Systems and methods for canine chronic kidney disease

Abstract

The subject matter disclosed herein relates to a method or system for identifying a dog's susceptibility to developing chronic kidney disease (CKD). The method may include, for example, receiving at least one of one or more biomarkers or demographic information of the dog. The method may also include processing at least one of the one or more biomarkers or demographic information of the dog using a predictive model. The predictive model may include a recurrent neural network. Additionally, the method may determine a probability risk score for the dog to develop CKD based on the processed one or more biomarkers or demographic information.

Description

CROSS-REFERENCE TO RELATED APPLICATIONS

[0001] This application claims priority to U.S. Provisional Patent Application No. 63 / 033,154, filed June 01, 2020, and U.S. Provisional Patent Application No. 63 / 038,552, filed June 12, 2020, the contents of each of which are hereby incorporated by reference in their entirety. [Technical Field]

[0002] The subject matter disclosed herein relates to methods and systems for determining a pet's susceptibility to developing chronic kidney disease (CKD). [Background technology]

[0003] Chronic kidney disease (CKD), also known as chronic kidney disease or chronic renal failure, is a progressive decline in kidney function over months or years. CKD can be caused by a variety of conditions and mechanisms and affects both humans and pets. The incidence of CKD in dogs or canines is estimated to be approximately 0.5-1.0% of dogs in the United States, but has been shown to approach 25% in some populations, including certain breeds with known predispositions. CKD in dogs or canines is generally considered to have a poorer prognosis and shorter survival times than other pets, such as felines. Summary of the Invention [Problem to be solved by the invention]

[0004] Given the increased risks associated with CKD in dogs or canines, there remains a need for systems and methods that aid in the early detection or diagnosis of CKD. Additionally, there remains a need for providing customized recommendations that help mitigate the health risks associated with CKD. [Means for solving the problem]

[0005] In certain non-limiting embodiments, the subject matter disclosed herein provides a computer system for identifying a dog's susceptibility to developing chronic kidney disease (CKD). The computer system may include a processor and memory storing code that, when executed by the processor, causes the computer system to receive at least one of one or more biomarkers for the dog, which may include information regarding at least one of urine specific gravity, creatinine, urinary protein, and blood urea nitrogen (BUN), or demographic information for the dog, which may include at least one of the dog's age or weight. The computer system may process at least one of the one or more biomarkers or demographic information for the dog using a predictive model. The predictive model may include a recurrent neural network. Additionally, the computer system may determine a probability risk score for the dog to develop CKD based on the processed one or more biomarkers. In certain other non-limiting embodiments, the one or more biomarkers may include information regarding amylase.

[0006] In certain non-limiting embodiments, the computer system can determine customized recommendations based on the dog's probability risk of developing CKD. The customized recommendations can include at least one of one or more therapeutic interventions, one or more dietary recommendations, one or more renal-sparing strategies, or one or more tests for disease progression. The dietary recommendations can include recommended use of one or more pet products, such as pet food products, or the use of one or more pet products, and / or recommended use of any combination of pet products or the use of any combination of pet products. Furthermore, the dietary recommendations can include recommended dietary changes, recommended feeding regimens, and / or recommended supplemental foods, such as nutritional or pharmaceutical supplements for the dog. In another example, the one or more renal-sparing strategies can include avoiding nonsteroidal anti-inflammatory drugs, aminoglycosides, or any combination thereof, and / or the one or more tests for disease progression can include testing serum parathyroid hormone levels. In some non-limiting embodiments, the customized recommendations can be sent to the dog's veterinarian, owner, or caregiver.

[0007] In certain non-limiting embodiments, the recurrent neural network may include a hidden layer architecture with three layers. The three layers may include a first layer with five nodes, a second layer with three nodes, and a third layer with three nodes. The recurrent neural network may undergo a 10-fold cross-validation process and / or may be trained for 8 or 18 epochs. The decision threshold for developing CKD using the recurrent neural network may be about 0.5. In another example, the decision threshold for developing CKD using the recurrent neural network may be about 0.3 to about 0.9. In yet another example, the decision threshold may be between about 0 and about 1. In some non-limiting embodiments, the recurrent neural network may be trained using a training dataset. The training dataset may include one or more biomarkers and multiple other canine demographic information. In certain non-limiting embodiments, the predictive model may further include a recurrent neural network with a long short-term memory (LSTM).

[0008] In certain non-limiting embodiments, the presently disclosed subject matter provides a method for identifying a dog's susceptibility to developing CKD. The method may include receiving at least one of the following: one or more biomarkers for the dog, including information about at least one of urine specific gravity, creatinine, urinary protein, and BUN; or at least one of the dog's demographic information, which may include at least one of the dog's age or weight. The method may also include processing the one or more biomarkers or at least one of the dog's demographic information using a predictive model. The predictive model may include a recurrent neural network. In addition, the method may include determining a probability risk score for the dog to develop CKD based on the processed one or more biomarkers.

[0009] In certain non-limiting embodiments, the subject matter disclosed herein provides a computer system for identifying a dog's susceptibility to developing CKD. The computer system includes a processor and memory storing code that, when executed by the processor, causes the computer system to receive one or more biomarkers for the dog, which may include information regarding at least one of urine specific gravity, creatinine, urinary protein, and BUN, or at least one of the dog's demographic information, which may include at least one of the dog's age or weight. The computer system can also process at least one of the one or more biomarkers or demographic information for the dog using a predictive model. The predictive model can include a recurrent neural network. The recurrent neural network can include a hidden layer architecture with three layers, including a first layer with five nodes, a second layer with three nodes, and a third layer with three nodes. In addition, the computer system can determine a probability risk score for the dog to develop CKD based on the processed one or more biomarkers.

[0010] In certain non-limiting embodiments, the presently disclosed subject matter provides a computer system for identifying a dog's susceptibility to developing CKD. The method may include receiving at least one of one or more biomarkers for the dog, which may include information regarding at least one of urine specific gravity, creatinine, urinary protein, and BUN, or at least one of the dog's demographic information, which may include at least one of the dog's age or weight. The method may include processing the at least one of the one or more biomarkers or demographic information for the dog using a predictive model. The predictive model may include a recurrent neural network. The recurrent neural network may include a hidden layer architecture with three layers, including a first layer with five nodes, a second layer with three nodes, and a third layer with three nodes. In addition, the method may include determining a probability risk score for the dog to develop CKD based on the processed one or more biomarkers. [Brief explanation of the drawings]

[0011] [Figure 1A] 1 is a distribution chart of a research dataset according to certain embodiments described herein. [Figure 1B] 1 is a distribution chart of a research dataset according to certain embodiments described herein. [Figure 1C] 1 is a distribution chart of a research dataset according to certain embodiments described herein. [Figure 1D] 1 is a distribution chart of a research dataset according to certain embodiments described herein. [Figure 1E] 1 is a distribution chart of a research dataset according to certain embodiments described herein. [Figure 1F] 1 is a distribution chart of a research dataset according to certain embodiments described herein. [Figure 1G] 1 is a distribution chart of a research dataset according to certain embodiments described herein. [Figure 1H]1 is a distribution chart of a research dataset according to certain embodiments described herein. [Figure 2A] 1 is an exemplary electronic health record (EHR) chart of a patient without CKD, according to certain embodiments described herein. [Figure 2B] 1 is an exemplary electronic health record (EHR) chart of a patient without CKD, according to certain embodiments described herein. [Figure 2C] 1 is an exemplary electronic health record (EHR) chart of a patient without CKD, according to certain embodiments described herein. [Figure 2D] 1 is an exemplary electronic health record (EHR) chart of a patient without CKD, according to certain embodiments described herein. [Figure 2E] 1 is an exemplary electronic health record (EHR) chart of a patient without CKD, according to certain embodiments described herein. [Figure 2F] 1 is an exemplary electronic health record (EHR) chart of a patient without CKD, according to certain embodiments described herein. [Figure 2G] 1 is an exemplary electronic health record (EHR) chart of a patient without CKD, according to certain embodiments described herein. [Figure 2H] 1 is an exemplary electronic health record (EHR) chart of a patient without CKD, according to certain embodiments described herein. [Figure 2I] 1 is an exemplary electronic health record (EHR) chart of a patient without CKD, according to certain embodiments described herein. [Figure 2J] 1 is an exemplary electronic health record (EHR) chart of a patient without CKD, according to certain embodiments described herein. [Figure 2K] 1 is an exemplary electronic health record (EHR) chart of a patient without CKD, according to certain embodiments described herein. [Figure 2L] 1 is an exemplary electronic health record (EHR) chart of a patient without CKD, according to certain embodiments described herein. [Figure 3A]1 is an exemplary electronic health record (EHR) chart of a patient with CKD according to certain embodiments described herein. [Figure 3B] 1 is an exemplary electronic health record (EHR) chart of a patient with CKD according to certain embodiments described herein. [Figure 3C] 1 is an exemplary electronic health record (EHR) chart of a patient with CKD according to certain embodiments described herein. [Figure 3D] 1 is an exemplary electronic health record (EHR) chart of a patient with CKD according to certain embodiments described herein. [Figure 3E] 1 is an exemplary electronic health record (EHR) chart of a patient with CKD according to certain embodiments described herein. [Figure 3F] 1 is an exemplary electronic health record (EHR) chart of a patient with CKD according to certain embodiments described herein. [Figure 3G] 1 is an exemplary electronic health record (EHR) chart of a patient with CKD according to certain embodiments described herein. [Figure 3H] 1 is an exemplary electronic health record (EHR) chart of a patient with CKD according to certain embodiments described herein. [Figure 3I] 1 is an exemplary electronic health record (EHR) chart of a patient with CKD according to certain embodiments described herein. [Figure 3J] 1 is an exemplary electronic health record (EHR) chart of a patient with CKD according to certain embodiments described herein. [Figure 3K] 1 is an exemplary electronic health record (EHR) chart of a patient with CKD according to certain embodiments described herein. [Figure 3L] 1 is an exemplary electronic health record (EHR) chart of a patient with CKD according to certain embodiments described herein. [Figure 4] FIG. 1 illustrates a computer system according to certain implementations described herein. [Figure 5]FIG. XX is a more detailed view of the server of FIG. XX, according to certain embodiments described herein. [Figure 6] FIG. 1 illustrates a user equipment in accordance with certain implementations described herein. [Figure 7] 1 is a chart showing model performance as a function of age according to certain embodiments described herein. [Figure 8] 1 is a chart showing model sensitivity as a function of visit number according to certain implementations described herein. [Figure 9] 1 is a chart showing model sensitivity as a function of time before diagnosis according to certain embodiments described herein. [Figure 10A] 1 is a chart illustrating the preprocessing of a dataset according to certain embodiments described herein. [Figure 10B] 1 is a chart illustrating the preprocessing of a dataset according to certain embodiments described herein. [Figure 10C] 1 is a chart illustrating the preprocessing of a dataset according to certain embodiments described herein. [Figure 10D] 1 is a chart illustrating the preprocessing of a dataset according to certain embodiments described herein. [Figure 10E] 1 is a chart illustrating the preprocessing of a dataset according to certain embodiments described herein. [Figure 11] 1 is a chart illustrating principal component analysis or factor analysis according to certain embodiments described herein. [Figure 12] Wrapper-based functional order chart according to certain implementations described herein [Figure 13] 1 is a chart showing the average best F1 score according to certain embodiments described herein. [Figure 14] 1 is a chart illustrating function selection according to certain implementations described herein. [Figure 15] 1 is a graph illustrating Bayesian Information Criterion during wrapper function selection, according to certain implementations described herein. [Figure 16]Graphs illustrating performance metrics according to certain implementations described herein. [Figure 17] Graphs illustrating performance metrics according to certain implementations described herein. [Figure 18] Graphs illustrating RNN and LSTM architectures, according to certain implementations described herein. [Figure 19] 1 illustrates cross-validation performance according to certain implementations described herein. [Figure 20] Decision Threshold Tables According to Certain Implementations Described Herein DETAILED DESCRIPTION OF THE INVENTION

[0012] There remains a need for systems and methods that can aid in the early diagnosis of CKD in dogs or canines. Accordingly, certain non-limiting embodiments can use a predictive model to process one or more biomarkers or demographic information of the dog. The processed one or more biomarkers or demographic information can be used to determine a probability risk score for the dog to develop CKD. Based on the probability risk score, customized recommendations can be determined to help mitigate health risks associated with CKD. For clarity, and not by way of limitation, the detailed description of the presently disclosed subject matter is divided into the following subsections: 1. Definition; 2. Biomarkers; 3. Predictive models; 4. Customized recommendations; and 5. Devices and Systems.

[0013] 1.Definition The terms used herein generally have their ordinary meaning in the art, within the context of this disclosure and within the specific context in which each term is used. Certain terms are discussed below or elsewhere herein to provide additional guidance to the practitioner in describing the methods and systems of the present disclosure and how to make and use them.

[0014] As used herein, the use of the word "a" or "an" in conjunction with the word "comprising" in the claims and / or specification may mean "one," but is also consistent with the meanings "one or more," "at least one," and "one or more." Furthermore, the terms "having," "including," "containing," and "comprising" are interchangeable, and those of skill in the art will recognize that these terms are open-ended terms.

[0015] The terms "comprises," "comprising," or any other variations thereof, are intended to cover non-exclusive inclusions, such that a process, method, article, system, or apparatus that includes a list of elements may include not only those elements, but may also include other elements not expressly listed or inherent in such process, method, article, or apparatus.

[0016] Terms such as "an embodiment," "an embodiment," "one embodiment," "in various embodiments," "certain embodiments," "some embodiments," "other," and "certain other embodiments" indicate that the described embodiment(s) may include a particular feature, structure, or characteristic, but not all embodiments necessarily include the particular feature, structure, or characteristic. Moreover, such phrases do not necessarily refer to the same embodiment. Furthermore, if a particular feature, structure, or characteristic is described in the context of one embodiment, it is believed to be within the knowledge of one of ordinary skill in the art to affect such feature, structure, or characteristic in the context of any other embodiment, whether or not explicitly stated.

[0017] The term "about" or "approximately" means within an acceptable error range for a particular value, as determined by one of ordinary skill in the art, which will depend, in part, on how the value is measured or determined, i.e., the limitations of the measurement system. For example, "about" can mean within 3 or more than 3 standard deviations, in accordance with practice in the art. Alternatively, "about" can mean a range of up to 20%, preferably up to 10%, more preferably up to 5%, and even more preferably up to 1% of a given value. Alternatively, particularly with respect to biological systems or processes, the term can mean within an order of magnitude, preferably within 5-fold, and more preferably within 2-fold, of a value. It is also understood that there are many values ​​disclosed herein, and that each value is also herein disclosed as "about" that particular value in addition to the value itself. For example, if the value "10" is disclosed, then "about 10" is also disclosed. It is also understood that each unit between two specified units is disclosed. For example, if 10 and 15 are disclosed, then 11, 12, 13, and 14 are also disclosed.

[0018] The term "effective treatment" or "effective amount" of a substance refers to an amount of treatment or substance sufficient to produce beneficial or desired results, including clinical results; therefore, "effective treatment" or "effective amount" depends on the context in which it is applied. In the context of administering a composition to reduce the risk of developing CKD and / or administering a composition to treat or delay the progression of CKD, an effective amount of a composition described herein is an amount sufficient to treat and / or ameliorate CKD and reduce the symptoms and / or likelihood of CKD. An effective treatment described herein is a treatment sufficient to treat and / or ameliorate CKD and reduce the symptoms and / or likelihood of CKD. The reduction can be a 10%, 20%, 30%, 40%, 50%, 60%, 70%, 80%, 90%, 95%, 98%, or 99% reduction in the severity of CKD symptoms or the likelihood of CKD. An effective amount can be administered in one or more administrations. The likelihood of effective treatment as described herein is the likelihood that the treatment will be effective, ie, sufficient to treat and / or ameliorate CKD and reduce symptoms.

[0019] As used herein, and as is well understood in the art, "treatment" refers to an approach for obtaining beneficial or desired results, including clinical results. For purposes of this subject matter, beneficial or desired clinical results include, but are not limited to, alleviation or amelioration of one or more symptoms, reduction in the extent of disease, stable (i.e., not worsening) disease, prevention of disease, reduction in the likelihood of disease onset, delay or slowing of disease progression, and / or improvement or palliation of the disease state. A reduction can be a 10%, 20%, 30%, 40%, 50%, 60%, 70%, 80%, 90%, 95%, 98%, or 99% reduction in the severity of a complication or symptom. "Treatment" can also mean prolonging survival as compared to expected survival if not receiving treatment.

[0020] The term "animal" or "pet" as used in accordance with the present disclosure refers to domestic animals, including, but not limited to, domestic dogs, domestic cats, horses, cows, ferrets, rabbits, pigs, rats, mice, gerbils, hamsters, goats, etc. Domestic dogs and cats are specific, non-limiting examples of pets. The term "animal" or "pet" as used in accordance with the present disclosure may further refer to wild animals, including, but not limited to, bison, elk, deer, venison, ducks, poultry, fish, etc.

[0021] The terms "pet product," "pet food," "pet food composition," "pet food product," and / or "finished pet food product" refer to a product or composition intended for use or consumption by an animal, such as a cat, dog, guinea pig, rabbit, bird, or horse. For example, and without limitation, the animal can be a "domestic" dog or canine. A "pet product," "pet food," or "pet food composition," or "pet food product," or "finished pet food product" includes any food, feed, snack, dietary supplement, liquid, beverage, treat, toy (chewable and / or consumable), meal, meal substitute, or meal replacement. In certain embodiments, a "pet product" can provide a particular health or nutritional benefit to the animal.

[0022] As used herein, the term "decision threshold" can refer to a predefined or predetermined value or level used to diagnose CKD. The "decision threshold" can range, for example, from about 0.0 to about 1. In certain embodiments, a "decision threshold" of 0.5 can be used. In certain embodiments, the value of the "decision threshold" can be derived by evaluating or balancing one or more of the F1 score, precision, accuracy, sensitivity, and / or specificity. The "decision threshold" can be, for example, a sliding scale in which a trade-off or balance between one or more of the F1 score, precision, accuracy, sensitivity, and / or specificity can be determined based on clinical need or application.

[0023] The term "biomarker" refers to a characteristic that is objectively measured and evaluated as an indicator of normal biological processes, pathogenic processes, or pharmacological responses to therapeutic intervention. The term "biomarker" can also refer to any substance, structure, or process that can be measured in the body or its products and that can influence or predict the incidence of an outcome or disease. For example, biomarkers can be analyzed or determined from urine or blood samples from dogs. Examples of biomarkers may include, but are not limited to, alkaline phosphatase, amylase, protein, BUN or urea levels, creatinine, phosphorus, calcium, urinary protein, potassium, glucose, hematocrit, hemoglobin, red blood cell (RBC) count, red blood cell distribution width (RDW), alanine aminotransferase, albumin, bilirubin, chloride, cholesterol, eosinophils, globulins, lymphocytes, monocytes, mean corpuscular hemoglobin (MCH), mean corpuscular hemoglobin concentration (MCHC), mean corpuscular volume (MCV), mean platelet volume (MPV), platelet count, segmented neutrophils, sodium, urinary pH level, and / or white blood cell count. In certain non-limiting embodiments, the one or more biomarkers may be obtained from the blood, urine, serum, plasma, or saliva of the dog or canine.

[0024] The term "training dataset" refers to a database of unique dogs or canines that can be used to train a predictive model. In some non-limiting embodiments, the "training dataset" can include demographic information such as the dog's age, weight, breed, and reproductive status. Age can include the pet's age at the time of the veterinarian visit and / or the pet's age at the time of initial diagnosis of CKD. In certain non-limiting embodiments, the "training dataset" can include one or more biomarkers.

