent-related puzzles, or puzzle feeders, based on their phenotypic age known or determined by any of the methods or systems described herein. Some embodiments of this document include increasing the opportunities for animals to interact with one or more people or one or more other animals based on their phenotypic age known or determined by any of the methods or systems described herein. Some embodiments of this document include reducing the opportunities for animals to interact with one or more people or one or more other animals based on their phenotypic age known or determined by any of the methods or systems described herein.

[0363] Development of biomarker-driven phenotypic aging models in dogs and felines

[0364] Biological age (BA) is a measure of healthy aging, assessing the accumulation of cellular damage, physiological changes, and loss of function over time. [1] Circulating and noncirculating biomarkers can be assessed in populations to estimate BA, and predictors of BA can be further utilized to assess mortality and morbidity risk. [2, 3]

[0365] Although research involving humans is developing and inspired by numerous surveys, similar studies are lacking in feline and canine models. Raj et al. developed an aging model using DNA methylation profiles of 130 felines, but did not assess the association between age and mortality in the modeled felines. [4] Similarly, common “markers of aging,” such as immunosenescence, are largely absent from the canine literature. [5]

[0366] Therefore, we collected biomarker data from large canine and feline cohorts to develop phenotypic aging models based on the methodology developed by Levine et al. [6], which were subsequently validated in US populations by Liu et al. [7,8]. The utility of biomarkers as predictors of BA and mortality, as well as whether phenotypic age is lower or higher than actual age, was also evaluated in canines and felines.

[0367] This example examines and presents the following research objectives: 1. Determine the phenotypic age of felines and canines, respectively; and 2. Calculate the phenotypic age acceleration.

[0368] method

[0369] Feline specimens

[0370] Eighty-three felines with a baseline age ranging from 0.1 to 19.5 years (mean ± SD = 5.7 ± 4.5 years) were followed up until death. Biomarkers were measured at baseline and at each follow-up. Our analysis sample was limited to a follow-up period of 3 years and included 721 felines with complete biomarker data at baseline. Biomarkers measured at baseline were used to predict the risk of death during the 3-year follow-up period. During the follow-up period, 109 felines (15.1%) died.

[0371] Statistical Methods—Analysis of Felines

[0372] Baseline characteristics are presented as mean and standard deviation by mortality status. A two-sample t-test was used to compare biomarkers between felines that died and those that did not die during the study period. Cox proportional hazards models were used to assess univariate associations between biomarkers and mortality risk at any point during the follow-up period. In addition, receiver operating characteristic (ROC) curves were used to assess all-cause mortality prediction and compared using the area under the curve (AUC).

[0373] Age and a total of 34 biomarkers were considered: platelets, red blood cells (RBC), hematocrit (HCT), hemoglobin, mean corpuscular volume (MCV), mean corpuscular hemoglobin (MCH), mean corpuscular hemoglobin concentration (MCHC), white blood cells (WBC), red blood cell distribution width (RDW), basophils, eosinophils, lymphocytes, monocytes, neutrophils, alanine aminotransferase (ALT), alkaline phosphatase (ALP), creatinine, blood urea nitrogen (BUN), BUN / creatine ratio, calcium, chloride, cholesterol, albumin, globulin, albumin / globulin ratio, glucose, phosphorus, magnesium, sodium, potassium, sodium / potassium ratio (Na / K ratio), bilirubin, total protein, and triglycerides.

[0374] Three methods are used to select key biomarkers for predicting mortality risk:

[0375] First, a stepwise backward selection method was used to select key biomarkers associated with mortality risk. The procedure resulted in the selection of nine variables: HCT, MCH, WBC, RDW, Bun / Creat, BUN, chloride, sodium, and animal age (in years). (Model 1) The stepwise procedure was performed in SAS.

[0376] Secondly, an elastic Cox regression model was used to select clinical biomarkers associated with mortality risk. Tenfold cross-validation was performed to select the parameter value lambda for the penalized regression, resulting in a lambda of 0.0812, one standard deviation (SD) higher than the lambda with the smallest mean squared error during cross-validation. Of the 34 biomarkers and animal age included in the elastic Cox model, age and seven biomarkers were selected: RBC, MCHC, MCH, MCV, RDW, chloride, and triglycerides. (Model 2) The elastic network model was performed in R.

