WO2011161186A1 - Method for in vitro diagnosing sepsis utilizing biomarker composed of more than two different types of endogenous biomolecules - Google Patents

Abstract

The present invention relates to an in vitro method for predicting a likelihood of an onset of a sepsis by a) selecting at least two distinct types of biomolecules, wherein said types of biomolecules are selected from the group consisting of: RNAs and/or their DNA counterparts, mRNAs and/or their DNA counterparts, non-coding RNAs and/or their DNA counterparts, microRNAs and/or their DNA counterparts, and endogenous target metabolites; b) measuring at least one parameter selected from the group consisting of presence or absence, qualitative and/or quantitative molecular pattern and/or molecular signature, level, amount, concentration and expression level of a plurality of biomolecules of each species in said sample using said two sets of distinct types of biomolecules and storing the obtained set of values as raw data in a database; after mathematical processing predict a class label of an obtained data set in order to calculate prognostic likelihoods.

Description

Method for in vitro diagnosing sepsis utilizing biomarker composed of more than two different types of endogenous biomolecules

The present invention relates to a method for in vitro diagnosing sepsis in accordance with claim 1 and a kit for carrying out the in vitro diagnosis method in accordance with claim 14.

BACKGROUND of the Invention

In classical patient screening and diagnosis, the medical practitioner uses a number of diagnostic tools for diagnosing a patient suffering from a certain disease. Among these tools, measurement of a series of single routine parameters, e.g. in a blood sample, is a common diagnostic laboratory approach. These single parameters comprise for example enzyme activities and enzyme concentration and/or detection of metabolic indicators such as glucose and the like. As far as such diseases are concerned which easily and unambiguously can be correlated with one single parameter or a few number of parameters achieved by clinical chemistry, these parameters have proved to be indispensible tools in modern laboratory medicine and diagnosis. Under the provision that excellently validated cut-off values can be provided, such as in the case of diabetes, clinical chemical parameters such as blood glucose can be reliably used in diagnosis.

In particular, when investigating pathophysiological states underlying essentially a well known pathophysiological mechanism, from which the guiding parameter is resulting, such as a high glucose concentration in blood typically reflects an inherited defect of an insulin gene, such single parameters have proved to be reliable biomarkers for "its" diseases.

However, in a number of pathophysiological conditions, such as cancer or demyelinating diseases such as multiple sclerosis or sepsis and sepsis related conditions, which share a lack of an unambiguously assignable single parameter or marker, differential diagnosis from blood or tissue samples is currently difficult to impossible.

Sepsis is a common cause of mortality and morbidity worldwide. In a recent evaluation of the epidemiology of severe sepsis in children in the United States, the estimated annual incidence of severe sepsis in newborns is 0.3 per 100 live births (Watson et al., 2003), with most mortality occurring within the first 48 hours of infection. (Weinschenk et al., 2000; Stoll et al., 2002).

Clinically, sepsis is a term used to describe symptomatic bacteremia, with or without organ dysfunction. Sustained bacteremia, in contrast to transient bacteremia, may result in a sustained febrile response that may be associated with organ dysfunction. Septicemia refers to the active multiplication of bacteria in the bloodstream, leading to an overwhelming infection. The pathophysiology of sepsis is complex and the roles of inflammation, coagulation, and suppressed fibrinolysis are emerging as important mechanisms in the pathophysiology of sepsis. These mediators of inflammation are often responsible for the clinically observable effects of the bacteremia in the host. Furthermore, impaired pulmonary, hepatic, or renal function may result from excessive release of inflammatory mediators during a septic process.

The mortality rate associated with sepsis, severe sepsis and septic shock are high and reported as 25 to 30% and 40 to 70% respectively (Bernard GR, Vincent JL, Laterre PF, et al., "Efficacy and safety of recombinant human activated protein C for severe sepsis", N. Engl. J. Med. 2001 ; 344: 699-709;Annane D, Aegerter P, Jars-Guincestre MC, Guidet B., "Current epidemiology of septic shock: the CUB-Rea Network", Am. J. Respir. Crit. Care Med. 2003; 168: 165-72).

In sepsis diagnosis, and particularly with respect to an early result in diagnosis and/or prognostic information on the bedside of intensive care patients, in the prior art a number of approaches were undertaken. So it is well established clinical laboratory routine by intensive care specialists to measure single parameters such as white blood cell count (WBC), C- reactive Protein (CRP) and procalcitonin (PCT), in a patient showing clinical suspicion signs of a sepsis or sepsis related disease.