[0025] The term "visit" refers to an encounter between a medical professional or provider, such as a veterinarian, and a dog. In certain embodiments, a medical record is generated during or after the visit. In certain embodiments, the amount of one or more biomarkers is determined during the visit. In certain embodiments, a diagnosis of CKD is made during the visit. The practitioner may visit at a hospital and / or at home or other location. Owner-accompanied dogs or canines may visit the practitioner in an outpatient clinic or clinic.

[0026] The term "urine specific gravity" (also known as urine SG or USG) measures the ratio of urine density compared to the density of water. It is a measure of the concentration of solutes in urine and provides information about the ability of the kidneys to concentrate urine.

[0027] The term "customized recommendations" means treatments, methods, or tests used to reduce / mitigate the risk of developing CKD or to reduce / manage the symptoms or effects of CKD. For example, "customized recommendations" can include one or more therapeutic interventions, one or more dietary recommendations, one or more kidney-sparing strategies, or one or more tests for disease progression.

[0028] 2. Biomarkers In certain non-limiting embodiments, one or more biomarkers or demographic information of the dog or canine may be used, in part, to determine a probabilistic risk score for developing CKD.

[0029] In certain embodiments disclosed herein, one or more biomarkers can be used to predict CKD based on one or more biological parameters associated with the development of CKD. Customized recommendations can then be tailored according to the risk of developing CKD indicated by the biomarkers.

[0030] In certain embodiments, BUN and urea measurements are interchangeable: BUN reflects only the nitrogen content of urea (molecular weight 28), while urea measurements reflect the entire molecule (molecular weight 60), so urea measurements are 2.14 (60 / 28) times larger than BUN measurements.

[0031] In certain non-limiting embodiments, biomarkers can include one or more of the following: urine specific gravity level in dog or canine urine samples; total creatinine level in dog or canine blood; creatinine level in dog or canine serum; creatinine in dog or canine plasma; creatinine in dog or canine urine samples; urinary protein in dog or canine urine samples; total urea in dog or canine blood; urea in dog or canine serum; urea in dog or canine plasma; urea in dog or canine urine samples; BUN or urea in dog or canine blood; white blood cell count (WBC) in dog or canine blood; urine pH in dog or canine urine samples.In certain non-limiting embodiments, changes in the level of biomarkers can be associated with an increased risk of developing CKD.

[0032] For each biomarker, an increase or decrease in the level of the biomarker can provide information regarding the likelihood that a dog or canine will develop CKD, depending on the particular biomarker. For example, in certain embodiments, a decrease in the level of urine specific gravity indicates an increased risk of developing CKD. In certain embodiments, an increase in the level of urine specific gravity indicates a decreased risk of developing CKD. In certain embodiments, a lower level of urine specific gravity compared to a predetermined reference value based on the average level of urine specific gravity in a population of dogs or canines in the dataset can indicate an increased risk of developing CKD. In certain embodiments, a higher level of urine specific gravity compared to a predetermined reference value based on the average level of urine specific gravity in a control population indicates a decreased risk of developing CKD. In certain embodiments, the average level of urine specific gravity in the dataset population is between about 1.00 and about 1.1, between about 1.01 and about 1.09, between about 1.02 and about 1.08, or between about 1.03 and about 1.07. In certain embodiments, the mean level of urine specific gravity in the control population is between about 1.001 and about 1.08. In certain embodiments, the predetermined reference value for urine specific gravity is about 100%, about 99%, about 98%, about 97%, about 96%, about 95%, about 94%, about 93%, about 92%, about 91%, about 90%, about 89%, about 88%, about 87%, about 86%, about 85%, about 80%, about 75%, about 70% or less, or any intermediate percentage or range, of the mean level of urine specific gravity in the control population of the dataset. In certain embodiments, the predetermined reference value for urine specific gravity is between about 99.9% and about 90%, between about 95% and about 90%, or between about 99% and about 92% of the mean level of urine specific gravity in the control population of the dataset. In certain embodiments, the predetermined reference value for urine specific gravity is between about 1.001 and about 1.08, between about 1.001 and about 1.07, between about 1.001 and about 1.06, between about 1.001 and about 1.05, or between about 1.001 and about 1.04. In certain embodiments, the hydration status of the dog or canine can be considered to regulate the urine specific gravity level.

[0033] In certain non-limiting embodiments, an increased level of creatinine may indicate an increased risk of developing CKD. A decrease or lowering of creatinine level may indicate a decreased risk of developing CKD. A higher creatinine level may indicate an increased risk of developing CKD. For example, the average creatinine level in the dataset may be set between about 0 mg / dL and about 3 mg / dL, between about 0.8 mg / dL and about 3 mg / dL, between about 1 mg / dL and about 2.8 mg / dL, or between about 1.2 mg / dL and about 2.2 mg / dL. In certain embodiments, the mean level of creatinine in the control population is between about 0.8 mg / dL and about 2.4 mg / dL, and in certain embodiments, the predetermined reference value for creatinine can be about 100%, about 105%, about 110%, about 115%, about 120%, about 125%, about 130%, about 140%, about 150%, about 200%, about 250%, about 300%, about 400%, about 500% or more, or any intermediate percentage or range, of the mean level of creatinine in the control population. In certain embodiments, the predetermined reference value for creatinine can be between about 100% and about 120%, between about 120% and about 150%, between about 150% and about 200%, or between about 200% and about 500% of the mean level of creatinine in the control population. In certain non-limiting embodiments, the predetermined reference value for creatinine can be between about 0 mg / dL and about 3 mg / dL, between about 1 mg / dL and about 2.4 mg / dL, between about 1 mg / dL and about 2 mg / dL, or between about 1.2 mg / dL and about 1.8 mg / dL.

[0034] In certain embodiments, a decreased level of urinary protein may indicate an increased risk of developing CKD. An increased level of urinary protein may indicate a decreased or increased risk of developing CKD. A decreased level of urinary protein may indicate an increased or decreased risk of developing CKD. In certain embodiments, a low level of urinary protein compared to a predetermined reference value based on the average level of urinary protein may indicate an increased risk of developing CKD. In certain embodiments, a high level of urinary protein compared to a predetermined reference value based on the average level of urinary protein in a control population of a dataset indicates a decreased risk of developing CKD. A higher level of urinary protein may indicate infection or kidney damage. In certain non-limiting embodiments, a history of previous episodes of elevated urinary protein may indicate an increased risk of early infection and / or kidney damage. In certain non-limiting embodiments, current elevated urinary protein indicates an increased risk of decreased kidney function and / or CKD. A dog or canine may exhibit elevated levels of urinary protein compared to a predetermined reference value. For example, a higher urinary protein level is found in a current sample from the dog or canid, or in recent medical records from the dog or canid (e.g., records made within about 1 week, about 2 weeks, about 3 weeks, about 4 weeks, about 5 weeks, about 10 weeks, about 3 months, or about 6 months before performing any one of the methods disclosed herein). In certain embodiments, the dog or canid exhibits elevated urinary protein levels compared to a predetermined reference value in the past. For example, a higher urinary protein level is found in a previous sample from the dog or canid, or in previous medical records from the dog or canid (e.g., records made more than about 1 week, about 2 weeks, about 1 month, about 2 months, about 3 months, or about 6 months before performing any one of the methods disclosed herein). In certain embodiments, the average level of urinary protein in the control population can be between about 0 mg / dL and about 50 mg / dL, between about 0 mg / dL and about 25 mg / dL, between about 0 mg / dL and about 10 mg / dL, or between about 0 mg / dL and about 5 mg / dL, hi certain embodiments, the average level of urinary protein in the control population is between about 48 and 50 mg / dL.In certain embodiments, the predetermined reference value for urinary protein can be at least about 100%, about 110%, about 120%, about 130%, about 140%, about 150%, about 160%, about 170%, about 180%, about 190%, about 200%, about 250%, about 300%, about 400%, about 500%, about 1000%, about 2000%, about 5000%, about 10000% or more, or any intermediate percentage or range, of the average level of urinary protein in a control population. In certain embodiments, the predetermined reference value for urinary protein can be between about 100% and about 200%, between about 200% and about 500%, or between about 200% and about 1000% of the average level of urinary protein in a control population. In certain embodiments, the predetermined reference value for urinary protein is between about 0.001 mg / dL and about 100 mg / dL, between about 1 mg / dL and about 80 mg / dL, between about 5 mg / dL and about 70 mg / dL, between about 10 mg / dL and about 60 mg / dL, or between about 20 mg / dL and about 50 mg / dL.

[0035] In certain embodiments, an increase in BUN or urea levels indicates an increased risk of developing CKD. In certain embodiments, a decrease in BUN or urea levels indicates a decreased risk of developing CKD. In certain embodiments, a high BUN or urea level compared to a predetermined reference value based on the average BUN or urea levels in a control population may indicate an increased risk of developing CKD. In certain embodiments, a low BUN or urea level compared to a predetermined reference value based on the average BUN or urea levels in a control population may indicate a decreased risk of developing CKD. In certain embodiments, the average BUN level in the control population is between about 5 mg / dL and about 100 mg / dL, between about 10 mg / dL and about 55 mg / dL, between about 15 mg / dL and about 40 mg / dL, or between about 20 mg / dL and about 30 mg / dL. In certain embodiments, the average BUN level in the control population is between about 17 mg / dL and about 55 or 56 mg / dL. In certain embodiments, the predetermined reference value for BUN or urea can be about 100%, about 105%, about 110%, about 115%, about 120%, about 125%, about 130%, about 140%, about 150%, about 200%, about 250%, about 300%, about 400%, about 500% or more, or any intermediate percentage or range, of the average level of BUN or urea in a control population. In certain embodiments, the predetermined reference value for BUN or urea can be between about 100% and about 120%, between about 120% and about 150%, between about 150% and about 200%, or between about 200% and about 500% of the average level of BUN or urea in a control population. In certain embodiments, the predetermined reference value for BUN is between about 10 mg / dL and about 100 mg / dL, between about 15 mg / dL and about 90 mg / dL, between about 17 mg / dL and about 56 mg / dL, between about 20 mg / dL and about 80 mg / dL, between about 30 mg / dL and about 70 mg / dL, or between about 40 mg / dL and about 60 mg / dL.

[0036] In certain non-limiting embodiments, a decrease in WBC levels may indicate an increased risk of developing CKD. In certain non-limiting embodiments, an increase in WBC levels indicates an increased or decreased risk of developing CKD. In certain non-limiting embodiments, a decrease in WBC levels indicates an increased or decreased risk of developing CKD. In certain non-limiting embodiments, WBC can be used in a predictive model to rule out other infections or in one or more predictive models to relate previous infections to future risk. For example, WBC can be used in a predictive model to understand dehydration levels and normalize the values ​​of other biomarkers. In some non-limiting embodiments, the predictive model can be generated by a machine learning algorithm, such as a recurrent neural network or LTSM, as described below. The predictive model can interpret the WBC count according to the current and / or previous values ​​of other biomarkers. In certain non-limiting embodiments, a high WBC level compared to a predetermined reference value based on the average WBC level in a control population may indicate an increased risk of developing CKD. In certain other non-limiting embodiments, a higher WBC level may indicate infection or kidney damage. A history of previous bouts of elevated WBC may indicate a higher risk of early infection and / or kidney damage. In another example, a current elevated WBC may indicate a higher risk of declining kidney function and / or CKD. In certain non-limiting embodiments, the dog or canine may exhibit elevated WBC levels compared to a predetermined baseline. The elevated WBC levels may be found in a current sample or medical record of the dog or canine (e.g., records created within about 1 week, about 2 weeks, about 3 weeks, about 4 weeks, about 5 weeks, about 10 weeks, about 3 months, or about 6 months prior to performing any one of the methods disclosed herein). In some non-limiting embodiments, the dog or canine exhibited elevated WBC levels compared to a predetermined baseline in the past. The elevated WBC levels may be found in a previous sample or medical record of the dog or canine (e.g., a record made more than about 1 week, about 2 weeks, about 1 month, about 2 months, about 3 months, or about 6 months prior to performing any one of the methods disclosed herein).In one specific, non-limiting embodiment, the mean level of WBC in a control population is about 1 x 10. 9 / L to approximately 60 × 10 9 / L, approximately 2 × 10 9 / L to approximately 50 × 10 9 / L, approximately 5 × 10 9 / L to approximately 30 × 10 9 / L, approximately 6 × 10 9 / L to approximately 20 × 10 9 / L, or approximately 8 × 10 9 / L to approximately 16 x 10 9 / L. In one particular embodiment, the mean level of WBC in a control population is between about 13.5 x 10 9 / L. In certain embodiments, the predetermined reference value for WBC can be about 100%, about 105%, about 110%, about 115%, about 120%, about 125%, about 130%, about 140%, about 150%, about 200%, about 250%, about 300%, about 400%, about 500% or more, or any intermediate percentage or range. In certain embodiments, the predetermined reference value for WBC can be between about 100% and about 120%, between about 120% and about 150%, between about 150% and about 200%, or between about 200% and about 500% of the average level of WBC in a control population. For example, in certain non-limiting embodiments, the predetermined reference value for WBC is about 2×10 9 / L to approximately 100 × 10 9 / L, approximately 5 × 10 9 / L to approximately 80 × 10 9 / L, approximately 10 × 10 9 / L to approximately 70 × 10 9 / L, approximately 20 x 10 9 / L to approximately 60 × 10 9 / L, or approximately 30 x 10 9 / L to approximately 50 × 10 9 / L. In certain non-limiting embodiments, a lower WBC level may indicate a reduced risk of developing CKD. In certain embodiments, the predetermined reference value for WBC may be about 100%, about 95%, about 90%, about 85%, about 80%, about 75%, about 70%, about 60%, about 50% or less of the average level of WBC in a control population, or any intermediate percentage or range. In certain embodiments, the predetermined reference value for WBC may be between about 100% and about 90%, between about 80% and about 60%, or between about 60% and about 40% of the average level of WBC in a control population.

[0037] In certain embodiments, a decrease in urinary pH level indicates an increased risk of developing CKD. In certain embodiments, an increase in urinary pH level indicates a decreased risk of developing CKD, and a decrease in urinary pH level may indicate an increased risk of developing CKD. In certain embodiments, a higher urinary pH level may indicate a decreased risk of developing CKD. In some non-limiting embodiments, the average urinary pH level in the control population of the dataset may be between about 4 and about 8.5, between about 5 and about 8, between about 5.2 and about 7.5, or between about 6 and about 7. In particular, the average urinary pH level in the control population may be between about 5.5 and about 7.5. In certain non-limiting embodiments, the predetermined reference value for urinary pH may be about 100%, about 95%, about 90%, about 85%, about 80%, about 75%, about 70%, about 60%, about 50% or less of the average urinary pH level in the control population, or any intermediate percentage or range. In certain embodiments, the predetermined reference value for urine pH can be between about 100% and about 80%, between about 80% and about 60%, or between about 60% and about 40% of the average level of urine pH in a control population. In certain embodiments, the predetermined reference value for urine pH can be between about 3 and about 8, between about 4 and about 7.5, between about 4.5 and about 7, between about 4.5 and about 6.5, between about 5 and about 6.5, or between about 5 and about 6. In certain embodiments, the dog or canine diet and handling of the dog or canine urine sample can contribute, in part, to adjusting the urine specific gravity level.

[0038] In certain non-limiting embodiments, an increase or decrease in the level of a biomarker can be detected in a current sample or in the dog's or canine's recent medical records (e.g., records made within about 1 week, about 2 weeks, about 3 weeks, about 4 weeks, about 5 weeks, about 10 weeks, about 3 months, or about 6 months before performing any one of the methods disclosed herein). In some non-limiting embodiments, the dog or canine has previously exhibited an increase or decrease in the level of a biomarker. For example, an increase or decrease in the level of urinary protein can be found in a previous sample from the dog or canine or in the dog's or canine's previous medical records (e.g., records made more than about 1 week, about 2 weeks, about 1 month, about 2 months, about 3 months, about 6 months, about 12 months, about 2 years, about 3 years, about 4 years, about 5 years ago, and / or records made at any time during, before, or after performing any one of the methods disclosed herein).

[0039] Generally, the range of mean levels of a biomarker may represent 50 to 100% of a healthy, normal population. For some biomarkers, the range of mean levels of the biomarker may represent 80 to 95%. Thus, approximately 5 to 25% of the population may have values ​​above the upper limit of the mean / normal range, and another approximately 5 to 25% of the population may have values ​​below the lower limit of the mean / normal range. In certain embodiments, the actual range and validity of a biomarker can be determined by each laboratory or test, depending on the machine and / or the dog or canine population tested to determine the mean / normal range. In addition, laboratory tests may be affected by sample handling and machine maintenance / calibration. Machine updates may also cause changes in the normal range. The mean level and / or predetermined reference value for each biomarker can be adjusted to account for any one of these factors.

[0040] Beyond the biomarkers described above, certain non-limiting embodiments may include one or more of the following biomarkers: phosphate and parathyroid hormone (PTH), symmetric dimethylarginine (SDMA), systolic blood pressure, potassium, total calcium, hyaluronic acid, death receptor 5, transforming growth factor β1, ferritin, beta globin, catalase, alpha globin, epidermal growth factor receptor pathway substrate 8, mucin isoform precursor, ezrin, delta globin, moesin, phosphoprotein isoforms, annexin A2, myoglobin, hemopexin, serine protease inhibitors, serpin peptidase inhibitors, CD14 antigen precursor, fibronectin isoform preprotein, angiotensinogen preprotein, complement component precursor, carbonic anhydrase, uromodulin precursor, complement factor H, complement component 4.BP, heparan sulfate proteoglycan 2, olfactomedin-4, leucine-rich α-2 glycoprotein, ring finger protein 167, α-interglobulin inhibitor H4, heparan sulfate proteoglycan 2, N-acyl singosine aminohydrolase, serine protease inhibitor clade A member 1, mucin 1, clusterin isoform 1, brain-enriched membrane-associated signaling protein 1, dipeptidase 1, fibronectin 1 isoform 5 preprotein, angiotensinogen preproprotein, carbonic anhydrase, uromodulin precursor, metalloprotease inhibitor 2, insulin-like growth factor binding protein 7, immunoglobulin A, immunoglobulin G1, immunoglobulin G2, alpha-1 antitrypsin, serum amyloid P component, hepatocyte and any combination thereof. See, for example, U.S. Patent Application Publication Nos. 2012 / 0077690, 2013 / 0323751, European Patent Application Publication Nos. 3,112,871, 2,462,445, and 3,054,301.

[0041] In certain non-limiting embodiments, the amount of a biomarker in a dog or canine can be detected and quantified by any means known in the art. In certain other non-limiting embodiments, the levels of creatinine, urinary protein, WBC, urea, and / or BUN can be determined by fluorescence or luminescence methods.

[0042] In certain embodiments, creatinine, urinary protein, WBC, urea and / or BUN levels can be determined by antibody-based detection methods, such as enzyme-linked immunosorbent assay (ELISA) or sandwich ELISA.

[0043] In other examples, urinary protein levels can be determined using urinary albumin antibodies, urinary specific gravity levels can be measured by refractometry, densitometry, and / or reagent strips, while in certain non-limiting embodiments, urinary pH levels can be measured by pH test strips, or a pH meter and pH probe, and WBC levels can be measured by flow cytometry.

[0044] In certain non-limiting embodiments, other detection methods can be used, such as other spectroscopic, chromatographic, labeling techniques, and / or quantitative chemistry methods. The level of the dog or canine-derived biomarker and / or predetermined reference value for the biomarker can be determined by the same methods.

