Method for diagnosing and prognosticating chronic heart failure

Abstract

The present application relates to a method for determining whether an object has heart failure or is at risk of developing heart failure, particularly heart failure with reduced left ventricular ejection fraction (HFREF) and heart failure with preserved left ventricular ejection fraction (HFPEF). The method comprises determining the level of a selected miRNA observed in a sample obtained from the object, and wherein a change in the level of the miRNA compared to a control indicates that the object has heart failure or is at risk of developing heart failure. The invention also includes a method for determining a change in the risk of death or disease progression to hospitalization and death based on a change in a selected miRNA in a sample obtained from the object, and a kit thereof.

Description

[0001] This application is a divisional application of Chinese Patent Application No. 201680034445.9, entitled "Methods for Diagnosis and Prognosis of Chronic Heart Failure", which is the national stage entry of PCT International Patent Application PCT / SG2016 / 050217 with a filing date of May 9, 2016.

[0002] Cross - reference to related applications

[0003] This application claims the priority benefit of Singapore Patent Application No. 10201503644Q, filed on May 8, 2015, the content of which is hereby incorporated by reference in its entirety for all purposes. Technical field

[0004] The present invention generally relates to the field of molecular biology. In particular, the present invention relates to the use of biomarkers for the detection and diagnosis of heart failure. Background art

[0005] Cardiovascular diseases, including heart failure, are a major health problem, accounting for approximately 30% of human deaths worldwide [1]. Heart failure is also the leading cause of hospitalization for adults over 65 years old globally [2]. Adults in middle age have a 20% risk of developing heart failure during their lifetime. Despite progress in treatment, the incidence and mortality of heart failure (about 50% over 5 years) remain high and consume approximately 2% of the healthcare budget in many economies [3 - 6]. Due to population aging, the prevalence of major risk factors such as diabetes and obesity has increased, and the initial survival rate of acute myocardial infarction and severe hypertension has increased, the prevalence of heart failure will increase.

[0006] Heart failure has traditionally been regarded as systolic dysfunction, and the left ventricular ejection fraction (LVEF) has been widely used to define systolic function, assess prognosis, and select patients for therapeutic intervention. However, it has been recognized that heart failure can occur in the presence of normal or near-normal EF: so-called "heart failure with preserved ejection fraction (HFPEF)", which accounts for a large proportion of clinical cases of heart failure [7-9]. Heart failure with severe dilation and / or significantly reduced EF: so-called "heart failure with reduced ejection fraction (HFREF)", is the type of heart failure that is best understood in terms of pathophysiology and treatment

[10] . There are some epidemiological differences between HFREF patients and HFPEF patients. The latter are typically older and more often female, less likely to have coronary artery disease (CAD) and more likely to have underlying hypertension [7, 8, 11]. Additionally, HFPEF patients do not obtain similar clinical benefits from angiotensin-converting enzyme inhibition or angiotensin receptor blockade as HFREF patients [12, 13]. The symptoms of heart failure can suddenly progress to "acute heart failure", leading to hospitalization, but it can also progress gradually.

[0007] Timely diagnosis, classification of heart failure subtypes - HFREF or HFPEF, and improved risk stratification are crucial for the management and treatment of heart failure. Accordingly, there is a need to provide methods for determining the risk of a subject developing heart failure. There is also a need to provide methods for classifying heart failure subtypes. SUMMARY OF THE INVENTION

[0008] In one aspect, there is provided a method for determining whether a subject has heart failure or is at risk of developing heart failure. In some instances, the method comprises the steps of: a) measuring the level of at least one miRNA from the "increased" miRNA list or at least one from the "decreased" miRNA list listed in Table 25 or Table 20 or Table 21 or Table 22 in a sample obtained from the subject. In some instances, the method further comprises the step of: b) determining whether the level of at least one miRNA from the miRNA list is different compared to a control, wherein a change in the level of the miRNA indicates that the subject has heart failure or is at risk of developing heart failure.

[0009] On the other hand, a method for determining whether an object has heart failure is provided. In some instances, heart failure is selected from heart failure with reduced ejection fraction (HFREF) and heart failure with preserved ejection fraction (HFPEF). In some instances, the method comprises the steps of: a) detecting the level of at least one miRNA listed in Table 9 in a sample obtained from the object. In some instances, the method further comprises the steps of: determining whether the level of at least one miRNA indicates that the object has heart failure with reduced ejection fraction (HFREF) or heart failure with preserved ejection fraction (HFPEF) or is at risk of developing it.

[0010] On the other hand, a method for determining the risk of change in the risk of death in a heart failure patient is provided. In some instances, the method comprises the steps of: a) detecting the level of at least one miRNA listed in Table 14 in a sample obtained from the object. In some instances, the method further comprises the steps of: b) measuring the level of at least one miRNA listed in Table 14. In some instances, the method further comprises the steps of: c) determining whether the level of at least one miRNA listed in Table 14 is different compared to the level of miRNA in a control population, wherein a change in the level of miRNA indicates that the risk of death of the object may change (change in observed (all-cause) survival rate) compared to the control population.

[0011] On the other hand, a method for determining the risk of change in the risk of a heart failure patient's disease progressing to hospitalization or death is provided. In some instances, the method comprises the steps of: a) detecting the level of at least one miRNA listed in Table 15 in a sample obtained from the object. In some instances, the method comprises the steps of: b) measuring the level of at least one miRNA listed in Table 15. In some instances, the method further comprises the steps of: c) determining whether the level of at least one miRNA listed in Table 15 is different compared to the level of miRNA in a control population, wherein a change in the level of miRNA indicates that the risk of the object's disease progressing to hospitalization or death may change (change in event-free survival rate) compared to the control population.

[0012] On the other hand, a method for determining the risk of developing heart failure in an object or determining whether an object has heart failure is provided. In some instances, the method comprises the steps of: (a) detecting the presence of miRNA in a sample obtained from the object. In some instances, the method further comprises the steps of: (b) measuring the levels of at least three miRNAs listed in Table 16 or Table 23 in the sample. In some instances, the method further comprises the steps of: (c) using a score based on the levels of the miRNAs measured in step (a) to predict the likelihood that the object will develop or has heart failure.

[0013] In another aspect, a method for determining the risk of developing heart failure in a subject or determining whether a subject has heart failure is provided. In some instances, the method comprises the steps of: (a) detecting the presence of miRNAs in a sample obtained from the subject. In some instances, the method further comprises the steps of: (b) measuring the levels of at least three miRNAs listed in Table 17 in the sample. In some instances, the method further comprises the steps of: (c) using a score based on the levels of the miRNAs measured in step (a) to predict the likelihood that the subject has or develops heart failure.

[0014] In another aspect, a method for determining the likelihood that a subject has heart failure with reduced left ventricular ejection fraction (HFREF) or heart failure with preserved left ventricular ejection fraction (HFPEF) is provided. In some instances, the method comprises the steps of: (a) detecting the presence of miRNAs in a sample obtained from the subject. In some instances, the method further comprises the steps of: (b) measuring the levels of at least three miRNAs listed in Table 18 in the sample. In some instances, the method comprises: (c) using a score based on the levels of the miRNAs measured in step (a) to predict the likelihood that the subject has heart failure with reduced left ventricular ejection fraction (HFREF) or heart failure with preserved left ventricular ejection fraction (HFPEF).

[0015] In another aspect, a method for determining the likelihood that a subject has heart failure with reduced left ventricular ejection fraction (HFREF) or heart failure with preserved left ventricular ejection fraction (HFPEF) is provided. In some instances, the method comprises the steps of: (a) detecting the presence of miRNAs in a sample obtained from the subject. In some instances, the method comprises the steps of: (b) measuring the levels of at least three miRNAs listed in Table 19 or Table 24 in the sample. In some instances, the method further comprises the steps of: (c) using a score based on the levels of the miRNAs measured in step (a) to predict the likelihood that the subject has heart failure with reduced left ventricular ejection fraction (HFREF) or heart failure with preserved left ventricular ejection fraction (HFPEF). BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The present invention will be better understood with reference to the detailed description when considered in conjunction with the non-limiting examples and the drawings, in which:

[0017] Figure 1 A schematic diagram showing an overview of the number of miRNAs identified from the studies described herein is shown.

[0018] Figure 2Histograms and skewness plots of N-terminal pro-brain natriuretic peptide hormone (NT-proBNP) and the natural logarithm of N-terminal pro-brain natriuretic peptide hormone (ln_NT-proBNP) levels are shown. Distributions of NT-proBNP levels (A to C) and ln_NT-proBNP levels (natural logarithm of NT-proBNP, D to F) for control subjects (A, D), heart failure with reduced left ventricular ejection fraction subjects (HFREF) (B, E), and heart failure with preserved left ventricular ejection fraction subjects (HFPEF) (C, F). The skewness of each plot is calculated and shown. Figure 2 It is shown that N-terminal pro-brain natriuretic peptide hormone (NT-proBNP) in all groups is positively skewed. In contrast, the natural logarithm of N-terminal pro-brain natriuretic peptide hormone (ln_NT-proBNP) levels has a lower skewness. Therefore, the natural logarithm levels of N-terminal pro-brain natriuretic peptide hormone are used for all analyses involving NT-proBNP.

[0019] Figure 3 Analysis results of the performance of the natural logarithm of N-terminal pro-brain natriuretic peptide hormone (ln_NT-proBNP) as a biomarker for heart failure are shown. In particular, (A) shows a box plot of ln_NT-proBNP (natural logarithm of NT-proBNP) levels. Each box plot represents the 25th, 50th, and 75th percentiles in the distribution. (B to D) show the receiver operating characteristic curves of ln_NT-proBNP for control versus heart failure (HFREF and HFPEF, B), HFREF versus HFPEF of the heart (C), control versus HFREF (D), and control versus HFPEF (E). AUC: area under the receiver operating characteristic curve, C: control (healthy), HF: heart failure, HFREF: heart failure with reduced left ventricular ejection fraction subjects, HFPEF: heart failure with preserved left ventricular ejection fraction subjects. Figure 3 A shows that in HFPEF, the loss of NT-proBNP test performance is more obvious. Figure 3 B to D show the natural logarithm of N-terminal pro-brain natriuretic peptide hormone (ln_NT-proBNP) that performs better in detecting HFREF than HFPEF.

[0020] Figure 4 An exemplary workflow for high-throughput miRNA RT-qPCR measurement is shown. Figure 4The steps shown include separation, multiplexing, multiplex RT, amplification, singleplex PCR, and synthesis of miRNA standard curves. Details of each step are as follows: Separation refers to the step of separating and purifying miRNA from plasma samples; spike-in miRNA refers to non-natural synthetic miRNA mimics (small single-stranded RNAs ranging from 22 to 24 bases in length) added to samples to monitor the efficiency of each step (including separation, reverse transcription, amplification, and qPCR); multiplex design refers to the miRNA assay that is deliberately divided into multiple multiplex groups (45 to 65 miRNAs per group) by computer to minimize non-specific amplification and primer-primer interactions during the RT and amplification processes; multiplex reverse transcription refers to combining different reverse transcription primer pools and adding them to different multiplex groups to generate cDNA; amplification refers to combining the PCR primer pool and adding it to each cDNA pool generated by a certain multiplex group, and performing optimized touch down PCR to amplify the amount of all cDNA in the group simultaneously; singleplex qPCR refers to distributing the amplified cDNA pools into different wells of a 384-well plate and then performing singleplex qPCR reactions; synthesis of miRNA standard curves refers to the synthetic miRNA standard curves measured together with the samples for interpolating absolute copy numbers in all measurements.

[0021] Figure 5 Bar plot results of principal component analysis are shown. Principal component analysis was performed on all 137 reliably detected mature miRNAs (Table 4) based on the log2-scaled expression levels (copies / mL). (A): Eigenvalues of the first 15 principal components. (B): Classification efficiency (AUC) of the first 15 principal components for separating controls (C) and heart failure (HF). (C): Classification efficiency (AUC) of the first 15 principal components for separating HFREF (heart failure with reduced ejection fraction) and HFPEF (heart failure with preserved ejection fraction). AUC: Area under the receiver operating characteristic curve. Figure 5 It is shown that multivariate assays may be needed to capture information in multiple dimensions for classifying HFREF and HFPEF.

[0022] Figure 6 Scatter plots of the top (AUC) principal components in heart failure subjects compared to controls are shown. Specifically, the top (AUC) principal components for discriminating between control (C, black circles) and heart failure (HF, white triangles) subjects are shown in A. The top two (AUC) principal components for differentiating between HFREF (heart failure with reduced ejection fraction, black circles) and HFPEF (heart failure with preserved ejection fraction, white triangles) subjects are shown in B. AUC: Area under the receiver operating characteristic curve. PC: Based on Figure 10Principal component number. Change: percentage of the change represented by the principal components calculated from the eigenvalues. Figure 6 It is shown that control, HFREF, and HFPEF subjects can be separated based on their miRNA profiles.

[0023] Figure 7 Venn diagrams showing the overlap of biomarkers that can be used to detect heart failure are presented. Comparisons between control (healthy) and different groups of heart failure patients (HF, HFREF, and HFPEF) are made by univariate analysis (t-test) and multivariate analysis (logistic regression) incorporating age and AF (atrial fibrillation or atrial flutter), hypertension, and diabetes. For the three comparisons: C vs. HF (HFREF and HFPEF), C vs. HFREF, and C vs. HFPEF, the number and overlap of miRNAs with p-values (after false discovery rate correction) below 0.01 in univariate analysis (A) and multivariate analysis (B) are shown. HF: Heart failure, HFPEF: Heart failure with preserved ejection fraction, HFREF: Heart failure with reduced ejection fraction, C: Control (healthy). Figure 7 It is shown that many miRNAs are found to be different between control and only one of the two heart failure subtypes, demonstrating a true difference in miRNA expression between the two subtypes.

[0024] Figure 8 Box plots and receiver operating characteristic curves of the top upregulated and downregulated miRNAs between healthy controls and heart failure patients are presented. Box plots and receiver operating characteristic (ROC) curves of the top (based on AUC) upregulated (A: ROC curve, C: box plot) and downregulated (B: ROC curve, D: box plot) miRNAs in all heart failure patients compared to control (healthy) subjects. The expression levels of miRNAs (copies / ml) are represented on a log2 scale. The box plots show the 25th, 50th, and 75th percentiles in the distribution of the expression levels. C: Control (healthy), HF: Heart failure. AUC: Area under the receiver operating characteristic curve. Figure 8 It is shown that combinations of multiple miRNAs can enhance the performance of heart failure diagnosis.

[0025] Figure 9A Venn diagram showing the overlap of biomarkers for detecting heart failure and classifying heart failure subtypes is presented. Comparison between HFREF and HFPEF was performed by univariate analysis (t-test) and multivariate analysis (logistic regression) incorporating age, sex, BMI (body mass index), and AF (atrial fibrillation or atrial flutter), hypertension (p-value, ln_BNP). miRNAs with p-values (after false discovery rate correction) less than 0.01 in univariate analysis (A) and multivariate analysis (B) were compared with miRNAs for detecting heart failure (C vs. HF or C vs. HFREF or C vs. HFPEF, Figure 5 )). HF: heart failure, HFPEF: heart failure with preserved ejection fraction, HFREF: heart failure with reduced ejection fraction, C: control (healthy) subjects.

[0026] Figure 10 Box plots and receiver operating characteristic (ROC) curves of upregulated and downregulated miRNAs with higher ranks in HFPEF patients compared to HFREF patients are shown. Box plots and receiver operating characteristic (ROC) curves of upregulated (A: ROC curve, C: box plot) and downregulated (B: ROC curve, D: box plot) miRNAs with higher ranks (based on AUC) in HFPEF patients compared to HFREF patients. The expression levels of miRNAs (copies / ml) are represented on a log2 scale. The box plots show the 25th, 50th, and 75th percentiles in the distribution of expression levels. HFPEF: heart failure with preserved ejection fraction, HFREF: heart failure with reduced ejection fraction, AUC: area under the receiver operating characteristic curve. Figure 10 It is shown that combining multiple miRNAs in a multivariate metric determination can provide more diagnostic power for subtype classification.

[0027] Figure 11 A line plot of overlapping miRNAs for detecting heart failure and classifying heart failure subtypes is presented. Based on the changes, 38 overlapping miRNAs among control, heart failure (HFREF or HFPEF), and HFREF, HFPEF ( Figure 7 , A) were divided into 7 groups. Two groups were defined as equal if the p-value (t-test) of the miRNA was higher than 0.01 after false discovery testing. The expression levels are based on a log2 scale and are normalized relative to a zero mean for each miRNA. HFPEF: heart failure with preserved ejection fraction, HFREF: heart failure with reduced ejection fraction, C: control (healthy). Figure 11 It is shown that, compared to healthy controls, unlike LVEF and NT-proBNP, the HFPEF subtype has a more specific miRNA profile than the HFREF subtype. Figure 11It is shown that miRNA can be complementary to NT-proBNP to provide better discrimination of HFPEF.

[0028] Figure 12 A scatter plot of the correlation analysis among all reliably detected miRNAs is shown. Based on the log2-scaled expression levels (copies / mL), Pearson linear correlation coefficients were calculated among all 137 reliably detected miRNA targets (Table 4). Each point represents a pair of miRNAs with a correlation coefficient higher than 0.5 (A, positive correlation) or lower than -0.5 (B, negative correlation). miRNAs with differential expression in HF and HFREF vs. HFPEF are represented in the horizontal dimension in black. HF: heart failure, HFPEF: heart failure with preserved ejection fraction, HFREF: heart failure with reduced ejection fraction, C: control (healthy). Figure 12 It is shown that, among all subjects, many miRNA pairs are similarly regulated.

[0029] Figure 13 A bar graph representing the drug treatments for HFREF and HFPEF is shown. The number of cases of different anti-HF drug treatments for 327 subjects classified into HFREF and HFPEF subtypes included in the prognostic analysis was summarized. The chi-square test was applied to compare the two subtypes for each treatment. *: p value < 0.05, **: p value < 0.01, ***: p value < 0.001. Figure 13 It is a summary of the treatment according to current clinical practice and is included in the clinical variables for analyzing prognostic markers.

[0030] Figure 14 shows the survival analysis of the subjects. Specifically, (A) shows the Kaplan-Meier plots of the clinical variables (Table 14) that significantly predict the observed survival based on univariate analysis (p value < 0.05). For categorical variables, the positive group (black) and the negative group (gray) were compared. For variables with a normal distribution, subjects with values above the median (black) and below the median (gray) were compared. The log-rank test was performed to test each variable between the two groups, and the p value is shown above each plot. (B) shows a bar graph representing the percentage of observed survival (OS) at 750 days after treatment.

[0031] Figure 15 shows the survival analysis of event-free survival. Specifically, (A) shows the Kaplan-Meier plots of clinical variables (Table 14) that significantly predict event-free survival based on univariate analysis (p-value < 0.05). For categorical variables, the positive group (black) and negative group (gray) are compared. For normally distributed variables, subjects with values above the median (black) and below the median (gray) are compared. The log-rank test is performed to test each variable between the two groups, and the p-value is shown above each plot. (B) shows a bar graph representing the percentage of event-free survival (EFS) at 750 days after treatment.

[0032] Figure 16 A Venn diagram showing the comparison between biomarkers for overall survival (OS) and event-free survival (EFS) is presented. Specifically, (A) shows the comparison between miRNAs that are significantly prognostic for OS identified by univariate and multivariate analysis using the CoxPH model. (B) shows the comparison between significant miRNAs for OS prognosis and EFS prognosis. miRNAs are identified using the CoxPH model by univariate or multivariate analysis. Figure 16 Shows different mechanisms of death and recurrent decompensated heart failure.

[0033] Figure 17 A Venn diagram showing the comparison between biomarkers for overall survival (OS) and event-free survival (EFS) is presented. Specifically, (A) shows the comparison between miRNAs that are significantly prognostic (for OS or for EFS) by the CoxPH model and miRNAs for detecting HF (either subtype). All miRNAs are identified by univariate or multivariate analysis. (B) shows the comparison between significant miRNAs for prognostic identification (for OS or for EPS) by the CoxPH model and miRNAs for classifying two HF subtypes. All miRNAs are identified by univariate or multivariate analysis. Figure 17 Shows that most prognostic markers are not found in the other two lists, indicating that separate miRNA sets can be used or combined to form a prognostic assay.