However, although, in accordance with guidelines of the Leitlinien der Deutschen Sepsis- Gesellschaft und der Deutschen Interdisziplinaren Vereinigung fur Intensiv- und Notfallmedizin, Prevention, Diagnose, Therapie und Nachsorge der Sepsis (AWMF- Leitlinien-Register Nr. 079/001 , Arbeitsgemeinschaft der Wissenschaftlichen Medizinischen Fachgesellschaften, AWMF-online, 16.03.2010), the use of PCT measurements are recommended as an early sepsis marker, particularly for excluding a severe sepsis, or confirmation of clinical diagnosis, the PCT detection is rather unreliable: According to AWMF procalcitonin concentrations of <0,5 ng/ml in serum render a severe sepsis or a septic shock as rather unlikely, whereas a PCT concentration beginning with a threshold of 2,0 ng/ml renders a severe sepsis or septic shock as highly likely. However, AWMF advises that traumatic stress caused by surgery and further effects can lead to a sepsis independent temporary increase of PCT (Meisner M, Tschaikowsky K, Hutzler A, Schick C, Schuttler J. Postoperative plasma concentrations of procalcitonin after different types of surgery. Intensive Care Med 1998;24(7):680-4.).

In summary, single markers such as e.g. CRP and/or PCT proved to be useful in clinical practice in many cases only mostly in context with other diagnostic tools such as general clinical diagnostic signs of sepsis, but are not reliable in routine screening for an onset of sepsis.

Vis-a-vie the prior art of the above mentioned single sepsis markers, it was a progress to use gene expression levels of a plurality of sepsis associated genes with the microarray or PCR technology.

Furthermore, non-prepublished EP09180820 discloses the use of a plurality of endogenous target metabolites for predicting a likelihood of an onset of a sepsis by measuring an oxysterol metabolic profile, and EP09167018 describes a method for predicting a likelihood of onset of an inflammation associated organ failure by means of investigating a metabolomics profile.

The content of EP09180820 and EP09167018, filed by the present applicant, is herewith in its entirety incorporated by reference.

The prior art discloses minimally invasive sample procurement method for obtaining blood cell RNA that can be analyzed by expression profiling, e.g., by array-based gene expression profiling. These methods can be used to identify patterns of gene expression that are diagnostic of sepsis, to identify subjects at risk for developing critical disorders and to custom design an array, e.g., a microarray, for the diagnosis or prediction of sepsis and systemic inflammation and or infection-related disorders or susceptibility to sepsis and related disorders.

While single transcripts or proteins are useful biomarkers in a variety of well characterized diseases such as troponin for early detection of myocardial infarction - the complexity of the host response in sepsis makes it unlikely that one single biomarker can adequately describe and stratify this complex syndrome. Multiple gene approaches are much more reliable then the above mentioned single parameters, however, are subject to complex mathematical and bioinformatics procedures. Nevertheless, these gene expression signatures are promising tools in sepsis and cancer diagnosis, but sometimes also have uncertainty limits what leads due to their underlying statistics, being restricted to one kind of nucleic acids this also sometimes affords unreliable results and validation problems.

Staring from the above mentioned prior art, it is the problem of the present invention to provide a use of biomarkers in diagnostics tools with the highest possible sensitivity and specificity for early diagnosis to identify diseased subjects, for use in patient pre-selection and stratification and for therapy control is a main goal in diagnostic development and still an urgent need in various complex diseases, in particular in sepsis. Sepsis is defined as the innate host response to infection.

There is an urgent need for reliable biomarkers for early diagnosis and for therapeutically monitoring of the disease, affecting 25 % of admissions to the intensive care units (ICU) with a mortality rate around 50 % [1 ].

A multitude of endogenous mediators of inflammation, coagulation, metabolism and cellular stress response are changed in expression and functional activity. Unique constituents of host response are attractive targets for therapy, diagnosis and/or monitoring of therapeutic interventions. Overall, sepsis is a significant health care problem with an estimated number of 750 000 cases per year in the U.S. [2-4]. Moreover, sepsis is the third leading cause of death in Western countries [2-4]. In the last decades, the incidence has increased despite recent advances in healthcare. Also, incidence is predicted to rise continuously due to more aggressive surgical interventions in older patients with numerous co-morbidities. While our understanding of molecular mechanisms of sepsis and organ failure has expanded, there is a need for translation into practice of intensive care, such as earlier diagnosis, more accurate identification of patients at risk and reliable monitoring of patients with sepsis and its sequelae, i.e. a need to identify and evaluate reliable biomarkers.