[0045] 3. Predictive Model Some non-limiting embodiments may be directed to a method or system for identifying a dog's susceptibility to developing CKD. The method may include processing, or having a system process, at least one of one or more biomarkers or demographic information of the dog using a predictive model. The predictive model may be trained using one or more machine learning techniques.

[0046] 3.1 Dataset For example, a predictive model can be trained using the following dataset:

[0047] [Table 1]

[0048] The dataset shown in Table 1 may include one or more biomarkers and / or demographic information for the dog. For example, the demographic information may include the pet's age or weight at the time of the visit, the pet's age when first diagnosed with CKD, and / or the dog or canine's sex. The one or more biomarkers shown in Table 1 may include BUN, creatinine, urinary protein, and urinary SG. One or more other biomarkers, such as amylase, or other demographic information may be included in the dataset.

[0049] In certain non-limiting embodiments, a dataset, which may be referred to as a training dataset, can include medical records of a plurality of dogs or canines. For example, the range of dogs or canines in the training dataset set forth in Table 1 is from about 15,128 to about 22,082. In another example, the range of dogs or canines in the testing dataset set forth in Table 1 is from about 7,430 to about 11,275. In other embodiments, the number of dogs or canines included in a dataset can be between 100 and 100,000, e.g., 55,885. The medical records can include, for example, the amount of one or more biomarkers and / or demographic information for the dog or canine. In certain embodiments, the medical records can include one or more visits for a dog or canine. For example, the training and testing datasets set forth in Table 1 range from about 11 to about 16 visits per dog or canine. However, in other non-limiting embodiments, the number of visits can range from 1 to 100 per dog or canid. In certain non-limiting embodiments, the medical record can include the dog's or canid's most recent two, three, four, or five visits at various times. In another non-limiting embodiment, the medical record can include records of the dog's or canid's first and last visits at various times.

[0050] In some non-limiting embodiments, the training dataset can be stratified, shaped, or arranged for the purpose of cross-validation. Cross-validation can be used to evaluate how the results of a predictive model can be generalized to an independent dataset. The dataset can be split or stratified, for example, into two or more divisions, and one or more subsets can be used to validate the predictive model with one or more different subsets. In certain embodiments, the training dataset is stratified into about 3, 4, 5, 6, 7, 8, 9, 10, 20, 30, 40, or 50 divisions.

[0051] In certain non-limiting embodiments, rather than stratifying for purposes of cross-validation, the dataset can be divided into subsets for one or more different predictive models. The subsets can correspond, for example, to dogs or canines diagnosed with CKD during a given visit, or 3 months, 6 months, 12 months, 2 years, 3 years, 4 years, or 5 years after a given visit. In other non-limiting embodiments, the training dataset can be divided into any other subsets.

[0052] In some non-limiting embodiments, when a dog or canine medical record or chart is missing values, amounts, or levels of one or more biomarkers and / or demographic information, the missing amounts, levels, or demographic information can be imputed. Missing data can be based on either: completely missing data at random (MCAR), where the probability that an instance has a missing value for a variable does not depend on either the known values ​​or the missing data; missing data at random (MAR), where the probability that an instance has a missing value for a variable can depend on the known values, but not on the value of the missing data itself; or missing data not at random (MNAR), where the probability that an instance has a missing value for a variable can depend on the value of that variable.

[0053] Missing values ​​can be imputed, which may mean that the missing values ​​are replaced with reasonable values. Imputation, in certain non-limiting embodiments, can be calculated using statistical methods or processes such as mean, median regression multiple, or ridge regression imputation. In mean or median imputation, missing components of a vector can be filled with the mean or median value of that component. In some non-limiting embodiments, a matrix factorization method or process can be used to impute missing values. For example, the matrix factorization method or process can include UV matrix factorization, soft imputation, iterative singular value decomposition (SVD) imputation, or a combination of biscaler and soft imputation.

[0054] In other non-limiting embodiments, imputation can be calculated using machine learning. For example, the value, amount, level, or demographic information of imputation can be determined using one or more of the following machine learning methods: k-nearest neighbor (KNN) imputation such as missingpy KNN or fancyimpute KNN; multiple or multivariate imputation by chained equations (MICE) imputation such as linear regression, ridge regression, or gradient boosting; and / or random forest algorithms such as missingpy missForest, sciblox MICE random forest, or related algorithms, or any other random variable in missing forest. Metrics used to measure imputation can include, for example, root mean square error (RMSE), mean absolute error (MAE) metric, and normalized RMSE.

[0055] The RMSE can be calculated, for example, using the following formula:

[0056]

number

[0057] where N is the number of missing values ​​and y i is the assigned value, and x iis the true value. MAE can be calculated, for example, using the following formula:

[0058]

number

[0059] where N is the number of missing values ​​and y i is the assigned value, and x i is the true value. The normalized RMSE can be calculated, for example, using the following formula:

[0060]

number

[0061] where X true can be the full data matrix, and X imp can be the imputed data matrix. var() can be the calculated variance for continuous missing values.

[0062] In certain non-limiting embodiments, when imputing all 34 features, including one or more biomarkers and / or demographic information, missForest and MICE imputation, particularly linear regression, were the best-performing imputations based on RMSE and MAE, with missForest outperforming in 75% of experiments. Meanwhile, when imputing six features, including one or more biomarkers and / or demographic information, such as urine protein, urine specific gravity, urine pH, BUN, creatinine, or WBC, MICE imputation performed better than missForest. For example, in the six-feature imputation, the best-performing imputations in terms of RMSE may be fancyipute MICE, sciblox MICE libear, and / or sciblox MICE boost. In another example, in the six-feature imputation, the best-performing imputations in terms of MAE may be fancyipute MICE, sciblox MICE libear, and / or sciblox MICE boost.

[0063] In some non-limiting embodiments, visit age can be beneficial for imputing six features when either a block of urinalysis values ​​(e.g., urinary protein, urinary specific gravity, urinary pH) or a block of blood values ​​(e.g., BUN, creatinine, WBC) is missing. Specifically, the dog or canine visit age can help improve the accuracy of imputation. The improvement in the block of blood values ​​can be greater than the block of urinalysis values ​​when visit age is taken into account. For example, the median MEA gain can be 1.1% for blood and 0.2% for urine.

[0064] The imputation method or process selected can be based on the amount of missing information in the dataset: for example, a dataset with 10% missing values ​​can use MICE linear regression, while a dataset with 20% or 30% missing values ​​can use missForest.

[0065] In certain non-limiting embodiments, the training dataset can be filtered by a set of inclusion and exclusion criteria. For example, the number of dog or canine visits (two or more, three or more, four or more, or five or more) can be used as an inclusion or exclusion criterion. In another example, the medical history of the dog or canine visit or the age at the visit can be used as an inclusion or exclusion criterion.

[0066] In certain non-limiting embodiments, the dataset may include a total of 55,885 dogs or canids. The dataset may also exclude information collected from dogs or canids under 1.5 years of age and over 22 years of age. The dataset may include a diverse population of dogs or canids, including mixed-breed dogs and / or over 280 pedigree breeds. One or more biomarkers and / or demographic information were selected from the dataset as a function of the CKD prediction model. In one example, 35 biomarkers and / or demographic information items were selected. The dataset may also include dogs or canids diagnosed with CKD and dogs or canids not diagnosed with CKD.

[0067] Figures 1A-1H show distribution charts of study datasets according to certain embodiments described herein. In particular, Figures 1A-1H show biomarker and demographic information for both dogs diagnosed with CKD, shown in Figures 1B, 1D, 1F, and 1H, and dogs not diagnosed with CKD, shown in Figures 1A, 1C, 1E, and 1G. The age T(0) in Figures 1A-1H can be the age at which the dog or canine was first or originally diagnosed with CKD. In some non-limiting embodiments, data collected more than 30 days after the first or original diagnosis are excluded, and / or an additional 30-day window is included to capture serum, blood, or urine laboratory data entered into the database immediately after the diagnostic visit.

[0068] The datasets shown in Figures 1A-1H may exclude dogs or canines that do not have a formal CKD diagnosis but have at least two data points suggestive of CKD, such as blood creatinine above normal and / or urine specific gravity below normal. Other data points suggestive of CKD may include one or more of the following terms appearing in the medical record: "CKD," "azotemia," "Royal Canin Veterinary diet Renal," or "Hill's prescription diet k / d." Data points suggestive of CKD may be based on blood or urine test results.

[0069] All other pets or canines that were not diagnosed with CKD or had at least two data points suggestive of CKD and had at least two years of data were included in the dataset and assigned a "CKD-free" status, as shown in Figures 1A, 1C, 1E, and 1G. For "CKD-free" dogs or canines, T(0) was set as the age at the last visit minus 2 years.

[0070] In some non-limiting embodiments, health records can be further filtered based on information content by requiring that dogs or canines have at least two visits with blood creatinine data. This resulted in a final study dataset of 55,915 dogs or canines, of which 22,558 dogs were diagnosed with CKD and the remaining 33,357 were "CKD-free." The "CKD-free" dogs or canines can be referred to as the control group. As shown in the graphs in Figures 1A-1H, graphs 102, 106, 110, and 114 may represent visit age, creatinine level, BUN, and urine specific gravity, respectively, for dogs or canines assigned the "CKD-free" label. Meanwhile, graphs 104, 108, 112, and 116 represent visit age, creatinine level, BUN, and urine specific gravity, respectively, for dogs or canines diagnosed with CKD. Dogs or canines diagnosed with CKD were older, had high BUN levels, and / or low urine specific gravity, as shown in Figures 1B, 1F, and 1H. The results support the quality of CKD diagnoses within the dataset and provide confidence in the data used to build the model.

[0071] Figures 2A-2L show exemplary "CKD-free" EHRs according to certain embodiments described herein. In particular, Figures 2A-2L show individual dog or canine observations for creatinine, blood urea nitrogen, and urine specific gravity as a function of time before T(0). Graphs 202, 208, 214, and 220 shown in Figures 2A, 2D, 2G, and 2J may show creatinine as a function of time before T(0). Graphs 204, 210, 216, and 222 shown in Figures 2B, 2E, 2H, and 2K may show BUN as a function of time before T(0). Graphs 206, 212, 218, and 224 shown in Figures 2C, 2F, 21, and 2L may show urine specific gravity as a function of time before T(0). In graphs 202, 204, and 206 shown in Figures 2A, 2B, and 2C, the age T(0) may be 9.7 years; in graphs 208, 210, and 212 shown in Figures 2D, 2E, and 2F, the age T(0) may be 6.2 years; in graphs 214, 216, and 218 shown in Figures 2G, 2H, and 2I, the age T(0) may be 9.5 years; and in graphs 220, 222, and 224 shown in Figures 2J, 2K, and 2L, the age T(0) may be 9.2 years.

[0072] Figures 3A-3L show exemplary EHRs for a patient with CKD, according to certain embodiments described herein. In particular, Figures 3A-3L show observations of an individual dog or canine for creatinine, blood urea nitrogen, and urine specific gravity as a function of time before T(0). Graphs 302, 308, 314, and 320 shown in Figures 3A, 3D, 3G, and 3J may show creatinine as a function of time before T(0). Graphs 304, 310, 316, and 322 shown in Figures 3B, 3E, 3H, and 3K may show BUN as a function of time before T(0). Graphs 306, 312, 318, and 324 shown in Figures 3C, 3F, 31, and 3L may show urine specific gravity as a function of time before T(0). In graphs 302, 304, and 306 shown in Figures 3A-3C, age T(0) may be 12.3 years, in graphs 308, 310, and 312 shown in Figures 3D-3F, age T(0) may be 13.7 years, in graphs 314, 316, and 318 shown in Figures 3G-3I, age T(0) may be 8.4 years, and in graphs 320, 322, and 324 shown in Figures 3J-3L, age T(0) may be 9.4 years.

[0073] In the samples shown in Figures 2A-2L and 3A-3L, "No CKD" is distinct from CKD. Significant heterogeneity exists within the latter group, and many changes may occur prior to the time of diagnosis. This helps demonstrate that predictive models must not only consider multiple factors at the time of diagnosis, but also include information at various time points prior to diagnosis.

[0074] To determine the CKD prediction model, the dataset was randomly divided into two parts. For example, as shown in Table 1, of the total 55,915 health records, 37,210 health records, or approximately 67% of the data, were used to build the CKD prediction model. The remaining 18,705 health records, or approximately 33% of the data, were used as a test dataset to evaluate model performance or cross-validation. The model-building dataset, also known as the training dataset, and the model-testing dataset can be kept separate throughout the analysis to eliminate bias during the testing phase. Prior to use, missing information in the blood and urine test datasets can be estimated using all available blood and / or urine data, but CKD status information is unavailable. In certain non-limiting embodiments, neural networks may require complete data, so missing information can be estimated. In some datasets, the prevalence of missing data can be approximately 10% for most blood chemistry measurements and / or approximately 60% for urine test results. The model-building dataset and the test dataset were kept separate to avoid information flow between the datasets.

[0075] 3.2 Types of Prediction Models In certain non-limiting embodiments, a predictive model for dog or canine CKD can include one or more machine learning algorithms. The inherently multifactorial nature of canine CKD presents an ideal setting for predictive models to add clinical value. For example, machine learning algorithms such as logistic regression or backpropagation neural networks can be supervised. In other examples, the machine learning algorithm can use unsupervised, semi-supervised, or reinforcement techniques such as Apriori algorithms or K-means clustering, or Q-learning algorithms or time-difference learning. In other non-limiting embodiments, any other suitable learning style can be used.

[0076] In such embodiments in which the predictive model utilizes a machine learning algorithm, the machine learning algorithm may be, for example, any of the following algorithms or methods: regression algorithms (e.g., ordinary least squares, logistic regression, stepwise regression, multivariate adaptive regression splines, locally estimated scatterplot smoothing, etc.), example-based methods (e.g., k-nearest neighbors, learning vector quantization, self-organizing maps, etc.), regularization methods (e.g., ridge regression, least absolute shrinkage and selection operator, elastic net, etc.), decision tree learning methods (e.g., classification and regression trees, iterative bisection3, C4.5, chi-squared automated interaction detection, decision strains, random forests, multivariate adaptive regression splines, gradient boosting machines, etc.), Bayesian methods (e.g., naive Bayes, AODE, Bayesian belief networks, etc.), kernel methods (e.g., support vector machines, radial basis functions, linear discriminant analysis, etc.), clustering methods (e.g., ensemble learning methods, etc.), and the like. The algorithms may include any one of the following: k-means clustering, expectation maximization, etc.), association rule learning algorithms (e.g., Apriori algorithm, Eclat algorithm, etc.), artificial neural network models (e.g., perceptrons, backpropagation, Hopfield networks, self-organizing maps, learning vector quantization, etc.), deep learning algorithms (e.g., restricted Boltzmann machines, deep belief networks, convolutional networks, stacked autoencoders, etc.), dimensionality reduction methods (e.g., principal component analysis, partial least squares regression, Sammon mapping, multidimensional scaling, projection pursuit, etc.), ensemble methods (e.g., boosting, bootstrap aggregation, AdaBoost, stacked generalization, gradient boosting machines, random forests, etc.), conditional random field algorithms, as well as any suitable type of algorithm.

[0077] In certain non-limiting embodiments, the predictive model may include one or more of a logistic regression algorithm, an artificial neural network algorithm (ANN), a recurrent neural network algorithm (RNN), a k-nearest neighbor algorithm (KNN), a KNN with dynamic time warping (KNN-DTW), a naive Bayes algorithm, a support vector machine algorithm (SVM), a random forest algorithm, an AdaBoost algorithm, and / or any combination thereof. In some non-limiting embodiments, a regularization algorithm may be used. For example, the regularization algorithm may help prevent overfitting.

[0078] In certain non-limiting embodiments, the predictive model may be an RNN, which includes an algorithm including an input layer, an output layer, and / or one or more hidden layers. The RNN may be, for example, a vanilla RNN, a long short-term memory (LSTM) RNN, and / or a gated recurrent unit (GRU) RNN. In some non-limiting embodiments, the RNN may include 1 to 50 hidden layers, such as 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 hidden layers. Each input layer, output layer, or hidden layer may include 1 to 500 nodes. Some non-limiting embodiments may include, for example, 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 nodes. The input, output, or hidden layers may each include the same or different numbers of nodes.

[0079] In certain implementations, one or more hidden layers may include an activation function. The activation function serves to determine the output of a particular node in one or more hidden layers. For example, the activation function may be a TanH function, a sigmoid function, a logistic function, a rectified linear unit (ReLU) function, a maxout function, or a Gaussian function. Any other activation function known in the art may also be used as part of the predictive model.

[0080] As described above, in certain non-limiting embodiments, the predictive model can be trained using KNN with dynamic time warping (DTW). One or more biomarkers and / or demographic information processed by the predictive model can be selected by a filtering method such as Pearson correlation coefficient. In some non-limiting embodiments, one or more biomarkers and / or demographic information can be selected by a top-down wrapper method, KNN-DTW, using K or between about 7 and 17 nearest neighbors. In certain other non-limiting embodiments, one or more biomarkers and / or demographic information can be selected by a bottom-up wrapper. In other non-limiting embodiments, the predictive model can be trained using a mixture of experts (MOE) approach, where an ensemble of predictors can be combined using simple or weighted voting.

[0081] In certain non-limiting embodiments, the classification algorithm can be trained using an RNN, such as a vanilla RNN, an LSTM RNN, or a GRU RNN, that includes an input layer, an output layer, and one or more hidden layers, each having one or more nodes. For example, the RNN can include three hidden layers, including a first layer including five nodes, a second layer including three nodes, and a third layer including three nodes. The RNN can utilize a cross-validation process, including a cross-validation process with about 1 to about 100 folds. Furthermore, the RNN can be trained for about 1 to 100 epochs. For example, the RNN can include a 10-fold cross-validation process and can be trained for 8 or 18 epochs.

[0082] The input layer of the predictive model may include one or more biomarkers and / or dog or canine demographic information. The output layer of the predictive model may include a softmax or normalized exponential function. A softmax function can help normalize the output of the predictive model into a probability distribution, where the number of probabilities is proportional to the exponent of the input values. In some non-limiting embodiments, binary cross-entropy can also be used to calculate the loss. Other non-limiting embodiments may utilize a regularization algorithm to prevent overfitting. The regularization algorithm may, for example, drop out approximately 5%, 10%, 15%, 20%, 25%, 30%, 35%, 40%, or any other percentage to avoid overfitting.

[0083] In certain non-limiting embodiments, the predictive model can include evaluation or validation of the predictive model. In some non-limiting embodiments, the evaluation or validation can be used to update the predictive model. For example, the predictive model can include 10-fold cross-validation. As part of the cross-validation, the dataset can be stratified into about 2, about 3, about 4, about 5, about 6, about 7, about 8, about 9, about 10, about 20, about 30, about 40, about 50 or more folds of the cross-validation, or any intermediate number of folds.

[0084] In some non-limiting embodiments, the performance of the predictive model can be characterized by an area under the curve (AUC) ranging from about 0.50 to about 0.99.

[0085] The predictive model can be used to determine a probability risk score for a dog to develop CKD. The probability risk score can be based, for example, on the probability that the dog or canine will develop CKD. The probability risk score can be any value between 0 and 100%, or between 0.0 and 1.0. Based on the determined probability risk score, the dog or canine can be determined to have a high risk of developing CKD with low or high certainty. The low or high certainty can be based on at least one of the precision, sensitivity, specificity, or F1 score of the probability risk score. For example, an accuracy of 95% or more can be said to indicate a high risk of developing CKD, while an accuracy of 25 to 50% can be said to indicate a low risk of developing CKD.

