Methods of assessing alzheimer's disease

Abstract

The present invention provides a method for assessing Alzheimer' s disease in a subject.

Description

[0001] Atty Dkt No.: RICE-231WO METHODS OF ASSESSING ALZHEIMER’S DISEASE CROSS-REFERNCE TO RELATED APPLICATIONS This application claims the benefit of priority to U.S. Provisional Application No. 63 / 463,808, filed May 3, 2023, the disclosure of which is incorporated herein by reference in its entirety. FIELD The present disclosure generally relates to lipidomics and methods of assessing Alzheimer’s disease based on a plurality of lipid biomarkers. BACKGROUND Bibliographic details of references referred to by author name in the subject specification are listed at the end of the specification. Reference to any prior art in this specification is not, and should not be taken as, acknowledgement or any form of suggestion that this prior art forms part of the common general knowledge in any country. Despite lipid profiling still lagging behind the advances made in the study and application of other types of biomarkers such as nucleic acid and protein biomarkers, lipid biomarkers are increasingly studied for their potential to act as important biomarkers. In particular, levels of certain plasma lipid species have been associated with metabolic health, and with cardiovascular, neurodegenerative, and inflammatory diseases. Alzheimer’s disease (AD) is one such progressive neurodegenerative disease for which there is no cure. AD is characterized by the extracellular deposition of amyloid-beta (Aβ) as amyloid plaques and the intracellular deposition of tau as neurofibrillary tangles, leading to progressive dementia in AD patients. Early identification of AD is pivotal because it allows for the preservation of a patient’s level of function and potential prevention or delay of disease progression. There is therefore a need for improved methods for assessing AD in a subject based on lipidomics. SUMMARY In one aspect of the invention, there is provided a method of determining an Alzheimer’s disease (AD) risk score of a subject, the method comprising: Atty Dkt No.: RICE-231WO (i) detecting in a biological sample from the subject a level of at least two lipid species selected from Table 4 and / or 6; and (ii) comparing the levels of the lipid species detected in the biological sample to reference lipid species levels to generate a risk score, wherein the risk score is determinative of the risk of developing AD in the subject. In an embodiment, the methods described herein further comprise obtaining a biological sample from a subject. In an embodiment, the methods described herein can be used to determine AD risk score at a time point between about 6 months and about 10 years before the clinical onset of one or more signs or symptoms of AD as described herein or known in the art. In another embodiment, the methods described herein can be used to determine AD risk score at a time point between about 6 months and about 7 years, between about 1 year and about 5 years, or between about 2 years and about 4 years before the clinical onset of one or more signs or symptoms of AD as described herein or known in the art. In another embodiment, the methods described herein can be used to determine AD risk score at a time point of about 6 months, about 1 year, about 2 years, about 3 years, about 4 years, about 5 years, 6 years, about 7 years, about 8 years, about 9 years or about 10 years before the clinical onset of one or more signs or symptoms of AD as described herein or known in the art. In another embodiment, the methods described herein can be used to determine AD risk score at a time point between about 6 months and about 10 years after the biological sample is obtained from a subject. In another embodiment, the methods described herein can be used to determine AD risk score at a time point between about 6 months and about 7 years, between about 1 year and about 5 years, or between about 2 years and about 4 years after the biological sample is obtained from a subject. In another embodiment, the methods described herein can be used to determine AD risk score at a time point of about 6 months, about 1 year, about 2 years, about 3 years, about 4 years, about 5 years, about 6 years, about 7 years, about 8 years, about 9 years or about 10 years after the biological sample is obtained from a subject. In another embodiment, the methods provided herein further comprise: (i) standardising each continuous variable; (ii) log transforming each lipid concentration; and (iii) further standardising the variables prior to their use as the predictors of AD. Atty Dkt No.: RICE-231WO In an embodiment, the method further comprises a step of confirming that the subject has, or is likely to develop Alzheimer’s disease based on the comparison of the lipid species in the biological sample to reference lipid species levels. In another aspect, there is provided a method of determining the presence of, or risk of developing, Alzheimer’s disease (AD) in a subject, the method comprising: (i) detecting in a biological sample from the subject a level of at least two lipid species selected from Table 4 and / or 6; and (ii) comparing the levels of the lipid species detected in the biological sample to reference lipid species levels, wherein the comparison is determinative of whether the subject has, or is likely to develop, AD. In an embodiment, a method provided herein further comprises a step of confirming that the subject has, or is likely to develop AD based on the comparison of the lipid species in the biological sample to reference lipid species levels. In another aspect, provided herein is panel or kit for use in a method described herein, wherein the kit comprises one or both of: (i) a set of stable labelled isotopes or non-physiological lipid standards for quantification of the lipid species; (ii) reference plasma samples for standardisation of the resulting lipid measures. In another aspect, there is provided a method of treating a subject at risk of developing Alzheimer’s disease (AD) or identified as having AD, wherein the method comprises: (i) detecting in a biological sample from the subject a level of at least two lipid species selected from Table 4 and / or 6; and (ii) comparing the levels of the lipid species detected in the biological sample to reference lipid species levels of the lipid species to generate a risk score; and (iii) treating the subject if the risk score confirms that the subject has or is likely to develop AD. In an embodiment, the treatment comprises at least one of cholinesterase inhibitors, for example, Galantamine, Rivastigmine and Donepezil, and Memantine, Aducanumab, Lecanemab and antidepressants. In an embodiment, the treatment comprises changes to diet. In an embodiment, the method comprises a further diagnostic step, optionally selected from the group consisting of one or more of brain scans, for example, computed tomography, magnetic resonance imaging, and positron emission topography, blood tests, Atty Dkt No.: RICE-231WO memory tests, physical evaluation, behavioural tests, cognitive tests, genetic risk scores, peptide biomarker, neuropsychological tests, and cerebrospinal fluid examination. In an embodiment, the AD risk score is used in conjunction with at least one further diagnostic step. In an embodiment, the AD risk score is used before and / or after at least one further diagnostic step. In an embodiment, the treatment comprises lifestyle changes including changes to diet, exercise regime, quitting or reducing smoking, and undertaking cognitive activities. In an embodiment, the methods provided herein further comprise detecting in a biological sample from the subject a level of at least 10 lipid species, at least 20 lipid species, at least 50 lipid species or at least 100 lipid species selected from Table 4 and / or Table 6. In an embodiment, the methods provided herein further comprise detecting in the biological sample from the subject a level of at least one additional lipid species not defined in Table 4 and / or 6. In an embodiment, the methods provided herein further comprise detecting in the biological sample from the subject a level of at least one additional lipid species defined in Table 2. In another aspect, there is provided herein a method of calculating an Alzheimer’s disease (AD) risk score of a subject comprising: (i) obtaining lipid profile data from a biological sample taken from the subject; (ii) normalising the numeric values of the lipid profile data against a reference sample; (iii) refining the discriminatory power of one or more lipid species by statistically weighting one or several of the numeric values associated therewith according to a predetermined equation; and (iv) normalising the resulting AD risk score of the subject to a reference population. In another aspect, provided herein is a method of calculating an AD risk score of a subject comprising: (i) obtaining lipid profile data from a biological sample taken from the subject; (ii) normalising the numeric values of the lipid profile data against a reference sample; Atty Dkt No.: RICE-231WO (iii) refining the discriminatory power of one or more lipid species by statistically weighting one or several of the numeric values associated therewith according to a predetermined equation; (iv) optionally adding one or more additional, statistically weighted, risk factors into the model, selected from the group consisting of: sex, age, fasting glucose, BMI, HDL-cholesterol, cholesterol, triglycerides, APOE4 status (defined as having one or two APOE4 alleles), statin status (defined as currently taking statin medication) , Omega-3 status (defined as currently taking omega 3 dietary supplements), ethnicity, family history of disease, smoking status, exercise levels, diet, cognitive activity, memory level and place of birth; and (v) normalising the resulting AD risk score of the subject to a reference population. In an embodiment, the methods provided herein further use the AD risk score to predict incident AD in a subject. In an embodiment, the methods provided herein further use the AD risk score to determine presence of AD in a subject. In an embodiment, the methods provided herein further comprise lipid profile data comprising two or more lipid species selected from the group set forth in one or more of Tables 2, 3, 4, 5, and 6. In an embodiment, the methods provided herein further comprise calculating a sum of the weighted lipids. In an embodiment, the reference lipid species levels and / or the reference sample described by the methods provided herein is obtained from a population. In an embodiment, the reference lipid species levels and / or the reference sample described by the methods provided herein is included in a predictive model. In another aspect, there is provided a method of determining an Alzheimer’s disease (AD) risk score of a subject, the method comprising a logistic regression model that uses an equation comprising: (i) input values comprising measured levels of two or more lipid species in a biological sample obtained from a subject, wherein the two or more lipid species include at least two or more lipid species selected from Table 4 and / or Table 6; and (ii) coefficient values that take into consideration risk factors selected from the group consisting of: sex, age, fasting glucose, BMI, HDL-cholesterol, cholesterol, triglycerides, APOE4 status, statin status, Omega-3 status, ethnicity, family history of Atty Dkt No.: RICE-231WO disease, smoking status, exercise levels, diet, cognitive activity, memory level and place of birth; wherein the input values are combined linearly using the coefficient values to predict an output value, and the output value is a risk score for determining risk of developing AD in the subject. In an embodiment, the risk score differentiates between subjects with high risk of developing AD or having AD, and those with low risk of developing AD. In an embodiment, the risk score differentiates between subjects that develop AD and subjects that do not develop AD. In an embodiment, the biological sample in the methods provided herein is selected from the group consisting of blood, plasma, serum, dried blood spots and dried plasma spots. In another aspect, there is provided a method of generating a lipid profile in a sample from a subject, the method comprising: (i) obtaining lipid profile data from a biological sample taken from the subject; and (ii) normalising the numeric values of the lipid profile data against a reference sample, wherein the discriminatory power of one or more lipid species can be refined by statistically weighting one or several of the numeric values associated therewith according to a predetermined equation; wherein the resulting lipid profile of the subject can be normalised to a reference population; and wherein the lipid profile can be used to calculate an Alzheimer’s disease (AD) risk score of a subject. In another aspect, there is provided a method of generating a lipid profile in a sample from a subject, the method comprising: (i) detecting in a biological sample from the subject a level of at least two lipid species selected from Table 4 and / or 6, wherein the lipid species detected in the biological sample can be compared to reference levels of the lipid species to generate a risk score such that the risk score is determinative of the Alzheimer’s disease (AD) risk of the subject. In another aspect, there is provided a method of calculating an Alzheimer’s disease (AD) risk score of a subject, the method comprising: Atty Dkt No.: RICE-231WO (i) refining a discriminatory power of one or more lipid species by statistically weighting one or several of the numeric values associated therewith according to a predetermined equation; and (ii) normalising the resulting AD risk score of the subject to a reference population, wherein lipid profile data can be obtained from a biological sample taken from the subject; and wherein the numeric values of the lipid profile data can be normalised against a reference sample. In an embodiment, the method further comprises use of an algorithm as defined herein, preferably in any one of Tables 9 to 12. In another embodiment, the algorithms can be prepared from any subset of the lipids defined in Table 4 and / or 6. In another aspect, there is provided a method of calculating an Alzheimer’s disease (AD) risk score of a subject, the method comprising: (i) comparing levels of at least two lipid species detected in a biological sample of the subject to reference lipid species levels to generate a risk score, wherein the risk score is determinative of the risk of developing AD in the subject, wherein the levels of at least two lipid species selected from Table 4 and / or 6 can be detected in the biological sample from the subject. In one embodiment, the detection of a plurality or population of lipid species in the subject according to a method described herein is performed using a high throughput lipidomics assay. In one embodiment, the detection of a plurality or population of lipid species in the subject according to a method described herein is performed using a high throughput lipidomics assay wherein the total run time is less than 15 minutes. In one embodiment, the detection of a plurality or population of lipid species in the subject according to a method described herein is performed using a high throughput lipidomics assay wherein the total run time is not more than 5 minutes. In one embodiment, the detection of a plurality or population of lipid species in the subject according to a method described herein is performed using a high throughput lipidomics assay wherein the total run time is less than 5 minutes. In one embodiment, the detection of a plurality or population of lipid species in the subject according to a method described herein is performed using a high throughput Atty Dkt No.: RICE-231WO lipidomics assay, wherein the assay is liquid chromatography with tandem mass spectrometry and the total run time is less than 15 minutes. In one embodiment, the detection of a plurality or population of lipid species in the subject according to a method described herein is performed using a high throughput lipidomics assay, wherein the assay is liquid chromatography with tandem mass spectrometry and the total run time is not more than 5 minutes. In one embodiment, the detection of a plurality or population of lipid species in the subject according to a method described herein is performed using a high throughput lipidomics assay, wherein the assay is liquid chromatography with tandem mass spectrometry and the total run time is less than 5 minutes. In one embodiment, the lipids analysed to generate an AD risk score or predict or confirm the presence of AD in a subject are selected based on their suitability for high throughput analysis, including of 5 minutes or less, including by liquid chromatography with tandem mass spectrometry. In another aspect, there is provided a lipidomics assay for AD risk or presence of AD in a subject according to a method described herein, the method comprising: (i) separating in a sample obtained from a subject the sample analytes by liquid chromatographic separation; (ii) analysing the analytes using a mass spectrometer; wherein the total run time is less than 15 minutes; and wherein the resulting data can accurately assign individual analytes to a single lipid species. In one embodiment, the run time of the lipidomics assay is not more than 5 minutes. In one embodiment, the biological sample of the lipidomics assay is at least one of plasma, serum, dried blood spots, and dried plasma spots. In one embodiment, the combined sensitivity and specificity of the AD risk score to identify a subject with AD is at least 75%. In one embodiment, the combined sensitivity and specificity of the AD risk score to identify a subject with AD is 77.7%. In one embodiment, the combined sensitivity and specificity of the AD risk score to identify a subject with AD is at least 80%. In one embodiment, the combined sensitivity and specificity of the AD risk score to identify a subject with AD is 84.8%. In one embodiment, the combined sensitivity and specificity of the AD risk score to identify a subject with AD is at least 89%. In one embodiment, the combined sensitivity and specificity of the AD risk score to identify a subject with AD is 89.9%. Atty Dkt No.: RICE-231WO In one embodiment, the combined sensitivity and specificity of the AD risk score to identify a subject who will develop AD is at least 70%. In one embodiment, the combined sensitivity and specificity of the AD risk score to identify a subject who will develop AD is 72.1%. In one embodiment, the combined sensitivity and specificity of the AD risk score to identify a subject who will develop AD is at least 74%. In one embodiment, the combined sensitivity and specificity of the AD risk score to identify a subject who will develop AD is 74.0%. In one embodiment, the combined sensitivity and specificity of the AD risk score to identify a subject who will develop AD is at least 76%. In one embodiment, the combined sensitivity and specificity of the AD risk score to identify a subject who will develop AD is 76.2%. In one embodiment, the AD risk score is used to stratify subjects into high risk and low risk. In one embodiment, there is provided a method of stratifying a group of human subjects for a treatment therapy, the method comprising: using an Alzheimer’s disease (AD) risk score determined using the methods according to any one of claims 1 to 22 to select subjects for the treatment therapy. In one embodiment, the treatment therapy is a clinical trial of a candidate therapy. Panels or kits or compositions for use in the methods described herein are also provided. In an embodiment, the kit comprises one or more or all of: (i) a set of stable-labelled isotopes or non-physiological lipid standards for quantification of the lipid species; (ii) reference plasma samples for standardisation of the resulting lipid measures. In another aspect, there is provided a system for diagnosing Alzheimer’s disease (AD), or the likelihood of developing AD in a subject not diagnosed with AD, comprising: a processor operable to execute programs; a memory associated with the processor; a database associated with said processor and said memory; and a program stored in the memory and executable by the processor, the program being operable for: a) detecting in a biological sample from the subject a level of at least two lipid species selected from Table 2and / or 4 and / or 6; and b) correlating the levels of the lipid species detected in the biological sample to reference lipid species levels to a likelihood that the subject has or will develop AD. Atty Dkt No.: RICE-231WO It will be appreciated that other aspects, embodiments, or examples, of the methods and / or lipid biomarkers may be provided according to any aspects, embodiments, or examples thereof as described below and herein. Any embodiment herein shall be taken to apply mutatis mutandis to any other embodiment unless specifically stated otherwise. The present invention is not to be limited in scope by the specific embodiments described herein, which are intended for the purpose of exemplification only. Functionally-equivalent products, compositions and methods are clearly within the scope of the invention, as described herein. The invention is hereinafter described by way of the following non-limiting Examples and with reference to the accompanying figures. BRIEF DESCRIPTION OF ACCOMPANING DRAWINGS Figure 1. Stepped, linear solvent gradient for Research Protocol LC-MS / MS. Flow rates and percentage solvent B is detailed in the table and visually depicted in the graph. Briefly, a flow rate of 0.4ml / minute at 10% solvent B increased to 45% solvent B over 2.7 minutes, then to 53% over 0.1 minutes, to 65% over 6.2 minutes, to 89% over 0.1 minute, to 92% over 1.9 minutes and finally to 100% over 0.1 minute. The solvent was then held at 100% B for 0.8 minutes. Equilibration was as follows, solvent was decreased from 100% B to 10% B over 0.1. Figure 2. Stepped, linear solvent gradient for Clinical Protocol LC-MS / MS. Flow rates and percentage solvent B is detailed in the table and visually depicted in the graph. Briefly, flow rate was initiated at 1 ml / minute, 0% B, maintained until 0.3 min, at which point % B was increased to 40% over 0.1 min. % B was further increased linearly to 48 % B at 2.5 min, then stepped to 58 % B at 2.51 min. % B was increased linearly to 60 % at 3.1 min, then stepped to 76 % B at 3.11, before increasing linearly to 80 % B at 4.4 min. At 4.41 min the % B was increased to 100 %, maintained until 5 min, then immediately dropped to 0% B at 5.01 for equilibration. Figure 3. An overview of the study design for the development of AD risk scores using the full lipidomic data and CLP. A-C. The full lipidomic data or CLP lipids were used for the generation of AD risk scores in the ADNI cohort using ridge and lasso models on baseline data. The following temporal validations on baseline, 12 months, and 24 months were separately performed to 1) classify prevalent AD from cognitively normal (CN); 2) Atty Dkt No.: RICE-231WO classify incident AD from CN and MCI. AD, Alzheimer’s disease; CN, cognitive normal; MCI, mild cognitive impairment. Figure 4. The discrimination ability of the ADAS-Cog based models to stratify prevalent AD from CN. Three models were developed using the ADAS-Cog score as the outcome, (A) a model using only the standard risk variables; (B) a complex model using the standard risk variables and the full lipidome (n=749); and (C) a complex model using the standard risk variables and the CLP2 lipids (n=258). The ability of the models to discriminate AD and CN was assessed by the Area under the Receiver Operator Curve (AUC). Figure 5. The discrimination ability of the ADAS-Cog based models to stratify incident AD prevalent AD from CN and MCI. Three models were developed using the ADAS-Cog score as the outcome, (A) a model using only the standard risk variables; (B) a complex model using the standard risk variables and the full lipidome (n=749); and (C) a complex model using the standard risk variables and the CLP2 lipids (n=258). The ability of the models to discriminate incident AD from a combination of CN and MCI was assessed by the Area under the Receiver Operator Curve (AUC).