[0034] Figure 18 shows the analysis of miRNAs with the maximum and minimum hazard ratios for overall survival (OS). In (A), univariate CoxPH models or multivariate CoxPH models including six additional clinical variables were constructed using miRNAs with the maximum hazard ratio (hsa-miR-503) and minimum hazard ratio (hsa-miR-150-5p) for overall survival (OS): sex, hypertension, BMI, ln_NT-proBNP, β-blocker, and warfarin for overall survival (OS). All normal variable levels (including BMI, ln_NT-proBNP, and miRNA expression levels (log2 scale)) were scaled to have one standard deviation. Based on the explanatory score values according to the CoxPH model, the top 50% of subjects (black) and the bottom 50% of subjects (gray) were compared. A log-rank test was performed to test between the two groups, and the p-value (B) for overall survival (OS) 750 days after treatment is shown.

[0035] Figure 19 shows the analysis of miRNAs with the maximum and minimum hazard ratios for EFS. In (A), univariate CoxPH models or multivariate CoxPH models including two additional clinical variables were constructed using miRNAs with the maximum hazard ratio (hsa-miR-331-5p) and minimum hazard ratio (hsa-miR-191-5p) for EFS: diabetes status and ln_NT-proBNP for EFS. All normal variable levels (including diabetes status, ln_NT-proBNP, and miRNA expression levels (log2 scale)) were scaled to have one standard deviation. Based on the explanatory score values according to the CoxPH model, the top 50% of subjects (black) and the bottom 50% of subjects (gray) were compared. A log-rank test was performed to test between the two groups, and the p-value (B) for EFS 750 days after treatment is shown.

[0036] Figure 20 shows representative results for generating a multivariate biomarker panel for heart failure detection. In (A), box plots show the diagnostic power (AUC) of the multivariate biomarker panel (number of miRNAs = 3 to 10) during the discovery and validation phases of heart failure detection in computerized two-fold cross-validation. The box plots show the 25th, 50th, and 75th percentiles in the AUC for classification of healthy and heart failure patients. A quantitative representation of the results for the discovery group (black) and validation group (gray) is shown in (B). Error bars represent the standard deviation of the AUC. To test the significance of the increase in AUC in the validation group when including more miRNAs in the panel, a right-tailed t-test was performed to compare all adjacent gray bars. *: p-value < 0.05; **: p-value < 0.01; ***: p-value < 0.001.

[0037] Figure 21 Shows the comparison between the multivariate miRNA score and NT-proBNP in HF detection using a two-dimensional plot. (A) Shows a two-dimensional plot of the NT-proBNP level (y-axis) and one of the 6-miRNA group scores (x-axis) for all subjects. The threshold for NT-proBNP (125) is indicated by a dashed line. False-positive and false-negative subjects according to NT-proBNP are boxed. (B) Shows a two-dimensional plot of the NT-proBNP level (y-axis) and 6-miRNA group score (x-axis) for false-positive and false-negative subjects classified according to NT-proBNP using a 125 pg / ml threshold. The threshold miRNA score (0) is indicated by a dashed line. Control subjects are represented by crosses; HFREF subjects are represented by filled circles, and HFPEF subjects are represented by open triangles. Figure 21 Verified the following hypothesis: the hypothesis that miRNA biomarkers carry different information from the N-terminal pro-brain natriuretic peptide hormone (NT-proBNP).

[0038] Figure 22 Shows the analysis of a multivariate biomarker panel for heart failure detection that combines miRNA with NT-proBNP. (A) Shows a series of box plots of the diagnostic power (AUC) of the multivariate biomarker panel (ln_NT-proBNP plus 2 to 8 miRNAs) during the discovery and validation phases of HF detection during computerized two-fold cross-validation. The box plots show the 25th, 50th, and 75th percentiles of the AUC for classifying healthy and HF patients. (B) Shows a quantitative representation of the results for the discovery group (black) and the validation group (gray) and ln-NT-proBNP itself (first column). Error bars represent the standard deviation of the AUC. To test the significance of the increase in AUC in the validation group when more miRNAs are included in the panel, a right-tailed t-test was performed to compare all adjacent gray bars. *: p-value < 0.05; **: p-value < 0.01; ***: p-value < 0.001. Thus, Figure 22 Shows that the classification efficiency is significantly improved when miRNA is combined with the N-terminal pro-brain natriuretic peptide hormone (NT-proBNP).

[0039] Figure 23 Shows a Venn diagram of the overlap of miRNAs selected for the multivariate HF detection panel with or without the N-terminal pro-brain natriuretic peptide hormone (NT-proBNP). Comparison between biomarkers selected for HF detection using miRNAs alone (Table 16) or using miRNAs in combination with NT-proBNP (Table 17) during the multivariate biomarker retrieval process. Significantly miRNAs (A) and non-significantly miRNAs (B) were compared separately. Figure 23It is shown that when using NT-proBNP, different miRNA lists can be used.

[0040] Figure 24 Representative results are shown for generating multi-miRNA panels for heart failure subtype stratification with and without NT-proBNP. (A) Shows the multivariate miRNA biomarker panel retrieval (3 to 10 miRNAs) for heart failure subtype classification. The AUC results for the discovery group (black bars) and the validation group (gray bars) are shown. (B) Shows the multivariate miRNA and NT-proBNP biomarker panel retrieval (ln_NT-proBNP plus 2 to 8 miRNAs) for heart failure subtype classification. The AUC results for the discovery group (black bars), the validation group (gray bars), and ln_NT-proBNP itself (first column) are shown. Error bars represent the standard deviation of the AUC. Right-tailed t-tests were performed to compare all adjacent gray bars. *: p-value < 0.05; **: p-value < 0.01; ***: p-value < 0.001. Figure 24 It is shown that even clearer classification can be achieved when using both miRNA and NT-proBNP.

[0041] Brief description of the attached tables

[0042] When considering the non-limiting examples and the attached tables in combination, the present invention will be better understood with reference to the detailed description, in which:

[0043] Table 1 is a summary of reported serum / plasma miRNA biomarkers for heart failure. Studies measuring cell-free serum / plasma miRNA or whole blood are included in the table. Only miRNAs verified by qPCR are shown. Upregulated: miRNAs with higher levels in HF patients compared to control (healthy) subjects. Downregulated: miRNAs with lower levels in HF patients compared to control (healthy) subjects. The numbers in "Study design" represent the number of samples used in the study. PBMC: peripheral blood mononuclear cells, AMI: acute myocardial infarction, HF: heart failure, HF: heart failure, HFPEF: heart failure with preserved left ventricular ejection fraction, HFREF: heart failure with reduced left ventricular ejection fraction, BNP: brain natriuretic peptide, C: control (healthy subjects).

[0044] Table 2 is a table listing the clinical information of the subjects included in the study. The clinical information of 546 subjects was included in the study. All plasma samples were stored at -80 °C before use. N.A.: not available, C: control (healthy subjects), PEF: heart failure with preserved left ventricular ejection fraction, REF: heart failure with reduced left ventricular ejection fraction.

[0045] Table 3 is a table listing the characteristics of healthy subjects and heart failure patients. The ejection fraction (left ventricular ejection fraction), ln_NT-proBNP, age, and body mass index are shown as the arithmetic mean ± standard deviation, and NT-proBNP is shown as the geometric mean. The percentages next to the variables represent the percentage of subjects with known values for the variables. HF: heart failure, HFPEF: heart failure with preserved ejection fraction, HFREF: heart failure with reduced ejection fraction, C: control (healthy) subjects. For the comparison of variables between control and heart failure (C vs. HF) and HFPEF vs. HFREF (HFREF vs. HFPEF), the t-test was used for normal variables and the chi-square test was used for categorical variables.

[0046] Table 4 is a table listing the sequences of 137 reliably detected mature miRNAs. 137 mature miRNAs were reliably detected in plasma samples. "Reliably detected" was defined as the concentration being higher than 500 copies / ml in at least 90% of plasma samples. The miRNAs were named according to the miRBase V18 version.

[0047] Table 5 is a table listing the miRNAs differentially expressed between control and all heart failure subjects. The comparison between control (healthy) and all heart failure subjects (both HFREF and HFPEF) was performed by univariate analysis (p-value, t-test) and multivariate analysis adjusted for age and AF (atrial fibrillation or atrial flutter), hypertension, and diabetes (p-value, logistic regression). The enhancement of the diagnostic performance of miRNAs for ln_NT-proBNP for heart failure was tested by logistic regression adjusted for age and AF (atrial fibrillation or atrial flutter), hypertension, and diabetes (p-value, ln_BNP). All p-values were adjusted for false discovery rate correction using the Bonferroni method. Only those miRNAs with a p-value of less than 0.01 for both the "p-value, t-test" and "p-value, logistic regression" tests are shown. Fold change: the miRNA expression level in heart failure subjects divided by the miRNA expression level in control subjects.

[0048] Table 6 is a table listing miRNAs differentially expressed between control and HFREF subjects. Comparison between control (healthy) and HFREF subjects (heart failure with reduced left ventricular ejection fraction) was performed by univariate analysis (p-value, t-test) and multivariate analysis (p-value, logistic regression) adjusted for age, AF (atrial fibrillation or atrial flutter), hypertension, and diabetes. Enhancement of HFREF discrimination of miRNAs for ln_NT-proBNP was tested by logistic regression adjusted for age and AF (atrial fibrillation or atrial flutter), hypertension, and diabetes (p-value, ln_BNP). All p-values were adjusted for false discovery rate correction using the Bonferroni method. Only those miRNAs with p-values < 0.01 for both the "p-value, t-test" test and the "p-value, logistic regression" test are shown. Fold change: miRNA expression level in HFREF subjects divided by miRNA expression level in control subjects.

[0049] Table 7 is a table listing miRNAs differentially expressed between control and HFPEF subjects. Comparison between control (healthy) and HFPEF subjects (heart failure with preserved left ventricular ejection fraction) was performed by univariate analysis (p-value, t-test) and multivariate analysis (p-value, logistic regression) adjusted for age and AF (atrial fibrillation or atrial flutter), hypertension, and diabetes. Enhancement of HFPEF diagnosis discrimination of miRNAs for ln_NT-proBNP was tested by logistic regression adjusted for age, AF (atrial fibrillation or atrial flutter), hypertension, and diabetes (p-value, ln_BNP). All p-values were adjusted for false discovery rate correction using the Bonferroni method. Only those miRNAs with p-values < 0.01 for both the "p-value, t-test" test and the "p-value, logistic regression" test are shown. Fold change: miRNA expression level in HFPEF subjects divided by miRNA expression level in control subjects.

[0050] Table 8 is a table listing the comparison between the current study and previously published reports. miRNAs not listed in Table 4 (expression level ≥ 500 copies / ml) are denoted as N.A. (not available), which may not be included in the study or below the detection limit. Top: miRNAs have higher expression levels in heart failure patients compared to control (healthy) subjects. Bottom: miRNAs have lower expression levels in heart failure patients compared to control (healthy) subjects. Those miRNAs with p-values below 0.01 after false discovery rate correction are denoted as no change. For hsa-miR-210, there are contradictions in the direction of change in different literature reports (denoted as top and bottom).

[0051] Table 9 is a table listing miRNAs that are differentially expressed between HFREF and HFPEF subjects. Comparison between HFREF (heart failure with reduced ejection fraction) and HFPEF subjects (heart failure with preserved ejection fraction) was performed by univariate analysis (p-value, t-test) and multivariate analysis adjusted for age, sex, BMI (body mass index), AF (atrial fibrillation or atrial flutter), and hypertension (p-value, logistic regression). The ability of miRNAs to discriminate HFREF from HFPEF classification of ln_NT-proBNP was tested by logistic regression adjusted for age, sex, BMI (body mass index), AF (atrial fibrillation or atrial flutter), and hypertension (p-value, ln_BNP). All p-values were adjusted for false discovery rate correction using the Bonferroni method. Only those miRNAs with p-value < 0.01 for the "p-value, t-test" test are shown. Fold change: miRNA expression level in HFPEF subjects divided by miRNA expression level in HFREF subjects.

[0052] Table 10 is a table listing the clinical information of the subjects included in the prognostic study. Clinical information of 327 subjects included in the prognostic study. All subjects were followed up for two years after recruitment into the SHOP cohort study. During the follow-up period, 49 patients passed away.

[0053] Table 11 is a table listing the treatments of the subjects included in the prognostic study. Drug treatments of 327 subjects included in the prognostic study; drug names, Me1: ACE inhibitor, Me2: angiotensin 2 receptor blocker, Me3: loop / thiazide diuretic, Me4: β-blocker, Me5: aspirin or Plavix, Me6: statin, Me7: digoxin, Me8: warfarin, Me9: nitrate, Me10: calcium channel blocker, Me11: spironolactone, Me12: fibrate, Me13: anti-diabetic agent, Me14: hydralazine, Me15: iron supplement.

[0054] Table 12 is a table listing the analysis of clinical variables for observed survival. Clinical parameters included in the analysis of observed survival, including drug treatments and other variables, were analyzed using the Cox proportional hazards model. The levels of age, BMI, LVEF, and ln_NT-proBNP were adjusted proportionally to have one standard deviation. In the multivariate analysis, all variables were included. The cells of those variables with p-value less than 0.05 are shaded in gray. ln(HR): natural logarithm of the hazard ratio (a positive value indicates: the higher the value of the variable, the higher the probability of death), SE: standard error.

[0055] Table 13 is a table listing the analysis of clinical variables for event-free survival. The clinical parameters used for the analysis of event-free survival were adjusted proportionally for age, BMI, LVEF, and ln_NT-proBNP levels to a Cox proportional hazards model with one standard deviation. Drug treatment was also included. In the multivariable analysis, all variables were included. The cells for those variables with a p-value < 0.05 are represented in gray. ln(HR): natural logarithm of the hazard ratio (a positive value indicates that the higher the value of the variable, the higher the probability of death), SE: standard error.

[0056] Table 14 is a table listing the miRNAs that significantly predict observed survival. The association of each miRNA with observed survival was analyzed using a Cox proportional hazards model in univariate and multivariable analyses, and the multivariable analysis included additional clinical variables: sex, hypertension, BMI, ln_NT-proBNP, β-blocker, and warfarin. All normally distributed variables including ln_NT-proBNP, BMI, and miRNA expression levels (log2 scale) were adjusted proportionally to have one standard deviation. Those with a p-value < 0.05 are represented as gray cells. ln(HR): natural logarithm of the hazard ratio (a positive value indicates that the higher the value of the variable, the higher the probability of death), SE: standard error.

[0057] Table 15 is a table listing the miRNAs that significantly predict event-free survival. The association of each miRNA with event-free survival was analyzed using a Cox proportional hazards model in univariate and multivariable analyses, and the multivariable analysis included additional clinical variables: diabetes and ln_NT-proBNP. All normally distributed variables including ln_NT-proBNP and miRNA expression levels (log2 scale) were adjusted proportionally to have one standard deviation. Those with a p-value < 0.05 are represented as gray cells. ln(HR): natural logarithm of the hazard ratio (a positive value indicates that the higher the value of the variable, the higher the probability of death), SE: standard error.

[0058] Table 16 is a table listing the miRNAs identified during the multivariable group retrieval process for heart failure detection. The miRNAs selected for assembling biomarker groups for heart failure detection with 6, 7, 8, 9, and 10 miRNAs are listed. The prevalence was defined by dividing the count of miRNAs in all groups by the total number of groups. Groups with the top 10% and bottom 10% AUC were excluded to avoid miscounting false discovery biomarkers due to inaccurate data fitting from subpopulations generated by the randomization process in cross-validation analysis. Only miRNAs used in more than 2% of the groups are listed. The changes in miRNAs in different heart failure subtypes were defined based on Tables 5 to 7.

[0059] Table 17 is a table listing miRNAs identified in combination with NT-proBNP during the multivariate group retrieval process for heart failure (HF) detection. The miRNAs selected for assembling biomarker groups for heart failure detection with ln_NT-proBNP and 3, 4, 5, 6, 7, and 8 miRNAs are listed. Popularity is defined by dividing the count of miRNAs in all groups by the total number of groups. Groups with the top 10% and bottom 10% AUC are excluded to avoid false discovery of biomarkers due to counting inaccurate data fitting subpopulations generated by the randomization process in cross-validation analysis. Only miRNAs used in more than 2% of the groups are listed. The significance of miRNAs as supplements to ln_NT-proBNP in differentiating different heart failure subtypes was determined based on logistic regression using the selected miRNAs and ln_NT-proBNP as predictors, where the p-value of significant miRNAs after FDR correction was <0.01.

[0060] Table 18 is a table listing miRNAs identified during the multivariate group retrieval process for HF subtype classification. The miRNAs selected for assembling biomarker groups for heart failure (HF) subtype classification with 6, 7, 8, 9, and 10 miRNAs are listed. Popularity is defined by dividing the count of miRNAs in all groups by the total number of groups. Groups with the top 10% and bottom 10% AUC are excluded to avoid false discovery of biomarkers due to counting inaccurate data fitting subpopulations generated by the randomization process in cross-validation analysis. Only miRNAs used in more than 2% of the groups are listed. The miRNA changes between HFREF and HFPEF subtypes were determined based on Table 9.

[0061] Table 19 is a table listing miRNAs identified in combination with NT-proBNP during the multivariate group retrieval process for HF subtype classification. The miRNAs selected for assembling biomarker groups for HF subtype classification with ln_NT-proBNP and 5, 6, 7, and 8 miRNAs are listed. Popularity is defined by dividing the count of miRNAs in all groups by the total number of groups. Groups with the top 10% and bottom 10% AUC are excluded to avoid false discovery of biomarkers due to counting inaccurate data fitting subpopulations generated by the randomization process in cross-validation analysis. Only miRNAs used in more than 2% of the groups are listed. The significance of miRNAs as supplements to ln_NT-proBNP was determined based on logistic regression using the selected miRNAs and ln_NT-proBNP as predictors, where the p-value of significant miRNAs after FDR correction was <0.01.

[0062] Table 20 is a table listing miRNAs identified for heart failure (HF) detection. Comparison between controls (healthy) and all heart failure subjects (both HFREF and HFPEF) was performed by univariate analysis (p-value, t-test) and multivariate analysis adjusted for age and AF (atrial fibrillation or atrial flutter), hypertension, diabetes (p-value, logistic regression). The enhancement of the diagnostic performance of miRNAs for heart failure against ln_NT-proBNP was tested by logistic regression adjusted for age and AF (atrial fibrillation or atrial flutter), hypertension, diabetes (p-value, ln_BNP). All p-values were adjusted for false discovery rate correction using the Bonferroni method. Only those miRNAs with p-values < 0.01 for both the "p-value, t-test" test and the "p-value, logistic regression" test are shown. Fold change: miRNA expression level in HF subjects divided by miRNA expression level in control subjects. Table 20 corresponds to Table 5, except that the miRNAs listed in Table 20 are not part of the miRNAs known in the art (i.e., as listed in Tables 1 and 8).

[0063] Table 21 is a table listing miRNAs identified for HFREF detection. Comparison between controls (healthy) and HFREF subjects (heart failure with reduced left ventricular ejection fraction) was performed by univariate analysis (p-value, t-test) and multivariate analysis adjusted for age, AF (atrial fibrillation or atrial flutter), hypertension, and diabetes (p-value, logistic regression). The enhancement of HFREF discrimination of miRNAs against ln_NT-proBNP was tested by logistic regression adjusted for age and AF (atrial fibrillation or atrial flutter), hypertension, and diabetes (p-value, ln_BNP). All p-values were adjusted for false discovery rate correction using the Bonferroni method. Only those miRNAs with p-values < 0.01 for both the "p-value, t-test" test and the "p-value, logistic regression" test are shown. Fold change: miRNA expression level in HFREF subjects divided by miRNA expression level in control subjects. Table 21 corresponds to Table 6, except that the miRNAs listed in Table 21 are not part of the miRNAs known in the art (i.e., as listed in Tables 1 and 8).

[0064] Table 22 is a table listing miRNAs identified for HFPEF detection. Comparisons between control (healthy) and HFPEF subjects (heart failure with preserved left ventricular ejection fraction) were made by univariate analysis (p-value, t-test) and multivariate analysis adjusted for age and AF (atrial fibrillation or atrial flutter), hypertension, and diabetes (p-value, logistic regression). The enhancement of miRNA for ln_NT-proBNP discrimination of HFPEF diagnosis was tested by logistic regression adjusted for age, AF (atrial fibrillation or atrial flutter), hypertension, and diabetes (p-value, ln_BNP). All p-values were adjusted for false discovery rate correction using the Bonferroni method. Only those miRNAs with p-value < 0.01 for both the "p-value, t-test" test and the "p-value, logistic regression" test are shown. Fold change: miRNA expression level in HFPEF subjects divided by miRNA expression level in control subjects. Table 22 corresponds to Table 7, except that the miRNAs listed in Table 22 are not part of the miRNAs known in the art (i.e., as listed in Table 1 and Table 8).