[0086] In certain non-limiting embodiments, a dog or canine may have low or high certainty that it will not develop CKD. For example, an accuracy of about 80% or less may indicate low certainty that the dog or canine is not at risk for developing CKD. An accuracy of about 80% or more may indicate high certainty that the dog or canine is not at risk for developing CKD. In some non-limiting embodiments, a dog or canine with low certainty that it is not at risk for developing CKD may be classified as having future CKD. In some non-limiting embodiments, a high probability risk score may indicate that the dog or canine will develop CKD with high predictive accuracy. For example, high predictive accuracy may be greater than about 95%.

[0087] In some non-limiting embodiments, a moderate probability risk score may indicate inconclusive or insufficient data to accurately predict the likelihood that a dog or canine will or will not develop CKD. A moderate probability score may be, for example, a score between 40% and 60%, or any other value that does not indicate a high or low probability. On the other hand, a low probability risk score may indicate that a dog or canine will not develop CKD with high predictive accuracy. For example, high predictive accuracy may be greater than about 95%.

[0088] The probability risk score may relate to the risk of the dog or canine developing CKD within about 0 months, about 3 months, about 6 months, about 9 months, about 12 months, 0 years, about 0.5 years, about 1 year, about 2 years, about 3 years, about 4 years, or about 5 years or more after the amount of the value of the one or more biomarkers is determined or the probability score is determined. In certain other examples, the probability score may indicate the risk of the dog or canine developing CKD within about 12 months or about 2 years after the amount of the value of the one or more biomarkers is determined or the probability score is determined.

[0089] In certain non-limiting embodiments, a predictive model, such as an RNN, can be used to process one or more biomarkers, such as creatinine, BUN, urine specific gravity, urine protein, or demographic information, such as weight or age. In one example, a predictive model near the time of diagnosis can display a sensitivity of approximately 91.4% and a specificity of 97.2%. The specificity of the predictive model may remain at approximately 97% even one or two years prior to diagnosis, but the sensitivity may decrease to approximately 69% and 45%, respectively. Predictive models can help provide earlier diagnosis of CKD, allowing for greater opportunities for intervention and improved patient outcomes.

[0090] To determine the CKD prediction model, the dataset was randomly divided into two parts. A total of 37,210 health records, or approximately 67% of the data, were used to build the CKD prediction model. The remaining 18,705 health records, or approximately 33% of the data, were used as a test dataset to evaluate model performance or cross-validation. The model-building dataset and the model-testing dataset, also known as the training dataset, can be kept separate throughout the analysis to eliminate bias during the testing phase. Prior to use, missing information in the blood and urine test datasets can be estimated using all available blood and / or urine data, but CKD status information is unavailable. In certain non-limiting embodiments, missing information can be estimated because neural networks may require complete data. The model-building dataset and the test dataset were kept separate to avoid information flow between the datasets.

[0091] In certain non-limiting embodiments, for a predictive model to perform well in early detection of CKD, the model predicting the build dataset can be augmented by adding a truncated version of the original health records. For example, the last K visits can be removed, with K ranging from 1 to the total number of visits minus 1. In some embodiments, truncating the data can be useful for enriching the dataset with health records that have a gap of up to two years between the last visit in the dataset and the time of diagnosis. The truncated data can be useful for training the predictive model using a larger subset of pets that have a large gap in data between the last available data point and the date of diagnosis.

[0092] In some non-limiting embodiments, one of the first steps in building a CKD prediction model is to select a limited set of model features. Features can include one or more biomarkers and / or demographic information. Feature selection can be performed, for example, by a top-down or bottom-up wrapper method using a standard recurrent neural network. The recurrent neural network initially tested can have two hidden layers: a first layer with three nodes and a second layer with seven nodes. In some embodiments, the RNN can include a TanH activation function in the hidden layer and a softmax to convert the output layer into a CKD probability risk score. In one specific non-limiting embodiment, backpropagation through time can be used to train the RMSprop gradient optimization algorithm. Model performance can be evaluated, for example, based on the F1 cross-entropy in a three-fold cross-validation setup. The F1 cross-entropy can be used as a metric to balance sensitivity and specificity, independent of the incidence of CKD.

[0093] The selected features can then be used to perform a full model architecture screening. Various RNN configurations with 1 to 5 hidden layers were then tested, with 3 to 200 nodes per layer. In some non-limiting embodiments, a 20% dropout was added to avoid overfitting. For example, evaluation was based on F1 scores in a 10-fold cross-validation setup. The predictive model configuration can be fine-tuned with respect to training time in the same cross-validation setup.

[0094] In certain non-limiting embodiments, unbiased model performance can be evaluated by applying the selected predictive model to a test dataset. Predictions were performed on all dogs or canines in the CKD and "no CKD" groups. Predictive model results were interpreted as crude model output, such as the probability of CKD diagnosis, and / or post-classification output, with "no CKD" and CKD assessed using a p cut-point of 0.5. Categorical results for the "CKD" and "no CKD" groups were used to calculate sensitivity estimates (i.e., the proportion of true positives, "CKD" status predicted as CKD) and specificity estimates (i.e., the proportion of true negatives, "no CKD" predicted as no CKD), respectively. Confidence intervals for the sensitivity and specificity estimates were calculated using normal approximation. The ability of the model to predict CKD prior to a definitive diagnosis can be evaluated by truncating the health records to various time points prior to the age at assessment (T0) for the CKD group and / or allowing the predictive model to look only at the truncated data.

[0095] As discussed above, in some non-limiting embodiments, a standard RNN with a 3-7 hidden layer structure can be used as a starting point for a CKD prediction model. The described prediction model can recognize both multifactorial and temporal aspects of CKD diagnosis. This prediction model is used with 35 candidate factors, including one or more biomarkers and / or demographic information. In certain non-limiting embodiments, the most important features, such as one or more biomarkers and / or demographic information, can be selected using a top-down and bottom-up feature selection strategy. In some non-limiting embodiments, the cross-entropy score improved with the addition of up to six features but then plateaued. The best feature set can be, for example, creatinine, blood urea nitrogen, urine specific gravity, urine protein, weight, and age. These six features can be used to determine or select an updated prediction model. In certain non-limiting embodiments, a three-layer RNN with a 5-3-3 structure can be trained for eight epochs and show the best performance.

[0096] 4. Customized Recommendations In certain non-limiting embodiments, the customized recommendation can be determined based on the probability risk of the dog or canine developing CKD. The customized recommendation can be sent to the dog's or canine's veterinarian, owner, or caregiver. Generally, the customized recommendation can provide a method, process, test, or regimen for treating, preventing, or reducing the dog's or canine's risk of developing CKD. For example, the customized recommendation can include at least one of one or more therapeutic interventions, one or more dietary changes, one or more renal-sparing strategies, and / or one or more tests for disease progression.

[0097] In certain non-limiting embodiments, the customized recommendation may include testing for disease progression, such as testing serum parathyroid hormone levels. If the probability risk score is low or indicates a high degree of certainty that the dog or canine is not at risk for developing CKD, the customized recommendation may include testing for CKD within one week or more, one month or more, or one year or more after the calculation of the probability risk score. For example, the customized recommendation may include testing for CKD one or two years after the original probability risk score is determined. In some non-limiting embodiments, if a moderate probability score is determined or indicates a low degree of certainty that the dog or canine is not at risk for developing CKD, the customized recommendation may include testing the dog or canine for CKD within six months after the original probability risk score is determined. In other non-limiting embodiments, if the probability risk score is moderate or high or indicates a low degree of certainty that the dog or canine is at risk for developing CKD, the customized recommendation may include testing the dog or canine for CKD within three months after the original probability risk score is determined.

[0098] If the probabilistic risk score indicates a high certainty of risk of developing CKD, customized recommendations may include identifying the underlying product, testing the dog or canine for CKD, and / or continuing with the International Renal Research Association (IRIS) staging system disclosed herein.

[0099] For example, if the dog or canine's probabilistic risk score indicates a high certainty of risk for developing CKD, the customized recommendations can include therapeutic interventions, which can include at least one of monitoring water consumption and toileting habits, providing dietary therapy, providing supplemental recommendations, providing a high-quality diet with adequate phosphorus levels that is not protein-restricted, considering providing fatty acid supplements, avoiding nephrotoxic medications, implementing a dental treatment regimen, and / or maintaining good oral hygiene.

[0100] In certain non-limiting embodiments, the customized recommendation can include diagnosing the presence of a co-morbidity in the dog or canine. In certain embodiments, the co-morbidity can include one or more of hyperthyroidism, diabetes, liver damage, underweight, heart murmur, arthritis, malaise, constipation, gastroenteritis, vomiting, inflammatory bowel disease, crystalluria, enteritis, urinary tract infection, upper respiratory tract disease, urinary tract dysfunction, obesity, inadequate elimination, cystitis, colitis, and / or any combination thereof. In particular, in some non-limiting embodiments, the co-morbidity can include hyperthyroidism, diabetes, liver damage, underweight, heart murmur, and / or any combination thereof.

[0101] In certain embodiments, the customized recommendations may include therapeutic interventions or renal-sparing strategies. For example, the therapeutic interventions or renal-sparing strategies may include one or more of: avoidance of nonsteroidal anti-inflammatory drugs or aminoglycosides, hemodialysis, renal replacement therapy, discontinuation of kidney-damaging compounds, kidney transplantation, delay or avoidance of kidney-damaging procedures, modification of diuretic administration, and / or any combination thereof. In certain other non-limiting embodiments, the therapeutic interventions or renal-sparing strategies may include one or more of: reducing phosphate intake, reducing protein intake, administering polyunsaturated fatty acids, administering phosphate binder therapy, administering potassium, reducing dietary sodium intake, administering alkali supplements, or any combination thereof. See, e.g., Jonathan D. Foster, Update on Mineral and Bone Disorders in Chronic Kidney Disease, Vet Clin North Am Small Anim Pract. 2016 Nov;46(6):1131-49.

[0102] In certain embodiments, the customized recommendations may include dietary recommendations, such as nutritional recommendations, dietary or nutritional modifications, dietary or nutritional regimens, nutritional products, and / or dietary or nutritional therapy. Dietary recommendations may include recommendations for the use of any pet product and / or recommendations for the consumption or use of any pet product, such as pet food. For example, the recommendations may include one or more of a low-phosphorus diet, a low-protein diet, a low-sodium diet, a potassium-supplemented diet, a polyunsaturated fatty acid (PUFA, e.g., long-chain omega-3 fatty acid)-supplemented diet, an antioxidant-supplemented diet, a vitamin B-supplemented diet, a liquid diet, a calcium-supplemented diet, a normal protein diet, or any combination thereof. In other certain embodiments, the dietary recommendations may include one or more pet products. For example, the pet products may help delay the onset, limit the progression, mitigate the effects, minimize the physiological burden, or prevent CKD. For example, the diet may include low protein, low phosphorus, an increased calcium to phosphorus ratio, increased energy density, and / or a neutral acid-base balance.

[0103] In certain embodiments, a low phosphorus diet may contain between about 0.01% and about 5%, between about 0.1% and about 2%, between about 0.1% and about 1%, between about 0.05% and about 2%, or between about 0.5% and about 1.5% phosphorus by weight based on the weight of the pet food. In certain non-limiting embodiments, a low phosphorus diet can include about 0.01%, about 0.05%, about 0.1%, about 0.2%, about 0.3%, about 0.4%, about 0.5%, about 0.6%, about 0.7%, about 0.8%, about 0.9%, about 1%, about 1.1%, about 1.2%, about 1.3%, about 1.4%, about 1.5%, about 1.6%, about 1.7%, about 1.8%, about 1.9%, about 2%, about 3%, about 4%, about 5% phosphate, or any intermediate percentage or range of phosphate, by weight based on the weight of the pet food. In some non-limiting embodiments, the low phosphorus diet includes about 0.1g / 1000kcal, about 0.2g / 1000kcal, about 0.3g / 1000kcal, about 0.4g / 1000kcal, about 0.5g / 1000kcal, about 0.6g / 1000kcal, about 0.7g / 1000kcal, about 0.8g / 1000kcal, about 0.9g / 1000kcal, about 1.0g / 1000kcal, about 1.1g / 1000kcal, about 1.2g / 1000kcal, about 1.3g / 1000kcal, about 1.4g / 1000kcal, about 1.5g / 1000kcal, about 1.6g / 1000kcal, about 1.7g / 1000kcal, about 1.8g / 1000kcal, about 1.9g / 1000kcal, about 2.0g / 1000kcal, about 2.1g / 1000kcal, about 2.2g / 1000kcal, about 2.3g / 1000kcal, about 2.4g / 1000kcal, about 2.5g / 1000kcal, about 2.6g / 1000kcal, about 2.7g / 1000kcal, about 2.8g / 1000kcal, about 2.9g / 1000kcal, about 3.0g / 1000kcal, about 3.1g / 1000kcal, about 3.2g / 1000kcal, about 3.3g / 1000kcal, about 3.4g / 1000kcal, about 3.5g / 1000kcal, about 3.6g / 000kcal, about 1.7g / 1000kcal, about 1.8g / 1000kcal, about 1.9g / 1000kcal, about 2.0g / 1000kcal, about 2.1g / 1000kcal, about 2.2g / 1000kcal, about 2.5g / 1000kcal, about 2.8g / 1000kcal, about 3.0g / 1000kcal, about 3.5g / 1000kcal, about 4g / 1000kcal, about 5g / 1000kcal, about 10g / 1000kcal, about 15g / 1000kcal, about 20g / 1000kcal, or any intermediate percentage or range of phosphate.In certain other non-limiting embodiments, the low phosphorus diet includes between about 0.1g / 1000kcal and about 0.5g / 1000kcal, between about 0.5g / 1000kcal and about 1.0g / 1000kcal, between about 1.0g / 1000kcal and about 2.0g / 1000kcal, between about 2.0g / 1000kcal and about 5.0g / 1000kcal, between about 0.01 The diet may include between about 0.1 g / 1000 kcal and about 0.1 g / 1000 kcal, between about 0.05 g / 1000 kcal and about 1.0 g / 1000 kcal, between about 0.1 g / 1000 kcal and about 1 g / 1000 kcal, between about 0.1 g / 1000 kcal and about 2 g / 1000 kcal, and between about 1 g / 1000 kcal and 2 g / 1000 kcal of phosphate. In certain non-limiting embodiments, the low phosphorus diet may include about 0.5% phosphate by weight based on the weight of the pet food (e.g., about 1.2 g / 1000 kcal for a dry kidney diet or about 1.0 g / 1000 kcal for a wet kidney diet). In other examples, a low phosphorus diet can include about 0.9 or 1% phosphate by weight based on the weight of the pet food (e.g., about 1.8 g / 1000 kcal for a dry maintenance diet, or about 2.3 g / 1000 kcal for a wet maintenance diet). A low phosphorus diet can also include between about 1.0 g / 1000 kcal and about 1.5 g / 1000 kcal of phosphorus.

[0104] In certain non-limiting embodiments, the calcium supplement can include between about 0.01% and about 5%, between about 0.1% and about 2%, between about 0.1% and about 1%, between about 0.05% and about 2%, or between about 0.5% and about 1.5% calcium by weight based on the weight of the pet food. In some non-limiting embodiments, the calcium supplement can include about 0.01%, about 0.05%, about 0.1%, about 0.2%, about 0.3%, about 0.4%, about 0.5%, about 0.6%, about 0.7%, about 0.8%, about 0.9%, about 1%, about 1.1%, about 1.2%, about 1.3%, about 1.4%, about 1.5%, about 1.6%, about 1.7%, about 1.8%, about 1.9%, about 2%, about 3%, about 4%, about 5% calcium, or any intermediate percentage or range, by weight based on the weight of the pet food. In some other non-limiting embodiments, the calcium supplement includes about 0.1 g / 1000 kcal, about 0.2 g / 1000 kcal, about 0.3 g / 1000 kcal, about 0.4 g / 1000 kcal, about 0.5 g / 1000 kcal, about 0.6 g / 1000 kcal, about 0.7 g / 1000 kcal, about 0.8 g / 1000 kcal, about 0.9 g / 1000 kcal, about 1.0 g / 1000 kcal, about 1.1 g / 1000 kcal, about 1.2 g / 1000 kcal, about 1.3 g / 1000 kcal, about 1.4 g / 1000 kcal, about 1.5 g / 1000 kcal, about 1.6 g / 1000 kcal, about 1.7 g / 1000 kcal, about 1.8 g / 1000 kcal, about 1.9 g / 1000 kcal, about 2.0 g / 1000 kcal, about 2.1 g / 1000 kcal, about 2.2 g / 1000 kcal, about 2.3 g / 1000 kcal, about 2.4 g / 1000 kcal, about 2.5 g / 1000 kcal, about 2.6 g / 1000 kcal, about 2.7 g / 1000 kcal, about 2.8 g / 1000 kcal, about 2.9 g / 1000 kcal, about 3.0 g / 1000 kcal, about 3.1 g / 1000 kcal, about 3.2 g / 1000 kcal, about 3.3 g / 1000 kcal, about 3.4 g / 1000 kcal, about 3.5 g / 1000 kcal, about 3.6 g / 1 1000kcal, about 1.7g / 1000kcal, about 1.8g / 1000kcal, about 1.9g / 1000kcal, about 2.0g / 1000kcal, about 2.1g / 1000kcal, about 2.2g / 1000kcal, about 2.5g / 1000kcal, about 2.8g / 1000kcal, about 3.0g / 1000kcal, about 3.5g / 1000kcal, about 4g / 1000kcal, about 5g / 1000kcal, about 10g / 1000kcal, about 15g / 1000kcal, about 20g / 1000kcal, or any intermediate percentage or range of calcium may be included.In certain other non-limiting embodiments, the calcium supplement includes between about 0.1g / 1000kcal and about 0.5g / 1000kcal, between about 0.5g / 1000kcal and about 1.0g / 1000kcal, between about 1.0g / 1000kcal and about 2.5g / 1000kcal, between about 2.5g / 1000kcal and about 5.0g / 1000kcal, between about 0. Calcium may be included between about 0.1g / 1000kcal and about 0.1g / 1000kcal, between about 0.05g / 1000kcal and about 1.0g / 1000kcal, between about 0.1g / 1000kcal and about 1g / 1000kcal, between about 0.1g / 1000kcal and about 2g / 1000kcal, between about 1g / 1000kcal and 2g / 1000kcal.

[0105] In certain non-limiting embodiments, the combination of a calcium-supplemented diet and a low-phosphorus diet can include a calcium-to-phosphorus ratio (Ca:P ratio) of between about 1 and about 2, between about 1.1 and about 1.4, between about 1.2 and about 1.4, between about 1.1 and about 1.3, between about 1.3 and about 1.8, between about 1.4 and about 1.6, between about 1.5 and about 1.8, or between about 1.6 and about 1.8. In some non-limiting embodiments, the combination of a calcium-supplemented diet and a low-phosphorus diet can include a calcium-to-phosphorus ratio (Ca:P ratio) of about 1, about 1.1, about 1.2, about 1.3, about 1.4, about 1.5, about 1.6, about 1.7, about 1.8, about 1.9, or about 2.0.