[0002] Atty Dkt No.: RICE-231WO BRIEF DESCRIPTION OF TABLES Table 1 lists internal standards and volumes used. Table 2 lists the Research Lipid List. Table 3 lists the internal standards used with Research Lipid List analyses. Table 4 lists the clinical lipid platform (CLP1) lipid list. Table 5 lists the internal standards used with the CLP1 and CLP2 list analyses. Table 6 lists the clinical platform subset (CLP2) lipid list. Table 7 details the predictive performances of Ridge and LASSO models model to classify AD from CN (AUC). Table 8 details the predictive performances of Ridge and LASSO models to classify incident AD from the combination of CN and MCI (AUC). Table 9 lists predictive models of AD risk developed using Ridge regression incorporating all the covariates together with the CLP1s lipids. Table 10: lists predictive models of AD risk developed using LASSO regression incorporating all the covariates together with the CLP1s lipids. Table 11: lists predictive models of AD risk developed using Ridge regression incorporating all the covariates together with the CLP2s lipids. Table 12: lists the predictive models of AD risk developed using LASSO regression incorporating all the covariates together with the CLP2s lipids. Table 13: The reclassification performances of lipidome-based AD risk model to classify incident AD from MCI (NRI and IDI). Table 14: The predictive performances of ADAS-Cog score to classify AD from CN. Table 15: The predictive performances of ADAS-Cog score to classify incident AD from the combination of CN and MCI. LIST OF ABBREVIATIONS Abbreviation Full description Atty Dkt No.: RICE-231WO Cer(d) Ceramide(d) Atty Dkt No.: RICE-231WO LPE(P) Lysoalkenylphosphatidylethanolamine Atty Dkt No.: RICE-231WO TG Triacylglycerol Lipid category Lipid class Lipid class / subclass Abbreviation Atty Dkt No.: RICE-231WO Glycerophosphoethanol- Phosphatidylethanolamine PE Gl r h h liid min E Atty Dkt No.: RICE-231WO Quinones and Prenol Lipids Ubiquinone Ubiquinone h dr in n The present disclosure describes the following various non-limiting embodiments, which relate to research undertaken into identifying and developing lipidomics based methods for assessing Alzheimer’s disease (AD) in a subject. Terms Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the art to which the disclosure belongs. As used herein the singular forms "a", "an" and "the" include plural aspects unless the context clearly dictates otherwise. Thus, for example, reference to "a lipid species" includes a single lipid species, as well as two or more lipid species, reference to "the disclosure" includes single and multiple aspects of the disclosure and so forth. Throughout this specification, unless the context requires otherwise, the word "comprise", or variations such as "comprises" or "comprising", will be understood to imply the inclusion of a stated element or integer or group of elements or integers but not the exclusion of any other element or integer or group of elements or integers. By "consisting of" is meant including, and limited to, whatever follows the phrase "consisting of". Thus, the phrase "consisting of" indicates that the listed elements are required or mandatory, and that no other elements may be present. By "consisting essentially of" is meant including any elements listed after the phrase, and limited to other elements that do not interfere with or contribute to the activity or action specified in the disclosure for the listed elements. The term "and / or", e.g., "X and / or Y" shall be understood to mean either "X and Y" or "X or Y" and shall be taken to provide explicit support for both meanings or for either meaning. As used herein, the term "about", unless stated to the contrary, is to be construed according to the context in which it is used. The naming convention for lipids used herein follows the guidelines established by the Lipid Maps Consortium and the shorthand notation of Liebisch et al. (see for example Fahy et al.2005. J. Lipid Res.46: 839–861; Fahy et al 2009. J. Lipid Res.50 (Suppl.): S9– S14; Liebisch et al. 2013. J. Lipid Res. 54: 1523–1530; Liebisch et al.2020. J Lipid Res. 2020 Dec; 61(12): 1539–1555). Lipids can be divided into primary categories including Atty Dkt No.: RICE-231WO fatty acyls (FA), glycerolipids (GL), glycerophospholipids (GP), sphingolipids (SP), sterol lipids (ST), prenol lipids (PR), saccharolipids (SL), and polyketides (PK), which can be divided into molecule classes, subclasses and species representing chemical structures. Fatty acyls (FA) are a diverse group of molecules synthesised by chain elongation of an acetyl-CoA primer with malonyl-CoA (or methylmalonyl-CoA) groups that may contain a cyclic functionality and / or are substituted with heteroatoms. Structures with a glycerol group are represented by two distinct categories: the glycerolipids (GL), which include acylglycerols but also encompass alkyl and 1Z-alkenyl variants, and the glycerophospholipids (GP), which are defined by the presence of a phosphate (or phosphonate) group esterified to one of the glycerol hydroxyl groups. Sterol lipids (ST) and prenol lipids (PR) share a common biosynthetic pathway via the polymerization of dimethylallyl pyrophosphate / isopentenyl pyrophosphate but otherwise differ in structure and function. Sphingolipids (SP) contain a long-chain base as their core structure. Saccharolipids” (SL) contain fatty acyl groups linked directly to a sugar backbone. Glycerophospholipids (GP) typically contain two fatty acid chains and in the absence of detailed characterisation are expressed as the sum composition of carbon atoms and double bonds (i.e. PC(38:6)). However, where an acyl chain composition has been determined the naming convention indicates this (i.e. PC(38:6) is changed to PC(16:0_22:6)). This is also extended into other lipid classes or subclasses. The present disclosure refers to lipid molecules using the numbering system X:Y. The number X represents the number of carbon atoms present in the chain. In the context of alkylglycerols, alkyacylglycerols or alkyldiacylglycerols, the number Y represents the number of double bonds present in the chain. For example, an alkylglycerol numbered as 16:0 contains a hydrocarbon group having a 16 carbon chain with no double bonds. As a further example, an alkylglycerol numbered as 18:1 contains a hydrocarbon group having an 18 carbon chain with 1 double bond. In the context of plasmalogens / plasmenyl phospholipids, the number Y in the first listed alkenyl chain (i.e. PE(P-X:Y / X:Y) represents the number of double bonds present in the alkenyl chain in addition to the vinyl ether group. For example, a plasmalogen numbered as PE(P-16:0 / 20:4) the 16:0 alkenyl group contains a hydrocarbon group having a 16 carbon chain with no double bonds other than the vinyl ether group (i.e. there is a double bond between the first 2 carbons and the remaining 14 carbons are saturated). As another example, a plasmalogen numbered as PE(P-18:1 / 20:4) the 18:1 alkenyl group contains a hydrocarbon group having an 18 carbon chain with 1 double bond in addition Atty Dkt No.: RICE-231WO to the vinyl ether group (i.e. there is a double bond between the first 2 carbon atoms, and there is one other double bond between 2 carbons out of the remaining 16 carbons). Where ether lipids contain one or more double bonds, the double bonds may be located at various positions in the hydrocarbon chains. For example, an alkylglycerol numbered as 18:1 may contain a mixture of species, e.g. with cis-n7 and cis-n9 double bonds. As another example, a plasmalogen (e.g PE(P)) numbered as 18:1 may contain a mixture of species, e.g. with cis-n7 and cis-n9 double bonds. As used herein, the term “plasmanyl” refers to phospholipids having an ether bond in the sn-1 position to an alkyl group. As used herein, the term “plasmenyl” refers to phospholipids having an ether bond in the sn-1 position to an alkenyl group. The plasmenyl phospholipids are also referred to as “plasmalogens”. A plasmalogen having a “16:0” alkenyl group is typically a molecule having an ether bond in the sn-1 position to an 16 carbon chain which contains a double bond between carbons 1 and 2 (i.e. typically a cis-vinyl ether group), and no other double bonds in the chain. A plasmalogen having an “18:0” alkenyl group is typically a molecule having an ether bond in the sn-1 position to an 18 carbon chain which contains a double bond between carbons 1 and 2 (i.e. typically a cis-vinyl ether group), and no other double bonds in the chain. A plasmalogen having an “18:1” alkenyl group is typically a molecule having an ether bond in the sn-1 position to an 18 carbon chain which contains a double bond between carbons 1 and 2 (i.e. typically a cis-vinyl ether group), and having one additional double bond, typically between carbons 7 and 8 (e.g. n7), between carbons 9 and 10 (e.g. n9), or between carbons 11 and 12 (e.g. n11), and typically a cis-double bond. A plasmalogen having an “18:2” alkenyl group is typically a molecule having an ether bond in the sn-1 position to an 18 carbon chain which contains a double bond between carbons 1 and 2 (i.e. typically a cis-vinyl ether group), and having two additional double bonds, typically between carbons 9 and 10, and between carbons 11 and 12, and typically cis-double bonds. A plasmalogen having an “18:2” acyl alkenyl group is typically a molecule having an ester bond in the sn-2 position to an 18 carbon chain which has two double bonds, typically between carbons 9 and 10, and between carbons 11 and 12, and typically cis- double bonds. Atty Dkt No.: RICE-231WO A plasmalogen having a “20:4” acyl alkenyl group is typically a molecule having a ester bond in the sn-2 position to a 20 carbon chain which has four double bonds, typically between carbons 5 and 6, carbons 8 and 9, carbons 11 and 12, and carbons 14 and 15, and typically cis-double bonds. As used herein, "acyl" refers to a group having a straight, branched, or cyclic configuration or a combination thereof, attached to the parent structure through a carbonyl functionality. Such groups may be saturated or unsaturated, aliphatic or aromatic, and carbocyclic or heterocyclic. Examples of a C1-C24acyl- group include acetyl, benzoyl-, nicotinoyl-, propionyl-, isobutyryl-, oxalyl-, and the like. Lower-acyl refers to acyl groups containing one to four carbons. An acyl group can be unsubstituted or substituted, for example with one or more groups selected from halogen, -OH, -NH2, -CN, –OC1-4alkyl and –CO2H. Additional examples or generally applicable substituents are illustrated by the specific compounds described herein. The term "aliphatic" as used herein, includes saturated, unsaturated, straight chain (i.e., unbranched), or branched, aliphatic hydrocarbons, which are optionally substituted with one or more functional groups. In some embodiments, the aliphatic may contain one or more functional groups such as double bond, triple bond, or a combination thereof. As will be appreciated by one of ordinary skill in the art, "aliphatic" is intended herein to include, but is not limited to, alkyl, alkenyl, alkynyl, or acyl moieties. Thus, as used herein, the term "alkyl" includes straight and branched saturated groups. An analogous convention applies to other generic terms such as "alkenyl", "alkynyl", "acyl" and the like. Furthermore, as used herein, the terms "alkyl", "alkenyl", "alkynyl", "acyl" and the like encompass both substituted and unsubstituted groups. As used herein, "alkenyl" refers to a straight or branched chain hydrocarbon containing, for example, from 2 to 30 carbons and containing at least one carbon-carbon double bond. In some embodiments, the alkenyl group contains 10 to 25, 14 to 22, or 16 to 20 carbon atoms. In some embodiments, the alkenyl group contains 15, 16, 17, 18, 19 or 20 carbon atoms. Representative examples of "alkenyl" include, but are not limited to, ethenyl, 2-propenyl, 2-methyl-2-propenyl, 3-butenyl, 4-pentenyl, 5-hexenyl, 2-heptenyl, 2-methyl-1-heptenyl, 3-decenyl, 3-undecenyl, 4-dodecenyl, 4-tridecenyl, 9-tetradecenyl, 8-pentadecenyl, 5-hexadecenyl, 8-heptadecenyl, 9-octadecenyl, 9-nonadecenyl and the like. Additional examples or generally applicable substituents are illustrated by the specific compounds described herein. As used herein, "alkyl" refers to a straight or branched chain hydrocarbon containing, for example, from 1 to 30 carbon atoms. In some embodiments, the alkyl group Atty Dkt No.: RICE-231WO contains 10 to 25, 14 to 22, or 16 to 20 carbon atoms. In some embodiments, the alkyl group contains 15, 16, 17, 18, 19 or 20 carbon atoms. Representative examples of alkyl include, but are not limited to, methyl, ethyl, n-propyl, iso-propyl, n-butyl, sec-butyl, iso- butyl, tert-butyl, n-pentyl, isopentyl, neopentyl, n-hexyl, 3-methylhexyl, 2,2- dimethylpentyl, 2,3-dimethylpentyl, n-heptyl, noctyl, n-nonyl, n-decyl, n-undecyl, n- dodecyl, n-tridecyl, n-tetradecyl, n-pentadecyl, n-hexadecyl, n-heptadecyl, n-octadecyl, n- nonadecyl and the like. Additional examples or generally applicable substituents are illustrated by the specific compounds described herein. As used herein, “acyl alkenyl” refers to a straight or branched chain hydrocarbon containing, for example, from 2 to 30 carbons and containing at least one carbon-carbon double bond, which is covalently bonded to an acyl group. The use of nomenclature 22:6 or 18:2 and the like in the context of an acyl alkenyl group refers to an acyl alkenyl group having 22 carbons or 18 carbons respectively, and having 6 or 2 double bonds respectively. An example of an acyl alkenyl group is: or alkyldiacylglycerols (as an acyl group), or as an acyl group in plasmanyl- or plasmenyl- phospholipids. Typically, when present in those species, there is no double bond between the carbons which are α- and β- to the acyl group. As used herein, “acyl alkyl” refers to a straight or branched chain hydrocarbon containing, for example, from 1 to 30 carbons, which is covalently bonded to an acyl group. The use of nomenclature 22:0 or 18:0 and the like in the context of an acyl alkyl group refers to an acyl alkyl group having 22 carbons or 18 carbons respectively. An example of an acyl alkyl group is: It will also be recognised that the compounds described herein may possess asymmetric centres and are therefore capable of existing in more than one stereoisomeric form. The disclosure thus also relates to compounds in substantially pure isomeric form at Atty Dkt No.: RICE-231WO one or more asymmetric centres e.g., greater than 90% ee, such as 95% or 97% ee or greater than 99% ee, as well as mixtures, including racemic mixtures, thereof. Such isomers may be naturally occurring or may be prepared by asymmetric synthesis, for example using chiral intermediates, or by chiral resolution. The present disclosure relates to derivatives of glycerol. Whilst glycerol is achiral, derivatives are typically chiral. Typically the glycerol utilised will have a stereochemical configuration corresponding to that found in nature. In some embodiments, the glycerol derivatives