The term ^'systemic inflammatory response syndrome^' (SIRS) was proposed by the consensus conference of experts in 1991 [7, 8] for the description of a non-specific inflammatory process following various insults such as trauma, burns, infections, pancreatitis and other diseases. SIRS represents the physiological derangements that are non-specific, but frequently present in patients with sepsis. The clinical phenotype of "sepsis" - resulting from the interaction between either danger associated molecular patterns (DAMP) in the case of sterile systemic inflammation or pathogen associated molecular patterns (PAMP) of the invading pathogen and the human host [9] - is diagnosed when the signs of symptoms of SIRS are evident and the cause is confirmed or strongly suggested as an infectious process. Addition of predisposing factors or biomarkers such as C- reactive protein (CRP) or procalcitonin (PCT) and a panel of cytokines may contribute to identify infection as the cause of SIRS. It is recognized that progression from SIRS to sepsis/severe sepsis and septic shock is associated with an incrementally higher mortality risk [1 , 10].

Biomarker identification and evaluation for diagnosis and monitoring of sepsis have been performed using different approaches. In patients with sepsis, the plasma levels of constituents of micro-organisms such as endotoxin from Gram-negative bacteria may serve as a marker of infection [1 1 ]. On the other hand, proteins derived from the host circulating in plasma such as TNF-oc, IL-6, or activated protein C could be monitored [12]. A protein marker identified by chance is the prohormone procalcitonin (PCT) [15], a marker of infection without any plausible link of action within the field of sepsis, but with clear association with the disease.

Systems biology approaches utilizing varying omics approaches such as genomics, proteomics and metabolomics are increasingly applied to research and diagnostics of complex diseases. These technologies may provide data and biological indicators, so-called (prognostic, predictive and pharmacodynamic) biomarkers with the potential to revolutionize clinical practice in diagnosis. The molecular sepsis signature. Deigner HP, Kohl M., Crit Care Med. 2009 Mar;37(3):1 137-8.

Recent advances in diagnostic tools e.g. in cancer diagnostics typically comprise multi- component tests utilizing several biomarkers of the same class of biomolecules such as several proteins, RNA or micro RNA species and the analysis of high dimensional data gives a deeper insight into the abnormal signaling and networking which has a high potential to identify previously not discovered marker candidates. However, methods according to the present state of the art utilize single biomolecules or sets of a single type of biomolecules for biomarkers sets such as several RNA, microRNA or protein molecules. See Garzon R, Volinia S, Liu CG, Fernandez-Cymering C, Palumbo T, Pichiorri F, Fabbri M, Coombes K, Alder H, Nakamura T, Flomenberg N, Marcucci G, Calin GA, Kornblau SM, Kantarjian H, Bloomfield CD, Andreeff M, Croce CM, MicroRNA signatures associated with cytogenetics and prognosis in acute myeloid leukemia, Blood. 2008; 1 1 1 (6):3183-9 and Ramaswamy S, Tamayo P, Rifkin R, Mukherjee S, Yeang CH, Angelo M, Ladd C, Reich M, Latulippe E, Mesirov JP, Poggio T, Gerald W, Loda M, Lander ES, Golub TR., Multiclass cancer diagnosis using tumor gene expression signatures. Proc Natl Acad Sci USA. 2001 ;98(26):15149-54. For miRNA in Cancer see WO2008055158.

Although these products and prototypes demonstrate significant progress for specific areas of diagnostics, there is still an urgent need for reliable and early diagnostics with high sensitivities and specificities in a number of complex diseases such as sepsis. These diagnostic tools and biomarkers are also being used for the selection of responders among patients, for an assessment of disease recurrence, the selection of therapeutic options, efficacy, drug resistance and toxicity. Despite some advances in the management of severe sepsis and septic shock, problems remain regarding the usefulness of the currently used definitions and the often encountered delays in diagnosis. The reliable diagnosis of sepsis still remains a challenge. The diagnostics of sepsis and SIRS so far relies on clinical criteria and biochemical parameters limited to protein markers such as procalcitonin (PCT), C reactive protein (CRP) and interleukins (Beger HG, Rau BM, "Severe acute pancreatitis: Clinical course and management World", J. Gastroenterol. 2007, 13, 5043-51 ).