[0065] Table 23 is a table listing miRNAs frequently selected during multivariate group retrieval for heart failure detection. miRNAs selected for assembling biomarker groups for heart failure detection with 6, 7, 8, 9, and 10 miRNAs are listed. Popularity is defined by dividing the count of miRNAs in all groups by the total number of groups. Groups with the top 10% and bottom 10% AUC were excluded to avoid counting false discovery biomarkers due to inaccurate data fitting from subpopulations generated by the randomization process in cross-validation analysis. Only miRNAs used in more than 2% of the groups are listed. The variation of miRNAs in different heart failure HF subtypes is defined based on Tables 20 to 22. Table 23 corresponds to Table 16, except that the miRNAs listed in Table 23 are not part of the miRNAs known in the art (i.e., as listed in Table 1 and Table 8).

[0066] Table 24 is a table listing miRNAs that are frequently selected in the multi-variable group retrieval for HF detection in combination with NT-proBNP. miRNAs selected for assembling biomarker groups for HF detection with ln_NT-proBNP and 3, 4, 5, 6, 7, and 8 miRNAs are listed. Popularity is defined by dividing the count of miRNAs in all groups by the total number of groups. Groups with the top 10% and bottom 10% AUCs are excluded to avoid miscounting false discovery biomarkers due to inaccurate data fitting from subpopulations generated by the randomization process in cross-validation analysis. Only miRNAs used in more than 2% of the groups are listed. The significance of miRNAs as supplements to ln_NT-proBNP in discriminating different HF subtypes is determined based on logistic regression using the selected miRNAs and ln_NT-proBNP as predictor variables, where the p-value of significant miRNAs after FDR correction is <0.01. Table 24 corresponds to Table 17, with the difference that the miRNAs listed in Table 24 are not part of the miRNAs known in the art (i.e., as listed in Tables 1 and 8).

[0067] Table 25 is a table listing microRNAs that can be specifically used for heart failure detection. To the inventors' knowledge, these miRNAs are only related to heart failure. The miRNAs listed in Table 25 are not part of the miRNAs known in the art (i.e., as listed in Tables 1 and 8).

[0068] Table 26 is a table listing exemplary biomarker groups for heart failure detection. Based on the provided biomarkers, examples of the formula, cut-off value, and performance of the groups are provided in the table.

[0069] Table 27 is a table listing exemplary biomarker groups for heart failure subtype detection. Based on the provided biomarkers, examples of the formula, cut-off value, and performance of the groups are provided in the table. Detailed Description

[0070] Timely diagnosis, accurate classification of heart failure subtypes (including but not limited to heart failure with reduced left ventricular ejection fraction (HFREF), heart failure with preserved left ventricular ejection fraction (HFPEF), etc.), and improved risk stratification are of great significance for the management and treatment of heart failure. An attractive approach is to use circulating biomarkers

[14] . Established circulating biomarkers in heart failure are cardiac natriuretic peptides, B-type natriuretic peptide (BNP) and its coreleased homolog, N-terminal pro-brain natriuretic peptide (NT-proBNP). Both have been shown to have diagnostic utility in acute heart failure and are independently associated with prognosis at all stages of heart failure, thus leading to their inclusion in all major international guidelines for heart failure diagnosis and management [14,15]. However, confounding factors including age, renal function, obesity, and atrial fibrillation do affect their diagnostic performance [16,17]. In asymptomatic left ventricular dysfunction, early symptomatic heart failure, and treated heart failure, the discriminative power of B-peptides is significantly reduced, with half of all stable HFREF cases showing BNP below 100 pg / ml and 20% having NT-proBNP below the value used to exclude heart failure in the acute symptomatic state

[18] . This loss of test performance is even more evident in the case of HFPEF

[19] . B-peptides reflect the transmural distending pressure and myocyte stretch of the following ventricles, which (depending on ventricular diameter as well as intraventricular pressure and wall thickness) are much lower in HFPEF with normal or reduced ventricular cavity volume and thickened ventricular walls compared to HFREF with typical dilated ventricles and eccentric remodeling

[20] . Therefore, there is an unmet need for biomarkers that supplement or replace B-type peptides in screening for heart failure in the early or partial treatment state of heart failure and in monitoring the status of the chronic phase of heart failure. This is especially true for HFPEF, where B-peptide levels are lower than in HFREF and are usually normal

[21] . Currently, the classification of heart failure subtypes depends on cardiologists' imaging and imaging interpretation. There is no biomarker-based test available for this purpose. Therefore, a minimally invasive method to improve heart failure diagnosis and HF subtype classification is desirable.

[0071] MicroRNAs (miRNAs) are small non-coding RNAs that play a central role in regulating gene expression, and dysregulation of miRNAs is involved in the pathogenesis of various diseases [22-26]. Since their discovery in 1993

[27] , miRNAs have been estimated to regulate more than 60% of all human genes

[28] , and many of them have been identified as key players in key cellular functions such as proliferation

[29] and apoptosis

[30] . The discovery of miRNAs in human serum and plasma has increased the possibility of using circulating miRNAs as biomarkers for the diagnosis, prognosis, and treatment decisions of many diseases [31-35]. An integrated multidimensional approach using miRNAs or using miRNAs in combination with BNP / NT-proBNP for the diagnosis of HF can improve diagnosis. Compared with using BNP / NT-proBNP alone, combining genomic markers such as miRNAs and protein markers such as BNP / NT-proBNP can enhance diagnostic power in HF. Recently, various attempts have been made to identify circulating cell-free miRNA biomarkers in serum or plasma to distinguish HF patients from healthy subjects [36-47] (Table 1).

[0072] Table 1. Summary of reported serum / plasma miRNA biomarkers for heart failure

[0073]

[0074]

[0075]

[0076] These studies have reported miRNA profiles differentially regulated in heart failure subjects. However, there is a lack of concordance between these published works. Among the 67 reported miRNAs, only three have been found to be upregulated in more than one report. In particular, hsa-miR-210 has been reported to be upregulated in HF in one report and downregulated in another (Table 1). The lack of concordance between studies may be due to multiple reasons, including the use of small sample sizes or variability in sample selection, such as disease stage and importantly, the controls used [32,48]. The pre-analytical processes, including experimental design and workflow, are crucial in biomarker identification and validation. Most studies to date have used high-throughput array platforms to screen a limited number of samples. This approach lacks sensitivity and reproducibility. It yields a small panel of targets (less than 10 miRNAs) identified for further validation. Most studies have not been confirmed in larger patient cohorts. Another widely adopted approach is based on screening reported candidate miRNAs using quantitative real-time polymerase chain reaction (qPCR). Evaluation of different technologies based on array and qPCR platforms has shown substantial differences in the performance of these platforms for miRNA measurement. This may contribute to the observed inconsistencies between studies

[49] . To date, there is no consensus on specific circulating serum / plasma miRNAs that can be used as heart failure biomarkers. None of the previously reported miRNA profiles are available for the classification of heart failure subtypes. Therefore, there is a need to establish a robust pre-specified technology platform for heart failure biomarker discovery and validation and ensure the reproducibility of the results.

[0077] In the present disclosure, a panel of circulating miRNAs was identified as potential heart failure biomarkers. These multivariate assays are defined by Food and Drug Agency (FDA) guidelines, cited as follows: “combines the values of multiple variables using an interpretation function to yield a single, patient-specific result (e.g., a “classification,” “score,” “index,” etc.), that is intended for use in the diagnosis of disease or other conditions, or in the cure, mitigation, treatment or prevention of disease, and provides a result whose derivation is non-transparent and cannot be independently derived or verified by the end user”. Thus, highly reliable quantitative data based on qPCR according to the MIQE (Minimum Information for Publication of Quantitative Real-Time PCR Experiments) guidelines is a prerequisite, and the use of mathematical and biostatistical tools of the prior art is crucial for simultaneously determining the interrelationships of these multiple variables.

[0078] There are multiple methods for miRNA measurement, including hybridization-based (microarray, northern blot, bioluminescence), sequencing-based, and qPCR-based

[50] . Due to the small size of miRNAs (about 22 nucleotides), the most robust technique that provides accurate, reproducible, and accurate quantitative results with the largest dynamic range is the qPCR-based platform

[51] ; currently, it is the gold standard commonly used to validate results from other technologies (e.g., sequencing and microarray data). A variant of this method is digital PCR

[52] , which is an emerging technology based on similar principles but has not been widely recognized and used.

[0079] In this study, 203 miRNAs were profiled in plasma from 338 patients with chronic heart failure (180 HFREF and 158 HFPEF) and 208 non-heart failure patients (control group) by qPCR. This is the largest cohort to date for miRNA screening in heart failure compared to any cohort reported in the literature. Figure 1 A summary of the number of miRNAs identified by the different proposed methods used in this study is described.

[0080] The inventors of the present disclosure have established a well-designed workflow with multiple layers of technical and sample controls. This is to ensure the reliability of the assay and to minimize the possible cross-over of contaminants and technical noise. For heart failure diagnostic biomarker discovery, 203 miRNAs were screened, and the inventors detected that 137 miRNAs were expressed in all plasma samples. Among them, 75 miRNAs were identified as significantly altered between heart failure (HFREF and / or HFPEF) and control. A list of 52 miRNAs was able to distinguish HFREF from control, and 68 were found to be significantly differentially expressed between HFPEF and control. Thus, the inventors have discovered a group of miRNAs that can distinguish HFREF from HFPEF. The inventors have also discovered a group of miRNAs that are dysregulated in heart failure compared to control.

[0081] Thus, in one aspect, a method for determining whether a subject has heart failure or is at risk of developing heart failure is provided. In some instances, the method comprises the steps of: a) measuring the level of at least one miRNA from the "increased" (higher than control) miRNA list or at least one from the "decreased" (lower than control) miRNA list listed in Table 25 or Table 20 or Table 21 or Table 22 in a sample obtained from the subject. In some instances, the method further comprises: b) determining whether the level of the miRNA is different compared to control, wherein an altered level of the miRNA indicates that the subject has heart failure or is at risk of developing heart failure.

[0082] Table 20. miRNAs for heart failure detection

[0083]

[0084]

[0085]

[0086] Table 21. miRNAs identified for HFREF detection

[0087]

[0088]

[0089] Table 22. miRNAs Identified for HFPEF Detection

[0090]

[0091]

[0092] Table 25. Specific Novel microRNAs for Heart Failure Detection

[0093]

[0094]

[0095] As used throughout this disclosure, the term "miRNA" refers to microRNA, small non-coding RNA molecules, and is found in plants, animals, and some viruses. miRNAs are known to function in RNA silencing and post-transcriptional regulation of gene expression. These highly conserved RNAs regulate gene expression by binding to the 3' untranslated region (3'-UTR) of specific mRNAs. For example, each miRNA is thought to regulate multiple genes, and thus it is predicted that there are hundreds of miRNA genes in higher eukaryotes. An miRNA can be at least 10 nucleotides and no more than 35 nucleotides covalently linked together. In some instances, an miRNA can be a molecule that is 10 to 33 nucleotides in length, or 15 to 30 nucleotides in length, or 17 to 27 nucleotides in length, or 18 to 26 nucleotides in length. In some instances, an miRNA can be a molecule that is 10, or 11, or 12, or 13, or 14, or 15, or 16, or 17, or 18, or 19, or 20, or 21, or 22, or 23, or 24, or 25, or 26, or 27, or 28, or 29, or 30, or 31, or 32, or 33, or 34, or 35 nucleotides in length, where the length excludes any optionally attached and / or extended sequences (e.g., a biotin extension). miRNAs regulate gene expression and are encoded by genes from which the miRNAs are transcribed from the DNA but the miRNAs are not translated into proteins (i.e., miRNAs are non-coding RNAs). As used throughout this disclosure, a measured miRNA can have at least 90%, 95%, 97.5%, 98%, or 99% sequence identity with an miRNA listed in any of the tables provided in this disclosure. Thus, in some instances, a measured miRNA has at least 90%, 95%, 97.5%, 98%, or 99% sequence identity with an miRNA listed in any one of Table 9, Table 14, Table 15, Table 16, Table 17, Table 18, Table 19, Table 20, Table 21, Table 22, Table 23, Table 24, or Table 25. As used herein, the terms "sequence identity" and "sequence identity" refer to the relationship between two or more polypeptide sequences or two or more polynucleotide sequences (i.e., a reference sequence and a given sequence to be compared to the reference sequence). Sequence identity is determined by comparing the given sequence to the reference sequence after the sequences are optimally aligned to yield the highest degree of sequence similarity, as determined by matches between such sequences of strings. After such alignment, sequence identity is determined position by position, e.g., if at a particular position, the nucleotide or amino acid residue is the same, then at that position, the sequences are "identical". Then, the total number of such position identities is divided by the total number of nucleotides or residues in the reference sequence to give the % sequence identity. Sequence identity can be readily calculated by methods known to those of skill in the art.

[0096] As used throughout this disclosure, the term "heart failure" or "HF" refers to a complex clinical syndrome in which the pumping function of the heart becomes insufficient (ventricular dysfunction) to meet the demands of the body's vital systems and tissues. The severity of heart failure can range from not severe (mild), which is manifested by no limitation of the subject's physical activity, to increasing severity, which is manifested by the subject being unable to perform any physical activity but without discomfort. Heart failure is a progressive chronic disease that worsens over time. In extreme cases, heart failure can lead to the need for a heart transplant. In some instances, a subject can be determined to be at risk of developing heart failure if the subject may have further heart failure, such as worsening to recurrent acute decompensated heart failure or death in those with known chronic heart failure.

[0097] As used throughout this disclosure, the terms "subject" and "patient" are used interchangeably and refer to an individual or mammal suspected of being affected by heart failure. A patient can be predicted (or determined or diagnosed) to be affected by heart failure, i.e., diseased; or can be predicted not to be affected by heart failure, i.e., healthy. A subject can also be determined to be affected by a specific form of heart failure. In some instances, a heart failure patient can be a subject in which the heart failure patient is a subject that has had a preliminary diagnosis of heart failure and / or has improved symptomatically after being treated for 3 to 5 days, has resolution of the bedside signs of heart failure, and is considered suitable for discharge. Thus, a subject can be further determined to have developed heart failure or a specific form of heart failure. It should be noted that a subject determined to be healthy, i.e., not suffering from heart failure or a specific form of heart failure, may have other undetected / unknown diseases. As used herein, the subjects of this disclosure can be any mammal, including both humans and other mammals, such as animals like dogs, cats, rabbits, mice, rats, or monkeys. In some instances, the subject can be a human. Thus, the miRNA from a subject can be human miRNA, or miRNA from other mammals, such as animal miRNA like mouse, monkey, or rat miRNA, or the miRNA included in a group can be human miRNA, or miRNA from other mammals, such as animal miRNA like mouse, monkey, or rat miRNA. As shown in Table 2 in the experimental section, the subjects of this disclosure can be of Asian descent or ethnicity. In some instances, the subject can include, but is not limited to, any Asian ethnicity, including Chinese, Indian, Malay, etc.

[0098] In another aspect, the term "control" or "control subject" as used in the context of the present invention can refer to a subject known to be affected by heart failure, i.e., a diseased subject (from which a sample is obtained) (positive control, e.g., good prognosis, poor prognosis), and / or a subject suffering from heart failure subtype HFPEF and / or heart failure subtype HFREF, and / or a subject known not to be affected by heart failure, i.e., a healthy subject (negative control). It can also refer to a subject known to be affected by other diseases / conditions (from which a sample is obtained). It should be noted that a control subject known to be healthy, i.e., not suffering from heart failure, may have other undetected / unknown diseases. Thus, in some instances, the control can be a non-heart failure subject (or sometimes referred to as a normal subject). The control subject can be any mammal, including both humans and other mammals, such as animals like rabbits, mice, rats, or monkeys. In some instances, the control is a human. In some instances, the control can be an individual subject or a group of subjects (from which a sample is obtained).

[0099] Those skilled in the art will understand that the methods described herein cannot be used to replace the role of a physician in diagnosing a condition in a subject. It is understood that a clinical diagnosis of heart failure in a subject requires a physician to analyze other symptoms and / or other information available to the physician. The methods as described herein are intended to provide support or additional information for a physician to make a final diagnosis of a patient / subject.

[0100] As used throughout this disclosure herein, the term "sample" refers to a body fluid or extracellular fluid. In some instances, the body fluid can include, but is not limited to, amniotic fluid, breast milk, bronchoalveolar lavage fluid, cerebrospinal fluid, colostrum, interstitial fluid, peritoneal fluid, pleural fluid, saliva, semen, urine, tears, cellular and acellular components of whole blood, including plasma, red blood cells, white blood cells, serum, etc. In some instances, the body fluid can be blood, serum plasma, and / or plasma.

[0101] In some instances, an increase in the level of an miRNA listed as "increased" in Table 20 or Table 25 indicates that the subject has heart failure or is at risk of developing heart failure, as compared to a control.

[0102] In some instances, a decrease in the level of an miRNA listed as "decreased" in Table 20 or Table 25 indicates that the subject has heart failure or is at risk of developing heart failure, as compared to a control.

[0103] As used herein, the term "miRNA level" or "level of miRNA" in the context of the present disclosure represents the determination of the miRNA expression level (or miRNA expression profile) in a sample or a measurement related to the miRNA expression level in a sample. The miRNA expression level can be generated by any convenient means known in the art, such as, but not limited to, nucleic acid hybridization (e.g., nucleic acid hybridization with a microarray), nucleic acid amplification (PCR, RT-PCR, qRT-PCR, high-throughput RT-PCR), ELISA for quantification, next-generation sequencing (e.g., ABI SOLID, Illumina Genome Analyzer, Roche / 454GS FLX), flow cytometry (e.g., LUMINEX), etc., which allow the analysis of miRNA expression levels and comparison between samples of a subject (e.g., potentially diseased) and a control subject (e.g., a reference sample). The sample material measured by the above means can be an original or processed sample or total RNA, labeled total RNA, amplified total RNA, cDNA, labeled cDNA, amplified cDNA, miRNA, labeled miRNA, amplified miRNA, or any derivative that can be generated from the aforementioned RNA / DNA species. By determining the miRNA expression level, each miRNA is represented numerically. The higher the value of an individual miRNA, the higher the (expression) level of the miRNA, or the lower the value of an individual miRNA, the lower the (expression) level of the miRNA. When a higher value of an individual miRNA is detected that is higher than and exceeds the control, the miRNA expression is referred to as "increased" or "upregulated". On the other hand, when a lower value of an individual miRNA is detected that is lower than the control, the miRNA expression is referred to as "decreased" or "downregulated".

[0104] As used herein, "miRNA (expression) level" refers to the expression level / expression profile / expression data of a single miRNA, or at least two miRNAs, or at least 3, or at least 4, or at least 5, or at least 6, or at least 7, or at least 8, or at least 9, or at least 10, or at least 11, or at least 12, or at least 13, or at least 14, or at least 15, or at least 16, or at least 17, or at least 18, or at least 19, or at least 20, or at least 21, or at least 22, or at least 23, or at least 24, or at least 25, or at least 26, or at least 27, or at least 28, or at least 29, or at least 30, or at least 31, or at least 32, or at least 33, or at least 34, or at least 35, or more, or up to the set of expression levels of all known miRNAs.

[0105] In some instances, a method for determining whether an object has heart failure or is at risk of having heart failure may include measuring changes in the levels of at least 2, at least 3, at least 4, at least 5, at least 6, at least 7, at least 8, at least 9, at least 10, at least 11, at least 2 to at least 20, at least 10 to at least 50, at least 40 to at least 66, or all of the miRNAs listed in Table 20. In some instances, a method for determining whether an object has heart failure or is at risk of having heart failure may include measuring changes in the levels of at least 2, at least 3, at least 4, at least 5, at least 6, at least 7, at least 8, at least 9, at least 10, at least 11, or all of the miRNAs listed in Table 25.

[0106] In some instances, in the methods described herein, an increase in the level of an miRNA listed as "increased" in Table 21, as compared to a control, may indicate that the object has heart failure with reduced ejection fraction (HFREF) or may be at risk of developing heart failure with reduced ejection fraction (HFREF). In some instances, in the methods described herein, a decrease in the level of an miRNA listed as "decreased" in Table 21, as compared to a control, may indicate that the object has heart failure with reduced ejection fraction (HFREF) or may be at risk of developing heart failure with reduced ejection fraction (HFREF). In some instances, a method for determining whether an object has HFREF or is at risk of having HFREF may include measuring changes in the levels of at least 2, or at least 3, or at least 4, or at least 5, or at least 6, or at least 7, or at least 8, or at least 9, or at least 10, or at least 11, or at least 2 to at least 20, or at least 10 to at least 43, or at least 15 to at least 43, or at least 30 to at least 43, or at least 40, or all of the miRNAs listed in Table 21.