utilised have the following stereochemical configuration: herein, the term “alkylglycerol” means a compound in which the R1group is a hydrocarbon chain, the R2and R3groups are each hydrogen. Although the term “alkyl” glycerol is used, it will be understood by those of skill in the art that the term encompasses species with hydrocarbon groups at the R1position which include unsaturation in the hydrocarbon chain. However, an alkylglycerol does not contain a double bond between carbons 1 and 2 of the hydrocarbon chain, e.g. proximal to the ether linkage. An alkylglycerol having a “16:0” group is typically a molecule having an ether bond in the sn-1 position to a 16 carbon saturated hydrocarbon chain, and no double bonds in the chain. An alkylglycerol having an “18:0” group is typically a molecule having an ether bond in the sn-1 position to an 18 carbon saturated hydrocarbon chain, and no double bonds in the chain. An alkylglycerol having an “18:1” group is typically a molecule having an ether bond in the sn-1 position to an 18 carbon hydrocarbon chain, which contains one double bond, typically between carbons 9 and 10 and typically a cis-double bond. As referred to herein, the term “alkylacylglycerol” means a compound of Formula 1 in which the R1group is a hydrocarbon chain, one of the R2and R3groups is hydrogen, and the other of the R2and R3groups is an acyl group, either an acyl alkyl group or an acyl alkenyl group. Although the term “alkyl” acylglycerol is used, it will be understood by those of skill in the art that the term encompasses species with hydrocarbon groups at the Atty Dkt No.: RICE-231WO R1position which include unsaturation in the hydrocarbon chain. However, an alkylacylglycerol does not contain a double bond between carbons 1 and 2 of the R1hydrocarbon chain, e.g. proximal to the ether linkage. As referred to herein, the term “alkyldiacylglycerol” means a compound of Formula 1 in which the R1group is a hydrocarbon chain, and the R2and R3groups are acyl groups, either acyl alkyl or acyl alkenyl. Although the term “alkyl” diacylglycerol is used, it will be understood by those of skill in the art that the term encompasses species with hydrocarbon groups at the R1position which include unsaturation in the hydrocarbon chain. However, an alkyldiacylglycerol does not contain a double bond between carbons 1 and 2 of the R1hydrocarbon chain, e.g. proximal to the ether linkage. The term “BMI” refers to body mass index, and is calculated by dividing the weight of an individual in kg by their height in metres squared. Reference to "two or more", incudes 2, 3, 4, 5, 6, 7, 8, 9 or 10 or more lipids. As used herein, the term “incident AD” or “incident Alzheimer’s disease (AD)” refers to AD developed at some point after a sample is taken, for example, after a blood sample was taken, such that at the time of sampling, subjects were classified as cognitively normal (CN) or having a mild cognitive impairment (MCI), and then developed AD in a follow-up period, for example as identified at a subsequent sampling during a longitudinal study. As used herein, the term “prevalent AD” refers to AD that is present at the time a sample is taken, such that at time of sample, subjects have AD. As used herein, the term “subject” is any animal. In an embodiment, the subject is a mammal. In an embodiment, the subject is a human. In an embodiment, the subject is an adolescent. In an embodiment, the subject is an adult. In an embodiment, the subject is a woman. In an embodiment, the subject has been diagnosed with AD, for example using a method as described herein. In an embodiment, the subject is at risk of AD. In an embodiment, the subject is cognitively normal. In an embodiment, the subject has minor cognitive impairment. As used herein, the term “biological sample” or “sample” refers to any type of suitable material obtained from the subject. The term encompasses a clinical sample, for example a tissue biopsy, tissue samples, preserved tissue samples, for example paraffin- embedded tissues), biological fluids, for example blood (whole blood and blood fractions), live cells, cells in culture, cell supernatants, and cell lysates derived therefrom. The sample can be used as obtained directly from the source or following at least one-processing step, for example formalin fixing and / or paraffin embedding, or for example a purification step. Atty Dkt No.: RICE-231WO It will be apparent to the skilled person that the sample can be prepared in any medium which does not interfere with the method of the disclosure. Typically, the sample comprises blood or a blood fraction. The skilled person will be aware of suitable sample selection and any necessary preparatory steps. A used herein, “risk score” is a way of stratifying subjects, for example subjects in a population, wherein a higher risk score reflects a higher likelihood and / or increased risk of developing AD and / or having AD. In an embodiment, a lower risk score indicates decreased likelihood of developing AD and / or having AD. Lipidomic assessment of Alzheimer’s Disease The present disclosure provides methods for assessing Alzheimer’s disease (AD) in a subject based on a level of a plurality or population of lipid species in the subject. In some examples, assessing AD includes one or more of: calculating an AD risk score; determining presence of AD; determining AD incidence; monitoring the risk of developing AD in individuals and predicting the likelihood that a subject may develop AD. In some examples, assessing AD is a single assessment. In some examples, assessing AD is done on two or more occasions, for example, assessment may be done at regular intervals to monitor risk reduction regimes, or irregular intervals as needed, for example to determine whether an intervention is lowering risk or a lack of intervention is increasing the risk of developing AD. In one example, the present disclosure provides a method of assessing AD in a subject, the method comprising: detecting in a biological sample from the subject a level of a plurality of lipid species comprising at least two lipid species selected from the group consisting of any one or more Tables, lipid classes, and / or lipid species according to any embodiments or examples thereof as described herein; comparing the levels or ratios of levels of the lipid species detected in the biological sample to reference levels or ratios of the lipid species; and determining the risk of developing AD or the presence or incidence of AD in the subject on the basis of the comparison. It will be appreciated that any of the methods provided herein can comprise detecting a level of a plurality or population of lipid species comprising at least two lipid species selected from any lipid species of the group consisting of any one or more of the Tables, lipid classes, and / or lipid species according to any embodiments or examples thereof as described herein. In some examples, the plurality or population of lipid species detected in the sample can comprise at least 3, 4, 5, 6, 7, 8, 9, 10, 15, 20, 25, 30, 35, 40, Atty Dkt No.: RICE-231WO 45, 50, 55, 60, 65, 70, 75, 80, 85, 90, 95, 100, 110, 120, 130, 150, 160, 170, 180, 190, 200, 210, 220, 230, 240, 250, 260, 270, 280, 290, or 300 lipid species. In some examples, the plurality or population of lipid species detected in the sample can comprise less than about 800, about 700, about 600, about 500, about 400, about 300, about 290, about 280, about 270, about 260, about 250, about 240, about 230, about 220, about 210, about 200, about 190, about 180, about 170, about 160, about 150, about 140, about 130, about 120, about 110, about 100, about 95, about 90, about 85, about 80, about 75, about 70, about 65, about 60, about 55, about 50, about 45, about 40, about 35, about 30, about 25, about 20, about 15, or about 10 lipid species. In some examples the plurality or population of lipid species detected in the sample can comprise a range selected from any two upper and / or lower amounts as previously described, such as 2 to 20, 2 to 100, 2 to 200, 2 to 235, 2 to 250, 2 to 269, 2 to 298, 2 to 300, 2 to 339, 2 to 400, 2 to 500, 2 to 600, 2 to 700, 2 to 791, 10 to 50, 10 to 100, 10 to 250, 10 to 500, 20 to 100, 100 to 200, 200 to 300, 300 to 400, 400 to 500, 500 to 600, 600 to 700, 25 to 300, or 50 to 250. The lipid species may be selected from one or more lipid classes comprising or consisting of any one or more of sphingolipids, glycerophospholipids, sterol lipids, fatty acyls, glycerolipids, and prenol lipids. In one example, the lipid species is selected from the group consisting of sphingolipids, glycerophospholipids, and fatty acyls. In one example, the lipid species are selected from one or more lipid classes or sub- classes comprising or consisting of any one or more of ceramide, neutral glycosphingolipids, acid glycosphingolipids, phosphosphingolipids, glycerophosphocholines, glycerophosphoethanolamines, glycerophosphoinositols, glycerophosphoserines, glycerophosphoglycerols, sterols, fatty esters, diradylglycerols, triradylglycerols, quinones, and hydroquinones. In another example, the one or more lipid classes or sub-classes comprise or consist of any one or more of ceramide, neutral glycosphingolipids, acid glycosphingolipids, phosphosphingolipids, glycerophosphocholines, glycerophosphoethanolamines, glycerophosphoinositols, and fatty esters. In one example, the lipid species are selected from one or more lipid classes or sub- classes comprising or consisting of any one or more of sulphatides, GM3gangliosides, alkyldiacylglycerols, ubiquinones, alkenylphosphatidylethanol-amines, ceramides, acylcarnitines and phosphatidylcholines. In one example, the lipid species are selected from one or more lipid classes or sub- classes comprising or consisting of any one or more of triacylglycerols, Atty Dkt No.: RICE-231WO alkyldiacylglycerols, lysophosphatidylcholines, GM3gangliosides, alkenylphosphatidylethanol-amines, sulphatides and ceramides. In one example, the lipid species are selected from one or more lipid classes or sub- classes comprising or consisting of any one or more of GM3gangliosides, sulphatides, ubiquinones, alkenylphosphatidylethanol-amines, ceramides, alkylphosphatidylethanolamines, acylcarnitines, triacylglycerols and alkyldiacylglycerols. In one example, the lipid species are selected from one or more lipid classes or sub- classes comprising or consisting of any one or more of triacylglycerols, alkylphosphatidylethanolamines, GM3 gangliosides, alkenylphosphatidylethanol-amines, acylcarnitines, lysophosphatidylcholines, dimethyl cholesteryl esters and sulphatides. In one example, the plurality of lipid species detected comprises at least two lipid species of Table 2. In some examples, the lipid species detected from Table 2 comprise at least 2, 3, 4, 5, 6, 7, 8, 9, 10, 15, 20, 25, 30, 35, 40, 45, 50, 55, 60, 65, 70, 75, 80, 85, 90, 95, 100, 110, 120, 130, 150, 160, 170, 180, 190, 200, 210, 220, 230, 240, 250, 260, 270, 280, 290, or 300 lipid species. In some examples, the lipid species detected from Table 2 comprise less than about 800, 700, 600, 500, 400, 300, 290, 280, 270, 260, 250, 240, 230, 220, 210, 200, 190, 180, 170, 160, 150, 140, 130, 120, 110, 100, 95, 90, 85, 80, 75, 70, 65, 60, 55, 50, 45, 40, 35, 30, 25, 20, 15, or 10 lipid species. In some examples, the lipid species detected from Table 2 comprise a range selected from any two previously described upper and / or lower amounts, such as 2 to 700, 10 to 500, 25 to 300, or 50 to 250 lipid species. In one example, the plurality of lipid species comprises the lipids set out at Table 4. In some examples, the lipid species detected from Table 4 comprise at least 2, 3, 4, 5, 6, 7, 8, 9, 10, 15, 20, 25, 30, 35, 40, 45, 50, 55, 60, 65, 70, 75, 80, 85, 90, 95, 100, 110, 120, 130, 150, 160, 170, 180, 190, 200, 210, 220, 230, 240, 250, 260, 270, 280, 290, or 300. In some examples, the lipid species detected from Table 4 comprise less than about 350, about 340, about 330, about 320, about 310 about 300, about 290, about 280, about 270, about 260, about 250, about 240, about 230, about 220, about 210, about 200, about 190, about 180, about 170, about 160, about 150, about 140, about 130, about 120, about 110, about 100, about 95, about 90, about 85, about 80, about 75, about 70, about 65, about 60, about 55, about 50, about 45, about 40, about 35, about 30, about 25, about 20, about 15, or about 10. In some examples, the lipid species detected from Table 4 comprise a range selected from any two previously described upper and / or lower amounts, such as 2 to 300, 10 to 290, 25 to 270, or 50 to 250. Atty Dkt No.: RICE-231WO In one example, the plurality of lipid species comprises the lipids set out at Table 6. In some examples, the lipid species detected from Table 6 comprise at least 2, 3, 4, 5, 6, 7, 8, 9, 10, 15, 20, 25, 30, 35, 40, 45, 50, 55, 60, 65, 70, 75, 80, 85, 90, 95, 100, 110, 120, 130, 150, 160, 170, 180, 190, 200, 210, 220, 230, 240, or 250. In some examples, the lipid species detected from Table 6 comprise less than about 270, about 260, about 250, about 240, about 230, about 220, about 210, about 200, about 190, about 180, about 170, about 160, about 150, about 140, about 130, about 120, about 110, about 100, about 95, about 90, about 85, about 80, about 75, about 70, about 65, about 60, about 55, about 50, about 45, about 40, about 35, about 30, about 25, about 20, about 15, or about 10. In some examples, the lipid species detected from Table 6 comprise a range selected from any two previously described upper and / or lower amounts, such as 2 to 250, 10 to 240, 25 to 200, or 50 to 150. In some examples, the lipid species analysed may form a subset of the total lipid species detected. As defined herein, normalising numeric values involves adjusting the numeric values obtained for a sample to notionally common scale determined, for example, by running a reference sample of known values simultaneously to a patient sample, thereby enabling comparison across samples taken and / or analysed at different time points and / or from different sources. As defined herein, a reference sample or reference level is a sample or level obtained from population of individuals with predetermined characteristics. For example, populations of healthy individuals; for example, populations of subjects not known to have Alzheimer’s disease; for example, populations of male subjects; for example populations of female subjects; for example, populations of subjects of a predetermined chronological age, for example, populations of subjects of known actual BMI; for example, populations of subjects with known APOE4 status; for example, populations of subjects with known or predetermined HDL-cholesterol, cholesterol, triglyceride, stat and / or omega 3 levels. A skilled person will understand a suitable reference sample from which to derive reference values. In some examples, the model used to generate an AD risk score comprises calculating the sum of the lipid species, wherein each lipid species is weighted. The