[0377] Third, we used the variables retained by Liu et al. to develop human phenotypic age (PhenoAge). [7, 8] Nine out of ten variables were available in our dataset and included: creatinine, glucose, MCV, albumin, RDW, lymphocytes, WBC, ALP, and animal age (animal_age_yrs). (Model 3).

[0378] The selected variables were then included in a parametric proportional hazards model based on the Gompertz distribution. [6] The results were combined to estimate phenotypic age (in years), as shown below:

[0379] Table 9: Univariate Associations Between Statistical and Biomarker Characteristics in Felines and Mortality

[0380]

[0381] Phenotypic Age Derivation

[0382] Assuming each animal is observed once, this derivation is valid for the current dataset.

[0383] We have 365.25 * 5 days (5-year mortality risk per animal). The complete model includes all covariates:

[0384] (1)

[0385] in For the coefficients of the complete model, where "rate" estimation, and , is a scaling estimate from Gompertz.

[0386] Univariate model, where age is the only covariate:

[0387] (2)

[0388] We have Set a known constant. and This is possible because we assume that each subject has only one observation, meaning we have a scalar.

[0389] (3)

[0390] (4)

[0391] (5)

[0392] in

[0393] result:

[0394] (6)

[0395] in, The complete model, and for and , .

[0396] For each measure of phenotypic age, we computed the Phenoage Accel, defined as the residual produced from the linear model when the linear model regresses the phenotypic age on the chronological age. Thus, the Phenoage Accel represents the phenotypic age after accounting for the chronological age (i.e., whether the animal is physiologically older [positive] or younger [negative] than expected, based on its age). [7,8] assessed the correlation between phenotypic age and chronological age, and the distribution of the Phenoage Accel—the residuals of the regression of phenotypic age on chronological age.

[0397] Animals were then divided into quintiles for PhenoageAccel, such that the highest quintile represents the animal with the highest mortality risk at that age, i.e., those animals with the highest phenotypic age relative to their actual age. Kaplan-Meier curves were then plotted for the top 20% and bottom 20% of animals. Next, receiver operating characteristic (ROC) curves were used to compare the 2-year mortality risk predictions using AUC between phenotypic age measures and actual age (by cohort; ages 0–5 years, 6–10 years, and 11–20 years).

[0398] canine samples

[0399] Eighty-three dogs, aged 0.1 to 16.5 years at baseline (mean ± SD = 5.2 ± 5.1 years), were followed until death. Biomarkers were measured at baseline and at each follow-up visit. Our analysis sample was limited to a 3-year follow-up period and included 709 dogs with complete biomarker data at baseline. Note: Dogs lacking biomarker data were excluded from further analysis. Biomarkers measured at baseline were used to predict mortality risk during the 3-year follow-up period. During the follow-up period, 89 dogs (12.6%) died.

[0400] Statistical Methods - Canine Analysis

[0401] The statistical analysis of dogs followed the same methodology as that of felines. However, the following differences were noted: First, the model derived using a stepwise backward selection method resulted in the selection of 10 variables: MCHC, MCH, RDW, ALT, albumin, chloride, sodium, eosinophils, lymphocytes, and animal age (in years) (Model 1). Second, the result of the tenfold cross-validation was a lambda of 0.1225. Of the 34 biomarkers and animal age included in the elastic Cox model, age and 5 biomarkers were selected: RBC, MCH, RDW, ALT, and Na / K ratio (Model 2). No changes were found in the third model (Model 3).

[0402] Table 10: Univariate Associations Between Canine Statistics and Biomarkers and Mortality Rate

[0403]

[0404] result

[0405] Baseline characteristics of the feline and canine cohorts are shown in Tables 11a and 11b, respectively. The mean age ± SD of the feline cohort was 5.1 ± 4.2 years, and 63.8% were female (Table 11a). The proportion of feline deaths during the study period differed significantly by sex (p = 0.0034), with a higher proportion of male deaths than female deaths (20.3% vs. 12.2%). Except for platelet count, WBC, monocytes, ALT, ALP, creatinine, Na / K ratio, and cholesterol, most biomarkers (n = 25 / 34 biomarkers) showed statistically significant differences between surviving and deceased cats at the end of follow-up.