[0107] As used herein, the terms "heart failure with reduced ejection fraction (HFREF)" or "heart failure with preserved ejection fraction (HFPEF)" refer to the same terms as commonly used in the art. For example, the term HFREF may also be referred to as systolic heart failure. In HFREF, the heart muscle does not contract effectively and less oxygen-rich blood is pumped out of the body. In contrast, the term "heart failure with preserved ejection fraction (HFPEF)" refers to diastolic heart failure. In HFPEF, the heart muscle contracts normally, but the ventricles do not relax as they should during ventricular filling or during ventricular relaxation.

[0108] In some instances, in the methods described herein, an increase in the level of miRNAs listed as "increased" in Table 22, compared to a control, may indicate that the subject has heart failure with preserved ejection fraction (HFPEF) or may be at risk of developing heart failure with preserved ejection fraction (HFPEF). In some instances, in the methods described herein, a decrease in the level of miRNAs listed as "decreased" in Table 22, compared to a control, may indicate that the subject has heart failure with preserved ejection fraction (HFPEF) or may be at risk of developing heart failure with preserved ejection fraction (HFPEF). In some instances, a method for determining whether a subject may have HFPEF or may be at risk of having HFPEF may include measuring the change in the level of at least 2, or at least 3, or at least 4, or at least 5, or at least 6, or at least 7, or at least 8, or at least 9, or at least 10, or at least 11, or at least 2 to at least 20, or at least 10 to at least 50, or at least 20 to at least 55, or at least 30 to at least 60, or at least 35 to at least 60, or at least 40 to at least 60, or at least 40 to at least 62, or all of the miRNAs listed in Table 22.

[0109] In another aspect, provided is a method for determining whether a subject has heart failure selected from heart failure with reduced ejection fraction (HFREF) and heart failure with preserved ejection fraction (HFPEF), the method comprising the steps of: a) detecting (or measuring) the level of at least one miRNA listed in Table 9 in a sample obtained from the subject; and b) determining whether it is different compared to a control, wherein an alteration in the level of the miRNA may indicate that the subject has heart failure with reduced ejection fraction (HFREF) or heart failure with preserved ejection fraction (HFPEF) or may be at risk of its development.

[0110] Table 9. miRNAs differentially expressed between HFREF and HFPEF subjects

[0111]

[0112]

[0113] In some instances, in the methods described herein, an increase in the level of miRNAs listed as "increased" in Table 9, compared to a control, may indicate that the subject has heart failure with reduced ejection fraction (HFREF) or heart failure with preserved ejection fraction (HFPEF). In some instances, in the methods described herein, a decrease in the level of miRNAs listed as "decreased" in Table 9, compared to a control, may indicate that the subject has heart failure with reduced ejection fraction (HFREF) or heart failure with preserved ejection fraction (HFPEF).

[0114] In some instances of a method for determining whether a subject has heart failure selected from heart failure with reduced ejection fraction (HFREF) and heart failure with preserved ejection fraction (HFPEF), the method may include measuring the change in the level of at least 2, or at least 3, or at least 4, or at least 5, or at least 6, or at least 7, or at least 8, or at least 9, or at least 10, or at least 11, or at least 2 to at least 20, or at least 10 to at least 39, or all of the miRNAs listed in Table 9.

[0115] In some instances of a method for determining whether a subject has heart failure selected from heart failure with reduced ejection fraction (HFREF) and heart failure with preserved ejection fraction (HFPEF), the control may be a subject with heart failure with reduced ejection fraction (HFREF) or heart failure with preserved ejection fraction (HFPEF). In some instances, the control may be a patient with heart failure with reduced ejection fraction (HFREF), and the differential expression of the miRNAs listed in Table 9 indicates that the subject has heart failure with preserved ejection fraction (HFPEF). In some instances, when the control is a patient with heart failure with preserved ejection fraction (HFPEF), the differential expression of the miRNAs listed in Table 9 indicates that the subject has heart failure with reduced ejection fraction (HFREF).

[0116] In addition, the inventors of the present disclosure also detected the use of these miRNAs as prognostic markers. That is, the methods of the present disclosure can be used to predict the likely risk of future death or hospitalization events or the prognosis determined by diagnosing a disease. Prognosis in patients with heart failure means predicting the likelihood of observed survival (survival without death) or event-free survival (survival without hospitalization or death). As used herein, the terms "observed survival" or "all-cause survival" or "all-cause mortality" or "all-cause death" refer to the observed survival rate of an object considering any cause of death. This term is contrasted with the term "event-free survival (EFS)", which means the absence of readmission due to recurrence of heart failure (i.e., the length of time after heart failure treatment during which readmission due to decompensated heart failure is avoided) and any cause of death.

[0117] The inventors of the present disclosure found that many miRNAs were found to be good predictors of the combination of observed (all-cause) survival (OS) (i.e., the observed survival rate due to all causes of death) or event-free survival (EFS) of readmission due to recurrent heart failure (i.e., the length of time after heart failure treatment during which readmission due to decompensated heart failure is avoided) and all-cause death in patients with chronic heart failure. Accordingly, the present disclosure can also be used in a method for predicting the prognosis of an object. Accordingly, in another aspect of the present disclosure, there is provided a method for determining the risk of change in the risk of death (or decrease in observed (all-cause) survival rate) in a patient with heart failure. In some examples, the method may comprise the steps of: a) detecting the level of at least one miRNA listed in Table 14 in a sample obtained from an object, and / or measuring the level of at least one miRNA listed in Table 14; and b) determining whether the level of at least one miRNA listed in Table 14 is different compared to the level of the miRNA in a control population, wherein a change in the level of the miRNA indicates that the risk of death (change in observed (all-cause) survival rate) of the object may be different compared to the control population.

[0118] Table 14. miRNAs that can be used to predict observed survival

[0119]

[0120]

[0121] As used herein, the term "hazard ratio" refers to a term commonly known in the art and relates to a ratio or estimate of the likelihood of "death" or "hospitalization" per unit time at a particular moment (where the known subject "survives" until that moment ("death" or "hospitalization")). The hazard ratio is used to measure the magnitude of the difference between two survival curves. A hazard ratio (HR) > 1 indicates a higher risk of having a short survival time, and a hazard ratio (HR) < 1 indicates a higher risk of having a longer survival time. As is known in the art, the hazard ratio can be calculated by the Cox proportional hazards (CoxPH) model.

[0122] In some instances, in the methods described herein, an increase in the level of an miRNA listed as "hazard ratio > 1" in Table 14, compared to a control, may indicate an increased risk of death (decreased observed (all-cause) survival rate) in a subject. In some instances, in the methods described herein, a decrease in the level of an miRNA listed as "hazard ratio > 1" in Table 14, compared to a control, may indicate a decreased risk of death (increased observed (all-cause) survival rate) in a subject.

[0123] In some instances, in the methods described herein, an increase in the level of an miRNA listed as "hazard ratio < 1" in Table 14, compared to a control, may indicate a decreased risk of death (increased observed (all-cause) survival rate) in a subject. In some instances, in the methods described herein, a decrease in the level of an miRNA listed as "hazard ratio < 1" in Table 14, compared to a control, may indicate an increased risk of death (decreased observed (all-cause) survival rate) in a subject.

[0124] In another aspect, a method for determining the risk of a change in the risk of a heart failure patient's disease progression to hospitalization or death (decreased event-free survival) is provided. In some instances, the method comprises the steps of: a) detecting the level of at least one miRNA listed in Table 15 in a sample obtained from a subject and / or measuring the level of at least one miRNA listed in Table 15, and b) determining whether the level of at least one miRNA listed in Table 15 is different compared to the level of the miRNA in a control population, wherein a change in the level of the miRNA indicates that the risk of the subject's disease progression to hospitalization or death may have changed (changed event-free survival) compared to the control population.

[0125] Table 15. miRNAs Predictive of Event-Free Survival

[0126]

[0127]

[0128] In some instances, in the methods described herein, an increase in the level of miRNAs listed as "hazard ratio > 1" in Table 15, compared to a control, may indicate an increased risk (decreased event - free survival) of the subject's disease progressing to hospitalization or death. In some instances, in the methods described herein, a decrease in the level of miRNAs listed as "hazard ratio > 1" in Table 15, compared to a control, may indicate a decreased risk (increased event - free survival) of the subject's disease progressing to hospitalization or death.

[0129] In some instances, in the methods described herein, an increase in the level of miRNAs listed as "hazard ratio < 1" in Table 15, compared to a control, may indicate an increased risk (increased event - free survival) of the subject's disease progressing to hospitalization or death. In some instances, in the methods described herein, a decrease in the level of miRNAs listed as "hazard ratio < 1" in Table 15, compared to a control, may indicate an increased risk (decreased event - free survival) of the subject's disease progressing to hospitalization or death.

[0130] In some instances, in the methods described herein, the control can be a control population or cohort of heart - failure subjects. In some instances, the control population can be a population or cohort of heart - failure patients, in which the microRNA expression levels and the risk of death or disease progression can be determined. In some instances, the microRNA expression level of the control population can be the average or median expression level of all subjects in the population (including the patient under discussion). In some instances, if 10% of the patients in the control population die within 5 years, the 5 - year death risk of the control population is 10%. In some instances, the control population includes heart - failure patients whose risk of death or disease progression needs to be determined using microRNA expression levels.

[0131] In some instances, in the methods described herein, a heart - failure patient can be a subject who has had a preliminary diagnosis of heart failure and / or has been treated for 3 to 5 days when symptoms improve (the bedside physical symptoms of heart failure subside) and is considered suitable for discharge. In some instances, the patient can be a stable compensated heart - failure patient who has not further deteriorated into recurrent acute decompensated heart failure that requires readmission or death.

[0132] In another aspect, a method for determining the risk of developing heart failure in a subject or determining whether a subject has heart failure is provided, comprising the steps of: (a) detecting the presence of miRNA in a sample obtained from the subject and / or measuring the levels of at least three miRNAs listed in Table 16 or Table 23 in the sample; and (b) using a score based on the levels of the miRNAs measured in step (a) to predict the likelihood that the subject has or is developing heart failure. In some instances, the method may further comprise measuring the level of at least one miRNA listed in the "not significant group" in Table 16 or Table 23, and wherein the at least one miRNA is hsa-miR-10b-5p.

[0133] As used herein throughout the present disclosure and with reference to all methods described herein, the term "score" refers to an integer or number that can be determined mathematically, for example, by using computational models known in the art (as an example, which may include but is not limited to SMV), and calculated using any one of a number of mathematical equations and / or algorithms known in the art for statistical classification purposes. Such scores are used to list one result from a spectrum of possible results. The relevance and statistical significance of such scores depend on the size and quality of the underlying data set used to establish the result spectrum. For example, a blind sample can be input into an algorithm, which then calculates a score based on the information provided by the analysis of the blind sample. This enables the generation of a score for the blind sample. Based on this score, a decision can be made, for example, about how likely it is that the patient from whom the blind sample was obtained has or does not have heart failure. The ends of the spectrum can be defined logically based on the data provided, or arbitrarily according to the requirements of the experimenter. In both cases, the spectrum needs to be defined before testing the blind sample. As a result, a score (e.g., the number "45") generated from such a blind sample can indicate that the corresponding patient has heart failure based on a spectrum defined on a scale of 1 to 50, where "1" is defined as not having heart failure and "50" is defined as having heart failure. Thus, the term "score" refers to a mathematical score that can be calculated using any one of a number of mathematical equations and / or algorithms known in the art for statistical classification purposes. Examples of such mathematical equations and / or algorithms can be, but are not limited to, (statistical) classification algorithms selected from the following: support vector machine algorithm, logistic regression algorithm, polynomial logistic regression algorithm, Fisher linear discriminant algorithm, quadratic classifier algorithm, perceptron algorithm, k-nearest neighbor algorithm, artificial neural network algorithm, random forest algorithm, decision tree algorithm, naive Bayes algorithm, adaptive Bayesian network algorithm, and ensemble learning methods that combine multiple learning algorithms. In another example, a classification algorithm is pre-trained using the expression levels of a control. In some examples, the classification algorithm compares the expression level of an object with the expression level of a control and returns a mathematical score that determines the likelihood that the object belongs to any one of the control groups. In some examples, the classification algorithm can compare the expression level of an object with the expression level of a control and return a mathematical score that determines the likelihood that the object belongs to any one of the control groups. Some examples of algorithms that can be used in the present disclosure are provided below.

[0134] Table 16. miRNAs for multivariate detection of heart failure

[0135]

[0136]

[0137] Table 23. miRNAs for heart failure detection

[0138]

[0139]

[0140] In some instances, the methods disclosed herein measure changes in the levels of at least 4, or at least 5, or at least 6, or at least 7, or at least 8, or at least 9, or at least 10, or at least 11, or at least 2 to at least 20, or at least 10 to at least 45, or at least 40 to at least 50, or all of the miRNAs listed in Table 16. In some instances, the methods disclosed herein measure at least 4, or at least 5, or at least 6, or at least 7, or at least 8, or at least 9, or at least 10, or at least 11, or at least 2 to at least 20, or at least 10 to at least 41, or all of the miRNAs listed in Table 23.

[0141] In some instances, the methods disclosed herein measure the level of at least one (or more) miRNA(s) (in a plasma sample of a subject). The measurements of at least one miRNA can be combined to generate a score for predicting heart failure or classifying HFREF and HFPEF subtypes. In some instances, the equation for generating the score can be Equation 1, which is as follows:

[0142] Equation 1-

[0143]

[0144] where log2 copy_miRNA i is the log-transformed copy number of the individual miRNA (copies / ml plasma); K i is a coefficient for weighting multiple miRNA targets; and B is a constant value that adjusts the scale of the predicted score.

[0145] Equation 1 herein shows the application of a linear model for predicting heart failure or classifying HFREF and HFPEF subtypes. The predicted score (unique for each subject) is a numerical value for making predictive or diagnostic decisions.

[0146] In the experimental section of the present disclosure, further statistical evaluation of the diagnostic utility of the identified miRNAs was performed. Then, a multivariate miRNA biomarker panel (HF group, HFREF, and HFPEF groups) was developed using repeated computer cross-validation via sequence forward floating search (SFFS)

[53] and support vector machine (SVM)

[54] . The inventors of the present disclosure found that some of the miRNAs in the biomarker panel consistently produced an AUC value (area under the curve) of ≥0.92 for HF detection in the receiver operating characteristic (ROC) plot (Figure 20, B) and an AUC ≥0.75 for subtype classification ( Figure 24 , A). When used in combination with NT-proBNP, the miRNA panel showed significantly improved discrimination and better classification accuracy for both purposes ( Figure 22 , B; and 24, B).

[0147] Accordingly, in another aspect, provided is a method for determining the risk of developing heart failure in a subject or determining whether a subject has heart failure, comprising the steps of: (a) detecting the presence of miRNAs in a sample obtained from the subject, and / or measuring the levels of at least two RNAs listed in Table 17 in the sample; and (b) using a score based on the levels of the miRNAs measured in step (a) to predict the likelihood that the subject has or has heart failure.

[0148] Table 17. miRNAs Used in Combination with NP-proBNP for Heart Failure Detection

[0149]

[0150]

[0151]

[0152] In some instances, the methods described herein may further comprise the step of determining the levels of brain natriuretic peptide (BNP) and / or N-terminal pro-brain natriuretic peptide (NT-proBNP). In some instances, both NT-proBNP and BNP are good markers for the prognosis and diagnosis of heart failure (e.g., chronic heart failure).

[0153] In some instances, the methods described herein may measure the altered levels of at least 3, or at least 4, or at least 5, or at least 6, or at least 7, or at least 8, or at least 9, or at least 10, or at least 11, or at least 2 at least 20, or at least 10 at least 45, or at least 40 at least 48, or all of the miRNAs listed in Table 17.

[0154] In some instances, in the methods disclosed herein, when BNP and / or NT-proBNP are used in combination with miRNA, alternatively, Formula 2 can be used. In Formula 2, the levels of BNP / NT-proBNP in the plasma sample are included in the linear model. In one instance, Formula 2 is as follows:

[0155] Formula 2-

[0156]

[0157] where log2 copy_miRNA i is the log-transformed copy number of the individual miRNA (copies / ml plasma); K i is the coefficient used to weight multiple miRNA targets; B is a constant value that adjusts the scale of the prediction score; BNP is a measure that is positively or negatively correlated with the levels of BNP and / or NT-proBNP in the sample.

[0158] Additionally, for predicting heart failure, the prediction score (which is unique for each subject) is a numerical value indicating the likelihood that the subject has heart failure. In some instances, the result of the method described herein (i.e., the prediction of likelihood or diagnosis) can be found in Formula 3. If the value is higher than a preset cut-off value, the subject will be diagnosed or predicted to have heart failure. If the value is lower than the preset cut-off value, the subject will be diagnosed or predicted not to have heart failure. Formula 3 is as follows

[0159] Formula 3-

[0160]

[0161] In some instances, for the classification of HFREF and HFPEF subtypes, the prediction score (which is unique for each subject) can be a numerical value indicating the likelihood that a heart failure subject has heart failure of the HFPEF subtype. In some instances, the diagnostic result can be found in Formula 4. If the value is higher than a preset cut-off value, the heart failure subject will be diagnosed as (or predicted to have) heart failure of the HFPEF subtype. If the value is lower than the preset cut-off value, the heart failure subject will be diagnosed as (or predicted to have) heart failure of the HFREF subtype by this test.

[0162] Formula 4-

[0163]

[0164] In another aspect, a method for determining the likelihood that a subject has heart failure with reduced ejection fraction (HFREF) or heart failure with preserved ejection fraction (HFPEF) is provided. In some instances, the method comprises the steps of: (a) detecting the presence of miRNA in a sample obtained from the subject, and / or measuring the levels of at least three miRNAs listed in Table 18 in the sample; and (b) using a score based on the levels of the miRNAs measured in step (a) to predict the likelihood that the subject has heart failure with reduced ejection fraction (HFREF) or heart failure with preserved ejection fraction (HFPEF).

[0165] Table 18. miRNAs for determining heart failure subtype classification

[0166]

[0167]

[0168] In some instances, the methods described herein can measure the altered levels of at least 4, at least 5, at least 6, at least 7, at least 8, at least 9, at least 10, at least 11, at least 2 to at least 20, at least 10 to at least 30, at least 40 to at least 45, or all of the miRNAs listed in Table 18.

[0169] In some instances, the score in the methods disclosed herein can be calculated by an equation provided herein. In some instances, the equation can be at least one of the equations including, but not limited to, Equation 1 and / or Equation 2. The results of the methods disclosed herein can be determined by an equation such as, but not limited to, Equation 3, Equation 4, etc.

[0170] In another aspect, a method for determining the likelihood that a subject has heart failure with reduced ejection fraction (HFREF) or heart failure with preserved ejection fraction (HFPEF) is provided, comprising the steps of: (a) detecting the presence of miRNA in a sample obtained from the subject, and / or measuring the levels of at least two RNAs listed in Table 19 or Table 24 in the sample; and (b) using a score based on the levels of the miRNAs measured in step (a) to predict the likelihood that the subject has heart failure with reduced ejection fraction (HFREF) or heart failure with preserved ejection fraction (HFPEF).

[0171] Table 19. miRNAs used in combination with NT-proBNP in classifying heart failure subtypes

[0172]

[0173]

[0174] Table 24. miRNAs Used in Combination with NT-proBNP in Classifying Heart Failure Subtypes

[0175]

[0176]

[0177] As described in the experimental section and in the accompanying drawings as Figure 22 and 24 illustrated by way of example, when the methods of the present disclosure are used in combination with additional steps for determining NT-proBNP, the methods provide surprisingly accurate predictions. Thus, in some instances, the methods may further include the step of determining the levels of brain natriuretic peptide (BNP) and / or N-terminal pro-brain natriuretic peptide hormone (NT-proBNP).

[0178] In some instances, the methods described herein may measure at least 3, or at least 4, at least 5, at least 6, at least 7, at least 8, at least 9, at least 10, at least 11, at least 2 to at least 20, at least 10 to at least 30, or all of the miRNAs listed in Table 19, or the altered levels of at least 4, at least 5, at least 6, at least 7, at least 8, at least 9, at least 10, at least 11, at least 2 to at least 20, at least 10 to at least 41, or all of the miRNAs listed in Table 24.