sum of the weighted lipid species is then assessed by reference to a reference sample or level in a reference population. In one example, the present disclosure provides a method of determining an AD risk score of a subject, the method comprising: detecting in a biological sample from the Atty Dkt No.: RICE-231WO subject a level of at least two lipid species selected from Table 4 and / or 6; and comparing the levels of the lipid species detected in the biological sample to reference lipid species levels, wherein the comparison is determinative of the risk of developing AD in the subject. In another example, the methods provided further comprise: standardising each continuous variable; log transforming each lipid concentration; and further standardising the variables prior to their use as the predictors of AD. In an example, any of the methods provided herein further comprise a step of confirming that the subject has, or is likely to develop Alzheimer’s disease based on the comparison of the lipid species in the biological sample to reference lipid species levels. In one example, the AD risk score is used to determine a subject’s risk of developing AD. In one example, the AD risk score is used to identify incident AD in a subject. In one example, the AD risk score is used to determine whether a subject has AD and / or to monitor their risk of developing AD. In one example the lipid species are selected from one or more lipid classes or sub- classes comprising or consisting of any one or more of AC, AC-OH, CE, Cer(d), Cer(m), C1P, COH, DE, DG, dimethyl-CE, FFA, GM3, Hex2Cer, Hex3Cer, HexCer, LPC, LPC(O), LPC(P), LPE, LPE(P), LPI, methyl-CE, methyl-DE, PC, OxSpecies (oxidised PC), PC(O), PC(P), PE, PE(O), PE(P), PI, PIP1, PS, S1P, SHexCer, SM, TG, TG(O) and Ubiquinone. In one example the lipid species are selected from one or more lipid classes or sub- classes comprising or consisting of any one or more of AC, AC-OH, CE, Cer(d), Cer(m), C1P, COH, DE, DG, dimethyl-CE, FFA, GM3, Hex2Cer, Hex3Cer, HexCer, LPC, LPC(O), LPC(P), LPE, LPE(P), LPI, methyl-CE, methyl-DE, PC, PC(O), PC(P), PE, PE(O), PE(P), PI, PIP1, PS, S1P, SHexCer, SM, TG, TG(O) and Ubiquinone. In one example the lipid species are selected from one or more lipid classes or sub- classes comprising or consisting of any one or more of SHexCer, GM3, TG(O), Ubiquinone, Cer, AC, PC(O), TG, CE, DE, AC, FA, Hex3Cer, PC, and LPC(O). In one example the lipid species are selected from one or more lipid classes or sub- classes comprising or consisting of any one or more of TG, TG(O), LPC(O), GM3, TG, PE(P), SHexCer, GM3, Cer(d), AC, HexCer, Ubiquinone, CE, dimethyl-CE, and PC(P). In one example the lipid species are selected from one or more lipid classes or sub- classes comprising or consisting of any one or more of GMD3, SHexCer, GM3, Atty Dkt No.: RICE-231WO Ubiquinone, PE(P), Cer(d), PE(O), AC, TG, TG(O), CE, DE, FA, Hex3Cer, PC, dimethyl- CE, CE, TG(O), and PC(P). In one example the lipid species are selected from one or more lipid classes or sub- classes comprising or consisting of any one or more of TG, PE(O), GM3, PE(P), AC, LPC(O), dimethyl-CE, SHexCer, Hex3Cer, Ubiquinone, Cer(d), TG(O), CE, AC, HexCer, DE, and methyl-CE. In one example, the lipid class or subclass SHexCer comprises or consists of any one or more lipid species of d18:1 / 16:0(OH) and d18:1 / 16:0. In one example, the lipid class or subclass GM3 comprises or consists of any one or more lipid species of d18:1 / 22:0, d18:1 / 24:1, d18:1 / 16:0, and d18:1 / 18:0. In one example, the lipid class or subclass TG(O) comprises or consists of any one or more lipid species of 52:2, 54:4, 54:3, 50:1, 50:2, and 52:0. In one example, the lipid class or subclass TG comprises or consists of any one or more lipid species of 50:3, 58:8, 52:4, 52:3, 50:2, 56:7, 50:4, 56:8, 48:3, 48:1, 51:2, 50:4, 52:2, 54:3, 56:9, 48:2, 54:6, 58:9, 58:10, 54:5, 54:7, 50:1, 48:0, 52:5, 49:1, 54:4, 48:1, 51:0, 51:1, 56:6, 52:1, 53:2, 50:0, 52:4, 53:2, 54:2, and 52:0. In one example, the lipid class or subclass PE(P) comprises or consists of any one or more lipid species of 20:0 / 18:2, 18:0 / 18:2, 20:0 / 20:4, 16:0 / 20:4, 18:0 / 18:1, 17:0 / 20:4, 17:0 / 22:6, 16:0 / 18:1, 16:0 / 18:2, 16:0 / 20:5, 16:0 / 22:4, 16:0 / 22:6, 17:0 / 22:6, 18:0 / 20:4, 18:0 / 20:5, 18:0 / 22:4, 18:0 / 22:6, 18:1 / 20:4, 18:1 / 20:5, 20:0 / 20:4, 20:0 / 18:2, and 16:0 / 20:4. In one example, the lipid class or subclass PE comprises or consists of any one or more lipid species of 18:0 / 22:6, 36:2, 16:0 / 22:6, 16:0 / 16:1, 16:0 / 18:1, 16:0 / 18:2, 18:0 / 18:1, 16:0 / 20:4, and 18:0 / 20:4. In one example, the lipid class or subclass PE(O) comprises or consists of any one or more lipid species of 34:1, 16:0 / 18:2, and 16:0 / 20:4. In one example, the lipid class or subclass Cer(d) comprises or consists of any one or more lipid species of 18:1 / 20:0, 18:1 / 22:0, 18:1 / 24:1, 19:1 / 24:1, 19:1 / 22:0, 16:1 / 22:0, 16:1 / 24:0, 16:1 / 24:1, 17:1 / 24:0, 18:1 / 16:0, 18:1 / 23:0, 18:1 / 24:0, 18:2 / 22:0, 18:2 / 23:0, 18:1 / 24:0, 18:2 / 22:0, 18:2 / 23:0, 18:2 / 24:0, 18:2 / 24:1, 19:1 / 22:0, 19:1 / 24:0, 19:1 / 24:1, 18:1 / 24:1 In one example, the lipid class or subclass AC comprises or consists of any one or more lipid species of 24:0, 14:2, 12:0, 14:0, 12:0, 14:1, 12:1, 13:0, 16:0, 26:0, 18:1, 16:1, 18:2, 26:1, and 13:0. In one example, the lipid class or subclass PC(O) comprises or consists of any one or more lipid species of 36:0, 34:2, 16:0 / 16:0, and 36:5. Atty Dkt No.: RICE-231WO In one example, the lipid class or subclass PC(P) comprises or consists of any one or more lipid species of 16:0 / 20:5, 16:0 / 20:4, 16:0 / 16:0, 16:0 / 18:2, 17:0 / 20:4, 16:0 / 18:1, 16:0 / 14:0, 16:0 / 18:1, 16:0 / 18:2, 17:0 / 20:4, 16:0 / 22:5, and 16:0 / 20:5. In one example, the lipid class or subclass PC comprises or consists of any one or more lipid species of 18:1 / 20:3, 16:0 / 18:1, 18:0 / 22:6, 34:2, 18:1 / 20:3, 18:2 / 18:2, 16:0 / 18:0, 36:2, 15:0 / 20:4, 18:0 / 20:4, 38:4, 33:2, 33:1, 31:0, 32:1, 32:2, 33:2, 16:0 / 18:2, 14:0 / 20:4, 34:5, 15:0 / 20:4, 35:5, 16:0 / 22:6, 15:0 / 22:6, 38:6, 16:0 / 20:4, 18:0 / 18:1, 36:2, 36:4, 16:1 / 20:4, 16:0 / 20:5, 15:0 / 22:6, 38:4, 18:0 / 20:4, 16:0 / 22:6, 40:8, 16:0 / 18:3, 16:0 / 16:0, 18:2 / 18:2, 18:0 / 18:1, 16:0 / 18:3, 36:4, 16:0 / 20:5, and 16:0 / 18:2. In one example, the lipid class or subclass LPC(O) comprises or consists of any one or more lipid species of 20:0, 24:2, 24:0, 22:1, 22:0, 22:5, and 19:0. In other examples, the lipid species may comprise or consist of any one or more combinations of an of the above lipid species or any classes or subclasses thereof. In one example, the present disclosure provides a method of determining the likelihood a subject will contract Alzheimer’s disease (AD), monitoring the risk of AD in an asymptomatic subject, and / or determining whether a subject has AD, the method comprising: detecting in a biological sample from the subject a level of a plurality of lipid species comprising at least two lipid species selected from any lipid species of the group consisting of any one or more Tables, lipid classes, and / or lipid species according to any embodiments or examples thereof as described herein; comparing the levels or ratios of levels of the lipid species detected in the biological sample to reference levels or ratios of the lipid species; and determining an Alzheimer’s risk score, and / or Alzheimer’s status of the subject on the basis of the comparison. In one example, the predicted Alzheimer’s score is compared to the actual Alzheimer’s status of the subject. In one example, the plurality of lipid species comprises the lipids set out in Tables 3, 4, 13, 14, 15 and 16. In one example, the plurality of lipid species comprises the lipids set out in any one or more of Tables 2, 3, 4, 5, 6, 13, 14, 15 and 16. In some examples, the lipid species detected from any one or more of Tables 2, 3, 4, 5, 6, 13, 14, 15 and 16 comprises at least 2, 3, 4, 5, 6, 7, 8, 9, 10, 15, 20, 25, 30, 35, 40, 45, 50, 55, 60, 65, 70, 75, 80, 85, 90, 95, 100, 110, 120, 130, 150, 160, 170, 180, 190, 200, 210, 220, 230, 240, 250, 260, 270, 280, 290, or 300 lipids. In some examples, the lipid species detected from any one or more of Atty Dkt No.: RICE-231WO Tables 2, 3, 4, 5, 6, 13, 14, 15 and 16 comprises less than about 300, about 290, about 280, about 270, about 260, about 250, about 240, about 230, about 220, about 210, about 200, about 190, about 180, about 170, about 160, about 150, about 140, about 130, about 120, about 110, about 100, about 95, about 90, about 85, about 80, about 75, about 70, about 65, about 60, about 55, about 50, about 45, about 40, about 35, about 30, about 25, about 20, about 15, or about 10 lipids. In some examples, the lipid species detected from any one or more of Tables 2, 3, 4, 5, 6, 13, 14, 15 and 16 comprise a range selected from any two previously described upper and / or lower amounts, such as 2 to 300, 10 to 290, 25 to 270, or 50 to 250. In some examples, the lipid species detected from any one or more of Tables 2, 3, 4, 5, 6, 13, 14, 15 and 16, in order of priority numbering in any one of Tables 13, 14, 15 and 16, comprises at least the first 2, 3, 4, 5, 6, 7, 8, 9, 10, 15, 20, 25, 30, 35, 40, 45, 50, 55, 60, 65, 70, 75, 80, 85, 90, 95, 100, 110, 120, 130, 150, 160, 170, 180, 190, 200, 210, 220, 230, 240, 250, 260, 270, 280, 290, or 300. In some examples, the lipid species detected from any one or more of Tables 2, 3, 4, 5, 6, 13, 14, 15 and 16, in order of priority numbering in any one of Tables 13, 14, 15 and 16, comprises less than about 300, about 290, about 280, about 270, about 260, about 250, about 240, about 230, about 220, about 210, about 200, about 190, about 180, about 170, about 160, about 150, about 140, about 130, about 120, about 110, about 100, about 95, about 90, about 85, about 80, about 75, about 70, about 65, about 60, about 55, about 50, about 45, about 40, about 35, about 30, about 25, about 20, about 15, or about 10 of the first listed lipids. In some examples, the lipid species detected from any one or more of Tables 2, 3, 4, 5, 6, 13, 14, 15 and 16, in order of priority numbering in any one of Tables 13, 14, 15 and 16, comprise a range selected from any two previously described upper and / or lower amounts, such as 2 to 300, 10 to 290, 25 to 270, or 50 to 250. For avoidance of doubt, “order of priority numbering” refers to lipids listed in row 1 of a given table being of higher priority than lipids listed from row 2 onwards, for example. The present disclosure also provides a method of calculating an Alzheimer’s risk score. The method may comprise obtaining lipid profile data from a biological sample taken from the subject, which may comprise lipid species according to any aspects, embodiments, or examples thereof as described herein. The method may also comprise standardising or normalising or scaling the numeric values of the lipid profile data against a reference data set. The method may also comprise refining the discriminatory power of one or more lipid species by statistically weighting some of the numeric values associated Atty Dkt No.: RICE-231WO therewith. The method may also comprise summating or similar the numeric values obtained from steps (above) to provide a composite AD risk score. In one example, there is provided a method of a calculating an AD risk score comprising: obtaining lipid profile data from a biological sample taken from the subject; standardising the numeric values of the lipid profile data against a reference data set; refining the discriminatory power of one or more lipid species by statistically weighting some of the numeric values associated therewith; calculating a composite AD risk score. In some examples, the AD risk score can be used to determine the risk of the subject developing AD. In some examples, the AD risk score can be used to determine a treatment plan or method of treating a subject for AD. In some examples, the AD risk score is standardised for age and / or gender of the subject. In some examples, the subject is a mammal. In some examples, the subject is a human. In some examples, the lipid profile data is generated form a liquid chromatography- mass spectrometry (LC-MS) protocol. In some examples, the LC-MS protocol has a total run time of less than about (in minutes) 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 3, 2, or 1. In one example the total run or elution time is less than about 5 minutes. High throughput lipidomics assay Illustrative methods capable of analysing lipid species include classical lipid extraction methods, mass spectrometry together with electrospray ionization and matrix- assisted laser desorption ionisation, with mass analysis such as quadruple and / or TOF (e.g.,. Quadrapole / TOF) or orbitrap mass analysers. Chromatographic methods are used for the separation of lipid mixtures such as gas chromatography, high pressure liquid chromatography (HPLC), ultra-high pressure liquid chromatography (UHPLC), capillary electrophoresis (CE). These may be used with mass spectrometry based detection systems or other detectors including optical detectors. Clinical mass spectrometry systems are used by clinical laboratories to provide lipid profiles and ratios upon request. Another suitable Atty Dkt No.: RICE-231WO technique for quantitative lipid analysis is one or two dimensional nuclear magnetic resonance (NMR). Two dimensional techniques such as heteronuclear single quantum coherence (HSQC) are suitable for lipid profiling through the ability to elucidate C-H bonds within a structure. Any technique capable of identifying individual lipid species in the sample can be used for collecting information on the lipid species. Typically MS is used coupled to a separation method such as various forms of chromatography. In one example, there is provided a high throughput lipidomics assay comprising: obtaining a biological sample separating the plasma sample analytes by liquid chromatographic separation analysing the analytes using a mass spectrometer; wherein the total run time is less than 15 minutes; and wherein the resulting measured data can accurately assign individual analytes to a single lipid species. In one example, at least two solvents are run. In one example, the solvents are A) 50 % H2O / 30 % acetonitrile / 20 % isopropanol (v / v / v) with 10 mM ammonium formate and B) 1 % H2O / 9 % acetonitrile / 90 % isopropanol ( v / v / v ) with 10 mM ammonium formate. In one example, the total run time is less than 10 minutes. In one example, the total run time is less than 8 minutes. In one example, the total run time is less than 7 minutes. In one example, the total run time is not more than 5 minutes. In one example, the biological sample is at least one of plasma, serum, dried blood spots, and dried plasma spots. In one example, the plasma sample is extracted using a single-phase BuOH / MeOH. In one example, the plasma sample is obtained from a subject. Alzheimer’s disease Alzheimer's disease (AD) is a