[0406] The mean age ± SD of the canine cohort was 4.9 ± 3.9 years, and 54.7% were female (Table 1b). There was no sex difference in the proportion of dogs that died during the study period compared to the feline cohort (p = 0.3281). Significantly smaller differences in biomarkers were observed between surviving and deceased dogs at the end of follow-up (n = 18 / 34 biomarkers).

[0407] Table 11a: Baseline characteristics of felines based on survival status at the end of follow-up

[0408]

[0409] Table 11b: Baseline characteristics of dogs based on survival status at the end of follow-up

[0410]

[0411] Univariate model

[0412] Tables 9 and 10 present the associations between individual variables and mortality rates for felines and dogs, respectively. In felines, the unadjusted risk of death increased by 35% with each year of age. Among single biomarkers, MCHC was associated with almost double the risk of death (HR: 1.82; 95% CI: 1.65, 2.01). In dogs, the unadjusted risk of death increased by 55% with each year of age. Magnesium (HR: 3.22; 95% CI: 1.66, 6.24) and creatinine (HR: 2.96; 95% CI: 1.11, 7.87) were the biomarkers most strongly associated with individual mortality rates.

[0413] Feline analysis

[0414] Survival analysis

[0415] Figure 44 The overall survival rate among felines is shown. As shown in Table 11a, 109 out of 721 felines (15.1%) died during the 3-year follow-up period.

[0416] Association between phenotypic age and mortality

[0417] like Figure 45A-45C , Figures 46A-46C and Figures 47A-47C As shown, felines with the highest phenotypic age relative to their actual age experienced a faster decline in survival rate during the 3-year follow-up, and the difference increased across age intervals, from 0–5 years to 11–20 years, with the 11–20-year-old felines showing the fastest decline in survival rate in the high-risk group (up to 20% of PhenoageAccel). In Model 3, which relies on the human Phenoage predictor, the difference in mortality between high and low PhenoageAccel was slightly reduced. Figures 45A-47C For each of these, the y-axis indicates survival rate, and the x-axis indicates follow-up time (in years). Similar age cohort trends were observed in all models.

[0418] Accelerated phenotypic aging

[0419] Figure 48A , Figure 48B , Figure 49A , Figure 49B , Figure 50A and Figure 50B The correlation between phenotypic age and chronological age is shown, along with the distribution of PhenoAgeAccel—the residuals of the regression of phenotypic age on chronological age. Phenotypic age and chronological age are moderately strongly correlated across models; this is partly due to the fact that age is a measure of phenotypic age. In Models 1 and 2, phenotypic age is moderately correlated with chronological age (r=0.7), which is partly due to the fact that the models include chronological age. The red line depicts the expected phenotypic age for each chronological age; points on the line depict people who are phenotypic older than expected, and points below the line depict people who are phenotypic younger than expected. PhenoAgeAccel follows a fairly normal distribution, but it has a wider range in all models than Liu et al. observed in humans. Phenotypic age is more strongly correlated with chronological age in Model 3 compared to Models 1 and 2 (r=0.81).

[0420] 2-year mortality rate

[0421] ROC curves of felines ( Figure 51The results showed that phenotypic age (AUC) of 0.953 in Model 1 and 0.956 in Model 2) was better than chronological age (0.829). Model 3 (0.874) did not predict 2-year mortality better than Model 1 and Model 2 relative to chronological age.

[0422] Canine Analysis

[0423] Survival analysis

[0424] Figure 52 The overall survival rate among the dogs is shown. As shown in Table 11b, 89 out of 709 dogs (12.6%) died during the 3-year follow-up period.

[0425] Association between phenotypic age and mortality

[0426] like Figure 53A -53C, Figure 54A -54C and Figure 55A As shown in –55C, dogs with the highest phenotypic age relative to their actual age experienced a faster decline in survival rate during the 3-year follow-up, and the difference increased across age intervals, from 0–5 years to 11–20 years, with the 11–20-year-old dogs showing the fastest decline in survival rate in the high-risk group (the highest 20% in the PhenoAgeAccel). The difference in mortality rate between high-risk and low-risk PhenoAgeAccel dogs was the best predictor of mortality in the 11–20-year-old age group, with high-risk dogs dying within one year across the model. Age-based trends were observed for all models. Figure 53A-55C In the diagram, the y-axis indicates the survival rate, and the x-axis indicates the follow-up time (in years).