[0106] In certain non-limiting embodiments, a low sodium diet can include between about 0.00001% and about 5%, between about 0.0001% and about 1%, between about 0.001% and about 0.1%, or between about 0.001% and about 0.05% sodium, by weight based on the weight of the pet food. In some non-limiting embodiments, a low sodium diet can include about 0.01%, about 0.05%, about 0.1%, about 0.2%, about 0.3%, about 0.4%, about 0.5%, about 0.6%, about 0.7%, about 0.8%, about 0.9%, about 1%, about 1.1%, about 1.2%, about 1.3%, about 1.4%, about 1.5%, about 1.6%, about 1.7%, about 1.8%, about 1.9%, about 2%, about 3%, about 4%, about 5% sodium, or any intermediate percentage or range of sodium, by weight based on the weight of the pet food. A low sodium diet can also include about 1 mg / kg / day, about 2 mg / kg / day, about 3 mg / kg / day, about 4 mg / kg / day, about 5 mg / kg / day, about 6 mg / kg / day, about 7 mg / kg / day, about 8 mg / kg / day, about 9 mg / kg / day, about 10 mg / kg / day, about 15 mg / kg / day, about 20 mg / kg / day, about 30 mg / kg / day, about 40 mg / kg / day, about 46 mg / kg / day, about 50 mg / kg / day, about 60 mg / kg / day, about 70 mg / kg / day, about 80 mg / kg / day, about 90 mg / kg / day, about 100 mg / kg / day, about 120 mg / kg / day, about 150 mg / kg / day, or any intermediate amount or range of sodium. In other non-limiting embodiments, a low sodium diet comprises sodium between about 1 mg / 1000 kcal and about 50 mg / 1000 kcal, between about 2 mg / 1000 kcal and about 20 mg / 1000 kcal, between about 5 mg / 1000 kcal and about 50 mg / 1000 kcal, between about 1 mg / 1000 kcal and about 10 mg / 1000 kcal, between about 0.1 mg / 1000 kcal and about 5 mg / 1000 kcal, between about 0.1 mg / 1000 kcal and about 10 mg / 1000 kcal, between about 0.1 mg / 1000 kcal and about 20 mg / 1000 kcal, between about 0.1 mg / 1000 kcal and about 40 mg / 1000 kcal, between about 10 mg / 1000 kcal and 20 mg / 1000 kcal.A low sodium diet can include, for example, about 0.4 to about 0.9 mmol / kg / day, or about 9.2 to about 20.7 mg / kg / day.

[0107] In certain non-limiting embodiments, the potassium supplement can include between about 0.00001% and about 5%, between about 0.0001% and about 1%, between about 0.001% and about 0.1%, or between about 0.001% and about 0.05% potassium supplement, by weight based on the weight of the pet food, in addition to the potassium present in the pet food. In other non-limiting embodiments, the potassium supplement can include about 0.1%, about 0.2%, about 0.3%, about 0.4%, about 0.5%, about 0.6%, about 0.7%, about 0.8%, about 0.9%, about 1%, about 1.1%, about 1.2%, about 1.3%, about 1.4%, about 1.5%, about 1.6%, about 1.7%, about 1.8%, about 1.9%, about 2%, about 3%, about 4%, about 5% or more of the potassium supplement, by weight based on the weight of the pet food, in addition to the potassium present in the pet food, or any intermediate percentage or range of the potassium supplement, by weight based on the weight of the pet food. In some non-limiting embodiments, the potassium supplemented diet can include about 1 mg / kg / day, about 2 mg / kg / day, about 3 mg / kg / day, about 4 mg / kg / day, about 5 mg / kg / day, about 6 mg / kg / day, about 7 mg / kg / day, about 8 mg / kg / day, about 9 mg / kg / day, about 10 mg / kg / day, about 15 mg / kg / day, about 20 mg / kg / day, about 30 mg / kg / day, about 40 mg / kg / day, about 50 mg / kg / day, about 60 mg / kg / day, about 70 mg / kg / day, about 80 mg / kg / day, about 90 mg / kg / day, about 100 mg / kg / day or more, or any intermediate amount or range of potassium supplemented diet, in addition to the potassium present in the pet food.In certain embodiments, the potassium supplemented diet may include between about 1 mg / 1000 kcal and about 10 mg / 1000 kcal, between about 2 mg / 1000 kcal and about 20 mg / 1000 kcal, between about 5 mg / 1000 kcal and about 50 mg / 1000 kcal, between about 1 mg / 1000 kcal and about 10 mg / 1000 kcal, between about 0.1 mg / 1000 kcal and about 5 mg / 1000 kcal, between about 0.1 mg / 1000 kcal and about 10 mg / 1000 kcal, between about 0.1 mg / 1000 kcal and about 20 mg / 1000 kcal, between about 0.1 mg / 1000 kcal and about 40 mg / 1000 kcal, between about 10 mg / 1000 kcal and 20 mg / 1000 kcal of potassium supplemented diet in addition to the potassium present in the pet food.

[0108] In certain non-limiting embodiments, the potassium supplement can include between about 0.01% and about 5%, between about 0.1% and about 2%, between about 0.1% and about 1%, between about 0.05% and about 2%, or between about 0.5% and about 1.5% potassium by weight based on the weight of the pet food. In other non-limiting embodiments, the potassium supplement can include about 0.01%, about 0.05%, about 0.1%, about 0.2%, about 0.3%, about 0.4%, about 0.5%, about 0.6%, about 0.7%, about 0.8%, about 0.9%, about 1%, about 1.1%, about 1.2%, about 1.3%, about 1.4%, about 1.5%, about 1.6%, about 1.7%, about 1.8%, about 1.9%, about 2%, about 3%, about 4%, about 5% potassium, or any intermediate percentage or range, by weight based on the weight of the pet food. In some non-limiting embodiments, the potassium supplemented diet includes about 0.1g / 1000kcal, about 0.2g / 1000kcal, about 0.3g / 1000kcal, about 0.4g / 1000kcal, about 0.5g / 1000kcal, about 0.6g / 1000kcal, about 0.7g / 1000kcal, about 0.8g / 1000kcal, about 0.9g / 1000kcal, about 1.0g / 1000kcal, about 1.1g / 1000kcal, about 1.2g / 1000kcal, about 1.3g / 1000kcal, about 1.4g / 1000kcal, about 1.5g / 1000kcal, about 1.6g / 1000kcal, about 1.7g / 1000kcal, about 1.8g / 1000kcal, about 1.9g / 1000kcal, about 2.0g / 1000kcal, about 2.1g / 1000kcal, about 2.2g / 1000kcal, about 2.3g / 1000kcal, about 2.4g / 1000kcal, about 2.5g / 1000kcal, about 2.6g / 1000kcal, about 2.7g / 1000kcal, about 2.8g / 1000kcal, about 2.9g / 1000kcal, about 3.0g / 1000kcal, about 3.1g / 1000kcal, about 3.2g / 1000kcal, about 3.3g / 1000kcal, about 3.4g / 1000kcal, about 3.5g / 1000kcal, about 3.6g / 1000kcal, about 1.7g / 1000kcal, about 1.8g / 1000kcal, about 1.9g / 1000kcal, about 2.0g / 1000kcal, about 2.1g / 1000kcal, about 2.2g / 1000kcal, about 2.5g / 1000kcal, about 2.8g / 1000kcal, about 3.0g / 1000kcal, about 3.5g / 1000kcal, about 4g / 1000kcal, about 5g / 1000kcal, about 10g / 1000kcal, about 15g / 1000kcal, about 20g / 1000kcal, or any intermediate percentage or range of potassium.Potassium supplementation also includes between about 0.1g / 1000kcal and about 0.5g / 1000kcal, between about 0.5g / 1000kcal and about 1.0g / 1000kcal, between about 1.0g / 1000kcal and about 2.5g / 1000kcal, between about 2.5g / 1000kcal and about 5.0g / 1000kcal, and between about 0.01g / 1000kcal. The potassium supplement may also include between about 0.1 g / 1000 kcal and about 0.1 g / 1000 kcal, between about 0.05 g / 1000 kcal and about 1.0 g / 1000 kcal, between about 0.1 g / 1000 kcal and about 1 g / 1000 kcal, between about 0.1 g / 1000 kcal and about 2 g / 1000 kcal, or between about 1 g / 1000 kcal and 2 g / 1000 kcal. In other non-limiting embodiments, the potassium supplement includes between about 2 g / 1000 kcal and about 2.5 g / 1000 kcal of potassium. In one specific embodiment, the potassium supplement includes about 2.1 g / 1000 kcal of potassium.

[0109] In certain non-limiting embodiments, a normal protein diet can include a protein level of between about 70g / 1000kcal and about 90g / 1000kcal, between about 70g / 1000kcal and about 75g / 1000kcal, between about 70g / 1000kcal and about 80g / 1000kcal, between about 80g / 1000kcal and about 90g / 1000kcal, or between about 85g / 1000kcal and about 90g / 1000kcal. In some non-limiting embodiments, a normal protein diet can include a protein level of about 73g / 1000kcal, about 74g / 1000kcal, or about 75g / 1000kcal.

[0110] In certain non-limiting embodiments, the low protein diet can include between about 0.0001% and about 20%, between about 0.001% and about 10%, between about 0.01% and about 5%, between about 0.05% and about 2%, or between about 0.01% and about 1% protein by weight based on the weight of the pet food. In some non-limiting embodiments, the low protein diet can include about 0.01%, about 0.05%, about 0.1%, about 0.2%, about 0.3%, about 0.4%, about 0.5%, about 0.6%, about 0.7%, about 0.8%, about 0.9%, about 1%, about 1.1%, about 1.2%, about 1.3%, about 1.4%, about 1.5%, about 1.6%, about 1.7%, about 1.8%, about 1.9%, about 2%, about 3%, about 4%, about 5%, about 10%, about 15%, about 20% protein, or any intermediate percentage or range, by weight based on the weight of the pet food. In other non-limiting embodiments, the low protein diet can include about 1 g / kg / day, about 2 g / kg / day, about 3 g / kg / day, about 4 g / kg / day, about 5 g / kg / day, about 6 g / kg / day, about 7 g / kg / day, about 8 g / kg / day, about 9 g / kg / day, about 10 g / kg / day, about 15 g / kg / day, about 20 g / kg / day, or any intermediate amount or range of protein. The low protein diet can also include between about 1 g / kg / day to about 20 g / kg / day, between about 1 g / kg / day to about 50 g / kg / day, between about 2 g / kg / day to about 30 g / kg / day, between about 2 g / kg / day to about 10 g / kg / day, between about 2 g / kg / day to about 8 g / kg / day, between about 5 g / kg / day to about 20 g / kg / day, or any intermediate amount or range of protein. The low protein diet can include about 4 to about 6 g / kg / day, or about 5 to about 5.5 g / kg / day.

[0111] In certain non-limiting embodiments, the polyunsaturated fatty acid (PUFA) supplement can include, by weight based on the weight of the pet food, between about 0.01% and about 30%, between about 0.1% and about 20%, between about 1% and about 10%, between about 0.1% and about 5%, or between about 1% and about 10% of the PUFA supplement in addition to the PUFAs present in the pet food. In some non-limiting embodiments, the PUFA supplemented diet can include about 0.1%, about 0.2%, about 0.3%, about 0.4%, about 0.5%, about 0.6%, about 0.7%, about 0.8%, about 0.9%, about 1%, about 1.1%, about 1.2%, about 1.3%, about 1.4%, about 1.5%, about 1.6%, about 1.7%, about 1.8%, about 1.9%, about 2%, about 3%, about 4%, about 5%, about 10%, about 15%, about 20%, about 25%, about 30% or more of the PUFA supplemented diet in addition to the PUFAs present in the pet food, by weight based on the weight of the pet food, or any intermediate percentage or range of the PUFA supplemented diet in addition to the PUFAs present in the pet food. In other non-limiting embodiments, the PUFA supplemented diet can include about 0.1 g / kg / day, about 0.5 g / kg / day, about 1 g / kg / day, about 1 g / kg / day, about 2 g / kg / day, about 3 g / kg / day, about 4 g / kg / day, about 5 g / kg / day, about 6 g / kg / day, about 7 g / kg / day, about 8 g / kg / day, about 9 g / kg / day, about 10 g / kg / day, about 15 g / kg / day, about 20 g / kg / day, about 30 g / kg / day, about 40 g / kg / day, about 50 g / kg / day, about 60 g / kg / day, about 70 g / kg / day, about 80 g / kg / day, about 90 g / kg / day, about 100 g / kg / day, or any intermediate amount or range of PUFA supplemented diet in addition to the PUFAs present in the pet food. In certain other non-limiting embodiments, the PUFA supplemented diet can include between about 0.1 g / kg / day to about 20 g / kg / day, between about 1 g / kg / day to about 100 g / kg / day, between about 2 g / kg / day to about 200 g / kg / day, between about 5 g / kg / day to about 150 g / kg / day, between about 10 g / kg / day to about 100 g / kg / day, between about 5 g / kg / day to about 50 g / kg / day, or any intermediate amount or range of PUFA supplemented diet in addition to the PUFAs present in the pet food.A PUFA supplemented diet can also include PUFA levels of between about 1g / 1000kcal and about 10g / 1000kcal, between about 1g / 1000kcal and about 5g / 1000kcal, between about 5g / 1000kcal and about 10g / 1000kcal, between about 1g / 1000kcal and about 3g / 1000kcal, between about 1g / 1000kcal and about 2g / 1000kcal, between about 2g / 1000kcal and about 4g / 1000kcal, between about 5g / 1000kcal and about 8g / 1000kcal, between about 7g / 1000kcal and about 10g / 1000kcal. In certain embodiments, the PUFA supplemented diet comprises a PUFA level of about 1 g / 1000 kcal, about 2 g / 1000 kcal, about 2.1 g / 1000 kcal, about 3 g / 1000 kcal, about 4 g / 1000 kcal, about 5 g / 1000 kcal, about 6 g / 1000 kcal, about 7 g / 1000 kcal, about 8 g / 1000 kcal, about 9 g / 1000 kcal, or about 10 g / 1000 kcal.

[0112] In certain non-limiting embodiments, the PUFA supplemented diet can include n-6 PUFAs such as vegetable oils, n-3 PUFAs such as fish oils, eicosapentaenoic acid (EPA), and / or docosahexaenoic acid (DHA).

[0113] In certain non-limiting embodiments, the antioxidant supplement can include between about 0.001% and about 5%, between about 0.01% and about 1%, between about 0.01% and about 2%, between about 0.1% and about 1%, or between about 1% and about 5% of the antioxidant present in the pet food, by weight based on the weight of the pet food. In some non-limiting embodiments, the antioxidant supplement comprises, by weight based on the weight of the pet food, in addition to the antioxidants present in the pet food, about 0.1%, about 0.2%, about 0.3%, about 0.4%, about 0.5%, about 0.6%, about 0.7%, about 0.8%, about 0.9%, about 1%, about 1.1%, about 1.2%, about 1.3%, about 1.4%, about 1.5%, about 1.6%, about 1.7%, about 1.8%, about 1.9%, about 2%, about 3%, about 4%, about 5% or more of the antioxidant supplement, or any intermediate percentage or range. In other non-limiting embodiments, the antioxidant supplemented diet can include about 1 mg / kg / day, about 2 mg / kg / day, about 3 mg / kg / day, about 4 mg / kg / day, about 5 mg / kg / day, about 6 mg / kg / day, about 7 mg / kg / day, about 8 mg / kg / day, about 9 mg / kg / day, about 10 mg / kg / day, about 15 mg / kg / day, about 20 mg / kg / day, about 30 mg / kg / day, about 40 mg / kg / day, about 50 mg / kg / day, about 60 mg / kg / day, about 70 mg / kg / day, about 80 mg / kg / day, about 90 mg / kg / day, about 100 mg / kg / day or more, or any intermediate amount or range of antioxidant supplements, in addition to the antioxidants present in the pet food. An antioxidant-supplemented diet can include, in addition to the antioxidants present in the pet food, an antioxidant supplement of between about 1 mg / kg / day and about 20 mg / kg / day, between about 1 mg / kg / day and about 100 mg / kg / day, between about 2 mg / kg / day and about 200 mg / kg / day, between about 5 mg / kg / day and about 150 mg / kg / day, between about 10 mg / kg / day and about 100 mg / kg / day, between about 5 mg / kg / day and about 50 mg / kg / day, or any intermediate amount or range. In certain non-limiting embodiments, the antioxidant can be one or more of vitamin E, vitamin C, taurine, carotenoids, flavanols, or any combination thereof.For example, the flavanols can be catechin, epicatechin, epigallocatechin gallate, procyanidins, tannins, or any combination thereof. Antioxidant supplements can also include plants with high flavanol concentrations, such as cocoa, grapes, and green tea.

[0114] In certain non-limiting embodiments, the B vitamin supplement can include vitamin B1 (thiamine), vitamin B2 (riboflavin), vitamin B3 (niacin or nicotinamide riboside), vitamin B5 (pantothenic acid), vitamin B6 (pyridoxine, pyridoxal, or pyridoxamine), vitamin B7 (biotin), vitamin B9 (folic acid), vitamin B12 (cobalamins, e.g., cyanocobalamin or methylcobalamin), or any combination thereof. In some non-limiting embodiments, the B vitamin supplement can include between about 0.001% and about 2%, between about 0.01% and about 1%, between about 0.05% and about 1%, between about 0.001% and about 0.1%, or between about 0.01% and about 0.2% B vitamins, by weight based on the weight of the pet food, in addition to the B vitamins present in the pet food. In other non-limiting embodiments, the B vitamin supplemented diet comprises about 0.1%, about 0.2%, about 0.3%, about 0.4%, about 0.5%, about 0.6%, about 0.7%, about 0.8%, about 0.9%, about 1%, about 1.1%, about 1.2%, about 1.3%, about 1.4%, about 1.5%, about 1.6%, about 1.7%, about 1.8%, about 1.9%, about 2% or more B vitamins, or any intermediate percentage or range, by weight based on the weight of the pet food, in addition to the B vitamins present in the pet food. In certain embodiments, a vitamin B supplemented diet can include about 1 mg / kg / day, about 2 mg / kg / day, about 3 mg / kg / day, about 4 mg / kg / day, about 5 mg / kg / day, about 6 mg / kg / day, about 7 mg / kg / day, about 8 mg / kg / day, about 9 mg / kg / day, about 10 mg / kg / day, about 15 mg / kg / day, about 20 mg / kg / day, about 30 mg / kg / day, about 40 mg / kg / day, about 50 mg / kg / day, about 60 mg / kg / day, about 70 mg / kg / day, about 80 mg / kg / day, about 90 mg / kg / day, about 100 mg / kg / day or more, or any intermediate amount or range, of a vitamin B supplement in addition to the B vitamins present in the pet food.In certain non-limiting embodiments, a vitamin B supplemented diet can include between about 1 mg / kg / day to about 20 mg / kg / day, between about 1 mg / kg / day to about 100 mg / kg / day, between about 2 mg / kg / day to about 200 mg / kg / day, between about 5 mg / kg / day to about 150 mg / kg / day, between about 10 mg / kg / day to about 100 mg / kg / day, between about 5 mg / kg / day to about 50 mg / kg / day, or any intermediate amount or range of vitamin B supplements in addition to the B vitamins present in the pet food.

[0115] In certain non-limiting embodiments, the dietary therapy can include one or more of a low-phosphorus diet, a calcium-supplemented diet, a potassium-supplemented diet, a normal protein diet, or any combination thereof. In some non-limiting embodiments, the dietary therapy can include feeding a dog or canine at risk of developing CKD a diet containing a phosphorus level of about 1.5 g / 1000 kcal, a calcium level of about 2 g / 1000 kcal, a Ca:P ratio of about 1.3, a potassium level of about 2.1 g / 1000 kcal, and a protein level of about 74 g / 1000 kcal. In other non-limiting embodiments, the dietary therapy can be any dietary modification or therapy known in the art.

[0116] In certain non-limiting embodiments, based on the customized recommendations, a medical professional or veterinarian can administer customized recommendations to the pet or canine.