[0179] In some instances, when compared to a control, the level of at least one miRNA measured in step (b) does not change in the subject. In such instances, the miRNAs whose levels do not change in the subject when compared to the control are the miRNAs listed as "not significant" in the respective table.

[0180] In some instances, the score in the methods disclosed herein can be calculated by Equation 2.

[0181] As will be appreciated by those skilled in the art, the classification algorithms used in any of the methods described herein can be pre-trained using the expression levels of a control. In instances where a pre-existing clinical data is used to pre-train the classification algorithm, the control can be at least one selected from a control without heart failure (normal) and patients with heart failure. The control can include a population of subjects with and / or without heart failure (i.e., without heart failure). Thus, in some instances of the methods disclosed herein, the control can include, but is not limited to, a control without heart failure and patients with heart failure, patients with HFPEF subtype heart failure, patients with HFREF subtype heart failure, etc.

[0182] The present disclosure discusses differential comparison of miRNA expression levels in establishing a miRNA profile, based on which it can be determined whether a subject is at risk of developing heart failure or whether the subject has heart failure. As disclosed herein, the methods disclosed herein require differential comparison of miRNA expression levels typically from different groups. In one example, the comparison is made between two groups. These comparison groups can be defined as but not limited to heart failure, no heart failure (normal). In the heart failure group, further subgroups can be found, such as but not limited to HFREF and HFPEF. Differential comparisons can also be made between these groups described herein. Thus, in some examples, the expression level of miRNA can be expressed as but not limited to concentration, log(concentration), threshold cycle / quantification cycle (Ct / Cq) number, 2 to the power of the threshold cycle / quantification cycle (Ct / Cq) number, etc.

[0183] In any of the methods described herein, the method may further include but not be limited to the steps of: obtaining a sample from a subject at different time points, monitoring the progression of heart failure, staging heart failure, measuring the miRNA level and / or NT-proBNP level in the subject (from whom the sample is obtained), etc.

[0184] In some examples, based on the current cohort described in the following experimental section, a biomarker panel including multiple miRNAs or a biomarker panel including multiple miRNAs and BNP / NT-proBNP can be developed. Prediction score calculation can be by methods known in the art, such as optimized with a linear SVM model. In some examples, biomarker panels consisting of different numbers of miRNA targets can be optimized by SFFS and SVM, where the AUC is optimized for heart failure prediction (Table 26) or heart failure subtype classification (Table 27). Exemplary formulas, cut-off values, and performance are provided in the tables.

[0185] Table 26. Exemplary biomarker panels for HF detection

[0186]

[0187]

[0188] As used in Table 26, the symbol "*" refers to "×" or the multiplication sign; "-" refers to a negative value; "+" refers to addition; "log2(BNP)" refers to the log2 value of BNP expression. The second column of Table 26 also shows exemplary formulas used for calculating a score in the method as used herein. In this formula, the unit of measurement for the microRNA is copies / ml plasma, while for NT-proBNP it is pg / ml plasma. It will be apparent to those skilled in the art that the coefficients and cut-off values in the formula must be adjusted according to the different detection systems used for measurement and / or the different units used to represent the microRNA expression level and BNP level / type. The adjustment of the formula will not exceed the skills of a person of ordinary skill in the art.

[0189] Accordingly, in another aspect, there is provided a method for determining the risk of developing heart failure in a subject or determining whether a subject has heart failure, comprising the steps of: (a) detecting the presence of miRNAs in a selected group listed in Table 26 in a sample obtained from the subject, and / or measuring the level of miRNAs in the selected group of Table 26 in the sample; and (b) assigning a score based on the level of miRNAs measured in step (a) to predict the likelihood that the subject has or has developed heart failure. In one example, the score is calculated based on the formula listed in Table 26. In one example, when a biomarker panel of two miRNAs is required, the method can detect and measure the level of the miRNAs listed as the "panel of 2 miRNAs" in Table 26. In some examples, when a biomarker panel of three miRNAs is required, the method can detect and measure the level of the miRNAs listed as the "panel of 3 miRNAs" in Table 26. In some examples, when a biomarker panel of four miRNAs is required, the method can detect and measure the level of the miRNAs listed as the "panel of 4 miRNAs" in Table 26. In some examples, when a biomarker panel of 5 miRNAs is required, the method can detect and measure the level of the miRNAs listed as the "panel of 5 miRNAs" in Table 26. In some examples, when a biomarker panel of 6 miRNAs is required, the method can detect and measure the level of the miRNAs listed as the "panel of 6 miRNAs" in Table 26. In some examples, when a biomarker panel of 7 miRNAs is required, the method can detect and measure the level of the miRNAs listed as the "panel of 7 miRNAs" in Table 26. In some examples, when a biomarker panel of 8 miRNAs is required, the method can detect and measure the level of the miRNAs listed as the "panel of 8 miRNAs" in Table 26. In some examples, the method can be performed together with an additional step of detecting and measuring the level of NTproBNP in its sample.

[0190] In some instances, for predicting heart failure, a prediction score (which is unique for each subject) is a numerical value indicating the likelihood that a subject has heart failure. In some instances, the result of the methods described herein (i.e., the prediction of likelihood or diagnosis) can be found in Equation 3. If the value is above a preset cutoff value, the subject will be diagnosed or predicted to have heart failure. If the value is below the preset cutoff value, the subject will be diagnosed or predicted not to have heart failure. Equation 3 is as follows:

[0191] Equation 3-

[0192]

[0193] Table 27. Exemplary biomarker sets for heart failure subtype classification

[0194]

[0195]

[0196] As used in Table 27, the symbol "*" refers to "×" or the multiplication sign; "-" refers to a negative value; "+" refers to addition; "log2(BNP)" refers to the log2 value of BNP expression. The second column of Table 27 also shows an exemplary equation used for calculating a score in the methods used herein. In this equation, the measurement unit for microRNA is copies / ml plasma, while for NT-proBNP it is pg / ml plasma. It will be apparent to those skilled in the art that the coefficients and cutoff values in the equation must be adjusted according to the different detection systems used for measurement and / or the different units used to represent the microRNA expression level and BNP level / type. The adjustment of the equation will not be beyond the skills of a person of ordinary skill in the art.

[0197] In another aspect, a method for determining the likelihood that a subject has heart failure with reduced ejection fraction (HFREF) or heart failure with preserved ejection fraction (HFPEF) is provided, comprising the steps of: (a) detecting the presence of miRNAs in a selected group listed in Table 27 in a sample obtained from the subject, and / or measuring the levels of miRNAs listed in the selected group of Table 27 in the sample; and (b) assigning a score based on the levels of the miRNAs measured in step (a) to predict the likelihood that the subject has heart failure with reduced ejection fraction (HFREF) or heart failure with preserved ejection fraction (HFPEF). In one example, the score is calculated based on the formula listed in Table 27. In one example, when a biomarker panel of two miRNAs is required, the method can detect and measure the levels of the miRNAs listed as the "panel of 2 miRNAs" in Table 27. In some examples, when a biomarker panel of three miRNAs is required, the method can detect and measure the levels of the miRNAs listed as the "panel of 3 miRNAs" in Table 27. In some examples, when a biomarker panel of four miRNAs is required, the method can detect and measure the levels of the miRNAs listed as the "panel of 4 miRNAs" in Table 27. In some examples, when a biomarker panel of 5 miRNAs is required, the method can detect and measure the levels of the miRNAs listed as the "panel of 5 miRNAs" in Table 27. In some examples, when a biomarker panel of 6 miRNAs is required, the method can detect and measure the levels of the miRNAs listed as the "panel of 6 miRNAs" in Table 27. In some examples, the method can be performed in conjunction with an additional step of detecting and measuring the level of NTproBNP in the sample.

[0198] In some examples, for classifying HFREF and HFPEF subtypes, the prediction score (which is unique for each subject) can be a numerical value indicating the likelihood that a heart failure subject has heart failure of the HFPEF subtype. In some examples, the result of the diagnosis can be found in Equation 4. If the value is above a preset cut-off value, the heart failure subject will be diagnosed as (or predicted to have) heart failure of the HFPEF subtype. If the value is below the preset cut-off value, the heart failure subject will be diagnosed as (or predicted to be) having heart failure of the HFREF subtype by this test.

[0199] Equation 4-

[0200]

[0201] In one example, the methods described herein can be implemented as a device capable of (or adapted to) performing all (or some) of the steps described in the present disclosure. Thus, in one example, the present disclosure provides a device adapted to (or capable of) performing the methods described herein.

[0202] In another aspect, a kit for use in (or adapted for use in, or when used in) any of the methods described herein is provided. In one example, the kit may comprise reagents for determining the expression of at least one gene listed in Table 9, or at least one gene listed in Table 14, or at least one gene listed in Table 15, or at least two genes listed in Table 16, or at least two genes listed in Table 17, or at least two genes listed in Table 18, or at least two genes listed in Table 19, or at least one gene listed in Table 20, or at least one gene listed in Table 21, or at least one gene listed in Table 22, or at least one gene listed in Table 23, or at least one gene listed in Table 24, or at least one gene listed in Table 25.

[0203] In some examples, the reagents may comprise probes, primers, or primer sets adapted or capable of determining the expression of at least one gene listed in Table 9, or at least one gene listed in Table 14, or at least one gene listed in Table 15, or at least two genes listed in Table 16, or at least two genes listed in Table 17, or at least two genes listed in Table 18, or at least two genes listed in Table 19, or at least one gene listed in Table 20, or at least one gene listed in Table 21, or at least one gene listed in Table 22, or at least one gene listed in Table 23, or at least one gene listed in Table 24, or at least one gene listed in Table 25.

[0204] In some examples, the kit may further comprise reagents for determining the levels of brain natriuretic peptide (BNP) and / or N-terminal pro-brain natriuretic peptide (NT-proBNP).

[0205] In some instances, the methods described herein may further comprise treating a subject predicted (or diagnosed as) having heart failure or a subtype of heart failure with at least one therapeutic agent for treating heart failure (or a subtype of heart failure). In some instances, the methods may further comprise treatments known to relieve and / or mitigate the symptoms of heart failure. In some instances, the methods described herein may further comprise administering drugs including but not limited to classes of drugs proven to improve the prognosis of heart failure (e.g., ACEI / ARB, angiotensin receptor blockers, loop / thiazide diuretics, β-blockers, mineralocorticoid antagonists, aspirin or Plavix, statins, digoxin, warfarin, nitrates, calcium channel blockers, spironolactone, fibrates, anti-diabetic agents, hydralazine, iron supplements, anticoagulants, antiplatelet agents, etc.).

[0206] The invention illustratively described herein may suitably be practiced in the absence of any one or more of the elements, limitations specifically disclosed herein. Thus, for example, the terms "comprising", "including", "containing", etc. should be construed broadly and without limitation. Additionally, the terms and expressions used herein have been used as terms of description and not of limitation, and there is no intention in the use of such terms and expressions of excluding any equivalents of the features shown and described or portions thereof, but it is recognized that various modifications are possible within the scope of the claimed invention. Accordingly, it should be understood that although the invention has been specifically disclosed by some preferred embodiments and optional features, modifications and variations of the invention embodied herein disclosed by those skilled in the art, and such modifications and variations are considered to be within the scope of the invention.

[0207] The invention has been described herein in broad and general terms. Each of the narrower categories and sub-groupings falling within the broad disclosure also form part of the invention. This includes the broad description of the invention with the proviso or negative limitation of removing any subject matter from the broad disclosure, whether or not that removed material is specifically recited herein.

[0208] Other embodiments are within the following claims and non-limiting examples. Additionally, where features or aspects of the invention are described in terms of a Markush group, those skilled in the art will recognize that the invention is also thereby described in terms of any individual member or subgroup of members of the Markush group.

[0209] Experimental section

[0210] Methods

[0211] Pre-analysis (sample collection and miRNA extraction): Plasma samples were stored frozen at -80 °C until use. The well-established TRI reagent was used Isolate total RNA from 200 μl of each plasma sample according to the manufacturer's protocol. Plasma contains trace amounts of RNA. To reduce RNA loss and monitor extraction efficiency, a rationally designed isolation enhancer agent (MS2) and a spiked control RNA (MiRXESTM) are added to the samples before isolation.

[0212] RT-qPCR: The isolated total RNA and synthetic RNA standards are converted to cDNA in an optimized multiplex reverse transcription reaction with a second set of spiked control RNAs to detect the presence of inhibitors and monitor RT-qPCR efficiency. Use Improm II Reverse transcriptase to perform reverse transcription according to the manufacturer's instructions. Then, the synthesized cDNA is subjected to a multiplex amplification step and quantified using a Sybr Green-based singleplex qPCR assay (compliant with MIQE) (MiRXESTM). Use Applied ViiA 7 384 Real-Time PCR System or CFX384 Touch Real-Time PCR Detection System to perform the qPCR reaction. Figure 2 Summarizes the overview and details of the miRNA RT-qPCR measurement workflow.

[0213] Data processing: Process the raw cycle threshold (Ct) values and determine the absolute copy number of the target miRNA in each sample by interpolation of a synthetic miRNA standard curve. Normalize the technical variations introduced during RNA isolation and RT-qPCR by spiked control RNA. To analyze single miRNAs, further normalize biological variations by a set of validated endogenous reference miRNAs that are stably expressed in all control and diseased samples.

[0214] Results

[0215] I. Characteristics of study participants

[0216] Conduct a well-designed clinical study (case-control study) to ensure the accurate identification of biomarkers for chronic heart failure (HF). A total of 338 chronic heart failure patients from the Singapore cohort (180 HFREF and 158 HFPEF) were used in this study and compared with 208 non-heart failure subjects matched for race, gender, and age (as the control group). Heart failure patients were recruited from the Singapore Heart Failure Outcomes and Phenotypes (SHOP) study

[55] . Patients were included if they had an initial diagnosis of acute decompensated heart failure (ADHF) or were diagnosed and managed for heart failure within 6 months of a known ADHF event. Controls without a history of significant coronary artery disease or heart failure were recruited through the ongoing Singapore Longitudinal Ageing Study (SLAS)

[56] . All patients and controls underwent a detailed clinical examination (including comprehensive Doppler echocardiography) to determine the presence of clinical heart failure. The LVEF was evaluated using the biplane Simpson's method recommended by the American Society of Echocardiography (ASE) guidelines. Patients with verified heart failure and an LVEF ≥ 50% were classified as HFPEF, while those with an LVEF ≤ 40% were classified as HFREF. HF patients with an LVEF between 40% and 50% were excluded. Assessments, including plasma samples, were deliberately made when the patients had been treated (usually 3 to 5 days), symptoms had improved (where the bedside physical symptoms of heart failure had resolved), and they were considered fit for discharge. This ensured the evaluation of biomarker performance in treated or "chronic" stage heart failure. Clinical characteristics and demographic information are provided in Table 2. All plasma samples were stored at -80 °C before use.

[0217] Table 2. Clinical information of the subjects included in the study

[0218]

[0219]

[0220]

[0221]

[0222]

[0223]

[0224]

[0225]

[0226]

[0227]

[0228]

[0229]

[0230]

[0231]

[0232] Plasma NT-proBNP was measured in all samples by electrochemiluminescence immunoassay (Elecsys proBNP II assay) on an automated Cobase 411 analyzer according to the manufacturer's instructions (Roche Diagnostics GmbH, Mannheim, Germany). Preliminary examination of the distribution in the control, HFREF, and HFPEF groups ( Figure 2 , A to C) showed that NT-proBNP levels were positively skewed (skewness / skew > 2) in all groups. Since the statistical methods to be applied require a non-skewed distribution (Student t-distribution or logistic distribution), the natural logarithm of NT-proBNP was calculated to generate a new variable with skewness close to zero: ln_NT-proBNP( Figure 2 , D to F). ln_NT-proBNP was used for all analyses involving NT-proBNP.

[0233] The characteristics of the subject groups are summarized in Table 3.

[0234] Table 3. Characteristics of healthy subjects and heart failure subjects

[0235]

[0236]

[0237] In addition to demographic variables including age, race, and sex, key clinical variables of HF including LVEF, ln_NT-proBNP, body mass index (BMI), atrial fibrillation or atrial flutter (AF), hypertension, and diabetes were recorded. The mean LVEF of HFPEF patients (60.7 ± 5.9) was similar to that of healthy control subjects (64.0 ± 3.7), while as expected and by patient selection and assignment, HFREF patients significantly had lower LVEF (25.9 ± 7.7). The Student t-test was used for comparison of numerical variables, and the chi-square test was used for comparison of categorical variables between controls and HF (C vs. HF, Table 3) and between HFPEF and HFREF (HFREF vs. HFPEF, Table 3). Generally, the older the HF patients were, the higher the prevalence of hypertension, AF, and diabetes was compared with the controls. There were differences in the distribution of sex, age, BMI, hypertension, and AF between HFREF and HFPEF patients. All these variables with different distributions were included in the discovery of miRNA biomarkers for HF detection or for HF subtype classification by multivariate logistic regression.

[0238] Ln_NT-proBNP was lower in HFPEF compared with HFREF, and some results were below the ESC-promoted NT-proBNP cutoff (<125 pg / ml) for diagnosing HF in the non-acute setting

[57] . In HFPEF, the loss of NT-proBNP test performance was evident ( Figure 3 , A). The performance of ln_NT-proBNP as a biomarker for HF diagnosis was detected by ROC analysis. In this study, overall, ln_NT-proBNP had an AUC (area under the ROC curve) of 0.962 for HF diagnosis. It performed better for detecting HFREF (AUC = 0.985) compared with HFPEF (AUC = 0.935) ( Figure 3 , B to D). For classifying HFREF and HFPEF subtypes, ln_NT-proBNP showed an AUC of only 0.706 ( Figure 3 , C).

[0239] II. MiRNA Measurement

[0240] Circulating cell-free miRNAs in the blood are derived from various organs and blood cells

[58] . Thus, miRNA level changes caused by heart failure may be partially masked by the presence of the same miRNAs that may be secreted from other sources due to other stimuli. Therefore, determining the differential expression levels of miRNAs found in heart failure and control groups can be challenging. Additionally, the abundance of most cell-free miRNAs in the blood is very low

[59] . Therefore, accurately measuring multiple miRNA targets from a limited volume of serum / plasma is crucial and highly challenging. To maximize the discovery of significantly altered miRNA expression and to identify a multivariate miRNA biomarker panel for the diagnosis of heart failure, as an alternative to using low-sensitivity or semi-quantitative screening methods (microarrays, sequencing), the inventors of this study chose to perform qPCR-based assays with a specially designed workflow ( Figure 4 ).

[0241] All qPCR assays (designed by MiRXESTM, Singapore) were performed at least twice for miRNA targets in single replicates and for synthetic RNA "spike-in" controls in at least quadruplicate. To ensure the accuracy of the results of the high-throughput qPCR study, this study was designed and established through multiple iterations a robust workflow for the discovery of circulating biomarkers (refer to "Methods" and Figure 4 ). In this new workflow, different "spike-in" controls were used to monitor and correct for technical variations during isolation, reverse transcription, amplification, and qPCR. All spike-in controls were non-natural synthetic miRNA mimics (small single-stranded RNAs between 22 and 24 bases in length) that were computationally designed to have unexpectedly low similarity to all known human miRNAs, thereby minimizing cross-hybridization with the primers used in the assay. Additionally, the miRNA assays were computationally divided into multiple multiplexes to minimize non-specific amplification and primer-primer interactions. Synthetic miRNAs were used to construct standard curves for interpolating absolute copy numbers in all measurements, thereby further correcting for technical variations. Predictably, through this highly robust workflow and multiple control levels, this study was able to identify low expression levels of miRNAs in circulation, and the methods of this study are highly reliable and ensure the reproducibility of the data.

[0242] Two hundred and three (203) miRNA targets were selected for this study based on prior knowledge of highly expressed plasma miRNAs (data not shown), and the expression levels of those miRNAs were quantitatively measured in all 546 plasma samples (HF and controls) using a highly sensitive qPCR assay (designed by MiRXESTM, Singapore).

[0243] In the current experimental design, total RNA including miRNA was extracted from 200 μl of plasma. The extracted RNA was reverse transcribed and amplified by touchdown amplification to increase the amount of cDNA without changing the total miRNA expression level ( Figure 4 ). Then, the increased cDNA was diluted for qPCR measurement. Simple calculations based on dilution effects revealed that miRNAs expressed at levels ≤500 copies / ml in serum would be quantified at levels approaching the detection limit (≤10 copies / well) of singleplex qPCR assays. At such concentrations, measurement would be a major challenge due to technical limitations (errors in pipetting and qPCR reactions). Thus, miRNAs expressed at concentrations ≤500 copies / ml were excluded from the analysis and were considered undetectable.