[0427] Accelerated phenotypic aging

[0428] Figure 56A , Figure 56B , Figure 57A , Figure 57B , Figure 58A and Figure 58BThe correlation between phenotypic age and chronological age is shown, along with the distribution of PhenoAgeAccel—the residuals of the regression of phenotypic age on chronological age. Phenopic age and chronological age are moderately correlated; this is partly due to the fact that age is a measure of phenotypic age. In Model 1 (r=0.66), Model 2 (r=0.73), and Model 3 (0.79), phenotypic age is moderately correlated with chronological age, partly due to the fact that the models include chronological age. The red line depicts the expected phenotypic age for each chronological age, with points on the line depicting phenotypic individuals who are phenotypic older than expected, and points below the line depicting phenotypic individuals who are phenotypic younger than expected. PhenoAgeAccel follows a fairly normal distribution, but in all models, it has a wider range than observed by Liu et al. in humans. Phenopic age is more strongly correlated with chronological age in Model 3 (r=0.79) compared to Model 1 and Model 2, similar to the feline analysis.

[0429] 2-year mortality rate

[0430] ROC curve for dogs ( Figure 59 The results showed that phenotypic age (AUC) was 0.964 in Model 1, 0.935 in Model 2, and 0.932 in Model 3) was better than actual age (0.855).

[0431] Figure 59 The receiver operational characteristic curves for the 2-year mortality rate are shown using a canine model.

[0432] Discussion and Conclusion

[0433] Identifying reliable biomarkers of aging in dogs and felines could contribute to healthy aging in these species. In this analysis, we used a method from Levine et al. to estimate Phenoage in cohorts of felines and dogs, assessing mortality risk based on Phenoage and its correlation with chronological age. Our stepwise regression and Elastic Cox regression models more accurately reflect predictors of mortality in canine and feline populations than previously proposed human population variables.

[0434] Consistent with previous studies on humans,[6-8] felines and dogs with the highest phenotypic age had a faster decline in survival over a 3-year follow-up relative to their actual age, and the difference increased across age intervals, with the fastest increase in mortality observed in the oldest age group (11-20 years) of both species.

[0435] Notably, Liu et al. reported that Phenoage was more strongly correlated with chronological age compared to other Phenoage measures derived in our study. However, in our analysis, most variables in Liu's Phenoage model were not associated with mortality risk. In animal populations, variable selection methods based on prior knowledge of aging-related variables appear to be more important than data-driven variable selection methods.

[0436] Tables 12 and 13 below show what is seen in animals if the model is constructed based on a human-based phenotypic age model (Liu et al., performed in a human population). Phenotypic age models for cats and dogs are unique and cannot be easily transferred from models constructed for humans. Many variables identified as predictors of human mortality according to the models developed here do not strongly predict mortality in cats and dogs.

[0437] Table 12: Association between Model 3 (using variables from the Liu et al. model) and mortality risk in felines.

[0438]

[0439] Table 13: Association between Model 3 (using variables from the Liu et al. model) and canine mortality risk

[0440]

[0441] Each reference cited above or in the following lists is incorporated herein by reference as if fully explained.

[0442] References

[0443] 1. Poganik, JR et al., Biological age is increased by stress and restored upon recovery. Cell Metab, 2023. 35(5): p. 807-820.e5.

[0444] 2. Lohman, T. et al., Predictors of Biological Age: The Implications for Wellness and Aging Research. Gerontol Geriatr Med, 2021. 7: p.23337214211046419.

[0445] 3. Husted, K.L.S. et al., A Model for Estimating Biological Age From Physiological Biomarkers of Healthy Aging: Cross-sectional Study. JMIR Aging, 2022. 5(2): p. e35696.

[0446] 4. Raj, K. et al., Epigenetic clock and methylation studies in cats. Geroscience, 2021. 43(5): p. 2363 - 2378.

[0447] 5. Jiménez, A.G., A revisiting of “the hallmarks of aging” in domestic dogs: current status of the literature. Geroscience, 2024. 46(1): p. 241 - 255.