[0117] 5. Devices, Systems, and Applications In certain non-limiting embodiments, the embodiments described herein provide a computer system or method for identifying a dog's susceptibility to developing CKD. Any of the steps or processes including "receiving," "processing," "determining," or "transmitting" can be performed by one or more of the devices or apparatuses shown in FIG.

[0118] FIG. 4 illustrates a computing system 400 configured to provide a neonatal mortality application capable of implementing embodiments of the present disclosure. As shown, computing system 400 may include multiple web servers 408, a predictive model server 412, and multiple user computers / devices (e.g., mobile / wireless devices) 402 (only two are shown for clarity), each connected to a communications network 406 (e.g., the Internet). Web server 408 may communicate with database 414 via a local connection (e.g., a storage area network (SAN) or network-attached storage (NAS)) or over the Internet (e.g., a cloud-based storage service). Web server 408 may be configured to directly access the data contained in database 414 or may be configured to interface with a database manager, which may be configured to manage the data contained in database 414. Account 416 is a data object that may store data associated with a user, such as the user's email address, password, contact information, billing information, animal information, etc.

[0119] Each user computer 402 may include conventional components of a computing device, such as a processor, system memory, a hard disk drive, a battery, input devices such as a mouse and keyboard, and / or an output device such as a monitor or graphic user interface, and / or a combined input / output device such as a touchscreen that can receive input as well as display output. Each web server 408 and predictive model server 412 includes a processor and system memory (not shown) and may be configured to manage content stored in database 414 using, for example, relational database software and / or a file system. The web servers 408 may be programmed to communicate with each other and with the user computers 402 and the predictive model server 412 using, for example, a network protocol such as the TCP / IP protocol. The predictive model server 412 may communicate directly with the user computers 402, for example, via communications network 406. The user computers 402 may be programmed to execute software 404, such as web browser programs and other software applications, and may access web pages and / or applications hosted by the web server 408, for example, by specifying a Uniform Resource Locator (URL) that can direct them to the web server 408.

[0120] In the implementation described below, each user may operate a user computer 402 that may be connected to a web server 408 via a communications network 406. Web pages may be displayed to the user via the user computer 402. The web pages may be transmitted from the web server 408 to the user's computer 402 and processed by a web browser program stored on the user's computer 402 for display via a display device and / or graphical user interface in communication with the user's computer 402.

[0121] In one example, the information and / or images displayed on the user's computer 402 may relate to customized recommendations or any information contained in the health record, including one or more biomarkers and / or demographic information, accessed via the online database. The user's computer 402 may access the pet's health information via the communications network 406, which in turn retrieves the pet's health information from a web server 408 connected to a database 414 and displays the information and / or images via a graphical user interface on the user's computer 402. The online information and / or images, and / or neonatal mortality application may be controlled by a username and password combination or other similar limited access / verification access method that allows a user to "log in" and access the information.

[0122] It should be noted that user computer 402 can be a personal computer, laptop, handheld computing device, smartphone, video game controller, home digital media player, networked television, set-top box, and / or other computing device having components suitable for communication with communications network 406. User computer 402 may also run other software applications configured to receive customized recommendations from a predictive model server, such as, but not limited to, text and / or image display software, media players, computer and video games, and / or widget platforms, among others.

[0123] Figure 5 shows a more detailed diagram of the predictive model server 412 of Figure 4. The predictive model server 512 may include, but is not limited to, a central processing unit (CPU) 502, a network interface 504, memory 520, and storage 530 that communicate via an interconnect 506. The predictive model server 512 may also include an I / O device interface 508 that connects to I / O devices 510 (e.g., keyboard, video, mouse, audio, touch screen, etc.). The predictive model server 512 may further include a network interface 504 configured to transmit data over the data communications network 406.

[0124] The CPU 502 can retrieve and execute programming instructions stored in memory 520 and generally control and coordinate the operation of other system components. Similarly, the CPU 502 can store and retrieve application data resident in memory 520. Typical CPUs 502 can include a single CPU, multiple CPUs, a single CPU with multiple processing cores, etc. The interconnect 506 can be used to transmit programming instructions and application data between the CPU 502, the I / O device interface 508, the storage 530, the network interface 504, and the memory 520.

[0125] Memory 520, typically random access memory, may typically be included, which stores software applications and data used by CPU 502 during operation. Storage 530, although illustrated as a single unit, may be a combination of fixed and / or removable storage devices configured to store non-volatile data, such as a fixed disk drive, floppy disk drive, random access memory, hard disk drive, non-transitory computer-readable medium, flash memory storage drive, tape drive, removable memory card, CD-ROM, DVD-ROM, Blu-Ray, HD-DVD, optical storage, network attached storage (NAS), cloud storage, or storage area network (SAN).

[0126] The memory 520 can store instructions and logic for executing the application platform 526, which can include images 528 and / or predictive model software 538. The storage 530 can store images and / or information 534 and other user-generated media and can include a database 532, which can be configured to store images and / or information 534 related to the application platform content 5236. The database 532 can be any type of storage device and / or can include one or more datasets described herein. The database 532 can store application content related to data related to the user-generated media or images. The database 532 can also include one or more biomarkers or demographic information, and / or customized recommendations.

[0127] A network computer is another type of computer system that can be used in conjunction with the disclosure provided herein. Network computers generally do not include a hard disk or other mass storage; executable programs are loaded from a network connection into memory 520 for execution by CPU 502. A Web TV system can also be considered a computer system, but may lack some of the functionality shown in Figure 5, such as certain input / output devices. A typical computer system will generally include at least a processor, memory, and an interconnect coupling the memory to the processor.

[0128] 6 illustrates a user computer or device 402 used to access the predictive model server 412 and display images and / or information related to the application platform 620. The user computer or device 602 may be, for example, a desktop computer, a laptop computer, a mobile device, or any other user device. The user computer 602 may include, but is not limited to, a central processing unit (CPU) 602, a network interface 604, an interconnect 606, memory 620, and storage 630. The user computer 602 may also include an I / O device interface 608 that connects I / O devices 610 (e.g., devices such as a keyboard, display, touchscreen, and mouse) to the user computer 602.

[0129] Similar to CPU 502, CPU 602 may be included, typically a single CPU, multiple CPUs, a single CPU with multiple processing cores, etc., and memory 620 may generally be included, typically random access memory. Interconnect 606 may be used to transmit programming instructions and application data between CPU 602, I / O device interface 608, storage 630, network interface 604, and memory 620. Network interface 604 may be configured to transmit data over communications network 406, for example, to stream or serve content from predictive model server 512. Storage 630, such as a hard disk drive or solid-state storage drive (SSD), may store non-volatile data. Storage 630 may include photographs 632, graphs 634, charts 636, documents 638, and other media 640. Illustratively, memory 620 may include application interface 622, which may itself display images 624, such as graphs or charts, and / or information 626, among others. The application interface 622 may provide one or more software applications that may allow users to access media items and other content hosted by the predictive model server 412.

[0130] All of the above terms are merely convenient labels applied to these physical quantities. Unless otherwise indicated, as will be apparent from the discussion that follows, it will be recognized that throughout the description, discussions utilizing terms such as "processing" or "computing" or "calculating" or "determining" or "displaying" or "analyzing" refer to the actions and processes of a computer system, server, or any other electronic computing device that manipulates and converts data represented as physical (electronic) quantities in the computer system's registers and memory into other data that are also represented as physical quantities in the computer system's memory, registers, or other such information storage, transmission, or display device.

[0131] The present embodiments also relate to apparatus for performing the operations herein. This apparatus may be specially constructed for the required purposes, or it may comprise a general-purpose computer selectively activated or reconfigured by a computer program stored in the computer. Such computer program may be stored on a computer-readable storage medium such as, but not limited to, read-only memory (ROM), random-access memory (RAM), EPROM, EEPROM, flash memory, magnetic or optical cards, floppy disks, optical disks, CD-ROMs, and any type of disk, including magneto-optical disks, or any type of medium suitable for storing electronic instructions, each coupled to the interconnection of a computer system.

[0132] The algorithms and displays presented herein are not inherently related to any particular computer or other apparatus. Various general-purpose systems can be used with programs in accordance with the teachings herein, or it may prove convenient to construct more specialized apparatus to perform the required method acts. The structure of a variety of these systems will be apparent from the above description. In addition, the present embodiments are not described with reference to any particular programming language; thus, various examples can be implemented using a variety of programming languages.

[0133] The algorithms and displays presented herein are not inherently related to any particular computer or other device. Various general-purpose systems can be used with applications in accordance with the teachings herein, or it may prove convenient to construct more specialized devices to perform the required method operations. The structure of these various systems will be apparent from the above description. In addition, the present embodiments are not described with reference to any particular programming language; thus, various examples may be implemented using a variety of programming languages. All preferred features and / or embodiments of the methods and diets / diet plans disclosed in this application apply to devices, systems, and applications. [Example]

[0134] The subject matter disclosed herein may be better understood by reference to the following: The following examples are illustrative only and should not be considered limiting in any way.

[0135] Example 1 To determine the performance of the CKD prediction model around the time of diagnosis, the prediction model was applied to 15,044 of 18,705 dogs or canines in the study dataset who were visited within 3 months prior to the time of evaluation (T0). The prediction model demonstrated a sensitivity of 91.4% (687 / 752) for dogs or canines classified as "CKD" and a specificity of 97.2% (13891 / 14292) for dogs or canines classified as "no CKD." Given that age was a function of the prediction model, sensitivity and specificity were also reported for age at the time of evaluation (T0).

[0136] 7 shows a graph 702 illustrating performance metrics according to certain embodiments described herein. Specificity, for example, was consistently above 98% up to age 8, then decreased, reaching 67.0% at age 15. Sensitivity, on the other hand, increased with age, reaching above 96% after age 12. When predicting older pets, where the prevalence of CKD is much higher, the predictive model can sacrifice some specificity (increased false positive rate) for improved sensitivity (reduced false negative rate), thus optimizing overall accuracy for each age group.

[0137] Figure 8 shows a graph 802 illustrating performance metrics according to certain embodiments described herein. To understand how patient history affected model performance, the sensitivity of the model can be examined as a function of the number of health record visits before a diagnosis was made. Sensitivity increases from 76.8% with two visits before diagnosis to 85.5% with four visits, and continues to improve beyond 92% with additional data. This demonstrates that historical information can contribute to the quality of CKD diagnoses.

[0138] FIG. 9 shows a graph 902 illustrating performance metrics according to certain embodiments described herein. Because CKD diagnosis benefits from historical health record information, the ability of a model to predict future CKD diagnoses can be evaluated. To evaluate a predictive model, health records for "CKD" can be truncated at various time points prior to diagnosis. For example, for a one-year-old prediction, all information between diagnosis and one year prior can be removed. The predictive model's ability to predict future CKD development can then be evaluated. As expected, sensitivity decreased with increasing time between prediction and diagnosis, but 69.1% of dogs developing CKD were correctly predicted one year prior to diagnosis, 44.9% two years prior to diagnosis, and 22.0% 3.5 years prior to the initial or original CKD diagnosis.

[0139] In certain non-limiting embodiments, assessing specificity by truncating the data set may not make sense given that the dog remains classified as "CKD-free" at all previous visits. The specificity of early CKD detection can be best understood by the distribution according to age, as shown in Figure 7, where the mean level can exceed 90% for dogs up to 12 years of age.

[0140] In some non-limiting embodiments, advanced computational modeling can be used to predict CKD risk based on current and / or historical health records. This dataset can include assessed and refined clinicopathology results. The predictive model can include one or more of the following six features: serum creatinine, blood urea nitrogen, urine protein and urine specific gravity, and patient age and weight. The predictive model can demonstrate a sensitivity of 76.8% at two visits, increasing to 85.5% at four visits. This highlights the value of regular proactive health monitoring and obtaining as complete clinicopathology data as possible. Specificity can continue to increase beyond 92% with additional data. In some non-limiting embodiments, specificity, as indicated by a low false positive rate, can be useful, especially when considering predictive testing and subsequent clinical decision-making and communication with owners.

[0141] Current predictive models, such as the 5-3-3RNN, may differ from previously described predictive models, such as the 7-3RNN, in that they include one or more of urinary protein, patient weight, or any other additional biomarker or demographic information. In certain non-limiting embodiments, renal protein loss and associated proteinuria may be a more common feature of canine kidney disease than canine or canine kidney disease. Patient weight has been determined to be a useful component of predictive models, likely due to the large range of weights that can be seen in canine patients and the somewhat variable prevalence of disease among dogs of different sizes. In certain non-limiting embodiments, both the challenges of maintaining adequate nutrition and lean body mass and the difficulty of providing the necessary nursing care to patients with CKD may adversely affect the prognosis and outcomes of larger patients.

[0142] In some non-limiting embodiments, an individual patient's clinicopathological variables may change to some extent over time and / or may be affected by factors such as changes in dietary intake, posture, muscle mass, and / or hydration status. Commonly utilized reference ranges for evaluating clinicopathological results may provide limited interpretation and may not accurately reflect the unique condition of an individual patient. The use of predictive modeling can help accurately diagnose CKD.

[0143] Example 2 In a specific, non-limiting embodiment, the prediction model may include an RNN. The RNN includes a 5-3-3 architecture, 10 splits, and 18 epochs. The RNN evaluates seven selected features, such as one or more biomarkers and / or demographic information. For example, the seven selected features may be BUN, urine sigma, visit age, creatinine, urine protein, weight, and / or amylase. The performance of the RNN was measured for random CKD pets at two-year stratification points. The prediction model achieved an area under the receiver operating characteristic (AUCROC) of 94.2%, an area under the precision-recall curve (AUCPR) of 91.6%, and an F1 score of 82.6%.

[0144] In some non-limiting embodiments, the dataset used to train the predictive model included approximately 306,757 visit records for approximately 39,442 unique dogs or canines. Of the 39,442 unique dogs or canines, approximately 26,514 were CKD-free and approximately 12,928 had or developed CKD. The dataset included 35 features, including demographic information and one or more biomarkers derived from blood chemistry, hematology, and / or urine levels. Examples of the one or more biomarkers may include, but are not limited to, alkaline phosphatase, amylase, protein, BUN or urea levels, creatinine, phosphorus, calcium, urinary protein, potassium, glucose, hematocrit, hemoglobin, red blood cell (RBC) count, red blood cell distribution width (RDW), alanine aminotransferase, albumin, bilirubin, chloride, cholesterol, eosinophils, globulins, lymphocytes, monocytes, mean corpuscular hemoglobin (MCH), mean corpuscular hemoglobin concentration (MCHC), mean corpuscular volume (MCV), mean platelet volume (MPV), platelet count, segmented neutrophils, sodium, urinary pH level, and / or white blood cell count. Demographic information may include, for example, the age and weight of the dog or canine, and the age at which the dog or canine was first or originally diagnosed with CKD.

[0145] The number of visits per dog or canid included in the dataset ranges from 1 to 40. For example, the average number of visits per dog or canid may be 7.7. The number of visits is biased between 1 and 15 visits, with a rapid increase from 1 to 4 visits, followed by a steady decrease from 5 to 40 visits. In certain non-limiting embodiments, the dataset may contain missing information for one or more of 35 features. For example, features with more than 60% missing values ​​include approximately 66.5% urinary protein, approximately 64.2% urinary specific gravity, approximately 70% potassium, approximately 70.4% chloride, approximately 60.6% eosinophils, approximately 70.2% sodium, and approximately 64.8% urinary pH. Any other feature, such as one or more biomarkers or demographic information, may be missing in between 0% and 100% of the data in the database. Missing information can be imputed.

[0146] 10A-10E illustrate the preprocessing 1002 of a dataset according to certain implementations described herein. For example, as part of the preprocessing, missing data in the database can be imputed using, for example, a random forest implementation. In another example, the data can be normalized to a minimum and maximum value of each feature ranging from 0 to 1. In yet another example, the preprocessing can include a pre-smoothing function by thresholding the minimum and maximum values ​​to predefine their limits. As shown in FIGS. 10A-10E, a bottom-up wrapper can be used to select features to smooth to maximize the cumulative explained variance calculated via principal component analysis (PCA). Thus, the values ​​shown in FIGS. 10A-10E range from 0 to 1 and are smoothed to maximize the PCA explained variance. When all features are smoothed, the explained variance can be 75%, which indicates a noisy dataset. By optimizing feature smoothing, the PCA explained variance can be 95%.

[0147] FIG. 11 illustrates principal component analysis or factor analysis according to certain embodiments described herein. For example, graph 1102 illustrates the projection of a dataset into two-dimensional space by performing 2-D PCA linear regression. Meanwhile, graph 1104 illustrates PCA explained variance and variance ratios, which can confirm the absence of noise in the dataset. In some non-limiting embodiments, graph 1108 may illustrate two-dimensional factor analysis, measuring the number of components against cross-validation scores. Graph 1106 may illustrate two-dimensional factor analysis. Graph 1110 may illustrate two-dimensional t-distribution stochastic neighbor embedding (t-SNE) nonlinear dimensionality reduction.

[0148] After the dataset is preprocessed, kernel density estimation can be used on each feature to calculate the true probability density function (PDF) at each bin and / or to find a normalized histogram that can be used to show the original distribution from which the data was originally sampled. After analyzing the data using PDFs, in one specific, non-limiting embodiment, the most discriminatory features may be BUN, urine specific gravity, creatinine, and visit age. For each of these features, the normalized distribution for CKD / health was found to vary in region and range. Other features, such as amylase, urine protein, and cholesterol, also appeared to be potentially useful distinguishing features.

[0149] In one specific, non-limiting embodiment, univariate feature ranking was applied to the preprocessed dataset. Feature ranking included, for example, fast filtering methods (including linear and / or quadratic types) based on signal-to-noise ratio (SNR) and embedded machine learning methods. The results of the fast filtering methods and / or machine learning methods can be combined to create an average ranking of features. The fast filtering methods can include, for example, one or more of the following: correlation coefficient f-score, class-conditional SNR, two-sample test statistic, symmetric divergence, and / or Fisher discriminant ratio. The machine learning methods can include, for example, one or more of randomized lasso, ridge regression, random forest, and / or recursive feature reduction. Based on one or more normalized weights of the fast filtering and / or machine learning methods, the top four features can be visit age, BUN, urine specific gravity, and creatinine. The next three features can be amylase, cholesterol, and / or urine protein. The remaining characteristics were ranked as follows: potassium, alkaline phosphatase, lymphocytes, MCV, weight, globulin, urine PH, hematocrit, phosphorus, platelet count, hemoglobin, alanine aminotransferase, RDW, MCH, chloride, segmented neutrophils, RBC count, MPV, total protein, white blood cell count, glucose, calcium, MCHC, albumin, sodium, eosinophils, monocytes, and / or bilirubin.

[0150] In some non-limiting embodiments, a predictive model can be designed to determine whether a given canid or dog will develop CKD within the next two years. To design a predictive model that precisely addresses the above, a pan-pet database can be constructed, which can be a superset of all possible visit trajectories. For a dog or canid with N visits, a trajectory can be temporarily defined according to a list of visit numbers. The last K visits can then be removed to order the reduced trajectories with any ordered subset of visit numbers. K can be any number of visits between 1 and N. If the original dataset is expanded to include all possible reduced trajectories of CKD and visits up to two years prior to diagnosis are removed, the resulting expanded dataset can be referred to as a pan-canid dataset.

[0151] The sample dataset, in certain non-limiting embodiments, can be a subset of a pan-canid dataset in which a single trajectory is selected for each CKD dog or canid. Multiple sampled datasets can be created using a random number generator with different seeds, allowing different trajectories (e.g., different visit numbers) to be selected for each pet's identity (e.g., sampling with replacement). Using the sampled datasets, the predictor can learn patterns of pets likely to be diagnosed with CKS at any time over the next two years.