[0244] Approximately 70% (n = 137) of the total miRNAs assayed were found to be highly expressed in all samples. These 137 miRNAs were detected in >90% of the samples (expression levels ≥500 copies / ml; Table 4). Compared to the publicly available data (Table 1), the inventors of this study detected far more miRNAs that had not previously been reported in heart failure, highlighting the importance of using a careful and well-controlled experimental design.

[0245] Table 4. Sequences of 137 reliably detected mature miRNAs

[0246]

[0247]

[0248]

[0249]

[0250] III. MiRNA Biomarkers

[0251] First, all measured miRNAs were examined to obtain targets that were detectable only in heart failure samples but not in control samples. miRNAs specifically secreted by the myocardium in heart failure patients are ideal biomarkers for detecting the disease. Since miRNAs in the blood circulation system are known to be contributed by different organs and / or cell types (including the myocardium), it is not surprising that these miRNAs may be represented in the plasma of both normal and heart failure patients. However, the differential expression of these miRNAs in plasma can still be used as a useful biomarker during the development of heart failure.

[0252] Global unsupervised analysis (principal component analysis, PCA) was initially performed on the expression levels of all detected plasma miRNAs (137, Table 4) in all 546 samples. The top 15 principal components (PCs) with eigenvalues greater than 0.7 were selected for further analysis, which accounted for a total of 85% of the variance ( Figure 5 , A). To detect differences between controls and heart failure, the AUC was calculated to classify these two groups in each selected PC ( Figure 5 , B). Multiple PCs were found to have an AUC significantly higher than 0.5, and the second PC even had an AUC of 0.79, indicating that the differences between these two populations were mainly attributed to the overall changes in miRNA expression profiles. Since the differences between control and heart failure subjects were found in multiple dimensions (PCs), it was not possible to represent all the information based on a single miRNA. Therefore, a multivariate assay containing multiple miRNAs was necessary for optimal classification. Similarly, multiple PCs had an AUC significantly higher than 0.5 for classification into either of the two heart failure subtypes: HFREF or HFPEF ( Figure 5 , C) (including the first PC (AUC = 0.6)), but the AUCs were less than those used for heart failure detection. Therefore, for the classification of HFREF and HFPEF, multivariate analysis was also required to capture multi-dimensional information.

[0253] The two groups of subjects (C and heart failure (HF)) were plotted on the space defined by the two main discriminant PCs used for HF detection, showing that the two groups were separated and located ( Figure 6 , A). The separation of the HFREF and HFPEF groups ( Figure 6 , B) was less obvious. Global analysis revealed that it was possible to separate control, HFREF, and HFPEF subjects using miRNA profiles. However, using only one or two dimensions was not statistically robust for classification.

[0254] The key step in identifying biomarkers was to directly compare the expression levels of each miRNA in normal and disease states and between disease subtypes. Univariate comparisons were performed using the Student t-test to evaluate the significance of between-group differences for individual miRNAs, and multivariate logistic regression was used to adjust for confounding factors including age, sex, BMI, AF, hypertension, and diabetes. All p-values were corrected for false discovery rate (FDR) estimation using the Bonferroni-type multiple comparison procedure

[60] . miRNAs with p-values below 0.01 were considered significant in this study.

[0255] Then, the expression of 137 plasma miRNAs was compared: A] between controls (healthy) and heart failure (either subtype alone or grouped together), B] between the two heart failure subtypes (i.e., HFREF and HFPEF).

[0256] A] Identify miRNAs differentially expressed between non-HF control subjects and HF patients

[0257] Plasma from patients clinically diagnosed with any heart failure subtype (HFREF or HFPEF) was grouped together and compared with plasma from healthy non-heart failure donors.

[0258] Initial comparisons were made using univariate analysis (Student t-test), where 94 miRNAs were found to be significantly altered in heart failure patients compared to controls (p-value < 0.01 after FDR)( Figure 7 , A). Further examination of these two subtypes separately found that 82 and 94 miRNAs were significantly altered in HFREF and HFPEF subjects, respectively, compared to controls( Figure 7 , A). In total, 101 unique miRNAs were identified by univariate analysis, of which 75% (n = 76) were significant for both subtypes( Figure 7 A).

[0259] Since control subjects were recruited from the community, clinical parameters may not match well with heart failure patients, including three risk factors for heart failure: AF, hypertension, and diabetes, of which such conditions were rarely present in control subjects. Additionally, there were slight differences in age between the analyzed groups. To adjust for these potential confounding factors, multivariate analysis (logistic regression) was performed to test the significance of miRNAs selected by univariate analysis. After multivariate analysis, a total of 86 out of 101 miRNAs remained significantly different between the test groups( Figure 7 , B). For all heart failure detections compared to controls, 75 out of 94 miRNAs were found to be significant in multivariate analysis (p-value < 0.01 after FDR) (Table 5), while for HFREF detections compared to controls, 52 out of 82 (Table 6) were significant, and for HFPEF detections compared to controls, 68 out of 94 (Table 7) were significant( Figure 7 B). After multivariate analysis, 36 miRNAs were found to be still significantly different between the control group and the two heart failure subtypes, while 16 were only significantly different between the control and the HFREF subtype, and 32 were only significantly different between the control and the HFPEF subtype( Figure 7 , B). In multivariate analysis, many miRNAs were found to differ between the control and only one of the two heart failure subtypes, indicating a true difference in miRNA expression between these two subtypes.

[0260] Table 5. miRNAs differentially expressed between control and all heart failure subjects

[0261]

[0262]

[0263]

[0264] Table 6. miRNAs differentially expressed between control and REF subjects

[0265]

[0266]

[0267]

[0268] Table 7. miRNAs differentially expressed between control and HFPEF subjects

[0269]

[0270]

[0271] Numerous miRNAs have been reported to be upregulated and downregulated in HF (Table 1). Intriguingly, the differentially expressed miRNAs identified in the present study were significantly different from these reports. The significant miRNAs in univariate and multivariate analyses are listed in Table 5 (C versus heart failure (HF)), Table 6 (C versus HFREF), and Table 7 (C versus HFPEF), where 37, 25, and 33 miRNAs were found to be upregulated and 38, 27, and 35 miRNAs were found to be downregulated in the three comparisons, respectively. The number of differentially expressed miRNAs verified by qPCR (101 in univariate analysis and 86 in both univariate and multivariate analyses) was significantly higher than that previously reported (Table 8, a total of 47). Each miRNA or combination from these 86 miRNAs can be used as a biomarker or as a component of a biomarker panel (multivariate assay) for the diagnosis of heart failure.

[0272] Table 8. Comparison between the present study and previously published reports

[0273]

[0274]

[0275] A total of 47 different miRNAs have been reported in the literature (Table 1). There are conflicting observations regarding the direction of change of hsa-miR-210 in heart failure patients (Table 8). In the present study, 22 out of the other 46 reported miRNAs were not measurable or below the limit of detection (N.A. in Table 8), leaving 24 miRNAs for comparison. Comparing the results (p-value < 0.01 after FDR in univariate analysis) with the 24 reported miRNAs, only 4 (hsa-miR-423-5p, hsa-miR-30a-5p, hsa-miR-22-3p, hsa-miR-21-5) out of these previously reported miRNAs were found to be consistently downregulated and four (hsa-miR-103a-3p, hsa-miR-30b-5p, hsa-miR-191-5, and hsa-miR-150-5p) were found to be consistently upregulated (Table 8). Interestingly, among the 8 dysregulated miRNAs, the direction of change was opposite to that previously reported, while seven of them remained unchanged (Table 8). Thus, most of the miRNAs previously reported to be differentially regulated in heart failure could not be confirmed in the present study. Instead, the present study identified over 70 novel miRNAs that were not previously reported and may be potential biomarkers for HF detection.

[0276] NT-proBNP / BNP is the most well-studied biomarker for heart failure to date and exhibits the best clinical performance. Therefore, the present study aimed to examine whether these significantly regulated miRNAs could provide additional information to NT-proBNP. The enhancement of NT-proBNP for miRNA detection of heart failure was tested by logistic regression adjusted for age, AF, hypertension, and diabetes (p-value, ln_BNP, Tables 5 to 7). Using a p-value < 0.01 after FDR correction as the criterion, 55 miRNAs (p-value, ln_BNP, Table 7) were found to have information complementary to ln_NT-proBNP for HFPEF detection but not for HFREF (p-value, ln_BNP, Table 6). Compared with HFPEF (AUC = 0.935, Figure 3 E), NT-proBNP alone had significantly better diagnostic performance for HFREF detection (AUC = 0.985, Figure 3 D). Combining any one or more of these 55 miRNAs with ln_NT-proBNP in a multivariate assay could potentially improve the detection of HFPEF.

[0277] The most upregulated (hsa-let-7d-3p, Figure 8 , A) and the most downregulated (hsa-miR-454-3p,Figure 8 The AUC values of the miRNAs in (B) were 0.78 and 0.85, respectively. Neither of the two miRNAs had been previously reported for use in detecting heart failure. Although the diagnostic power of a single miRNA may not be clinically useful, combining multiple miRNAs in a multivariate manner can well improve the performance of heart failure diagnosis.

[0278] B] Identify miRNAs differentially expressed between HFREF and HFPEF

[0279] Univariate analysis (Student t-test) showed that 40 miRNAs were significantly altered (p-value < 0.01 after FDR) between HFREF and HFPEF subjects, of which 10 miRNAs had higher expression levels in HFPEF than in HFREF, and 30 miRNAs had higher expression levels in HFREF than in HFPEF (Table 9).

[0280] It was expected that the background clinical characteristics would be different between the two heart failure subtypes (Table 3). Compared with HFREF patients, HFPEF patients were more commonly female, had a higher BMI, were older, and more often had AF or hypertension. After multivariate analysis (logistic regression) adjusted for these characteristics, only 18 of the 40 miRNAs remained significant (p-value < 0.01 after FDR) (p-value, logistic regression, Table 9). However, since the differences between the two subtypes were due to the natural occurrence and manifestation of the disease rather than biased sample selection in the study, all 40 miRNAs in Table 9 (univariate analysis) could be used to classify heart failure subtypes.

[0281] The most upregulated miRNA in all heart failures (hsa-miR-223-5p, Figure 10 , A) and the most downregulated miRNA (hsa-miR-185-5p, Figure 10 , B) had only moderate AUC values for discriminating heart failure from controls: 0.68 and 0.69, respectively. This is the first report of using circulating cell-free miRNAs from blood (plasma / serum) to classify heart failure patients into two clinically relevant subtypes. Combining multiple miRNAs in a multivariate assay provides more diagnostic power for subtype classification.

[0282] Most of the miRNAs differentially expressed between HFREF and HFPEF (38 out of 40 in univariate analysis; 17 out of 18 in multivariate analysis) were also found to be different from controls, reflecting different degrees of dysregulation between the two heart failure subtypes ( Figure 11 ). To further examine the 38 overlapping miRNAs found to be altered in either HF subtype and between the two subtypes in univariate analysis (Figure 9 , A), It was divided into 6 groups based on the relationship of its expression levels in three object groups (control, HFREF, and HFPEF) Figure 11 ). If the p-value (FDR) for comparison between two groups is higher than 0.01, the relationship is defined as equal (denoted as "="), while if the p-value (after FDR correction) is lower than 0.01, the relationship is defined by the direction of change (denoted as higher ">", or lower "<").

[0283] Graded changes from control to HFREF to HFPEF were found in most miRNAs, among which 21 miRNAs gradually decreased (C > HFREF > HFPEF, Figure 11 ), 5 miRNAs gradually increased (C < HFREF < HFPEF, Figure 11 ). In addition, 5 miRNAs were found to be lower only in the HFPEF subtype (C = HFREF > HFPEF, Figure 11 ), while 2 were found to be higher only in the HFPEF subtype (C = HFREF < HFPEF, Figure 11 ), and there was no difference between HFREF and control. Compared with the control, only 3 miRNAs had more distinct levels in the HFREF subtype than in HFPEF (C < HFPEF < HFREF or C = HFPEF > HFREF or C = HFPEF < HFREF, Figure 11 ). Different from LVEF and NT-proBNP, HFPEF had a more distinct miRNA profile than the HFREF subtype compared with healthy controls. This indicates that miRNAs can supplement NT-proBNP to better distinguish HFPEF.

[0284] Analysis of all detectable miRNAs revealed a large number of positive correlations with each other (Pearson correlation coefficient > 0.5, Figure 12 ), especially among those miRNAs that were altered in HF patients and different between the two heart failure subtypes (miRNAs are represented in black on the x-axis, towards the right hand side of the x-axis, Figure 12 ). Changes in plasma miRNA levels were attributed to heart failure (HFREF and / or HFPEF). These observations suggest that many pairs of miRNAs are regulated similarly in all subjects. Therefore, miRNA panels can be assembled by replacing one or more specific miRNAs with other miRNAs to systematically optimize diagnostic performance. All significantly altered miRNAs are crucial for the development of multivariate index diagnostic assays for heart failure detection or heart failure subtype classification.

[0285] IV. Plasma miRNAs as prognostic markers

[0286] Patients with heart failure were sampled when they were admitted according to the index at the time of recruitment into the SHOP cohort study, when symptoms improved after 3 to 5 days of treatment (where the bedside physical symptoms of HF resolved) and they were considered suitable for discharge. This ensured that the assessment of biomarker performance in this study was relevant to the subacute or "chronic" phase of HF. This study evaluated the prognostic performance of circulating miRNAs for mortality and readmission for heart failure. 327 patients with heart failure (176 HFREF and 151 HFPEF) were followed up for 2 years (Table 10), during which 49 died (15%).

[0287] Table 10. Clinical information of the subjects included in the prognostic study

[0288]

[0289]

[0290]

[0291]

[0292]

[0293]

[0294]

[0295]

[0296] In all study cases, 115 were readmitted for heart failure during the follow-up period (Table 10) and 49 died. miRNAs were evaluated as potential markers (i.e., predictors) of both overall survival (OS) and event-free survival (EFS) (a composite of all-cause death and / or readmission) in decompensated heart failure.

[0297] The summary of anti-heart failure drug treatments prescribed to the study participants is presented in Table 11. Comparing the treatments of HFREF and HFPEF, it was found that the prescribing frequencies of half of the relevant drugs were different ( Figure 13 ). Notably, those drug classes that have been shown to improve the prognosis of HFREF (ACEI / ARB, β-blockers, and mineralocorticoid antagonists) were prescribed more often to HFREF patients than to HFPEF patients. Treatments were according to current clinical practice and were included among the clinical variables for analyzing prognostic markers.

[0298] Table 11. Treatments of the subjects included in the prognostic study

[0299]

[0300]

[0301]

[0302]

[0303]

[0304]

[0305]

[0306]

[0307]

[0308] Survival analysis was performed using Cox proportional hazards (CoxPH) modeling, and variables were accounted for individually (univariate analysis) or simultaneously (multivariate analysis) in the same model. To have better comparison between different hazard ratios (HRs), all normally distributed variables (including miRNA expression levels (log2 scale)), clinical variables (such as BMI, ln_NT-proBNP, LVEF, age), and multivariate scores generated by combining multiple variables were scaled proportionally to have one standard deviation. Then, the hazard ratio (HR) was used as an indicator of the prognostic power of these variables. A p-value < 0.05 was considered statistically significant. Patients were classified as high-risk and low-risk according to the presence or absence of categorical variables and the above or below median levels of normally distributed continuous variables. The Kaplan-Meier plot (KM plot) was used to illustrate survival over time between different risk groups by comparison between curves using the log-rank test. The between-group survival at 750 days (OS750) and / or the EFS at 750 days (EFS750) were also compared.

[0309] All clinical variables were initially evaluated to predict overall survival (OS). In univariate analysis, five variables (age, hypertension, ln_NT-proBNP, nitrates, and hydralazine) were found to be positively associated with the risk of death, and two variables (BMI and β-blockers) were found to be negatively associated with the risk of death (Table 12). Interestingly, there was no difference in overall survival between HFREF and HFPEF patients. The KM plots of the groups of subjects defined by these significant parameters are shown in Figure 14, A, and OS750 is shown in Figure 14, B. All parameters were able to define high-risk and low-risk groups, with ln_NT-proBNP being the most significant (p-value = 7.2E-07, HR = 2.36 (95% CI: 1.69 - 3.30)). Based on the level of ln_NT-proBNP, the OS750 of the low-risk group was 92.4%, and the value of the high-risk group was only 66.0%. In multivariate analysis including all clinical variables, six variables (gender, hypertension, BMI, ln_NT-proBNP, β-blockers, and warfarin) were found to be significant. These six variables were then combined with each of the 137 miRNAs in a CoxPH model to identify prognostic miRNA markers for overall survival.

[0310] Table 12. Analysis of Clinical Variables for Observed Survival

[0311]

[0312]

[0313] A similar analysis was performed for event-free survival (EFS), and in univariate analysis, seven variables (AF, hypertension, diabetes, age, ln_NT-proBNP, nitrates, and hydralazine) were found to be positively associated with the risk of readmission for decompensated heart failure (Table 13). The KM plots of the groups of subjects defined by these significant parameters are shown in Figure 15, A, and EFS750 is shown in Figure 15B , B. Again, there was no difference in event-free survival between HFREF and HFPEF, and ln_NT-proBNP was the most significant predictor of event-free survival (p-value = 1.5E-09, HR = 1.79 (95% CI: 1.47 - 2.17)). The median ln_NT-proBNP was associated with 65.1% of EFS750, and the upper median level was associated with only 34.1% of EFS750. By multivariate analysis, only two variables, diabetes and ln_NT-proBNP, were found to be significant. Subsequently, these variables were combined with each of the 137 miRNAs to identify prognostic miRNA markers for event-free survival.

[0314] Table 13. Analysis of Clinical Variables for Event-Free Survival (EFS)

[0315]

[0316] To identify miRNA biomarkers for predicting overall survival, 137 miRNAs were each tested by univariate CoxPH model and in a multivariate CoxPH model that included 6 additional predictive clinical variables. In total, 40 miRNAs had a p-value less than 0.05. Thirty-seven (37) were significant in univariate analysis and 29 were significant in multivariate analysis (Table 14). Eleven miRNAs found to be significant in univariate analysis did not improve the predictive performance of clinical parameters (multivariate analysis), and 3 miRNAs were significant only when combined with clinical variables ( Figure 16 , A). Except for hsa-miR-374b-5p (p-value = 0.25), 2 miRNAs had a p-value less than 0.1 in univariate analysis (Table 14).

[0317] The miRNA with the highest hazard ratio (HR) for mortality in univariate (HR = 1.90 (95% CI: 1:36 - 2.65, p-value = 0.00014)) and multivariate analysis (HR = 1.79 (95% CI: 1.23 - 2.59, p-value = 0.0028)) was hsa-miR-503. For both univariate analysis (HR = 0.52 (95% CI: 0.40 - 0.67, p-value = 1.3E-7)) and multivariate analysis (HR = 0.59 (95% CI: 0.45 - 0.78, p-value = 0.00032)), Hsa-miR-150-5p had the lowest HR (i.e., the expression level was negatively correlated with risk) (Table 14). The KM plots of the two miRNAs are shown in Figure 18, A. A good separation between the two risk groups could be observed. Based on a single miRNA, the high-risk and low-risk groups had a difference of approximately 21.3% (Hsa-miR-503) or 17.8% (has-miR-150-5p) in OS750 (Figure 18, B). With the addition of 6 clinical variables, the combined score provided better risk prediction, with a difference of 25.3% for hsa-miR-503 + 6 clinical variables and 22.4% for has-miR-150-5p + 6 clinical variables (Figure 18, B). Any one or more of the 40 miRNAs (Table 14) can be used as prognostic markers / groups for the death risk of patients with chronic HF.