[0448] 6. Levine, M.E. et al., An epigenetic biomarker of aging for lifespan and healthspan. Aging (Albany NY), 2018. 10(4): p. 573 - 591.

[0449] 7. Liu, Z. et al., A new aging measure captures morbidity and mortality risk across diverse subpopulations from NHANES IV: A cohort study. PLoS Med, 2018. 15(12): p. e1002718.

[0450] 8. Liu, Z. et al., Correction: A new aging measure captures morbidity and mortality risk across diverse subpopulations from NHANES IV: A cohort study. PLoS Med, 2019. 16(2): p. e1002760.

Claims

1. A system for generating a multi-component aging index for individual companion animals based on at least one of measuring digital biomarkers, conventional biomarkers, and subjective assessment methods to predict phenotypic age and phenotypic age acceleration / deceleration in dogs and cats. The system may also include determining at least one of sex, sterilization status, and life stage.

2. The system according to claim 1, further comprising at least one of the following: Wearable devices are used to measure physical activity, including walking, running, resting, jumping, sleep duration, sleep quality, and sleep patterns. Subjective assessments conducted via at least one of the pet parent questionnaire and the veterinary questionnaire. Clinical characteristics, including actual age, weight, BCS, BFI, temperature, respiratory rate, and heart rate. Environmental sensors include sensors for detecting location and sensors for location-based behaviors and activities, such as proximity to pet parents, playtime and frequency of feeding, location of eating, drinking, urination and defecation, body posture, posture estimation, tail position, body position, and motion tracking over time. Repeatability measures of clinical, digital, and biological data Eating, drinking, urination, and defecation patterns Emotional health, cognitive health, fear, anxiety, stress, dementia, and signs of social interaction with humans and other animals, and Veterinary evaluation for at least one of the following: gastrointestinal diseases, genitourinary diseases, kidney diseases, skin diseases, respiratory diseases, neurological diseases, muscle diseases, eye diseases, hearing diseases, cardiovascular diseases, cancer, oral health, endocrine disorders, infectious diseases, immune function, inflammation, orthopedic diseases, mobility impairment, and pain.

3. The system according to claim 1 or 2, wherein, The biomarker panel includes two or more conventional biomarkers selected from the following: CBC / chemical parameters, fecal microbiome, fecal metabolites, urinary microbiome, urinary metabolites, blood metabolites, and blood biomarkers, including albumin, creatinine, glucose, lymphocyte percentage, mean cell volume, erythrocyte distribution width, alkaline phosphatase, white blood cell count, SDMA, circulating peptides including Aβ42, post-circulating biotin, immunoglobulins, immunoglobulin M, growth hormone (GH) / insulin-1 (IGF-1), and DNA biomarkers, SNPs, and genetic variations.

4. The system according to any of the preceding claims, wherein, The biomarker panel includes three or more conventional biomarkers selected from the following: CBC / chemical parameters, fecal microbiome, fecal metabolites, urinary microbiome, urinary metabolites, blood metabolites, and blood biomarkers, including albumin, creatinine, glucose, lymphocyte percentage, mean cell volume, erythrocyte distribution width, alkaline phosphatase, white blood cell count, SDMA, circulating peptides including Aβ42, post-circulating biotin, immunoglobulins, immunoglobulin M, growth hormone (GH) / insulin-1 (IGF-1), and DNA biomarkers, SNPs, and genetic variations.

5. The system according to any of the preceding claims, wherein, The biomarker panel includes four or more conventional biomarkers selected from the following: CBC / chemical parameters, fecal microbiome, fecal metabolites, urinary microbiome, urinary metabolites, blood metabolites, and blood biomarkers, including albumin, creatinine, glucose, lymphocyte percentage, mean cell volume, erythrocyte distribution width, alkaline phosphatase, white blood cell count, SDMA, circulating peptides including Aβ42, post-circulating biotin, immunoglobulins, immunoglobulin M, growth hormone (GH) / insulin-1 (IGF-1), and DNA biomarkers, SNPs, and genetic variations.