neurodegenerative disease that progressively worsens and is the leading cause of dementia. Although currently available disease- modifying agents are not capable of reversing the initial pathological changes, it may be possible to prevent or delay the development of dementia by early identification or the determination of a risk score that indicates the likelihood that an individual will develop AD. Diagnosing the disease or risk of disease early also provides the individual and their carers have time to plan for the future, and to access to treatments that can help manage Atty Dkt No.: RICE-231WO symptoms. Early intervention is an optimal strategy because the patient’s level of function can be preserved for longer. The present invention provides for method of determining an AD risk score of a subject, the method comprising: (i) detecting in a biological sample from the subject a level of at least two lipid species selected from Table 4 and / or 6; and (ii) comparing the levels of the lipid species detected in the biological sample to reference lipid species levels, wherein the comparison is determinative of the risk of developing AD in the subject. Depending on the stage of progression, which is informed by methods for determining a risk score described herein, the patient may display one or more symptoms associated with AD that are known in the art and / or described as follows. At a clinical level, AD may present a number of cognitive symptoms including mental decline, difficulty thinking and understanding, depression, hallucination, or paranoia, confusion in the evening, delusion, disorientation, forgetfulness, making things up, mental confusion, difficulty concentrating, inability to create new memories, inability to do simple maths, or inability to recognise common things. Behavioural symptoms may also be present and include aggression, agitation, difficulty with self-care, irritability, meaningless repetition of words, personality changes, lack of restraint, or wandering and getting lost. Loss of loss of appetite or restlessness may also be present. Depending on the stage of progression, one or more cognitive tests may be used in addition to the methods of the invention to measure and evaluate cognitive, or ‘thinking’, functions such as memory, concentration, visual-spatial awareness, problem solving, counting and language skills. Particular cognitive tests that may be used include the following: Mini-Mental Status Examination (MMSE) The MMSE is the most common test for the screening of dementia. It assesses skills such as reading, writing, orientation and short-term memory. Alzheimer’s Disease Assessment Scale-Cognitive (ADAS-Cog) This 11-part test is more thorough than the MMSE and can be used for people with mild symptoms. It is considered the best brief examination for memory and language skills. Neuropsychological Testing Atty Dkt No.: RICE-231WO A variety of tests will be used and may include tests of memory such as recall of a paragraph, tests of the ability to copy drawings or figures and tests of reasoning and comprehension. Brain imaging techniques Various brain-imaging techniques are sometimes used to show brain changes and to rule out other conditions such as tumour, infarcts (strokes – dead areas of brain tissue) and hydrocephalus (fluid on the brain); these include: (a) Computed tomography (CT or CAT) scan This technique involves taking many X-rays from different angles in a very short period of time. These images are then used to create a 3-dimensional image of the brain. CT scans are mainly used to rule out other causes of dementia such as stroke, brain tumour, multiple sclerosis or haemorrhage. They can show certain changes that are characteristic of Alzheimer's disease or other causes of dementia. (b) Magnetic Resonance Imaging (MRI) This technique uses powerful magnets and radiowaves to produce very clear 3- dimensional images of the brain. Currently MRI is the radiological test of choice. As well as ruling out treatable causes of dementia, MRI can reveal patterns of brain tissue loss, which can be used to discriminate between different forms of dementia such as AD and frontotemporal dementia. (c) Positron Emission Tomography (PET) and Single-Photon Emission Computerized Tomography (SPECT) In both of these tests, a small amount of radioactive material is injected into the patient and detectors in the scanner detect emissions from the brain. PET provides visual images of activity in the brain. SPECT is used to measure blood flow to various regions of the brain. Once an AD risk score or assessment is completed using the methods of the invention, the patient may be treated with a suitable AD treatment depending on the stage or risk of AD as understood by a skilled person. Treatment as used herein refers to therapeutic treatment and also involves ameliorating a symptom associated with a disease. In the context of AD, therapeutic treatment can be measured by an increase or recovery in any one or more of the group consisting of cognitive function; short term memory; ability to acquire new information; semantic memory; apathy; language, executive or visuoconstructional problems or apraxia; long term memory; irritability and aggression; or exhaustion. Treatment can also be measured via reduction in the presence of pathogenic protein or a reduction in the particular forms of pathogenic protein such as protein Atty Dkt No.: RICE-231WO aggregates or deposits. The presence and reduction of the pathogenic protein that can be visualised or detected by imaging techniques or biochemical techniques known in the art. For example, in relation to AD, treatment may relate to a reduction in a soluble or insoluble isoforms of amyloid beta (Aβ) peptide or a reduction in the number of amyloid beta (Aβ) plaques. Alternatively, the outcome of the treatment may be determined by neuropsychological or cognitive testing. A subject in need of treatment may be one that exhibits impaired memory function, cognitive function or subclinical or clinical symptoms of a neurodegenerative disease. The selection of a subject for treatment may involve a screening step for identifying whether the subject is displaying impaired cognitive function, memory function or a clinical manifestation of a neurodegenerative disease. A subject in need of treatment may be one that is identified as having early, intermediate or late stage disease and in the case of Alzheimer's disease may be identified as having either diffuse Aβ oligomers or plaques. Suitable non-limiting treatments that may be utilised for treatment of AD include medication, including any one or more of cholinesterase inhibitors, Memantine, Aducanumab, Lecanemab and antidepressants; and / or lifestyle changes, including any one or more of changes to diet, changes to exercise, quitting or reducing smoking and undertaking cognitive activities. A patient’s responsiveness to treatment may be determine using methods known in the art. Improved memory may be determined by memory tests, typically a test administered by a clinical professional. Standardised neuropsychological tests of cognition that could be administered to test the effectiveness of the treatment include any of the following tests or one or more of its components: Neuropsychological Test Battery, Alzheimer's Disease Assessment Scale-cognitive sub scale (ADAS-cog), Mini-Mental State Examination, Severe Impairment Battery, Disability Assessment Scale for Dementia, Clinical Dementia Rating Scale Sum of Boxes, Alzheimer's Disease Cooperative Study Clinical Global Impression of Change, Wechsler Memory Scale Visual Immediate, Wechsler Memory Scale Verbal Immediate, Rey Auditory Verbal Learning Test, Wechsler Memory Digit Span, Controlled Word Association Test, Category Fluency Test, Wechsler Memory Scale Visual Delayed, Wechsler Memory Scale Verbal Delayed, Rey Auditory Verbal Learning Test, Wechsler Memory Scale, Stroop Task, Wisconsin Card Sorting Task, Trail Making Test, or any other tests of memory and executive function alone or in combination. Atty Dkt No.: RICE-231WO EXAMPLES Example 1: High Throughput Lipidomic Assay A chromatographic gradient was devised to achieve the best possible separation of lipid species while maintaining a total runtime no greater than 5 minutes. Lipid Extraction Aliquots of 10 uL plasma were mixed with 100 uL of butanol:methanol, 1:1, containing 10 mM ammonium formate, and a mixture of internal standards at known concentrations. Samples were vortexed thoroughly, bath sonicated for 1 hour at 25 °C, then centrifuged at 14,000 rpm for 10 min at 20 °C. A 90 uL aliquot was removed for analysis discarding any precipitate. Internal Standard Addition To calculate lipid concentration within the sample, a mixture of internal standards is added to the extraction solution, such that it is included in each sample at known concentrations. Internal standards consist of stable-isotope-labelled and non-physiological species, with a total of 30 used. A mixture of these standards is prepared in-house from commercially available sources. Table 1: Internal Standards Compound pmol / sample (10 uL) Atty Dkt No.: RICE-231WO Compound pmol / sample (10 uL) LPC(18:1) d7 (IS) 100 Data analysis Peak area was calculated using Agilent’s MassHunter Quant QQQ software package, via integration of the chromatogram. Lipid concentration was calculated using the ratio of peak area between analyte and assigned internal standard, along with the internal standard’s known concentration. Research Protocol LC-MS / MS Extracted plasma samples were separated and analysed using an Agilent 1290 Infinity liquid chromatography (LC) system coupled to an Agilent 6495C QQQ mass spectrometer equipped with Jet Stream ionisation source. The LC system utilised a ZORBAX eclipse plus C18 column (2.1x100mm 1.8mm, Agilent), with the thermostat set at 45°C. The running solvents consisted of: A) 50 % H2O / 30 % acetonitrile / 20 % isopropanol (v / v / v) with 10 mM ammonium formate and B) 1 % H2O / 9 % acetonitrile / 90 % isopropanol (v / v / v) with 10 mM ammonium formate. A stepped, linear solvent gradient was used, detailed in Figure 1. Atty Dkt No.: RICE-231WO The gradient started with a flow rate of 0.4ml / minute at 10% B and increased to 45% B over 2.7 minutes, then to 53% over 0.1 minutes, to 65% over 6.2 minutes, to 89% over 0.1 minute, to 92% over 1.9 minutes and finally to 100% over 0.1 minute. The solvent was then held at 100% B for 0.8 minutes (total 11.9 minutes). For equilibration the solvent was decreased from 100% B to 10% B over 0.1 minute and held for an additional 0.9 minutes. Flow rate was then switched to 0.6 ml / minute for 1 minute before returning to 0.4 ml / minute over 0.1 minutes. Solvent B was held at 10% B for a further 0.9 minutes at 0.4ml / minutes for a total cycle time of 15 minutes. Mass spectrometer conditions used were: gas temperature, 150 C; gas flow rate 17 L / min; nebulizer 20 psi; sheath gas temperature 200 C; capillary voltage 3500 V; and sheath gas flow 10 L / min. Clinical Protocol LC-MS / MS Additional modifications were made to solvent flow rate, column temperature, column length and packing material, and MS source conditions. The gradient was shortened to 5 minutes, and the length and packing material of the C18 column was modified. The result was a chromatographic gradient lasting 5 minutes in total, consisting of 5 steps of variable length, and which achieved good separation across elution bands of high abundance analytes. Extracted plasma samples were separated and analysed using an Agilent 1290 Infinity II LC system coupled to an Agilent Ultivo QQQ mass spectrometer equipped with Jet Stream ionisation source. The LC system utilised an Agilent Poroshell 1290 EC-C18 column (2.1x30mm 1.9mm, Agilent), with the thermostat set at 40°C. The running solvents consisted of: A) 50 % H2O / 30 % acetonitrile / 20 % isopropanol (v / v / v) with 10 mM ammonium formate and B) 1 % H2O / 9 % acetonitrile / 90 % isopropanol (v / v / v) with 10 mM ammonium formate. A stepped, linear solvent gradient was used, detailed in Figure 2. At 0 min, the gradient flow rate was initiated at 1 ml / minute, 0% B, maintained until 0.3 min, when B was increased to 40% over 0.1 min. B was further increased linearly to 48 % B at 2.5 min, then stepped to 58 % B at 2.51 min. B was increased linearly to 60 % at 3.1 min, then stepped to 76 % B at 3.11, before increasing linearly to 80 % B at 4.4 min. B was increased at 4.41 minutes to 100 %, maintained at 100% until 5 min, then immediately dropped to 0% B at 5.01 min, for equilibration at 0 % B for 1 minute until 6 min. Total cycle time was 6 minutes; total analysis cycle time was 5 minutes. Mass spectrometer conditions used were: gas temperature 200 C; gas flow rate 9 L / min; Atty Dkt No.: RICE-231WO nebulizer 20 psi; sheath gas temperature 250 C; capillary voltage 5500 V; and sheath gas flow 10 L / min. Good separation of high abundance analytes was achieved for many lipid species. Where two or more dissimilar, low correlation lipid structures displayed the same transition measurement in the spectrometer, these species were removed from analysis, as the data regarding their measured concentration could not be accurately assigned to a single lipid species. Where the concentration of structurally similar co-eluting species (for example, isomeric and isobaric species) was observed to be highly correlated in human plasma, the species remained suitable for transition measurement despite containing multiple isomeric compounds. Such was the case, for example, for many sn1 / sn2 positional isomers of glycerophospholipid structures. Example 2: Development of Clinical Lipidomics Platform (CLP) List: To develop the Clinical Lipidomics Platform 1 List (Table 4), a series of 791 mass transitions representing different lipid species were measured by mass spectrometer. These transitions were previously determined to be accurately quantifiable via the use of the Research Protocol LC-MS / MS totalling 15 minutes in length (described above). A series of experiments determined which of these species remain accurately quantifiable when the Clinical Protocol LC-MS / MS totalling 5 minutes in length (described above) is run to support high sample throughput. Analysed species (Table 2) which do not fit the criteria for accurate measurement following modifications to the chromatography conditions are discarded from the CLP1 List (Table 4). A summary of this process is provided below. Peak Overlap A series of identical plasma extracts pooled from multiple individuals were analysed using the Research Protocol LC-MS / MS, then again using the Clinical Protocol LC-MS / MS Conditions. In both cases, lipids from the Research Lipid List (Table 2) were analysed.30 internal standards were also measured (Table 3). Table 2: Research Lipid List Lipid Precursor Product Atty Dkt No.: RICE-231WO Lipid Precursor Product Lipid Class Compound Name Number (m / z) (m / z) Atty Dkt No.: RICE-231WO Lipid Precursor Product Lipid Class Compound Name Number (m / z) (m / z) Atty Dkt No.: RICE-231WO Lipid