[0318] To predict event - free survival, 13 miRNAs were found to be significant (p - value < 0.05) in univariate analysis, of which 4 were positively correlated with the risk of readmission for decompensated heart failure after treatment and 9 were negatively correlated (Table 15). In a multivariate analysis including 2 additional clinical variables in the CoxPH model, miRNAs were not found to be significant for EFS prediction. However, the miRNA most positively correlated with EFS (hsa - miR - 331 - 5p, HR = 1.27 (95% CI: 1.09 - 1.49, p - value = 0.0025)) and the most negatively correlated miRNA (hsa - miR - 30e - 3p, HR = 0.80 (95% CI: 0.69 - 0.94, p - value = 0.0070)) in univariate analysis also had a certain level of significance in multivariate analysis, with p - values of 0.15 and 0.14 respectively (Table 15). Kaplan - Meier plots of high - risk and low - risk groups of EFS defined by any one miRNA with and without additional clinical variables are shown in Figure 19, A, and EFS750 is shown in Figure 19, B. Based on a single miRNA (hsa - miR - 331 - 5p or hsa - miR - 30e - 3p), EFS750 for the high - risk group was approximately 40% and for the low - risk group was approximately 60%, while with the addition of 2 clinical variables, the values were 33% and 66% (Figure 19, B). Any one or more of the 13 miRNAs (Table 15) can be used as a prognostic marker / group for the risk of readmission for decompensated HF in patients with chronic HF.

[0319] Compared with overall survival (n = 43), fewer miRNAs were identified as predictive of event - free survival (n = 13) and only 3 overlapped ( Figure 16 , B). The results suggest different mechanisms for death and recurrent decompensated heart failure. An important issue to note is that the definition of event - free survival in this study involves a less well - defined clinical variable - namely hospitalization, which can be biased according to the patient or clinician on a case - by - case basis. Nevertheless, 53 miRNAs can still be valuable prognostic markers in patients with chronic heart failure.

[0320] Then, the 53 prognostic markers were compared with 101 markers used for HF detection ( Figure 17 , A) or 40 markers used for heart failure subtype classification ( Figure 17 B). Some overlap was observed, but a large majority of the prognostic markers were still not found in the other two lists, indicating that independent miRNA groups should be used or combined to form a multivariate assay for prognosis.

[0321] V. Multivariate Biomarker Groups for HF Detection

[0322] As described above, compared to the use of a single miRNA, a group consisting of a combination of multiple miRNAs may be used to provide better diagnostic power.

[0323] An important criterion for assembling such a multivariate group is to include at least one miRNA from a specific list for each heart failure subtype to ensure coverage of all heart failure subgroups. However, the miRNAs defining two heart failure subtypes overlap ( Figure 7 ). Also, a large number of heart failure-related or -unrelated miRNAs are positively correlated ( Figure 12 ), which makes it challenging to select the optimal miRNA combination for heart failure diagnosis.

[0324] Given the complexity of the task, the inventors of the present study decided to use the sequence forward floating search algorithm

[53] to identify the miRNA group with the highest AUC. The existing linear support vector machine (a well-utilized and recognized modeling tool for constructing variable groups) was also used to assist in the selection of the miRNA combination

[54] . This model generates a score according to a linear formula that takes into account the expression level of each member and its weighting coefficient. These linear models can be easily applied to clinical practice.

[0325] A key requirement for the success of such a process is the availability of high-quality data. Quantitative data on all detected miRNAs in a large number of well-defined clinical samples not only improves the accuracy and precision of the results but also ensures the consistency of the identified biomarker group for further clinical applications using qPCR.

[0326] To ensure the authenticity of the results, multiple (>80 times) hold-out validations (two-fold cross-validation) were performed to test the performance of the identified biomarker group in an independent validation sample group (the remaining half of the samples in each fold) based on the discovery group (half of the samples in each fold). In the case of a large number of clinical samples (546), the problem of data overfitting in modeling was minimized because only 137 candidate features needed to be selected from them, and 273 samples were used as the discovery group in each fold with a sample-to-feature ratio greater than 2. During the cross-validation process, the samples were matched in terms of subtype, gender, and race. Also, a process was carried out to optimize biomarker groups with 3, 4, 5, 6, 7, 8, 9, or 10 miRNAs, respectively.

[0327] A box plot showing the results (AUC of the biomarker sets in both the discovery and validation phases) is shown in Fig. 20, A. The AUC values are quite close in different discovery groups (box size < 0.01), and as the number of miRNAs in the group increases, it approaches unity (AUC = 1.0). When there are 4 or more miRNAs, the box size in the validation phase indicating the spread of the values is quite small (≤ 0.01 AUC value). As predicted, the AUC value decreases with each validation group searched (0.02 to 0.05 AUC).

[0328] A more quantitative representation of the results is shown in Fig. 20, B. Although the AUC always gradually increases in the discovery phase as the number of miRNAs in the biomarker set increases, when the number of miRNAs is greater than 8, the AUC value does not improve significantly further in the validation phase. Although the difference between the biomarker sets of 6 miRNAs and 8 miRNAs is statistically significant, the improvement in the AUC value is less than 0.01. Therefore, biomarker sets with 6 or more miRNAs with an AUC value of approximately 0.93 should be useful for detecting heart failure.

[0329] To detect the composition of the multivariate biomarker sets, this study calculated the occurrence of miRNAs in all sets containing 6 to 10 miRNAs, excluding the sets with the top 10% and bottom 10% AUC. This was done to avoid miscounting false discovery biomarkers due to fitting inaccurate data from the subsets generated by the randomization process in the cross-validation analysis. Excluding these miRNAs selected in less than 2% of the sets, a total of 51 miRNAs were selected in the discovery process (Table 16), and the expression of 42 of these miRNAs was also found to be significantly altered in HF (Tables 5 to 7). The other 9 miRNAs, although not altered in heart failure, were found to significantly improve the AUC value because 39% of the sets included at least one of these miRNAs from the list and the most frequently selected miRNA (hsa-miR-10b-5p) that occurred in 35% of the sets. Without direct quantitative measurement of all miRNA targets, these miRNAs would not have been selected in high-throughput screening studies (microarray, sequencing) and would have been excluded for further qPCR validation.

[0330] When comparing the identity of the selected miRNAs for the multivariate groups and single miRNAs as diagnostic markers, they are not necessarily the same. For example, the most upregulated (hsa-let-7d-3p) miRNA is not in the list, while the most upregulated (hsa-miR-454-3p) is used in only 24.2% of the sets. Therefore, it is not possible to simply combine the best single miRNAs identified to form the best biomarker set, and a set of miRNAs providing complementary information gives the best results.

[0331] None of these miRNAs were randomly selected, as 7 of them were present in more than 30% of the groups, but it was also difficult to find miRNAs crucial for good biomarkers, as the two most frequently selected miRNAs, hsa-miR-551b-3p and hsa-miR-24-3p, were present in only 59.7% and 57.3% of the groups, respectively. As discussed, many of these miRNAs are associated ( Figure 11 ), and they can substitute or replace each other in biomarkers. In summary, a biomarker panel with at least 6 miRNAs from the frequently selected list (Table 16) should be used to detect heart failure.

[0332] To compare the miRNA biomarker with NT-proBNP, a panel of six miRNA biomarkers was selected to calculate the combined miRNA score for all subjects, and it was plotted against the NT-proBNP levels from the same subjects ( Figure 21 , A). Overall, the inventors of this study observed a positive correlation, with the Pearson correlation coefficient between the miRNA score and ln_NT-proBNP being 0.61 (p-value = 8.2E-56). Applying the recommended NT-proBNP cut-off value (125 pg / mL, dashed line), 35 healthy subjects were misclassified as heart failure patients (false positives, FP, NT-proBNP > 125), and the NT-proBNP levels of 23 heart failure patients were below the cut-off value (false negatives, FN). It can be predicted that most of the false negatives (FN) were HFPEF subjects (n = 20). The false positive (FP) and false negative (FN) subjects related to NT-proBNP were selected, and the results were plotted against the miRNA score ( Figure 21 , B). Based on the individual plots, most of the false positive (FP) and false negative (FN) subjects could be correctly reclassified by the miRNA score with 0 as the cut-off value (dashed line). The results verified the hypothesis that the miRNA biomarker carries different information from NT-proBNP. The next step was to explore a multivariate biomarker panel that includes both miRNAs and NT-proBNP.

[0333] The same biomarker identification process (multiple two-fold cross-validations) was performed, where NT-proBNP was pre-fixed as one of the predictors and the ln_NT-proBNP levels and miRNA expression levels (log2 scale) were used to build a classifier using support vector machines. Since the AUC did not increase significantly when using more than 8 miRNAs to predict heart failure (Figure 20), this process was performed to optimize the biomarker panel with 2, 3, 4, 5, 6, 7, or 8 miRNAs (together with NT-proBNP).

[0334] The classifier built during the discovery phase approaches perfect separation (AUC = 1.00) as the number of miRNAs increases. Performance drops slightly during the validation phase ( Figure 22 , A). Nevertheless, the AUC for the group containing NT-proBNP during the validation phase (average AUC > 0.96) is always higher than that for the group of miRNA-only biomarkers (average AUC < 0.94). The quantitative results ( Figure 22 , B) show that when the number of miRNAs is greater than 4, there is no further significant improvement in the AUC value during the validation phase, and there is only a slight increase (0.001 AUC) between the groups of 4 and 5 miRNA biomarkers. Therefore, when combined with NT-proBNP, a group of biomarkers with four or more miRNAs that provides an AUC value of approximately 0.98 can be used for heart failure detection. By combining miRNAs and NT-proBNP, the classification efficiency is significantly improved relative to NT-proBNP (AUC = 0.962, Figure 22 , B).

[0335] Exclude the groups with the top 10% and bottom 10% AUCs and examine the composition of the multivariate biomarker groups containing 3 to 8 miRNAs (Table 17). A total of 49 miRNAs were selected during the discovery process, 14 of which had a prevalence greater than 10% (Table 17). Forty-two (42) of these also carried additional information on NT-proBNP (p-value < 0.01 after FDR in logistic regression). Similarly, 46% of the groups included at least one of the 13 miRNAs that were found to be non-significant in addition to NT-proBNP.

[0336] Although more than half of the significant miRNAs (Table 17, significant list) were also frequently selected when searching for miRNA-only biomarker groups (Table 16, significant list) ( Figure 23 , A), the ranking of prevalence was different. When searching for miRNA-based biomarker groups, some of the highly selected miRNAs that were combined with NT-proBNP were not even selected (hsa-miR-17-5p (11.6%) and hsa-miR-25-3p (11.0%)). In addition, there were only two miRNAs that overlapped between the non-significant lists (Table 16 and Table 17, non-significant list) ( Figure 23 , B). In summary, the evidence suggests that a different list of miRNAs should be used together with NT-proBNP compared to the list used to construct miRNA-only biomarker groups.

[0337] VI. Multivariate Biomarker Groups for HF Subtype Classification

[0338] Perform a next attempt to identify a multivariate biomarker panel for differentiating HFREF from HFPEF. Again, all quantitative data on 137 miRNAs of 338 heart failure patients were used. Due to the sample size limitation, multiple (>50) four-fold cross-validations were performed, where all subjects were randomly divided into four groups, and a classifier was constructed using three groups (the discovery group) to further predict the last group (the validation group). In this way, 253 to 254 subjects were used in the discovery phase, thus ensuring the same size for each subgroup (HFREF or HFPEF), similar to the number of candidate features (137) selected to minimize overfitting. This process was performed again to separately optimize biomarker panels of 3, 4, 5, 6, 7, 8, 9, or 10 miRNAs and biomarker panels of 2, 3, 4, 5, 6, 7, or 8 miRNAs plus NT-proBNP.

[0339] Quantitative results showed that when the miRNA-only biomarker panel contained more than 5 miRNAs, the AUC value did not improve further ( Figure 24 , A). The miRNA biomarker panel achieved an AUC of approximately 0.76, which was better than that of NT-proBNP (AUC = 0.706). Counting all groups of 6 to 10 miRNAs (excluding the top 10% and bottom 10% AUC), 46 miRNAs were frequently selected (in 2% of the groups), of which 22 were found to be significant in the t-test comparing HFREF with HFPEF, while 24 were not significant (Table 18). Since two miRNAs were present in more than 80% of the groups (hsa-miR-30a-5p (94.6%) and hsa-miR-181a-2-3p (83.7%), Table 18), the diversity of the groups used for heart failure subtype classification was lower than that of those used for heart failure detection.

[0340] Compared with the miRNA-only groups, the biomarker panels composed of both miRNA and NT-proBNP required fewer miRNAs, as the AUC value did not improve when more than 4 miRNAs were included ( Figure 24, B). Compared with the group with only miRNA, even more distinct classification can be achieved (AUC is about 0.82). Moreover, miRNA and NT-proBNP can carry complementary information for the classification of heart failure subtypes. For the composition of the group detecting 5 to 8 miRNAs plus NT-proBNP, 31 miRNAs were frequently selected (in >2% of the groups), among which 14 were found to be significant in the logistic regression with ln_NT-proBNP, while 17 were not (Table 19). Two different miRNAs were found in more than 80% of the groups: hsa-miR-199b-5p (91.5%) and hsa-miR-191-5p (74.9%). Although the most frequently selected non-significant miRNA for the groups with only miRNA and miRNA plus NT-proBNP is the same (hsa-miR-199b-5p), significant differences can be found in terms of identity and ranking between the remaining significant and non-significant lists.

[0341] References

[0342] 1. Organization, W.H., Annex Table 2: Deaths by cause, sex and mortality stratum in WHO regions, estimates for 2002. The world health report 2004 - changing history, 2004.

[0343] 2. Alla, F., F. Zannad, and G. Filippatos, Epidemiology of acute heart failure syndromes. Heart Fail Rev, 2007. 12(2): p. 91 - 5.

[0344] 3. Stewart, S., et al., More'malignant' than cancer? Five-year survival following a first admission for heart failure. Eur J Heart Fail, 2001. 3(3): p. 315 - 22.

[0345] 4. Pritchard, C.C., et al., Blood cell origin of circulating microRNAs: a cautionary note for cancer biomarker studies. Cancer Prev Res (Phila), 2012. 5(3): p. 492-7.

[0346] 5. McDonald, J.S., et al., Analysis of circulating microRNA: preanalytical and analytical challenges. Clin Chem, 2011. 57(6): p. 833-40.

[0347] 6. Nishimura, J. and Y. Kanakura, [Paroxysmal nocturnal hemoglobinuria (PNH)]. Nihon Rinsho, 2008. 66(3): p. 490-6.

[0348] 7. Owan, T.E., et al., Trends in prevalence and outcome of heart failure with preserved ejection fraction. N Engl J Med, 2006. 355(3): p. 251-9.

[0349] 8. Bhatia, R.S., et al., Outcome of heart failure with preserved ejection fraction in a population-based study. N Engl J Med, 2006. 355(3): p. 260-9.

[0350] 9. Yancy, C. W., et al., Clinical presentation, management, and in-hospital outcomes of patients admitted with acute decompensated heart failure with preserved systolic function: a report from the Acute Decompensated Heart Failure National Registry (ADHERE) Database. J Am Coll Cardiol, 2006. 47(1): p. 76-84.

[0351] 10. Borlaug, B. A. and M. M. Redfield, Diastolic and systolic heart failure are distinct phenotypes within the heart failure spectrum. Circulation, 2011. 123(18): p. 2006-13; discussion 2014.

[0352] 11. Hogg, K., K. Swedberg, and J. McMurray, Heart failure with preserved left ventricular systolic function; epidemiology, clinical characteristics, and prognosis. J Am Coll Cardiol, 2004. 43(3): p. 317-27.

[0353] 12. Yusuf, S., et al., Effects of candesartan in patients with chronic heart failure and preserved left-ventricular ejection fraction: the CHARM-Preserved Trial. Lancet, 2003. 362(9386): p. 777-81.

[0354] 13. Massie, B.M., et al., Irbesartan in patients with heart failure and preserved ejection fraction. N Engl J Med, 2008. 359(23): p. 2456 - 67.

[0355] 14. Writing Committee, M., et al., 2013 ACCF / AHA guideline for the management of heart failure: a report of the American College of Cardiology Foundation / American Heart Association Task Force on practice guidelines. Circulation, 2013. 128(16): p. e240 - 327.

[0356] 15. Troughton, R.W., et al., Effect of B - type natriuretic peptide - guided treatment of chronic heart failure on total mortality and hospitalization: an individual patient meta - analysis. Eur Heart J, 2014. 35(23): p. 1559 - 67.

[0357] 16. Christenson, R.H., et al., Impact of increased body mass index on accuracy of B - type natriuretic peptide (BNP) and N - terminal proBNP for diagnosis of decompensated heart failure and prediction of all - cause mortality. Clin Chem, 2010. 56(4): p. 633 - 41.

[0358] 17. Richards, M., et al., Atrial fibrillation impairs the diagnostic performance of cardiac natriuretic peptides in dyspneic patients: results from the BACH Study (Biomarkers in ACute Heart Failure). JACC Heart Fail, 2013. 1(3): p. 192-9.

[0359] 18. Masson, S., et al., Direct comparison of B-type natriuretic peptide (BNP) and amino-terminal proBNP in a large population of patients with chronic and symptomatic heart failure: the Valsartan Heart Failure (Val-HeFT) data. Clin Chem, 2006. 52(8): p. 1528-38.

[0360] 19. Komajda, M., et al., Factors associated with outcome in heart failure with preserved ejection fraction: findings from the Irbesartan in Heart Failure with Preserved Ejection Fraction Study (I-PRESERVE). Circ Heart Fail, 2011. 4(1): p. 27-35.

[0361] 20. Kraigher-Krainer, E., et al., Impaired systolic function by strain imaging in heart failure with preserved ejection fraction. J Am Coll Cardiol, 2014. 63(5): p. 447-56.

[0362] 21. Richards, A.M., J.L. Januzzi, Jr., and R.W. Troughton, Natriuretic Peptides in Heart Failure with Preserved Ejection Fraction. Heart Fail Clin, 2014. 10(3): p. 453 - 470.

[0363] 22. Liang, H., et al., The origin, function, and diagnostic potential of extracellular microRNAs in human body fluids. Wiley Interdiscip Rev RNA, 2014. 5(2): p. 285 - 300.

[0364] 23. Cortez, M.A., et al., MicroRNAs in body fluids--the mix of hormones and biomarkers. Nat Rev Clin Oncol, 2011. 8(8): p. 467 - 77.

[0365] 24. Bronze - da - Rocha, E., MicroRNAs Expression Profiles in Cardiovascular Diseases. Biomed Res Int, 2014. 2014: p. 985408.

[0366] 25. Kim, K.M. and S.G. Kim, Autophagy and microRNA dysregulation in liver diseases. Arch Pharm Res, 2014.

[0367] 26. Tan, L., J.T. Yu, and L. Tan, Causes and Consequences of MicroRNA Dysregulation in Neurodegenerative Diseases. Mol Neurobiol, 2014.

[0368] 27. Lee, R.C., R.L. Feinbaum, and V. Ambros, The C. elegans heterochronic gene lin-4 encodes small RNAs with antisense complementarity to lin-14. Cell, 1993. 75(5): p. 843-54.

[0369] 28. Friedman, R.C., et al., Most mammalian mRNAs are conserved targets of microRNAs. Genome Res, 2009. 19(1): p. 92-105.

[0370] 29. Hayashita, Y., et al., A polycistronic microRNA cluster, miR-17-92, is overexpressed in human lung cancers and enhances cell proliferation. Cancer Res, 2005. 65(21): p. 9628-32.

[0371] 30. Jovanovic, M. and M.O. Hengartner, miRNAs and apoptosis: RNAs to die for. Oncogene, 2006. 25(46): p. 6176-87.

[0372] 31. Grasso, M., et al., Circulating miRNAs as biomarkers for neurodegenerative disorders. Molecules, 2014. 19(5): p. 6891-910.

[0373] 32. Sayed, A.S., et al., Diagnosis, Prognosis and Therapeutic Role of Circulating miRNAs in Cardiovascular Diseases. Heart Lung Circ, 2014. 23(6): p. 503-510.

[0374] 33. de Candia, P., et al., Serum microRNAs as Biomarkers of Human Lymphocyte Activation in Health and Disease. Front Immunol, 2014.5: p. 43.

[0375] 34. Sheinerman, K.S. and S.R. Umansky, Circulating cell-free microRNA as biomarkers for screening, diagnosis and monitoring of neurodegenerative diseases and other neurologic pathologies. Front Cell Neurosci, 2013.7: p. 150.

[0376] 35. Dorval, V., P.T. Nelson, and S.S. Hebert, Circulating microRNAs in Alzheimer's disease: the search for novel biomarkers. Front Mol Neurosci, 2013.6: p. 24.

[0377] 36. Vogel, B., et al., Multivariate miRNA signatures as biomarkers for non-ischaemic systolic heart failure. Eur Heart J, 2013.34(36): p. 2812-22.

[0378] 37. Endo, K., et al., MicroRNA 210 as a biomarker for congestive heart failure. Biol Pharm Bull, 2013.36(1): p. 48-54.