6. The system according to any of the preceding claims, wherein, The biomarker panel includes five or more conventional biomarkers selected from the following: CBC / chemical parameters, fecal microbiome, fecal metabolites, urinary microbiome, urinary metabolites, blood metabolites, and blood biomarkers including albumin, creatinine, glucose, lymphocyte percentage, mean cell volume, erythrocyte distribution width, alkaline phosphatase, white blood cell count, SDMA, circulating peptides including Aβ42, post-circulating biotin, immunoglobulins, immunoglobulin M, growth hormone (GH) / insulin-1 (IGF-1), and DNA biomarkers, SNPs, and genetic variations.

7. The system according to any of the preceding claims, wherein, The multi-component aging index for individual companion animals is also based on measurements of epigenetic modifications, including the DNA methylome.

8. The system according to any one of claims 3 to 7, wherein, The DNA biomarkers include one or more single nucleotide polymorphisms (SNPs).

9. The system according to any one of claims 2 to 8, wherein, Wearable devices are collar-mounted motion sensors (CMAS).

10. The system according to any one of claims 2 to 9, further comprising one or more of a smart bed, a smart room, and a smart toy, and optionally one or more of a camera, a computer vision unit, and an audio unit.

11. A method for mitigating phenotypic aging in companion animals in need, comprising determining an index using the system described in any of the preceding claims, and providing the companion animal with customized health, diet, and / or nutritional measures based thereon.

12. The method according to claim 11, wherein, The companion animal is an overweight animal.

13. The method according to claim 11, wherein, The companion animal is a malnourished animal.

14. The method according to any one of claims 11 to 13, wherein, The companion animal is an animal that has difficulty maintaining a healthy weight.

15. The method according to any one of claims 11 to 14, wherein, The companion animal is an older animal, or one of them. The companion animal is of unknown age and has been determined to have experienced accelerated phenotypic age or accelerated aging by analysis of at least one conventional biomarker from the companion animal.

16. The method according to any one of claims 11 to 15, wherein, The personalized health measures include diets that improve phenotypic aging.

17. A system comprising: (a) Biosensors, including: (i) A solid support, comprising an inner cavity and an outer surface; (i) A strip operatively linked to the outer surface of the solid support; (ii) A circuit located within the cavity, the circuit including at least a first position sensor and at least a first motion sensor; (b) at least one computer storage memory; and (c) Controller; Each of the sensors is in electrical communication with the controller.

18. A method for determining whether a subject's age is accelerating or decelerating, comprising: (a) Measure one or a combination of the subject’s activity measures over a period of time; (b) Determine the subject's action ability score relative to a control subject of the same age; (c) If the mobility score is equal to or higher than the control mobility score for the subject's age, the subject is classified as active; or if the mobility score is lower than the control mobility score for the subject's age, the subject is classified as inactive.

19. The method according to claim 18, wherein, The subjects were companion animals.

20. The method according to claim 18 or 19, wherein, The activity measures are: proximity to pet parents, play, timing and frequency of feeding, location of eating, drinking, urination and defecation, body posture such as via posture estimation, tail position, body position, movement tracking over time, and a combination of repeated measures of selected clinical, numerical and biological data, eating, drinking, urination, defecation, and signs of emotional / cognitive health such as fear, anxiety, stress, dementia, and social interactions with humans and other animals, tail movements, barking, jumping, scratching, animal speed in a certain direction, or sleep.

21. The method according to any one of claims 18 to 20, wherein, The time period is no less than approximately one week.

22. The method according to any one of claims 18 to 21, wherein, The time period is no less than approximately 26 weeks.

23. The method according to any one of claims 18 to 22, wherein, The measurement steps are performed by the system according to claim 17.

24. The method according to any one of claims 18 to 23, wherein, The measurement steps are performed by one or more weight sensors positioned above, below, or near the subject's bedding.

25. The method according to claim 24, wherein, The weight sensor is a component of one or more biosensors positioned under the subject's bedding, wherein the biosensor includes a top outer surface and a bottom outer surface, and wherein the sensor is capable of measuring the subject's weight based on compression of the sensor between the top outer surface and the bottom outer surface.

26. The method of claim 25, wherein, The biosensor includes at least one additional sensor for measuring activity metrics other than sleep.

27. The method according to any one of claims 18 to 26, further comprising the step of determining acceleration or deceleration of phenotypic age based on the characteristic that the subject is active or inactive.