Precursor Product Lipid Class Compound Name Number (m / z) (m / z) Atty Dkt No.: RICE-231WO Lipid Precursor Product Lipid Class Compound Name Number (m / z) (m / z) Atty Dkt No.: RICE-231WO Lipid Precursor Product Lipid Class Compound Name Number (m / z) (m / z) Atty Dkt No.: RICE-231WO Lipid Precursor Product Lipid Class Compound Name Number (m / z) (m / z) Atty Dkt No.: RICE-231WO Lipid Precursor Product Lipid Class Compound Name Number (m / z) (m / z) Atty Dkt No.: RICE-231WO Lipid Precursor Product Lipid Class Compound Name Number (m / z) (m / z) Atty Dkt No.: RICE-231WO Lipid Precursor Product Lipid Class Compound Name Number (m / z) (m / z) Atty Dkt No.: RICE-231WO Lipid Precursor Product Lipid Class Compound Name Number (m / z) (m / z) Atty Dkt No.: RICE-231WO Lipid Precursor Product Lipid Class Compound Name Number (m / z) (m / z) Atty Dkt No.: RICE-231WO Lipid Precursor Product Lipid Class Compound Name Number (m / z) (m / z) Atty Dkt No.: RICE-231WO Lipid Precursor Product Lipid Class Compound Name Number (m / z) (m / z) Atty Dkt No.: RICE-231WO Lipid Precursor Product Lipid Class Compound Name Number (m / z) (m / z) Atty Dkt No.: RICE-231WO Lipid Precursor Product Lipid Class Compound Name Number (m / z) (m / z) Atty Dkt No.: RICE-231WO Lipid Precursor Product Lipid Class Compound Name Number (m / z) (m / z) Atty Dkt No.: RICE-231WO Lipid Precursor Product Lipid Class Compound Name Number (m / z) (m / z) Atty Dkt No.: RICE-231WO Lipid Precursor Product Lipid Class Compound Name Number (m / z) (m / z) Atty Dkt No.: RICE-231WO Lipid Precursor Product Lipid Class Compound Name Number (m / z) (m / z) Atty Dkt No.: RICE-231WO Lipid Precursor Product Lipid Class Compound Name Number (m / z) (m / z) Atty Dkt No.: RICE-231WO Lipid Precursor Product Lipid Class Compound Name Number (m / z) (m / z) Atty Dkt No.: RICE-231WO Lipid Precursor Product Lipid Class Compound Name Number (m / z) (m / z) Atty Dkt No.: RICE-231WO Lipid Precursor Product Lipid Class Compound Name Number (m / z) (m / z) Atty Dkt No.: RICE-231WO Lipid Precursor Product Lipid Class Compound Name Number (m / z) (m / z) Table 3: Internal standards Atty Dkt No.: RICE-231WO Lipid Precursor Product Compound Name Number (m / z) (m / z) Comparison of the resulting extracted ion chromatograms was used to determine whether the peak of interest overlapped with any additional peaks after reducing the length Atty Dkt No.: RICE-231WO of the chromatographic separation to 5 minutes for all 791 monitored transitions. A binary value was thus assigned to each lipid species (peak overlap = y / n). Where peak overlap occurred between the peaks of highly correlated lipid species, such as the sn1 and sn2 positional isomers of glycerophospholipids, but no other species, these compounds were combined into a single compound and carried through to the next step of analysis. Correlation was determined through the use of a correlation matrix calculated using previously generated data resulting from a large cohort study. Additional Peak Parameters: A series of 24 identical plasma extracts pooled from multiple individuals were analysed using the Clinical Protocol LC-MS / MS Conditions. The resulting dataset was used to determine the Mean Peak Area (the integrated chromatogram peak area for the target analyte) and the Measurement CV% (the peak area coefficient of variation across all replicate measurements) for each of the monitored transitions Lipid List Refinement The following steps were undertaken to refine the lipid list: 1) From the list of 791 lipid transitions analysed, peaks determined to have merged with a neighbouring peak upon reduction in HPLC runtime were eliminated from further analysis unless they were highly correlated compounds, in which case they were combined and carried through to the next step. 2) Peaks that displayed a mean peak area <250 counts were eliminated from further analysis. 3) Peaks that displayed peak area CV% > 20% were eliminated from further analysis. 4) Some minor additions were made to ensure good representation of all lipid classes. This process of elimination resulted in the reduction of the 791 lipid species measured in the Research Lipid List (Table 2) to a list of 339 analytes (CLP1 list; Table 4) and 30 internal standards (Table 5). In order to utilise previously generated large cohort datasets, in the following examples a subset of CLP1 is sometimes used, termed CLP1s, which contains 292 of the above species (indicated ‘y’ in Table 4) which map directly on to older datasets. Table 4: CLP1 List Atty Dkt No.: RICE-231WO Lipid Precursor Product Lipid Class Compound Name CLP1s Number (m / z) (m / z) Atty Dkt No.: RICE-231WO Lipid Precursor Product Lipid Class Compound Name CLP1s Number (m / z) (m / z) Atty Dkt No.: RICE-231WO Lipid Precursor Product Lipid Class Compound Name CLP1s Number (m / z) (m / z) Atty Dkt No.: RICE-231WO Lipid Precursor Product Lipid Class Compound Name CLP1s Number (m / z) (m / z) Atty Dkt No.: RICE-231WO Lipid Precursor Product Lipid Class Compound Name CLP1s Number (m / z) (m / z) Atty Dkt No.: RICE-231WO Lipid Precursor Product Lipid Class Compound Name CLP1s Number (m / z) (m / z) Atty Dkt No.: RICE-231WO Lipid Precursor Product Lipid Class Compound Name CLP1s Number (m / z) (m / z) Atty Dkt No.: RICE-231WO Lipid Precursor Product Lipid Class Compound Name CLP1s Number (m / z) (m / z) Atty Dkt No.: RICE-231WO Lipid Precursor Product Lipid Class Compound Name CLP1s Number (m / z) (m / z) Atty Dkt No.: RICE-231WO Lipid Precursor Product Lipid Class Compound Name CLP1s Number (m / z) (m / z) Atty Dkt No.: RICE-231WO Lipid Precursor Product Lipid Class Compound Name CLP1s Number (m / z) (m / z) Table 5: Internal standards Lipid Precursor Product Compound Name Atty Dkt No.: RICE-231WO Lipid Precursor Product Compound Name Number (m / z) (m / z) Example 3: Further Refinement to Generate Clinical Lipidomics Platform 2 (CLP2) List To produce a list of lipid species in which the resulting chromatographic peaks of interest could be most readily integrated via software automation without the need for manual intervention, multiple reference peaks, or complex peak picking algorithms, the CLP1 list was further refined such that any lipid species for which the chromatographic peak of interest eluted within a 0.5 minute retention time window of any other chromatographic peaks sharing the same transition, were omitted from the CLP2 list (Table 6). This was achieved using the same dataset as above. The resulting CLP2 list contains 269 lipid species (Table 6) excluding 30 internal standards (Table 5). Table 6: CLP2 Lipid List Lipid Precursor Product Li id Cl C m nd N m Atty Dkt No.: RICE-231WO Lipid Precursor Product Lipid Class Compound Name Number (m / z) (m / z) Atty Dkt No.: RICE-231WO Lipid Precursor Product Lipid Class Compound Name Number (m / z) (m / z) Atty Dkt No.: RICE-231WO Lipid Precursor Product Lipid Class Compound Name Number (m / z) (m / z) Atty Dkt No.: RICE-231WO Lipid Precursor Product Lipid Class Compound Name Number (m / z) (m / z) Atty Dkt No.: RICE-231WO Lipid Precursor Product Lipid Class Compound Name Number (m / z) (m / z) Atty Dkt No.: RICE-231WO Lipid Precursor Product Lipid Class Compound Name Number (m / z) (m / z) Atty Dkt No.: RICE-231WO Lipid Precursor Product Lipid Class Compound Name Number (m / z) (m / z) Atty Dkt No.: RICE-231WO Lipid Precursor Product Lipid Class Compound Name Number (m / z) (m / z) Atty Dkt No.: RICE-231WO Example 4: Development of Alzheimer’s disease (AD) risk score Statistical modelling was used to condense large volumes of lipidomic data into a single risk score, an Alzheimer’s disease (AD) score (Figure 3), which can be easily interpreted by clinician and patient alike. Statistical modelling is used to predict an individual’s AD risk from their lipid profile, with some additional transformation. The development and subsequent calculation of an AD risk score was made using a large longitudinal clinical cohort: the Alzheimer's Disease Neuroimaging Initiative (ADNI; n=1,517) (Figure 3A). In this ADNI study, there were a total of 4,730 longitudinal plasma samples from 1,517 participants of ADNI -1, -GO and -2 cohorts examined from baseline up to the 13th time point (10-year follow up period), with three major time points of baseline, 12 months, and 24 months that include the largest number of individuals (Figure 3A). Values were compared and calculated utilising all measured lipids (the Research Lipid List, Table 2), or a selection of lipids suited to high throughput measurement (the CLP1s, 324 species, Table 4 ‘y’). The calculation of these risk scores was further demonstrated using a second, truncated list featuring only the lipids which can be easily integrated without the use of peak picking algorithms (CLP2s, 258 species, Table 6 ‘y’). This list was created to support automated data processing. The normalised lipidomic data was utilised on AD subjects (n=243) and cognitive normal individuals (CN; n=408) at baseline to build the classification model. Ridge and Lasso regression models were created to stratify AD from CN, optimising the C-statistic using the R package ‘glmnet v4.1-4’. To avoid over-fitting, five models were created from an external 5-fold cross-validation framework (Figure 1B). In detail, the dataset was evenly separated into five groups. The models were trained on the 4 / 5th of the cohort, then used to predict AD status in the remaining 1 / 5th. This process was repeated five times. These models were adjusted for age, sex, BMI, APOE ε4 status, HDL-C, total cholesterol, triglycerides, fasting status, cohort (a categorical variable indicating ADNI 1, GO, and 2 phases), omega-3 status, and statin status. The validation of the derived AD risk scores using the full lipidome (n=749 lipid species) and the CLP lipid lists (CLP1s and CLP2s) was performed across three major time points in ADNI. The models from Ridge and Lasso were separately used to stratify AD from CN across baseline, 12 months and 24 months. Further, the models were also used to predict incident (diagnosed during the 10-year follow-up period) AD (Figure 3C). Atty Dkt No.: RICE-231WO The predictive models of AD risk The models are described as a list of the covariates and lipids used in each model together with their beta-coefficients (Tables 9 to 12). The lipid species are ranked by the absolute value of their beta-coefficients which defined the relative contribution they made to the model. Lasso models with the CLP1s and CLP2s lipid lists were developed and the number of lipids and performance of these models was compared (Tables 9 to 12). It is recognised that the lipids at the top of these lists will contribute more to the model and so a subset of the lipids could be selected from the top 20 lipids, the top 50 lipids, the top 100 lipids, the top 200 lipids and used to make a model containing fewer lipids. Performance of AD risk models using the Research Lipid List and the CLP1s and CLP2s Lists to classify prevalent AD from CN in ADNI at baseline, 12 months and 24 months The risk models were temporally validated across different time points. To assess the discrimination ability of the models, the metric of the “receiver operating characteristic area under the curve” (ROC-AUC) was employed. Using the baseline data, the models discriminated prevalent AD from CN with AUC being 87.6%, 81.7%, and 80.7% when either the Research Lipid List, CLP1s, or CLP2s lists were used, respectively (Table 7). AUC is an effective and combined measure of sensitivity and specificity that describes the inherent validity of the models. The results demonstrated that the model on the Research Lipid List performed slightly better than the models on CLP1 and CLP2. Similar results were observed consistently across 12 months and 24 months. Lasso models with either the Research Lipid List, CLP1s, or CLP2s lists were also developed and the number of lipids and performance of these models was compared (Table 7). Consistent with the findings in ridge model, the Lasso model from different lipidomic sets displayed similar discrimination ability in classifying AD from CN, with AUC=80.8% for full Research Lipid List, 77.2% for CLP1 and 76.1% for CLP2 at baseline. Similar findings were observed for 12 month and 24 month data. We note that 71 lipid species out of 749 (full lipidome), 50 out of 324 (CLP1), and 58 out of 258 (CLP2) lipid species were selected in the Lasso models. Therefore, when fewer lipid species are selected for use in the model, a similar level of discrimination can be maintained as when more lipids are used. Table 7: The predictive performances of ridge and Lasso model to classify AD from CN (AUC) Atty Dkt No.: RICE-231WO AUC (%) Lipid set Baseline 12 months 24 months Lipid List (ridge, n=749 species) 87.6 79.4 85.1 CLP1s List (ridge, n=324 species) 81.7 75.0 81.0 CLP2s List (ridge, n=258 species) 80.7 74.4 80.5 Research Lipid List (Lasso, n=71 species) 80.8 73.5 80.8 CLP1s List (Lasso, n=50 species) 77.2 71.2 76.3 CLP2s List (Lasso, n=58 species) 76.1 70.1 75.4 Temporal validation of AD-CN scores using the Research Lipid List and the CLP1s and CLP2s Lists to discriminate incident AD from the combination of CN and MCI The risk models derived from different lipid sets were further validated for their ability to predict incident AD (327 cases presented between 6 months and 10 years; average time to presentation was 3.0 years) using either the baseline, 12 months or 24 months lipidomics data. Table 8 demonstrates that the models using either of the machine learning methods (Ridge / Lasso) and each of the different lipid panels (Research Lipid List, CLP1 and CLP2) achieve a similar level of discrimination. For instance, the model built using the Research Lipid List could distinguish incident AD from CN and MCI using the baseline lipidomic data with the AUC=76.3%, which is only 2% higher in AUC that the performance of the models built using CLP1 and CLP2. Table 8: The predictive performances of ridge and Lasso model to classify incident AD from the combination of CN and MCI (AUC) AUC (%) Lipid set Baseline 12 months 24 months Research Lipid List (ridge, n=749 species) 76.3 72.5 76.8 Atty Dkt No.: RICE-231WO AUC (%) Lipid set Baseline 12 months 24 months (ridge, n=324 species) 73.8 71.8 74.7 CLP2s List (ridge, n=258 species) 73.8 71.1 74.3 Research Lipid List (Lasso, n=71 species) 75.7 73.3 77.4 CLP1s List (Lasso, n=50 species) 74.1 73.0 75.4 CLP2s List (Lasso, n=58 species) 73.7 71.6 74.5 Conclusions In summary, the AD risk model derived from clinical lipid panels (CLP1 and CLP2) showed a similar ability to predict both prevalent and incident AD as the model developed from the Research Lipid List. Additionally, the model built using Lasso regression shows that when utilising only a limited number of lipids or subset of the Research Lipid List, CLP1 and CLP2 lists, the model retains a reasonable prediction ability for future AD risk. These models have utility in clinical risk assessment and management of individuals at the early stage of AD risk, as well as for AD diagnosis.