[0152] In one specific, non-limiting embodiment, a 3-7 RNN can be used to generate baseline performance for a predictive model using a dataset. For example, the 3-7 RNN can be trained with 10-fold cross-validation and 16 epochs. This RNN can then be used to demonstrate the performance of the top four features, which can be visit age, BUN, creatinine, and / or urine specific gravity, as well as the performance of all 35 features. Performance of the top four features using the 3-7 RNN resulted in 76.2% sensitivity, 93.9% specificity, 88.2% accuracy, and an F1 score of 80.6%. Meanwhile, performance of all 35 features using the 3-7 RNN resulted in 77.2% sensitivity, 95.1% specificity, 89.3% accuracy, and an F1 score of 82.3%.

[0153] To reduce the total computational cost by a factor of six, we can train the RNN 3-7 configuration with 3-fold cross-validation and 10 epochs. We then demonstrate the performance of the top four features, which can be visit age, BUN, creatinine, and / or urine specific gravity, as well as the performance of all 35 features using this RNN. Performance of the top four features using this RNN yielded 76.3% sensitivity, 92.5% specificity, 87.3% accuracy, and an F1 score of 79.5%. Meanwhile, performance of all 35 features using the RNN with the 3-7 configuration yielded 75.1% sensitivity, 95.0% specificity, 88.6% accuracy, and an F1 score of 80.9%.

[0154] Another comparison can be performed using an RNN with 10-fold cross-validation and 16 epochs (LSTM3-7). Next, we demonstrate the performance of the top four features using this LSTM3-7, which can be visit age, BUN, creatinine, and / or urine specific gravity. Performance of the top four features using this LSTM3-7 yielded a sensitivity of 75.7%, a specificity of 94.3%, an accuracy of 88.3%, and an F1 score of 80.7%. For further comparison, we can use an RNN with a 3-5-3 configuration trained with 10-fold cross-validation and 20 epochs using the top six features, such as BUN, creatinine, urine protein, urine specific gravity, urine pH, and WBC. Performance of the top six features using the 3-5-3 configuration RNN yielded a sensitivity of 75.9%, a specificity of 93.0%, an accuracy of 87.5%, and an F1 score of 79.6%.

[0155] FIG. 12 illustrates a wrapper-based feature ordering chart 1202 according to certain embodiments described herein. In certain non-limiting embodiments, supervised feature selection can be performed using a top-down wrapper or a bottom-up wrapper method. In other embodiments, supervised feature selection can be performed using both a top-down wrapper and a bottom-up wrapper method. The predictive model, for example, can be an RNN with a 3-7 architecture and trained using 3-fold cross-validation and 10 training epochs. For example, several top-down and bottom-down wrapper feature selection experiments can be performed by varying the bootstrap samples and randomness within the RNN cross-validation. As shown in FIG. 12, the results can be combined to create an average wrapper-based feature selection. In particular, the average wrapper-based features are ordered by their mean position (POS), sorted from the most selected feature to the least selected feature. In some non-limiting embodiments, 30 different experiments with varying randomness can be included. The arrows in FIG. 12 indicate the quantiles for the ordering of the selected features from each different wrapper experiment. As shown in FIG. 12, the top seven functions may be BUN, urine specific gravity, visit age, creatinine, urine protein, weight, and / or amylase.

[0156] FIG. 13 shows the average best F1 score 1302 according to certain embodiments described herein. In certain non-limiting embodiments, the average best F1 score of several bootstrap sampling experiments using top-down and bottom-up wrapper methods can be graphed. As shown in FIG. 13, the F1 score can begin to decline when fewer than seven features are used. Thus, the predictive model can utilize seven features, including a combination of one or more biomarkers and / or demographic information. The seven features can be, for example, BUN, urine specific gravity, visit age, creatinine, urine protein, weight, and / or amylase.

[0157]

[00130] Figure 14 shows a chart 1402 illustrating feature selection according to certain embodiments described herein. In particular, Figure 14 illustrates feature selection in a top-down wrapper method with an RNN predictor. As shown in Figure 14, the top seven selected features include BUN, urine specific gravity, visit age, creatinine, urine protein, weight, and / or amylase.

[0158] Figure 15 shows a graph 1502 illustrating Bayesian Information Criterion (BIC) during wrapper feature selection, according to certain embodiments described herein. In particular, Figure 15 shows the BIC for the dataset feature selection shown in Figure 14. As shown in Figure 15, the inflection point corresponds to the point at which seven features were selected. For example, the seven features may include BUN, urine specific gravity, visit age, creatinine, urine protein, weight, and / or amylase.

[0159] FIG. 16 shows a graph 1602 illustrating performance metrics according to certain implementations described herein. Specifically, FIG. 16 illustrates top-down feature selection for wrappers, including a 10-fold cross-validation performance metric. This metric allowed for optimizing the F1 measure through a grid search at various thresholds in one or more steps. Performance metrics such as AUCPR, AUCROC, sensitivity, or NPV can be calculated during the wrapper feature selection process. FIG. 16 illustrates that the optimal number of selected features may be seven.

[0160] FIG. 17 shows a graph 1702 illustrating performance metrics according to certain implementations described herein. Specifically, FIG. 17 illustrates top-down feature selection for wrappers, including a 10-fold cross-validation performance metric. This metric allowed for optimizing the F1 measure through a grid search at various thresholds in one or more steps. Performance metrics such as AUCPR, AUCROC, sensitivity, or NPV can be calculated during the wrapper feature selection process. Similar to FIG. 16, FIG. 17 illustrates that the optimal number of selected features may be seven.

[0161] In certain non-limiting embodiments, the RNN architecture can be optimized. For example, for an RNN-LSTM, various configurations of 1 to 5 hidden layers and 3 to 200 nodes per layer can be tested. The TanH activation function can be used in the hidden layers, and softmax is used in the output layer. Given a binary classification of no CKD or CKD, the softmax can be sigmoidized. Binary cross-entropy can also be used for loss calculation, and / or a 20% dropout can be considered to avoid overfitting. Backpropagation through time can be used for training using the RMSprop gradient descent optimization algorithm. Furthermore, in some other non-limiting embodiments, LSTM cell structures can be tested to address gradient vanishing.

[0162] FIG. 18 shows graphs illustrating RNN and LSTM architectures according to certain embodiments described herein. As shown in FIG. 10, the F1 measure or score varies as a function of the total number of nodes. The best-performing configurations after 10-fold cross-validation were, for example, a three-layer vanilla RNN (5-3-3) with a 3-9 architecture and / or a two-layer RNN-LSTM. A Pareto front can then be used to find the optimum for both the F1 score and the number of nodes or neurons in the RNN. As shown in graph 1802, the best-performing RNN may include a 5-3-3 architecture, while in graph 1804, the best-performing LSTM may include a 3-9 architecture. As shown in graph 1806, the RNN or vanilla RNN with a 5-3-3 architecture had an F1 score of 0.82, an AUCPR of 0.91, and an AUCROC of 0.94. On the other hand, the LSTM with 3-9 architecture had an F1 score of 0.819, an AUCPR of 0.907, and an AUCROC of 0.938, as shown in graph 1808.

[0163] Other vanilla RNNs were tested, including, but not limited to, 5-3-3, 9-3-0, 5-5-0, 3-7-0, 5-5-10, 7-3-0, 8-4-0, 3-9-0, 7-5-2, 20-0-0, 30-0-0, 2-6-3, 5-9-0, 4-2-4, 5-5-5-5, 3-3-3-3, 3-3-3, 7-9-4-8, 5-4-6-3, 4-8-4, and 9-3-6-9. As mentioned above, RNNs with 1 to 5 layers were tested, with each layer having between about 1 and about 250 nodes. While an RNN with a 5-3-3 architecture was selected, in certain other implementations, any other RNN may be selected.

[0164] Additionally, other LSTMs were tested, including, but not limited to, 3-9-0, 5-5-0, 3-3-3, 7-3-0, 2-4-4, 2-6-3, 4-8-4, 3-3-5, 7-13-0, 3-3-3-3, 6-4-6, 3-7-0, 9-3-0, 10-10-0, 5-9-0, 8-4-0, 7-3-7-3, 8-4-8, 20-20-0, 3-9-6-5, and 5-3-3-5-6. As noted above, LSTMs with 1 to 5 layers were tested, with each layer having between about 1 and about 250 nodes. While an LSTM with a 3-9-0 architecture was selected, any other LSTM may be selected in certain other implementations.

[0165] 19 shows cross-validation performance 1902 according to certain implementations described herein. In particular, an RNN with a 5-3-3 architecture can be used. The RNN can include 10 folds and 18 epochs. The RNN can yield a sensitivity of 79.2%, a specificity of 94%, an accuracy of 89.2%, and an F1 score of 82.6%.

[0166] In one specific, non-limiting implementation, temperature scaling can be used. Due to the nature of RNN output using softmax or sigmoid functions, probabilities occupy neighborhoods close to the boundary between 0 and 1 and can be rescaled. Temperature scaling can be used, for example, as a single parameter random variable for Platt scaling. The temperature scaling parameter can be determined by minimizing the negative log-likelihood (e.g., cross-entropy loss) using the following formula: sigmoid(logit(p i ) / T), where p i is equal to the initial neural network prediction for each iPet. In some non-limiting embodiments, the selected temperature parameter T was found to be equal to 1.297 using 10-fold cross-validation of the model.

[0167] In one specific, non-limiting embodiment, the prevalence of CKD in the dataset was determined to be 33%. Given that current diagnostics may miss 50% of "at-risk" pets, the decision threshold can be re-optimized to a 5% prevalence (which may be slightly higher than most dog breeds, typically older dogs). In some non-limiting embodiments, the predictive model is not retrained using a dataset with a 5% CKD class fraction. Rather, the predictive model's decision threshold can be selected for a 5% prevalence.

[0168] The calculation of the decision threshold can be performed by 100 iterations of undersampling the calibrated predicted probability of CKD class, e.g., 5% obtained from a random sample of 1,400 diseased dogs or canines. The average performance metric during the grid search decision threshold ranges from 0.05 to 0.95 in steps of 0.05. The selected decision threshold can depend on the metric selected for evaluating the threshold. For example, the metric can be F1 score, AUCPR, or geometric mean. In one specific, non-limiting embodiment, a decision threshold of 0.9 corresponds to the maximum F1 score, with an average F1 score of 0.6825. However, in other embodiments, a decision threshold of 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, or 0.08 can be used.

[0169] Example 3 In a specific, non-limiting embodiment, the prediction model may include an RNN. The RNN includes a 5-3-3 architecture, 10 splits, and 18 epochs. The RNN evaluates six selected features, such as one or more biomarkers and demographic information. For example, the six selected features may be BUN, urine sg, visit age, creatinine, urine protein, and / or weight. Compared to Example 2, Example 3 does not use amylase in the prediction model. Amylase was removed because cholesterol has little or no effect on the prediction model. Feature selection performed using top-down and bottom-up wrappers supported the removal of amylase from the prediction model of Example 3. By removing amylase, the prediction model of Example 3 relies on six features instead of seven. By doing so, the performance of the prediction model may decrease by 0.2% in F1 score. However, overall, the predictive model of Example 3 achieved a sensitivity of 79.1%, a specificity of 93.8%, an AUCROC of 94.2%, an AUCPR of 91.6%, and an F1 score of 82.4%.

[0170] Amylase is a calcium-dependent enzyme that hydrolyzes complex carbohydrates via alpha-1,4-linkages to form maltose and glucose. Amylase is filtered by the renal tubules and can be reabsorbed (inactivated) by the renal tubular epithelium. Active enzyme does not appear in the urine. Small amounts of amylase can be taken up by Kupffer cells in the liver. For example, in healthy dogs, 14% of amylase can bind to globulin. Because of this polymerization, canine amylase can have a variable (high) molecular weight and is usually unable to be filtered by the kidney. In dogs with renal disease, this polymerized (macroamylase) amylase is found in high concentrations (5-62% of total amylase activity) and may contribute to the hyperamylasemia seen in CKD.

[0171] In some non-limiting embodiments, the performance of the predictive model of Example 3 was boosted to achieve a sensitivity of 84.1%, a specificity of 94.2%, an AUCROC of 95.6%, an AUCPR of 93.8%, and an F1 score of 85.8%. The performance improvement was achieved by modifying the training of the predictive model. In particular, after subsampling, one or more bootstrap samples with different seeds can be formed, and a training subset can be selected using an RNN cross-validated predictor. The training subset that yields the best F1 score can be selected. For example, an RNN with a 5-3-3 configuration or architecture with 10 folds and 8 epochs can be selected.

[0172] In certain non-limiting embodiments, a decision threshold of 0.5 can be used to optimize the true positive rate (e.g., sensitivity or recall) over the false positive rate (e.g., specificity). In some non-limiting embodiments, the selected temperature parameter T was found to be equal to 1.296 by minimizing the negative log-likelihood using quasi-Newtonian Broyden-Fletcher-Goldfarb-Shanno (BFGS) optimization.

[0173] 20 illustrates a decision threshold table 2002 according to certain implementations described herein. For example, the table illustrates decision thresholds ranging from 0.1 to 0.9 in steps of 0.1. The G-mean shown in FIG. 13 may represent the geometric mean of sensitivity and specificity. The decision threshold may be selected based on one or more of the F1 score, precision, accuracy, sensitivity, specificity, and G-mean.

[0174] Sensitivity can be the percentage of true positives for disease. A highly sensitive test may be useful for ruling out disease with a negative test, but does not necessarily determine it. Specificity, on the other hand, is the percentage of true negatives without disease, and (if specificity is high) may be useful for determining a positive test, but does not rule it out. In the setting of a highly sensitive and specific test, sensitivity is easily understood (if the test is not positive, the disease may not be present), but specificity is confusing because it focuses on not having disease, rather than focusing on having it. A highly specific test may have a low false positive rate, and a highly sensitive test may have a low false negative rate. Sensitivity and specificity may be on a continuum, with perfect sensitivity (approaching 100%) leading to a loss of specificity, and vice versa. A receiver operating curve (ROC) may be a statistical and graphical illustration of a process that shows the balance between sensitivity and specificity. A similar continuum can be found when describing the sensitivity and specificity of classification and / or diagnostic criteria.

[0175] Positive predictive value (PPV) can illustrate the above point. PPV is the ratio of true positives to the number of positive tests and can be a measure of the accuracy or performance of a diagnostic test. Negative predictive value (NPV) is the inverse, the ratio of the number of true negatives to the number of negative tests. Both PPV and NPV can be highly dependent on the prevalence of CKD.

[0176] While the subject matter disclosed herein and its advantages have been described in detail, it should be understood that various changes, substitutions, and alterations can be made therein without departing from the spirit and scope of the present disclosure, as defined by the appended claims. Furthermore, the scope of the present application is not intended to be limited to the particular embodiments of the processes, machines, manufacture, compositions of matter, means, methods, and steps described herein. As one skilled in the art will readily recognize from the disclosure of the presently disclosed subject matter, processes, machines, manufacture, compositions of matter, means, methods, or steps, any existing or future-developed means or steps that perform substantially the same function or achieve substantially the same results as the corresponding embodiments described herein can be utilized in accordance with the presently disclosed subject matter. Accordingly, the appended claims are intended to include within their scope such processes, machines, manufacture, compositions of matter, means, methods, or steps.