28. A method for determining the phenotypic age of a subject, comprising: (a) Measure one or a combination of the subject’s activity measures over a period of time; (b) Determine the subject's action ability score relative to a control subject of the same age; (c) If the mobility score is equal to or higher than the control mobility score for the subject's age, the subject is classified as healthy; or if the mobility score is lower than the control mobility score for the subject's age, the subject is classified as unhealthy; and / or (d) Determine the age of the subject based on the mobility score.

29. The method according to claim 28, wherein, The subjects were companion animals.

30. The method according to claim 28 or 29, wherein, The activity metric is at least one of the following: Closeness to pet parents, playtime, feeding timing and frequency, location of eating, drinking, urination and defecation, body posture, posture estimation, tail position, body position, and movement tracking over time. Repeatability measures of at least one of clinical data, digital data, and biological data, or Patterns of at least one of eating, drinking, urinating, and defecating, as well as emotional health, cognitive health, fear, anxiety, stress, dementia, and social interactions with humans, interactions with other animals, tail movements, barking, jumping, scratching, the animal's speed in a certain direction, or signs of sleep.

31. The method according to any one of claims 28 to 30, wherein, The time period is no less than approximately one week.

32. The method according to any one of claims 28 to 31, wherein, The time period is no less than approximately 26 weeks.

33. The method according to any one of claims 28 to 32, wherein, The measurement steps are performed by the system according to claim 17.

34. The method according to any one of claims 28 to 33, wherein, The measurement steps are performed by one or more weight sensors positioned above, below, or near the subject's bedding.

35. The method according to claim 34, wherein, The weight sensor is a component of one or more devices positioned under the subject's bedding, wherein the device includes a top outer surface and a bottom outer surface, and wherein the sensor is capable of measuring the subject's weight based on compression of the sensor between the top outer surface and the bottom outer surface.

36. The method according to claim 35, wherein, The device includes at least one additional sensor for measuring activity metrics other than sleep.

37. A computer program product encoded on a computer-readable storage medium, wherein, The computer program product includes instructions, the instructions being used for: (a) Receiving data from one or more biosensors on the subject; (b) Calculate a capability score based on the data; (c) Determine the subject’s activity level or age based on the mobility score.

38. The computer program product of claim 37, further comprising the step of associating the mobility score with the health of the subject.

39. The computer program product of claim 37 or 38, further comprising instructions for selecting treatment for the subject based on the subject's phenotypic age or health status.

40. A biosensor, comprising: A top outer surface and a bottom outer surface, separated by a height, define an inner cavity, the inner cavity comprising at least one of the following: (i) Gyroscope; (ii) At least the first pressure sensor; (iii) At least a first temperature sensor; (iv) at least the first accelerometer; and (v) Controller. Each of (i), (ii), (iii) and (iv) communicates electrically with the controller via circuitry.

41. The biosensor according to claim 40, wherein, The height is no more than about 3 inches.

42. The biosensor according to claim 40 or 41, wherein, The top and bottom outer surfaces comprise a flexible material selected from rubber, latex, vinyl, or polyurethane, or combinations thereof.

43. The biosensor according to any one of claims 40 to 42, wherein, The pressure sensor includes at least one compression spring that can be operatively connected to the circuit.

44. The biosensor according to any one of claims 40 to 43 further comprises a UV spectrophotometer.

45. The biosensor according to any one of claims 40 to 44, further comprising a pH meter.

46. ​​The biosensor according to any one of claims 40 to 45 further includes a WiFi and Bluetooth communication antenna having a charging port.

47. The biosensor according to any one of claims 40 to 46 further comprises an amperometric hydrogen sulfide sensor.

48. A smart bed comprising a frame, a base, and sensing equipment.

49. A smart room comprising at least one of: (1) a biosensor according to any one of claims 39 to 46, and (2) a smart bed according to claim 48.

50. A system comprising at least one of: (1) a biosensor according to any one of claims 40 to 47; (2) a smart bed according to claim 48; and (3) a smart room according to claim 49; The system further includes at least one computer storage memory; and a controller; wherein... The biosensor, the smart bed, and / or the sensors in the smart room are in electrical communication with the controller.

Images