[0003] Atty Dkt No.: RICE-231WO Table 9: Predictive models of AD risk developed using Ridge regression incorporating all the covariates together with the CLP1s lipids Ridge Rank # Lipid # Lipid / Covariate Ridge (ABS) Atty Dkt No.: RICE-231WO Ridge Rank # Lipid # Lipid / Covariate Ridge (ABS) Atty Dkt No.: RICE-231WO Ridge Rank # Lipid # Lipid / Covariate Ridge (ABS) Atty Dkt No.: RICE-231WO Ridge Rank # Lipid # Lipid / Covariate Ridge (ABS) Atty Dkt No.: RICE-231WO Ridge Rank # Lipid # Lipid / Covariate Ridge (ABS) Atty Dkt No.: RICE-231WO Ridge Rank # Lipid # Lipid / Covariate Ridge (ABS) Atty Dkt No.: RICE-231WO Ridge Rank # Lipid # Lipid / Covariate Ridge (ABS) Atty Dkt No.: RICE-231WO Ridge Rank # Lipid # Lipid / Covariate Ridge (ABS) Atty Dkt No.: RICE-231WO Ridge Rank # Lipid # Lipid / Covariate Ridge (ABS) Atty Dkt No.: RICE-231WO Ridge Rank # Lipid # Lipid / Covariate Ridge (ABS) Atty Dkt No.: RICE-231WO Ridge Rank # Lipid # Lipid / Covariate Ridge (ABS) Table 10: Predictive models of AD risk developed using LASSO regression incorporating all the covariates together with the CLP1s lipids LASSO Rank # Lipid # Lipid / Covariate LASSO Atty Dkt No.: RICE-231WO LASSO Rank # Lipid # Lipid / Covariate LASSO (ABS) Atty Dkt No.: RICE-231WO LASSO Rank # Lipid # Lipid / Covariate LASSO (ABS) 36 43 Cer(d18:1 / 24:1) 0007 0007 Atty Dkt No.: RICE-231WO LASSO Rank # Lipid # Lipid / Covariate LASSO (ABS) 69 22 CE(17:0) 0000 0000 Atty Dkt No.: RICE-231WO LASSO Rank # Lipid # Lipid / Covariate LASSO (ABS) Atty Dkt No.: RICE-231WO LASSO Rank # Lipid # Lipid / Covariate LASSO (ABS) Atty Dkt No.: RICE-231WO LASSO Rank # Lipid # Lipid / Covariate LASSO (ABS) Atty Dkt No.: RICE-231WO LASSO Rank # Lipid # Lipid / Covariate LASSO (ABS) Atty Dkt No.: RICE-231WO LASSO Rank # Lipid # Lipid / Covariate LASSO (ABS) 232 219 SM(34:3) 0000 0000 Atty Dkt No.: RICE-231WO LASSO Rank # Lipid # Lipid / Covariate LASSO (ABS) Atty Dkt No.: RICE-231WO LASSO Rank # Lipid # Lipid / Covariate LASSO (ABS) Table 11: Predictive models of AD risk developed using Ridge regression incorporating all the covariates together with the CLP2s lipids. Atty Dkt No.: RICE-231WO Ridge Rank # Lipid # Lipid / Covariate Ridge (ABS) Atty Dkt No.: RICE-231WO Ridge Rank # Lipid # Lipid / Covariate Ridge (ABS) Atty Dkt No.: RICE-231WO Ridge Rank # Lipid # Lipid / Covariate Ridge (ABS) Atty Dkt No.: RICE-231WO Ridge Rank # Lipid # Lipid / Covariate Ridge (ABS) Atty Dkt No.: RICE-231WO Ridge Rank # Lipid # Lipid / Covariate Ridge (ABS) Atty Dkt No.: RICE-231WO Ridge Rank # Lipid # Lipid / Covariate Ridge (ABS) Atty Dkt No.: RICE-231WO Ridge Rank # Lipid # Lipid / Covariate Ridge (ABS) Atty Dkt No.: RICE-231WO Ridge Rank # Lipid # Lipid / Covariate Ridge (ABS) Atty Dkt No.: RICE-231WO Ridge Rank # Lipid # Lipid / Covariate Ridge (ABS) Table 12: Predictive models of AD risk developed using LASSO regression incorporating all the covariates together with the CLP2s lipids. LASSO Rank # Lipid # Lipid / Covariate LASSO Atty Dkt No.: RICE-231WO LASSO Rank # Lipid # Lipid / Covariate LASSO (ABS) Atty Dkt No.: RICE-231WO LASSO Rank # Lipid # Lipid / Covariate LASSO (ABS) Atty Dkt No.: RICE-231WO LASSO Rank # Lipid # Lipid / Covariate LASSO (ABS) 71 18 CE(16:0) 0000 0000 Atty Dkt No.: RICE-231WO LASSO Rank # Lipid # Lipid / Covariate LASSO (ABS) Atty Dkt No.: RICE-231WO LASSO Rank # Lipid # Lipid / Covariate LASSO (ABS) Atty Dkt No.: RICE-231WO LASSO Rank # Lipid # Lipid / Covariate LASSO (ABS) Atty Dkt No.: RICE-231WO LASSO Rank # Lipid # Lipid / Covariate LASSO (ABS) 201 185 SM(d18:1 / 16:0) 0000 0000 Atty Dkt No.: RICE-231WO LASSO Rank # Lipid # Lipid / Covariate LASSO (ABS) Example 5: Evaluation of the reclassification ability of the lipidome based Alzheimer’s disease risk score In Example 4, the Receiver Operating Characteristic - Area Under Curve (AUC) was employed as an evaluation metric to access the discrimination ability of the models. This section introduces the Net Reclassification Improvement (NRI) metric and Integrated Atty Dkt No.: RICE-231WO Discrimination Improvement (IDI) to evaluate the reclassification ability of the lipidome- based AD risk score (described in Example 4), compared to traditional risk scores derived from factors such as age, sex, BMI, APOE4, statins, and Omega3. The development and subsequent calculation of an AD risk score was based on AD subjects (n=243) and cognitively normal individuals (CN; n=408) within Alzheimer's Disease Neuroimaging Initiative (ADNI) baseline data (n=1,393). To assess the impact of incorporating lipidomic data into the AD risk scores, two models were derived: a basic model and a complex model. The basic model included traditional risk factors (age, sex, BMI, APOE4, statins, and Omega3) for predicting prevalent AD. In contrast, the complex model, detailed in Example 4, utilised all measured lipids (as listed in Table 2) in addition to the traditional risk factors. The reclassification ability of lipidome based AD risk score To critically evaluate the effectiveness of the lipidome based AD risk score in re- stratifying incident AD from MCI individuals, we analysed both categorical and continuous NRI based on the raw counts of the reclassifications, which have direct clinical implications. Relative to the basic model, the lipidome based AD risk model was able to reclassify individuals: at the cut-off point of 0.5, 31% of all AD cases initially labelled as low-risk were moved up to high-risk (as shown in Table 13), while 5% of all AD cases initially labelled as high-risk were moved down to low-risk. By incorporating both upward (to higher risk categories) and downward (to lower risk categories) reclassifications, the lipidome-based AD risk model provided a net reclassification improvement of 26% for incident AD cases (the proportions of cases moving up minus those moving down). Overall, the lipidome based AD risk score yielded a category NRI of 0.14 (95% CI:0.07- 0.21, p=8.0x10-5) which included both reclassification of cases and controls, at the cut-off point of 0.5, and a continuous NRI of 0.55 (95% CI: 0.41-0.69, p<1.0x10-5). Furthermore, there was an increase in the risk differences between the lipidome based AD risk score and basic score with an Integrated Discrimination Index (IDI) of 0.11 (95% CI: 0.08-0.14, p<1.0x10-5). Table 13: The reclassification performances of lipidome-based AD risk model to classify incident AD from MCI (NRI and IDI). Raw counts and NRI stats Events Non-events n 327 413 Downward reclassification n 16 13 Unchanged classification n 209 337 Upward reclassification n 102 63 Proportion upward 0.31 0.15 Proportion downward 0.05 0.03 NRI (events or non-events) 0.26 -0.12 Total NRI P value Atty Dkt No.: RICE-231WO Category NRI 0.14 [0.07 - 0.21] 8.0x10-5Continuous NRI 0.55 [0.41 - 0.69] <1.0x10-5IDI 0.11 [0.08 - 0.14] <1.0 x10-5Conclusion In summary, the NRI metric efficiently demonstrated the addition of the lipidome data improved classification of AD risk, and was especially useful for classification and reclassification of AD individuals. The IDI metric demonstrated an improvement in the discrimination slope supporting the enhanced risk differentiation between events and non-events as a result of the addition of the lipidome data. Example 6: Development of ADAS-Cog risk score Following a similar approach to the development of AD risk scores described in Example 4, statistical modelling was used to condense a large volume of lipidomic data into a single risk score: an ADAS-Cog score. Unlike the classification model used for the AD risk score, this model predicts an individual’s Alzheimer's Disease Assessment Scale- Cognitive Subscale (ADAS-Cog) score from their lipid profile, incorporating some additional transformations. The ADAS-Cog score was developed using a large longitudinal clinical cohort: the Alzheimer's Disease Neuroimaging Initiative (ADNI; n=1,517). In this ADNI study, there were a total of 4,730 longitudinal plasma samples from 1,517 participants of ADNI -1, - GO and -2 cohorts examined from baseline up to the 13th time point (10-year follow up period). Values were compared and calculated utilising all measured lipids (the Research Lipid List, Table 2), or a selection of lipids suited to high throughput measurement (the CLP2s, 258 species, Table 6 ‘y’). Normalised lipidomic data from ADNI baseline data (n=1,393) was used to construct the ADAS-Cog risk score. Penalised ridge regression models were employed, optimising the Root Mean Square Error (RMSE) using the R package ‘glmnet v4.1-4’. To avoid over-fitting, five models were created from an external 5-fold cross-validation framework. In detail, the dataset was evenly separated into five groups. The models were trained on the 4 / 5th of the cohort, then used to predict ADAS-Cog score in the remaining 1 / 5th. This process was repeated five times. These models were adjusted for age, sex, BMI, APOE ε4 status, HDL-C, total cholesterol, triglycerides, fasting status, cohort (a categorical variable indicating ADNI 1, GO, and 2 phases), omega-3 status, and statin status. The derived models were termed as ‘complex model’. In parellel, ‘basic models’ were derived using a similar approach but did not include the lipidomic measures as predictors. The ADAS-Cog risk scores, using the full lipidome (n=749 lipid species) and the CLP2 (n=258 lipid species), were validated at the ADNI baseline. The derived scores were then utilised to predict both prevalent AD and incident AD diagnosed during the 10-year follow-up period. Performance of AD risk models using the Research Lipid List and the CLP2 List to classify prevalent AD from CN in ADNI at baseline Atty Dkt No.: RICE-231WO To assess the discrimination ability of the models, the metric of the “receiver operating characteristic area under the curve” (ROC-AUC) was employed. In addition, the Net Reclassification Improvement (NRI) metric was applied to evaluate the improved reclassification ability of the lipidome-based ADAS-Cog score, compared to the basic model-derived score. The basic model was able to differentiate prevalent AD from CN with an AUC of 77.7%. In comparison, the complex model exhibited a significantly improved discrimination ability, with AUCs of 89.9% and 84.8%, when using either the research lipid panel (n=749) or the CLP2 (n=258), respectively (Figure 14; Table 14). Additionally, to access the reclassification ability of the complex models, the metric of categorical NRI was computed using a cut-off point of 0.5. Relative to the basic model, the ADAS-Cog scores utilising the full research lipid panel resulted in a categorical NRI of 0.23 (95% CI:0.15-0.30, P<1.0x10-5) (Table 14). However, the ADAS-Cog scores derived from the CLP2 lipid panel demonstrated a reduced reclassification ability, with a categorical NRI of 0.09 (95% CI:0.02-0.17, p=1.5x10-2). Table 14: The predictive performances of ADAS-Cog score to classify AD from CN. AUC(%) Categorical NRI Lipid set Basic Complex Event Non-Event Total NRI Total NRI model model (Estimate) (Estimate) (Estimate) (P value) Research 0.235 CLP2s List 77.7 84.8 0.09-2(n=258) 0.10 -0.01 [0.02 – 0.17] 1.5x10 Validation of ADAS-Cog risk scores using the Research Lipid List and the CLP2s Lists to discriminate incident AD from the combination of CN and MCI Two metrics, the area under the curve (AUC) and the categorical Net Reclassification Improvement (NRI), were employed to evaluate the models’ performance in stratifying incident AD (AD, n=327) from the combined group of cognitively normal (CN, n=408) and Mild Cognitive Impairment (MCI, n=413) individuals. The basic model distinguished incident AD from the CN and MCI combination with an AUC of 72.1%. In contrast, the complex model exhibited a significantly improved discrimination ability, with AUCs of 76.2% and 74.0% for the research lipid panel and the CLP2, respectively (Figure 5; Table 15). Additionally, to access the reclassification ability of the complex models, the metric of categorical NRI was calculated using the cut-off point of 0.5. Relative to the basic model, the ADAS-Cog scores utilising the full research lipid panel yielded a categorical NRI of 0.1 (95% CI:0.04-0.16, p=9.6x10-4) (Table 15). The ADAS-Cog scores based on CLP2 lipid panel showed a diminished reclassification ability with a categorical NRI of 0.08 (95% CI:0.03-0.13, p=2.5x10-3). Table 15: The predictive performances of ADAS-Cog score to classify incident AD from the combination of CN and MCI. Atty Dkt No.: RICE-231WO AUC(%) Categorical NRI Research Lipid List 72.1 76.2 0.12 -0.02 0.10-5749) [ 9.6x10 (n= 0.04 – 0.16] CLP2s List 72.1 74 0.08-3(n=258) .0 0.09 -0.01 [0.03 -0.13] 2.5x10 traditional risk factors, demonstrated enhanced performance in both discrimination and reclassification abilities for predicting prevalent and incident AD, in comparison with the model that employs traditional risk factors alone. Additionally, the model built on CLP2 list retained a substantial predictive ability for future AD risk. These models are useful in the clinical risk assessment, monitoring, and management of individuals in the early stages of AD risk, as well as for the diagnosis of AD. REFERENCES 1. Fahy et al.2005. J. Lipid Res.46: 839–861 2. Fahy et al 2009. J. Lipid Res.50 (Suppl.): S9–S14 3. Liebisch et al.2013. J. Lipid Res.54: 1523–1530 4. Liebisch et al.2020. J Lipid Res.2020 Dec; 61(12): 1539–1555 5. ADNI dataset http: / / adni.loni.usc.edu /