[0379] 38. Zhang, R., et al., Elevated plasma microRNA-1 predicts heart failure after acute myocardial infarction. Int J Cardiol, 2013. 166(1): p. 259-60.

[0380] 39. Fukushima, Y., et al., Assessment of plasma miRNAs in congestive heart failure. Circ J, 2011. 75(2): p. 336-40.

[0381] 40. Corsten, M.F., et al., Circulating MicroRNA-208b and MicroRNA-499 reflect myocardial damage in cardiovascular disease. Circ Cardiovasc Genet, 2010. 3(6): p. 499-506.

[0382] 41. Matsumoto, S., et al., Circulating p53-responsive microRNAs are predictive indicators of heart failure after acute myocardial infarction. Circ Res, 2013. 113(3): p. 322-6.

[0383] 42. Goren, Y., et al., Serum levels of microRNAs in patients with heart failure. Eur J Heart Fail, 2012. 14(2): p. 147-54.

[0384] 43. Xiao, J., et al., MicroRNA-134 as a potential plasma biomarker for the diagnosis of acute pulmonary embolism. J Transl Med, 2011. 9: p. 159.

[0385] 44. Tijsen, A.J., et al., MiR423-5p as a circulating biomarker for heart failure. Circ Res, 2010. 106(6): p. 1035-9.

[0386] 45. Zhao, D.S., et al., Serum miR-210 and miR-30a expressions tend to revert to fetal levels in Chinese adult patients with chronic heart failure. Cardiovasc Pathol, 2013. 22(6): p. 444-50.

[0387] 46. Goren, Y., et al., Relation of reduced expression of MiR-150 in platelets to atrial fibrillation in patients with chronic systolic heart failure. Am J Cardiol, 2014. 113(6): p. 976-81.

[0388] 47. Ellis, K.L., et al., Circulating microRNAs as candidate markers to distinguish heart failure in breathless patients. Eur J Heart Fail, 2013. 15(10): p. 1138-47.

[0389] 48. Redova, M., J. Sana, and O. Slaby, Circulating miRNAs as new blood-based biomarkers for solid cancers. Future Oncol, 2013. 9(3): p. 387-402.

[0390] 49. Mestdagh, P., et al., Evaluation of quantitative miRNA expression platforms in the microRNA quality control (miRQC) study. Nat Methods, 2014.

[0391] 50. Cissell, K.A. and S.K. Deo, Trends in microRNA detection. Anal Bioanal Chem, 2009. 394(4): p. 1109-16.

[0392] 51. Tsongalis, G.J., et al., MicroRNA analysis: is it ready for primetime? Clin Chem, 2013. 59(2): p. 343-7.

[0393] 52. Hindson, B.J., et al., High-throughput droplet digital PCR system for absolute quantitation of DNA copy number. Anal Chem, 2011. 83(22): p. 8604-10.

[0394] 53. Xiong, M., X. Fang, and J. Zhao, Biomarker identification by feature wrappers. Genome Res, 2001. 11(11): p. 1878-87.

[0395] 54. Saeys, Y., I. Inza, and P. Larranaga, A review of feature selection techniques in bioinformatics. Bioinformatics, 2007. 23(19): p. 2507-17.

[0396] 55. Santhanakrishnan, R., et al., The Singapore Heart Failure Outcomes and Phenotypes (SHOP) study and Prospective Evaluation of Outcome in Patients with Heart Failure with Preserved Left Ventricular Ejection Fraction (PEOPLE) study: rationale and design. J Card Fail, 2013. 19(3): p. 156 - 62.

[0397] 56. Santhanakrishnan, R., et al., Growth differentiation factor 15, ST2, high - sensitivity troponin T, and N - terminal pro brain natriuretic peptide in heart failure with preserved vs. reduced ejection fraction. Eur J Heart Fail, 2012. 14(12): p. 1338 - 47.

[0398] 57. McMurray, J.J., et al., ESC Guidelines for the diagnosis and treatment of acute and chronic heart failure 2012: The Task Force for the Diagnosis and Treatment of Acute and Chronic Heart Failure 2012 of the European Society of Cardiology. Developed in collaboration with the Heart Failure Association (HFA) of the ESC. Eur Heart J, 2012. 33(14): p. 1787 - 847.

[0399] 58. Etheridge, A., et al., Extracellular microRNA: a new source of biomarkers. Mutat Res, 2011. 717(1-2): p. 85-90.

[0400] 59. Li, Y. and K. V. Kowdley, Method for microRNA isolation from clinical serum samples. Anal Biochem, 2012. 431(1): p. 69-75.

[0401] 60. Benjamini, Y., and Hochberg, Y., Controlling the false discovery rate: A practical and powerful approach to multiple testing. Journal of the Royal Statistical Society, 1995. 57(289-300). Sequence Listing <110> Agency for Science, Technology and Research, Singapore National University Hospital, Singapore National University of Singapore <120> Method for the Diagnosis and Prognosis of Chronic Heart Failure <130> 9869SG3818 <150> SG 10201503644Q <151> 2015-05-08 <160> 137 <170> PatentIn version 3.5 <210> 1 <211> 24 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-125a-5p <400> 1 ucccugagac ccuuuaaccu guga 24 <210> 2 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-134 <400> 2 ugugacuggu ugaccagagg gg 22 <210> 3 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-let-7b-3p <400> 3 cuauacaacc uacugccuuc cc 22 <210> 4 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-34b-3p <400> 4 caaucacuaa cuccacugcc au 22 <210> 5 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-101-5p <400> 5 caguuaucac agugcugaug cu 22 <210> 6 <211> 23 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-550a-5p <400> 6 agugccugag ggaguaagag ccc 23 <210> 7 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-576-5p <400> 7 auucuaauuu cuccacgucu uu 22 <210> 8 <211> 23 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-181b-5p <400> 8 aacauucauu gcugucggug ggu 23 <210> 9 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-197-3p <400> 9 uucaccaccu ucuccaccca gc 22 <210> 10 <211> 21 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-369-3p <400> 10 aauaauacau gguugaucuu u 21 <210> 11 <211> 21 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-126-5p <400> 11 cauuauuacu uuugguacgc g 21 <210> 12 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-375 <400> 12 uuuguucguu cggcucgcgu ga 22 <210> 13 <211> 21 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-379-5p <400> 13 ugguagacua uggaacguag g 21 <210> 14 <211> 23 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-579 <400> 14 uucauuuggu auaaaccgcg auu 23 <210> 15 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-106b-3p <400> 15 ccgcacugug gguacuugcu gc 22 <210> 16 <211> 21 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-497-5p <400> 16 cagcagcaca cugugguuug u 21 <210> 17 <211> 23 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-199a-5p <400> 17 cccaguguuc agacuaccug uuc 23 <210> 18 <211> 23 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-19b-3p <400> 18 ugugcaaauc caugcaaaac uga 23 <210> 19 <211> 23 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-20a-5p <400> 19 uaaagugcuu auagugcagg uag 23 <210> 20 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-424-5p <400> 20 cagcagcaau ucauguuuug aa 22 <210> 21 <211> 20 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-144-3p <400> 21 uacaguauag augauguacu 20 <210> 22 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-154-5p <400> 22 uagguuaucc guguugccuu cg 22 <210> 23 <211> 23 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-191-5p <400> 23 caacggaauc ccaaaagcag cug 23 <210> 24 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-30d-5p <400> 24 uguaaacauc cccgacugga ag 22 <210> 25 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-30e-3p <400> 25 cuuucagucg gauguuuaca gc 22 <210> 26 <211> 23 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-10a-5p <400> 26 uacccuguag auccgaauuu gug 23 <210> 27 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-374c-5p <400> 27 auaauacaac cugcuaagug cu 22 <210> 28 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-495 <400> 28 aaacaaacau ggugcacuuc uu 22 <210> 29 <211> 17 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-1275 <400> 29 gugggggaga ggcuguc 17 <210> 30 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-1 <400> 30 uggaauguaa agaaguaugu au 22 <210> 31 <211> 21 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-23a-3p <400> 31 aucacauugc cagggauuuc c 21 <210> 32 <211> 21 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-27a-3p <400> 32 uucacagugg cuaaguuccg c 21 <210> 33 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-122-5p <400> 33 uggaguguga caaugguguu ug 22 <210> 34 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-133a <400> 34 uuuggucccc uucaaccagc ug 22 <210> 35 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-146b-5p <400> 35 ugagaacuga auuccauagg cu 22 <210> 36 <211> 23 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-20b-5p <400> 36 caaagugcuc auagugcagg uag 23 <210> 37 <211> 21 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-27b-3p <400> 37 uucacagugg cuaaguucug c 21 <210> 38 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-30b-5p <400> 38 uguaaacauc cuacacucag cu 22 <210> 39 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-let-7e-3p <400> 39 cuauacggcc uccuagcuuu cc 22 <210> 40 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-337-3p <400> 40 cuccuauaug augccuuucu uc 22 <210> 41 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-363-3p <400> 41 aauugcacgg uauccaucug ua 22 <210> 42 <211> 23 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-421 <400> 42 aucaacagac auuaauuggg cgc 23 <210> 43 <211> 23 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-335-5p <400> 43 ucaagagcaa uaacgaaaaa ugu 23 <210> 44 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-518b <400> 44 caaagcgcuc cccuuuagag gu 22 <210> 45 <211> 23 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-103a-3p <400> 45 agcagcauug uacagggcua uga 23 <210> 46 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-660-5p <400> 46 uacccauugc auaucggagu ug 22 <210> 47 <211> 21 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-192-5p <400> 47 cugaccuaug aauugacagc c 21 <210> 48 <211> 23 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-199b-5p <400> 48 cccaguguuu agacuaucug uuc 23 <210> 49 <211> 23 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-19a-3p <400> 49 ugugcaaauc uaugcaaaac uga 23 <210> 50 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-493-5p <400> 50 uuguacaugg uaggcuuuca uu 22 <210> 51 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-377-3p <400> 51 aucacacaaa ggcaacuuuu gu 22 <210> 52 <211> 23 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-500a-5p <400> 52 uaauccuugc uaccugggug aga 23 <210> 53 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-125b-5p <400> 53 ucccugagac ccuaacuugu ga 22 <210> 54 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-let-7i-5p <400> 54 ugagguagua guuugugcug uu 22 <210> 55 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-299-3p <400> 55 uaugugggau gguaaaccgc uu 22 <210> 56 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-15b-5p <400> 56 uagcagcaca ucaugguuua ca 22 <210> 57 <211> 21 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-21-3p <400> 57 caacaccagu cgaugggcug u 21 <210> 58 <211> 23 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-106a-5p <400> 58 aaaagugcuu acagugcagg uag 23 <210> 59 <211> 23 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-221-3p <400> 59 agcuacauug ucugcugggu uuc 23 <210> 60 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-22-3p <400> 60 aagcugccag uugaagaacu gu 22 <210> 61 <211> 21 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-23b-3p <400> 61 aucacauugc cagggauuac c 21 <210> 62 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-25-3p <400> 62 cauugcacuu gucucggucu ga 22 <210> 63 <211> 23 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-29b-3p <400> 63 uagcaccauu ugaaaucagu guu 23 <210> 64 <211> 21 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-33a-5p <400> 64 gugcauugua guugcauugc a 21 <210> 65 <211> 23 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-423-5p <400> 65 ugaggggcag agagcgagac uuu 23 <210> 66 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-124-5p <400> 66 cguguucaca gcggaccuug au 22 <210> 67 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-532-5p <400> 67 caugccuuga guguaggacc gu 22 <210> 68 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-200b-3p <400> 68 uaauacugcc ugguaaugau ga 22 <210> 69 <211> 21 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-222-3p <400> 69 agcuacaucu ggcuacuggg u 21 <210> 70 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-199a-3p <400> 70 acaguagucu gcacauuggu ua 22 <210> 71 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-451a <400> 71 aaaccguuac cauuacugag uu 22 <210> 72 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-1226-3p <400> 72 ucaccagccc uguguucccu ag 22 <210> 73 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-127-3p <400> 73 ucggauccgu cugagcuugg cu 22 <210> 74 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-374b-5p <400> 74 auauaauaca accugcuaag ug 22 <210> 75 <211> 21 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-4732-3p <400> 75 gcccugaccu guccuguucu g 21 <210> 76 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-487b <400> 76 aaucguacag ggucauccac uu 22 <210> 77 <211> 21 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-551b-3p <400> 77 gcgacccaua cuugguuuca g 21 <210> 78 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-23c <400> 78 aucacauugc cagugauuac cc 22 <210> 79 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-183-5p <400> 79 uauggcacug guagaauuca cu 22 <210> 80 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-29c-3p <400> 80 uagcaccauu ugaaaucggu ua 22 <210> 81 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-425-3p <400> 81 aucgggaaug ucguguccgc cc 22 <210> 82 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-484 <400> 82 ucaggcucag uccccucccg au 22 <210> 83 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-485-3p <400> 83 gucauacacg gcucuccucu cu 22 <210> 84 <211> 23 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-93-5p <400> 84 caaagugcug uucgugcagg uag 23 <210> 85 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-92a-3p <400> 85 uauugcacuu gucccggccu gu 22 <210> 86 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-140-5p <400> 86 cagugguuuu acccuauggu ag 22 <210> 87 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-15a-5p <400> 87 uagcagcaca uaaugguuug ug 22 <210> 88 <211> 23 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-10b-5p <400> 88 uacccuguag aaccgaauuu gug 23 <210> 89 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-130b-3p <400> 89 cagugcaaug augaaagggc au 22 <210> 90 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-24-3p <400> 90 uggcucaguu cagcaggaac ag 22 <210> 91 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-133b <400> 91 uuuggucccc uucaaccagc ua 22 <210> 92 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-186-5p <400> 92 caaagaauuc uccuuuuggg cu 22 <210> 93 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-193a-5p <400> 93 ugggucuuug cgggcgagau ga 22 <210> 94 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-23a-5p <400> 94 gggguuccug gggaugggau uu 22 <210> 95 <211> 23 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-454-3p <400> 95 uagugcaaua uugcuuauag ggu 23 <210> 96 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-501-5p <400> 96 aauccuuugu cccuggguga ga 22 <210> 97 <211> 23 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-18b-5p <400> 97 uaaggugcau cuagugcagu uag 23 <210> 98 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-223-5p <400> 98 cguguauuug acaagcugag uu 22 <210> 99 <211> 23 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-30c-5p <400> 99 uguaaacauc cuacacucuc agc 23 <210> 100 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-26a-5p <400> 100 uucaaguaau ccaggauagg cu 22 <210> 101 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-146a-5p <400> 101 ugagaacuga auuccauggg uu 22 <210> 102 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-452-5p <400> 102 aacuguuugc agaggaaacu ga 22 <210> 103 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-148a-3p <400> 103 ucagugcacu acagaacuuu gu 22 <210> 104 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-194-5p <400> 104 uguaacagca acuccaugug ga 22 <210> 105 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-29c-5p <400> 105 ugaccgauuu cuccuggugu uc 22 <210> 106 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-196b-5p <400> 106 uagguaguuu ccuguuguug gg 22 <210> 107 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-345-5p <400> 107 gcugacuccu aguccagggc uc 22 <210> 108 <211> 23 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-503 <400> 108 uagcagcggg aacaguucug cag 23 <210> 109 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-627 <400> 109 gugagucucu aagaaaagag ga 22 <210> 110 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-let-7d-3p <400> 110 cuauacgacc ugcugccuuu cu 22 <210> 111 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-30a-5p <400> 111 uguaaacauc cucgacugga ag 22 <210> 112 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-654-3p <400> 112 uaugucugcu gaccaucacc uu 22 <210> 113 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-598 <400> 113 uacgucaucg uugucaucgu ca 22 <210> 114 <211> 21 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-671-3p <400> 114 uccgguucuc agggcuccac c 21 <210> 115 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-132-3p <400> 115 uaacagucua cagccauggu cg 22 <210> 116 <211> 21 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-142-5p <400> 116 cauaaaguag aaagcacuac u 21 <210> 117 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-let-7b-5p <400> 117 ugagguagua gguugugugg uu 22 <210> 118 <211> 23 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-17-5p <400> 118 caaagugcuu acagugcagg uag 23 <210> 119 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-185-5p <400> 119 uggagagaaa ggcaguuccu ga 22 <210> 120 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-486-5p <400> 120 uccuguacug agcugccccg ag 22 <210> 121 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-99b-5p <400> 121 cacccguaga accgaccuug cg 22 <210> 122 <211> 21 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-128 <400> 122 ucacagugaa ccggucucuu u 21 <210> 123 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-16-5p <400> 123 uagcagcacg uaaauauugg cg 22 <210> 124 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-32-5p <400> 124 uauugcacau uacuaaguug ca 22 <210> 125 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-382-5p <400> 125 gaaguuguuc gugguggauu cg 22 <210> 126 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-532-3p <400> 126 ccucccacac ccaaggcuug ca 22 <210> 127 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-181a-2-3p <400> 127 accacugacc guugacugua cc 22 <210> 128 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-139-5p <400> 128 ucuacagugc acgugucucc ag 22 <210> 129 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-21-5p <400> 129 uagcuuauca gacugauguu ga 22 <210> 130 <211> 17 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-1280 <400> 130 ucccaccgcu gccaccc 17 <210> 131 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-331-5p <400> 131 cuagguaugg ucccagggau cc 22 <210> 132 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-150-5p <400> 132 ucucccaacc cuuguaccag ug 22 <210> 133 <211> 21 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-101-3p <400> 133 uacaguacug ugauaacuga a 21 <210> 134 <211> 23 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-200c-3p <400> 134 uaauacugcc ggguaaugau gga 23 <210> 135 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-205-5p <400> 135 uccuucauuc caccggaguc ug 22 <210> 136 <211> 22 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-505-3p <400> 136 cgucaacacu ugcugguuuc cu 22 <210> 137 <211> 23 <212> RNA <213> Homo sapiens <220> <221> misc_feature <223> hsa-miR-136-5p <400> 137 acuccauuug uuuugaugau gga 23

Claims

1. Use of a reagent for measuring the levels of at least three miRNAs in a plasma sample obtained from a subject in the preparation of a diagnostic agent for a method of determining whether the subject has heart failure, wherein the at least three miRNAs include hsa-miR-454-3p, hsa-miR-24-3p, and hsa-miR-551b-3p, and the method comprises the following steps: (a) Measure the levels of at least hsa-miR-454-3p, hsa-miR-24-3p, and hsa-miR-551b-3p in a plasma sample obtained from the subject; and (b) Determine whether it is different compared to a control, where an increased level of miR-24-3p and decreased levels of hsa-miR-454-3p and hsa-miR-551b-3p indicate that the subject has heart failure, where the subject is a human subject; and where the control is a human subject without heart failure.

2. Use of a reagent for measuring the levels of at least three miRNAs in a plasma sample obtained from a subject in the preparation of a diagnostic agent for a method of determining whether the subject has heart failure, wherein the at least three miRNAs include hsa-miR-454-3p, hsa-miR-24-3p, and hsa-miR-551b-3p, and the method comprises the following steps: (a) Measure the levels of at least hsa-miR-454-3p, hsa-miR-24-3p, and hsa-miR-551b-3p in a plasma sample obtained from the subject; and (b) Use a score based on the levels of the miRNAs measured in step (a) to predict the likelihood that the subject has heart failure, where the score is calculated based on Equation 1 and Equation 3, or Equation 2 and Equation 3, as follows: Equation 1 - where log2 copy_number_miRNA i is the log-transformed copy number of an individual miRNA, with the unit of the copy number being copies / ml plasma; K i is a coefficient for weighting multiple miRNA targets; and B is a constant value for adjusting the scale of the prediction score; Equation 2 - where, log2 copy_miRNA i is the log-transformed copy number of individual miRNAs, and the unit of the copy number is copies / ml plasma; K i is a coefficient for weighting multiple miRNA targets; B is a constant value for adjusting the scale of the prediction score; BNP is a measure that is positively or negatively correlated with the level of BNP and / or NT-proBNP in the sample; Equation 3 – 3. The use according to claim 2, wherein an increase in the level of miR-24-3p and a decrease in the levels of hsa-miR-454-3p and hsa-miR-551b-3p indicate that the subject has heart failure.

4. The use according to any one of claims 1-3, wherein step (a) further comprises measuring the level of at least one additional miRNA listed in Table 16 or Table 23 in the plasma sample obtained from the subject, wherein the miRNAs listed in Table 16 or Table 23 are as follows: Table 16 ; or Table 23 5. The use according to claim 1 or 2, wherein the subject is a subject of the Asian race.

Images