[0177] Patents, patent applications, publications, product descriptions, and protocols are cited throughout this application, the disclosures of which are incorporated herein by reference in their entireties for all purposes. Preferred embodiments of the present invention will be described below in detail. Embodiment 1 1. A computer system for identifying a susceptibility of a dog to developing chronic kidney disease (CKD), the computer system comprising: processor; and When executed by the processor, the computer system: (a)(i) one or more biomarkers of the dog, the one or more biomarkers comprising information regarding at least one of urine specific gravity, creatinine, urine protein, and blood urea nitrogen (BUN); or (ii) demographic information about the dog, the demographic information including at least one of the dog's age or weight; receiving at least one of (b) processing at least one of the one or more biomarkers or the dog's demographic information using a predictive model comprising a recurrent neural network; and (c) determining a probability risk score for the dog to develop CKD based on the processed one or more biomarkers or demographic information; Memory for storing code 2. A computer system comprising: Embodiment 2 The computer system includes: determining customized recommendations based on the dog's probability risk of developing CKD; 2. A computer system as described in embodiment 1. Embodiment 3 The computer system includes: causing the customized recommendations to be transmitted to a user device of the dog's veterinarian, owner, or caregiver; 3. A computer system according to embodiment 1 or 2. Embodiment 4 The customized recommendation: (a) one or more therapeutic interventions; (b) one or more dietary recommendations; (c) one or more renal-sparing strategies; or (d) one or more tests for disease progression 4. The computer system of claim 2 or 3, comprising at least one of: Embodiment 5 (i) the one or more renal-sparing strategies include avoidance of nonsteroidal anti-inflammatory drugs, aminoglycosides, or any combination thereof; and / or (ii) the one or more tests for disease progression include testing serum parathyroid hormone levels; 5. A computer system as described in embodiment 4. Embodiment 6 6. The computer system of any one of claims 1 to 5, wherein the recurrent neural network comprises a hidden layer architecture having three layers, including a first layer having five nodes, a second layer having three nodes, and a third layer having three nodes. Embodiment 7 7. The computer system of any one of embodiments 1 to 6, wherein the recurrent neural network undergoes a 10-fold cross-validation process and is trained for 8 or 18 epochs. Embodiment 8 8. The computer system of any one of embodiments 1 to 7, wherein the one or more biomarkers include information about amylase. Embodiment 9 9. The computer system of any one of embodiments 1 to 8, wherein the recurrent neural network is trained using a training dataset comprising the one or more biomarkers and the demographic information of a plurality of other dogs. Embodiment 10 10. The computer system of any one of embodiments 1 to 9, wherein the predictive model further comprises the recurrent neural network having a long short-term memory (LSTM). Embodiment 11 11. The computer system of any one of embodiments 1 to 10, wherein the decision threshold for the onset of CKD using the recurrent neural network is from about 0 to about 1. Embodiment 12 11. The computer system of any one of embodiments 1 to 10, wherein the decision threshold for the onset of CKD using the recurrent neural network is about 0.5. Embodiment 13 The computer system includes: imputing one or more missing values ​​from one or more biomarkers of said dog or one or more demographic information of said dog; 13. A computer system as described in any one of embodiments 1 to 12. Embodiment 14 14. The computer system of embodiment 13, wherein the imputation is linear regression. Embodiment 15 14. The computer system of embodiment 13, wherein the assignment is based on the age of the dog. Embodiment 16 14. The computer system of embodiment 13, wherein the imputation is based on the number of missing values. Embodiment 17 1. A method for identifying susceptibility to developing chronic kidney disease (CKD) in a dog, comprising: (a)(i) one or more biomarkers of the dog, the one or more biomarkers comprising information regarding at least one of urine specific gravity, creatinine, urine protein, and blood urea nitrogen (BUN); or (ii) demographic information about the dog, the demographic information including at least one of the dog's age or weight; receiving at least one of: (b) processing at least one of the one or more biomarkers or the dog's demographic information using a predictive model, wherein the predictive model comprises a recurrent neural network; and (c) determining a probability risk score for the dog to develop CKD based on the processed one or more biomarkers or demographic information. A method comprising: Embodiment 18 determining a customized recommendation based on the dog's probability risk of developing CKD. 18. The method of embodiment 17, further comprising: Embodiment 19 sending the customized recommendations to a user device of the dog's veterinarian, owner, or caregiver. 19. The method of embodiment 17 or 18, further comprising: Embodiment 20 The customized recommendation: (a) one or more therapeutic interventions; (b) one or more dietary recommendations; (c) one or more renal-sparing strategies; or (d) one or more tests for disease progression 20. The method of embodiment 18 or 19, comprising at least one of: Embodiment 21 (i) the one or more renal-sparing strategies include avoidance of nonsteroidal anti-inflammatory drugs, aminoglycosides, or any combination thereof; and / or (ii) the one or more tests for disease progression include testing serum parathyroid hormone levels; 21. The method of embodiment 20. Embodiment 22 22. The method of any one of embodiments 17 to 21, wherein the recurrent neural network comprises a hidden layer architecture having three layers, including a first layer having 21 nodes, a second layer having 3 nodes, and a third layer having 3 nodes. Embodiment 23 23. The method of any one of embodiments 17 to 22, wherein the recurrent neural network undergoes a 10-fold cross-validation process and is trained for 8 or 18 epochs. Embodiment 24 24. The method of any one of embodiments 17 to 23, wherein the one or more biomarkers comprise information about amylase. Embodiment 25 The method of any one of embodiments 17 to 24, wherein the recurrent neural network is trained using a training dataset comprising the one or more biomarkers and the demographic information of a plurality of other dogs. Embodiment 26 26. The method of any one of embodiments 17 to 25, wherein the predictive model further comprises the recurrent neural network equipped with long short-term memory (LSTM). Embodiment 27 27. The method of any one of embodiments 17 to 26, wherein the decision threshold for the onset of CKD using the recurrent neural network is from about 0.0 to about 1.0. Embodiment 28 28. The method of any one of embodiments 17 to 27, wherein the decision threshold for the onset of CKD using the recurrent neural network is about 0.5. Embodiment 29 imputing one or more missing values ​​from one or more biomarkers of said dog or one or more demographic information of said dog. 29. The method of any one of embodiments 17 to 28, further comprising: Embodiment 30 30. The method of any one of embodiments 17 to 29, wherein said imputation is a linear regression. Embodiment 31 30. The method of any one of embodiments 17 to 29, wherein the substitution is based on the age of the dog. Embodiment 32 30. The method of any one of embodiments 17 to 29, wherein said imputation is based on the number of missing values. Embodiment 33 1. A computer system for identifying a susceptibility of a dog to developing chronic kidney disease (CKD), the computer system comprising: processor; and When executed by the processor, the computer system: (a)(i) one or more biomarkers of the dog, the one or more biomarkers comprising information regarding at least one of urine specific gravity, creatinine, urine protein, and blood urea nitrogen (BUN); or (ii) demographic information about the dog, the demographic information including at least one of the dog's age or weight; receiving at least one of (b) processing at least one of the one or more biomarkers or the dog's demographic information using a predictive model, the predictive model comprising a recurrent neural network comprising a hidden layer architecture having three layers, the hidden layer architecture including a first layer having five nodes, a second layer having three nodes, and a third layer having three nodes; and (c) determining a probability risk score for the dog to develop CKD based on the processed one or more biomarkers or demographic information; Memory for storing code 2. A computer system comprising: Embodiment 34 The computer system includes: determining customized recommendations based on the dog's probability risk of developing CKD; 34. A computer system as described in embodiment 33. Embodiment 35 The computer system includes: causing the customized recommendations to be transmitted to a user device of the dog's veterinarian, owner, or caregiver; 35. A computer system as described in embodiment 33 or 34. Embodiment 36 The customized recommendation: (a) one or more therapeutic interventions; (b) one or more dietary recommendations; (c) one or more renal-sparing strategies; or (d) one or more tests for disease progression 36. The computer system of embodiment 34 or 35, comprising at least one of: Embodiment 37 (i) the one or more renal-sparing strategies include avoidance of nonsteroidal anti-inflammatory drugs, aminoglycosides, or any combination thereof; and / or (ii) the one or more tests for disease progression include testing serum parathyroid hormone levels; 37. A computer system as described in embodiment 36. Embodiment 38 38. The computer system of any one of embodiments 33 to 37, wherein the one or more biomarkers include information about amylase. Embodiment 39 The computer system of any one of embodiments 33 to 38, wherein the recurrent neural network is trained using a training dataset comprising the one or more biomarkers and the demographic information of a plurality of other dogs. Embodiment 40 40. The computer system of any one of embodiments 33 to 39, wherein the predictive model further comprises the recurrent neural network having a long short-term memory (LSTM). Embodiment 41 41. The computer system of any one of embodiments 33 to 40, wherein the decision threshold for the onset of CKD using the recurrent neural network is from about 0.0 to about 1.0. Embodiment 42 42. The computer system of any one of embodiments 33 to 41, wherein the decision threshold for the onset of CKD using the recurrent neural network is about 0.5. Embodiment 43 43. The computer system of any one of embodiments 33 to 42, wherein the recurrent neural network undergoes a 10-fold cross-validation process and is trained for 8 or 18 epochs. EMBODIMENT 44 The computer system includes: imputing one or more missing values ​​from one or more biomarkers of said dog or one or more demographic information of said dog; 44. A computer system as described in any one of embodiments 33 to 43. Embodiment 45 45. The computer system of embodiment 44, wherein the imputation is linear regression. Embodiment 46 45. The computer system of embodiment 44, wherein the assignment is based on the age of the dog. Embodiment 47 45. The computer system of embodiment 44, wherein the imputation is based on the number of missing values. Embodiment 48 1. A method for identifying susceptibility to developing chronic kidney disease (CKD) in a dog, comprising: (a)(i) one or more biomarkers of the dog, the one or more biomarkers comprising information regarding at least one of urine specific gravity, creatinine, urine protein, and blood urea nitrogen (BUN); or (ii) demographic information about the dog, the demographic information including at least one of the dog's age or weight; receiving at least one of: (b) processing at least one of the one or more biomarkers or the dog's demographic information using a predictive model, the predictive model comprising a recurrent neural network, the recurrent neural network comprising a hidden layer architecture having three layers, including a first layer having five nodes, a second layer having three nodes, and a third layer having three nodes; and (c) determining a probability risk score for the dog to develop CKD based on the processed one or more biomarkers or demographic information. A method comprising: Embodiment 49 determining a customized recommendation based on the dog's probability risk of developing CKD. 49. The method of embodiment 48, further comprising: Embodiment 50 sending the customized recommendations to a user device of the dog's veterinarian, owner, or caregiver. 50. The method of embodiment 48 or 49, further comprising: Embodiment 51 The customized recommendation: (a) one or more therapeutic interventions; (b) one or more dietary recommendations; (c) one or more renal-sparing strategies; or (d) one or more tests for disease progression 51. The method of embodiment 49 or 50, comprising at least one of: Embodiment 52 (i) the renal-sparing strategy comprises avoidance of nonsteroidal anti-inflammatory drugs, aminoglycosides, or any combination thereof; and / or (ii) the testing for disease progression includes testing serum parathyroid hormone levels; 52. The method of embodiment 51. Embodiment 53 53. The method of any one of embodiments 48 to 52, wherein the one or more biomarkers comprise information about amylase. EMBODIMENT 54 The method of any one of embodiments 48 to 53, wherein the recurrent neural network is trained using a training dataset comprising the one or more biomarkers and the demographic information of a plurality of other dogs. Embodiment 55 55. The method of any one of embodiments 48 to 54, wherein the predictive model further comprises the recurrent neural network equipped with long short-term memory (LSTM). Embodiment 56 The method of any one of embodiments 48 to 55, wherein the decision threshold for the onset of CKD using the recurrent neural network is from about 0.0 to about 1.0. Embodiment 57 The method of any one of embodiments 48 to 56, wherein the decision threshold for the onset of CKD using the recurrent neural network is about 0.5. Embodiment 58 58. The method of any one of embodiments 48 to 57, wherein the recurrent neural network undergoes a 10-fold cross-validation process and is trained for 8 or 18 epochs. Embodiment 59 imputing one or more missing values ​​from one or more biomarkers of said dog or one or more demographic information of said dog. 59. The method of any one of embodiments 48 to 58, further comprising: Embodiment 60 60. The method of embodiment 59, wherein said imputation is linear regression. Embodiment 61 60. The method of embodiment 59, wherein said substitution is based on the age of the dog. Embodiment 62 60. The method of embodiment 59, wherein said imputation is based on the number of missing values.

Claims

1. 1. A computer system for identifying susceptibility to the development of chronic kidney disease (CKD) in dogs, the computer system comprising: a processor; and When executed by the processor, the computer system: (a)(i) one or more biomarkers of the dog, the one or more biomarkers comprising information regarding urine specific gravity, creatinine, urine protein, and blood urea nitrogen (BUN); and (ii) receiving demographic information about the dog, the demographic information including at least the dog's age, weight, and size; (b) processing the one or more biomarkers and the dog's demographic information using a predictive model comprising a recurrent neural network; and (c) determining a probability risk score for the dog to develop CKD based on the processed one or more biomarkers and demographic information; Memory for storing code Including, A computer system wherein the one or more biomarkers or the dog's demographic information are selected by a top-down or bottom-up wrapper method using the recurrent neural network.

2. The computer system includes: determining customized recommendations based on the dog's probability risk of developing CKD; 10. The computer system of claim 1.

3. The computer system includes: causing the customized recommendations to be transmitted to a user device of the dog's veterinarian, owner, or caregiver; 3. The computer system of claim 2.

4. The customized recommendation: (a) one or more therapeutic interventions; (b) one or more dietary recommendations; (c) one or more renal-sparing strategies; or (d) one or more tests for disease progression 4. The computer system of claim 2, further comprising at least one of:

5. (i) the one or more renal-sparing strategies include avoidance of nonsteroidal anti-inflammatory drugs, aminoglycosides, or any combination thereof; and / or (ii) the one or more tests for disease progression include testing serum parathyroid hormone levels; 5. The computer system of claim 4.

6. 6. The computer system of claim 1, wherein the recurrent neural network comprises a hidden layer architecture having three layers, including a first layer having five nodes, a second layer having three nodes, and a third layer having three nodes.

7. 7. The computer system of claim 1, wherein the recurrent neural network undergoes a 10-fold cross-validation process and is trained for 8 or 18 epochs.

8. The computer system of claim 1 , wherein the one or more biomarkers include information about amylase.

9. 9. The computer system of claim 1, wherein the recurrent neural network is trained using a training dataset comprising the one or more biomarkers and the demographic information of a plurality of other dogs.

10. 10. The computer system of claim 1, wherein the predictive model further comprises the recurrent neural network having a long short-term memory (LSTM).

11. 11. The computer system of claim 1, wherein the decision threshold for the onset of CKD using the recurrent neural network is from about 0 to about 1.

12. 11. The computer system of claim 1, wherein the decision threshold for the onset of CKD using the recurrent neural network is about 0.

5.

13. The computer system includes: imputing one or more missing values ​​from one or more biomarkers of said dog or one or more demographic information of said dog; 13. A computer system according to any one of claims 1 to 12.

14. 14. The computer system of claim 13, wherein the imputation is a linear regression.

15. 14. The computer system of claim 13, wherein the assignment is based on the age of the dog.

16. The computer system of claim 13 , wherein the imputation is calculated based on the number of missing values.

17. 1. A method for identifying susceptibility to developing chronic kidney disease (CKD) in a dog, comprising: (a)(i) one or more biomarkers of the dog, the one or more biomarkers comprising information regarding urine specific gravity, creatinine, urine protein, and blood urea nitrogen (BUN); and (ii) receiving demographic information about the dog, the demographic information including at least the dog's age, weight, and size; (b) processing the one or more biomarkers and the dog's demographic information using a predictive model, wherein the predictive model comprises a recurrent neural network; and (c) determining a probability risk score for the dog to develop CKD based on the processed one or more biomarkers and demographic information. Including, The method, wherein the one or more biomarkers or the dog's demographic information are selected by a top-down or bottom-up wrapper method using the recurrent neural network.

18. determining a customized recommendation based on the dog's probability risk of developing CKD.

20. The method of claim 17, further comprising:

19. sending the customized recommendations to a user device of the dog's veterinarian, owner, or caregiver.

20. The method of claim 18, further comprising:

20. The customized recommendation: (a) one or more therapeutic interventions; (b) one or more dietary recommendations; (c) one or more renal-sparing strategies; or (d) one or more tests for disease progression 20. The method of claim 18 or 19, comprising at least one of:

21. (i) the one or more renal-sparing strategies include avoidance of nonsteroidal anti-inflammatory drugs, aminoglycosides, or any combination thereof; and / or (ii) the one or more tests for disease progression include testing serum parathyroid hormone levels; 21. The method of claim 20.

22. 22. The method of any one of claims 17 to 21, wherein the recurrent neural network comprises a hidden layer architecture having three layers, including a first layer having 21 nodes, a second layer having 3 nodes, and a third layer having 3 nodes.

23. 23. The method of any one of claims 17 to 22, wherein the recurrent neural network undergoes a 10-fold cross-validation process and is trained for 8 or 18 epochs.

24. 24. The method of any one of claims 17 to 23, wherein the one or more biomarkers comprise information about amylase.

25. 25. The method of any one of claims 17 to 24, wherein the recurrent neural network is trained using a training dataset comprising the one or more biomarkers and the demographic information of a plurality of other dogs.

26. 26. The method of any one of claims 17 to 25, wherein the predictive model further comprises the recurrent neural network with long short-term memory (LSTM).

27. 27. The method of any one of claims 17 to 26, wherein the decision threshold for the onset of CKD using the recurrent neural network is from about 0.0 to about 1.

0.

28. 28. The method of any one of claims 17 to 27, wherein the decision threshold for the onset of CKD using the recurrent neural network is about 0.

5.

29. imputing one or more missing values ​​from one or more biomarkers of said dog or demographic information of said dog.

29. The method of any one of claims 17 to 28, further comprising:

30. 30. The method of claim 29, wherein the imputation is a linear regression.

31. 30. The method of claim 29, wherein the assignment is based on the age of the dog.

32. 30. The method of claim 29, wherein the imputation is calculated based on the number of missing values.

33. 1. A computer system for identifying susceptibility to the development of chronic kidney disease (CKD) in dogs, the computer system comprising: a processor; and When executed by the processor, the computer system: (a)(i) one or more biomarkers of the dog, the one or more biomarkers comprising information regarding urine specific gravity, creatinine, urine protein, and blood urea nitrogen (BUN); and (ii) receiving demographic information for the dog, the demographic information including at least the age, weight, and size of the dog; (b) processing the one or more biomarkers and the dog's demographic information using a predictive model, the predictive model comprising a recurrent neural network comprising a hidden layer architecture having three layers, the hidden layer architecture including a first layer having five nodes, a second layer having three nodes, and a third layer having three nodes; and (c) determining a probability risk score for the dog to develop CKD based on the processed one or more biomarkers and demographic information; Memory for storing code Including, A computer system wherein the one or more biomarkers or the dog's demographic information are selected by a top-down or bottom-up wrapper method using the recurrent neural network.

34. The computer system includes: determining customized recommendations based on the dog's probability risk of developing CKD; 34. The computer system of claim 33.

35. The computer system includes: causing the customized recommendations to be transmitted to a user device of the dog's veterinarian, owner, or caregiver; 35. The computer system of claim 34.

36. The customized recommendation: (a) one or more therapeutic interventions; (b) one or more dietary recommendations; (c) one or more renal-sparing strategies; or (d) one or more tests for disease progression 36. A computer system according to claim 34 or 35, comprising at least one of:

37. (i) the one or more renal-sparing strategies include avoidance of nonsteroidal anti-inflammatory drugs, aminoglycosides, or any combination thereof; and / or (ii) the one or more tests for disease progression include testing serum parathyroid hormone levels; 37. The computer system of claim 36.

38. 38. The computer system of any one of claims 33 to 37, wherein the one or more biomarkers include information regarding amylase.

39. 39. The computer system of any one of claims 33 to 38, wherein the recurrent neural network is trained using a training dataset comprising the one or more biomarkers and the demographic information of a plurality of other dogs.

40. 40. The computer system of any one of claims 33 to 39, wherein the predictive model further comprises the recurrent neural network having long short-term memory (LSTM).

41. 41. The computer system of any one of claims 33 to 40, wherein the decision threshold for the onset of CKD using the recurrent neural network is from about 0.0 to about 1.

0.

42. 42. The computer system of any one of claims 33 to 41, wherein the decision threshold for the onset of CKD using the recurrent neural network is about 0.

5.

43. 43. The computer system of any one of claims 33 to 42, wherein the recurrent neural network undergoes a 10-fold cross-validation process and is trained for 8 or 18 epochs.

44. The computer system includes: imputing one or more missing values ​​from one or more biomarkers of said dog or one or more demographic information of said dog; 44. A computer system according to any one of claims 33 to 43.

45. 45. The computer system of claim 44, wherein the imputation is a linear regression.

46. 45. The computer system of claim 44, wherein the assignment is based on the age of the dog.

47. 45. The computer system of claim 44, wherein the imputation is calculated based on the number of missing values.

48. 1. A method for identifying susceptibility to developing chronic kidney disease (CKD) in a dog, comprising: (a)(i) one or more biomarkers of the dog, the one or more biomarkers comprising information regarding urine specific gravity, creatinine, urine protein, and blood urea nitrogen (BUN); and (ii) receiving demographic information about the dog, the demographic information including at least the dog's age, weight, and size; (b) processing the one or more biomarkers and the dog's demographic information using a predictive model, the predictive model comprising a recurrent neural network, the recurrent neural network comprising a hidden layer architecture having three layers, including a first layer having five nodes, a second layer having three nodes, and a third layer having three nodes; and (c) determining a probability risk score for the dog to develop CKD based on the processed one or more biomarkers and demographic information. Including, The method, wherein the one or more biomarkers or the dog's demographic information are selected by a top-down or bottom-up wrapper method using the recurrent neural network.

49. determining a customized recommendation based on the dog's probability risk of developing CKD.

49. The method of claim 48, further comprising:

50. sending the customized recommendations to a user device of the dog's veterinarian, owner, or caregiver.

50. The method of claim 49, further comprising:

51. The customized recommendation: (a) one or more therapeutic interventions; (b) one or more dietary recommendations; (c) one or more renal-sparing strategies; or (d) one or more tests for disease progression 51. The method of claim 49 or 50, comprising at least one of:

52. (i) the renal-sparing strategy comprises avoidance of nonsteroidal anti-inflammatory drugs, aminoglycosides, or any combination thereof; and / or (ii) the testing for disease progression includes testing serum parathyroid hormone levels; 52. The method of claim 51.

53. 53. The method of any one of claims 48 to 52, wherein the one or more biomarkers comprise information regarding amylase.

54. 54. The method of any one of claims 48 to 53, wherein the recurrent neural network is trained using a training dataset comprising the one or more biomarkers and the demographic information of a plurality of other dogs.

55. 55. The method of any one of claims 48 to 54, wherein the predictive model further comprises the recurrent neural network with long short-term memory (LSTM).

56. 56. The method of any one of claims 48 to 55, wherein the decision threshold for the onset of CKD using the recurrent neural network is from about 0.0 to about 1.

0.

57. 57. The method of any one of claims 48 to 56, wherein the decision threshold for the onset of CKD using the recurrent neural network is about 0.

5.

58. 58. The method of any one of claims 48 to 57, wherein the recurrent neural network undergoes a 10-fold cross-validation process and is trained for 8 or 18 epochs.

59. imputing one or more missing values ​​from one or more biomarkers of said dog or demographic information of said dog.

59. The method of any one of claims 48 to 58, further comprising:

60. 60. The method of claim 59, wherein the imputation is a linear regression.

61. 60. The method of claim 59, wherein the assignment is based on the age of the dog.

62. 60. The method of claim 59, wherein the imputation is calculated based on the number of missing values.

Images