Claims

Atty Dkt No.: RICE-231WO CLAIMS 1. A method of determining an Alzheimer’s disease (AD) risk score of a subject, the method comprising: (i) detecting in a biological sample from the subject a level of at least two lipid species selected from Table 4 and / or 6; and (ii) comparing the levels of the lipid species detected in the biological sample to reference lipid species levels to generate a risk score, wherein the risk score is determinative of the risk of developing AD in the subject.

2. The method according to claim 1 further comprising: (i) standardising each continuous variable; (ii) log transforming each lipid concentration; and (iii) further standardising the variables prior to their use as the predictors of AD.

3. The method according to claim 1 or 2, wherein the AD risk score is used to determine risk of developing AD in the subject at a time point between about 6 months and about 10 years before the onset of one or more clinical signs or symptoms of AD.

4. The method according to any one of claims 1 to 3, wherein the AD risk score is used to determine risk of developing AD in the subject between about 6 months and about 10 years after the biological sample is obtained from the subject.

5. A method of treating a subject at risk of developing Alzheimer’s disease (AD) or identified as having AD, wherein the method comprises: (i) detecting in a biological sample from the subject a level of at least two lipid species selected from Table 4 and / or 6; and (ii) comparing the levels of the lipid species detected in the biological sample to reference lipid species levels to generate a risk score; and (iii) treating the subject if the risk score confirms that the subject has or is likely to develop AD.

6. The method according to claim 5, wherein the treatment comprises one or more of:Atty Dkt No.: RICE-231WO (i) medication, including any one or more of cholinesterase inhibitors, Memantine, Aducanumab, Lecanemab and antidepressants; and (ii) lifestyle changes, including any one or more of changes to diet, changes to exercise, quitting or reducing smoking and undertaking cognitive activities.

7. The method according to any one of claims 1 to 6, wherein the method comprises detecting in a biological sample from the subject a level of at least 10 lipid species, at least 20 lipid species, at least 50 lipid species or at least 100 lipid species selected from Table 4 and / or 6.

8. The method according to any one of claims 1 to 7, wherein the method further comprises detecting in the biological sample from the subject a level of at least one additional lipid species not defined in Table 4 and / or Table 6.

9. The method according to claim 8, wherein the at least one additional lipid species is defined in Table 2.

10. A method of calculating an Alzheimer’s disease (AD) risk score of a subject comprising: (i) obtaining lipid profile data from a biological sample taken from the subject; (ii) normalising the numeric values of the lipid profile data against a reference sample; and (iii) refining the discriminatory power of one or more lipid species by statistically weighting one or several of the numeric values associated therewith according to a predetermined equation; and (iv) normalising the resulting AD risk score of the subject to a reference population.

11. A method of calculating an Alzheimer’s disease (AD) risk score of a subject comprising: (i) obtaining lipid profile data from a biological sample taken from the subject; (ii) normalising the numeric values of the lipid profile data against a reference sample;Atty Dkt No.: RICE-231WO (iii) refining the discriminatory power of one or more lipid species by statistically weighting one or several of the numeric values associated therewith according to a predetermined equation; (iv) optionally adding one or more additional, statistically weighted, risk factors into the model, taken from the list of: sex, age, fasting glucose, BMI, HDL-cholesterol, cholesterol, triglycerides, APOE4 status, statin status, Omega-3 status, ethnicity, family history of disease, smoking status, exercise levels, diet, cognitive activity, memory level and place of birth; and (v) normalising the resulting AD risk score of the subject to a reference population.

12. The method according to any one of claims 10 or 11, wherein the lipid profile data comprises two or more lipid species selected from the group set forth in one or more of Tables 2, 3, 4, 5, and 6.

13. The method according to any one of claims 10 to 12, further comprising calculating a sum of weighted lipids.

14. The method according to any one of claims 1 to 13, wherein the reference lipid species levels are obtained from a population.

15. The method according to any one of claims 1 to 14, wherein the reference lipid species levels are included in a predictive model.

16. A method of determining an Alzheimer’s disease (AD) risk score of a subject, the method comprising a logistic regression model that uses an equation comprising: (i) input values comprising measured levels of two or more lipid species in a biological sample obtained from a subject, wherein the two or more lipid species include at least two or more lipid species selected from Table 4 and / or Table 6; and (ii) coefficient values that take into consideration risk factors selected from the group consisting of: sex, age, fasting glucose, BMI, HDL-cholesterol, cholesterol, triglycerides, APOE4 status, statin status, Omega-3 status, ethnicity, family history of disease, smoking status, exercise levels, diet, cognitive activity, memory level and place of birth;Atty Dkt No.: RICE-231WO wherein the input values are combined linearly using the coefficient values to predict an output value, and the output value is a risk score for determining risk of developing AD in the subject.

17. The method according to any one of claims 1 to 16, wherein the biological sample is selected from the group consisting of blood, plasma, serum, dried blood spots and dried plasma spots.

18. A method of generating a lipid profile in a sample from a subject, the method comprising: (i) obtaining lipid profile data from a biological sample taken from the subject; and (ii) normalising the numeric values of the lipid profile data against a reference sample, wherein the discriminatory power of one or more lipid species can be refined by statistically weighting one or several of the numeric values associated therewith according to a predetermined equation; wherein the resulting lipid profile of the subject can be normalised to a reference population; and wherein the lipid profile can be used to calculate an Alzheimer’s disease (AD) risk score of a subject.

19. A method of calculating an Alzheimer’s disease (AD) risk score of a subject, the method comprising: (i) refining a discriminatory power of one or more lipid species by statistically weighting one or several of the numeric values associated therewith according to a predetermined equation; and (ii) normalising the resulting AD risk score of the subject to a reference population, wherein lipid profile data can be obtained from a biological sample taken from the subject; and wherein the numeric values of the lipid profile data can be normalised against a reference sample.Atty Dkt No.: RICE-231WO 20. The method according to any one of claims 1 to 19, wherein the AD risk score is used to predict incident AD in a subject.

21. The method according to any one of claims 1 to 19, wherein the AD risk score is used to determine presence of AD in a subject.

22. The method according to claim 21, wherein the combined sensitivity and specificity of the AD risk score to identify a subject with AD is at least 75%.

23. The method according to claim 21 or 22, wherein the combined sensitivity and specificity of the AD risk score to identify a subject with AD is at least 80%.

24. The method according to any one of claims 21 to 23, wherein the combined sensitivity and specificity of the AD risk score to identify a subject with AD is at least 89%.

25. The method according to claim 20, wherein the combined sensitivity and specificity of the AD risk score to identify a subject who will develop AD is at least 70%.

26. The method according to claim 20 or 25, wherein the combined sensitivity and specificity of the AD risk score to identify a subject who will develop AD is at least 74%.

27. The method according to any one of claims 20, 25, or 26, wherein the combined sensitivity and specificity of the AD risk score to identify a subject who will develop AD is at least 76%.

28. The method according to any one of claims 1 to 27, wherein the AD risk score is used to stratify subjects into high risk and low risk.

29. A method of stratifying a group of human subjects for a treatment therapy, the method comprising: using an Alzheimer’s disease (AD) risk score determined using the methods according to any one of claims 1 to 28 to select subjects for the treatment therapy.Atty Dkt No.: RICE-231WO 30. The method of claim 29, wherein the treatment therapy is a clinical trial of a candidate therapy.

31. A panel or kit for use in a method according to any one of claims 1 to 30, wherein the kit comprises one or both of: (i) a set of stable-labelled isotopes or non-physiological lipid standards for quantification of the lipid species; (ii) reference plasma samples for standardisation of the resulting lipid measures.

32. A system for diagnosing Alzheimer’s disease (AD), or the likelihood of developing AD in a subject not diagnosed with AD, comprising: a processor operable to execute programs; a memory associated with the processor; a database associated with said processor and said memory; and a program stored in the memory and executable by the processor, the program being operable for: a) detecting in a biological sample from the subject a level of at least two lipid species selected from Table 2and / or 4 and / or 6; and b) correlating the levels of the lipid species detected in the biological sample to reference lipid species levels to a likelihood that the subject has or will develop AD.

Images