Besides antibiotics therapy, the treatment of sepsis is still limited to preventive measures and symptomatic supportive strategies. Current diagnostics in clinical routine is limited to a) clinical information b) use of basic biochemical as outlined below in the definitions.

Or, by the use of unspecific biomarkers like C-reactive protein (CRP) and procalcitonin (PCT) which have low sensitivities and specificities (Carrol et al., 2002; Toikka et al., 2000; Gendrel et al., 1997). Moreover, the levels of these biomarker are also raised in many non sepsis- associated diseases like in small-cell lung cancer, medullary thyroid cancer and inhalation injury (Becker et al., 1993).

In classical patient screening and diagnosis, the medical practitioner uses a number of diagnostic tools for diagnosing a patient suffering from a certain disease. Among these tools, measurement of a series of single routine parameters, e.g. in a blood sample, is a common diagnostic laboratory approach. These single parameters comprise for example enzyme activities and enzyme concentration and/or detection.

As far as such diseases are concerned which easily and unambiguously can be correlated with one single parameter or a few number of parameters achieved by clinical chemistry, these parameters have proved to be indispensable tools in modern laboratory medicine and diagnosis. However, in pathophysiological conditions, such as cancer or sepsis which share a lack of an unambiguously assignable single parameter or marker, differential diagnosis from blood or tissue samples is currently difficult to impossible.

Although solely nucleic acid-based diagnosis of sepsis from blood cells has been explored recently (Yadav et al., 2008; Andrade et al., 2008), these approaches, however, suffer from several serious limitations:

The required sample size of usually several ml of blood is a problem for continuous monitoring of a critically ill subject; alternatives applying amplification of transcripts are lengthy and prone to error. The whole procedure affords numerous steps and due to laborious sample preparation and nucleic acid isolation, transcription and array or PCR analysis still takes at least several hours and a large technological effort.

Currently used diagnostic methods thus require time and appropriate equipment with high costs and frequently unsatisfying sensitivities. However this used diagnostic means have major limitations either to reduced area under the curve (AUC) and/or delay of diagnosis or increased costs due to equipment required. Accordingly these procedures do not allow a timely assessment of an acute and rapidly evolving disease and overall the situation is far from satisfying and from providing a rapid and reliable diagnosis of sepsis.

Therefore, there is still an urgent need for an early, rapid and reliable diagnosis of sepsis, differentiation of sepsis and SIRS or any other state of health providing the unspecific clinical symptoms listed in the definition of sepsis as well as sepsis severity, ideally requiring only minute amounts of blood; there is an urgent need for timely treatment and early diagnosis of sepsis as well as, an urgent need for therapy monitoring. Further, there is an urgent need for early sepsis biomarkers enabling early and reliable diagnosis, risk stratification of sepsis per se, sepsis duration and severity.

To meet the above needs, the prior art provides a number of biomarker approaches to early prognostic evaluations of sepsis, none of which however, has lead to any clinical relevance, let alone to true reliability: Xu, Ping-bo et al. "A metabolomic approach to early prognostic evaluation of experimental sepsis", Journal of Infection (2008) 56, 474-481 , describe the use of linolenic acid, linoleic acid, oleic acid, stearic acid, docosahexanenoic acid and docosapentaenoic acid in experimental rat sepsis with a predictive accuracy of 94%. No human data is provided.

Moyer, E. et al., "Multiple Systems Organ Failure: VI. Death Predictors in the Trauma-Septic State - The Most Critical Determinants", The Journal of Trauma (1981 ) 21 , 862 - 869 disclose the use of plasma amino acids, urea, lactate, ornithine, gucagon, glucose, alpha- aminobutyrate and a few proteins for predicting death in patients in a trauma-septic state with an accuracy of 99%, 9 days before demise.

WO 2006/071583 A2 describes methods and compositions for determining therapy regimens in systemic inflammatory response syndromes (SIRS), sepsis, severe sepsis, septic shock and/or multiple organ dysfunction syndrome by means of biomarkers. The biomarkers of WO 2006/071583 A2 are selected from the group consisting of at least one of matrix metalloproteinase 9 (MMP-9), interleukins-1 β, interleukin-6, interleukin-8, interleukin-8_6-77 , interleukin-10, interleukin-22, interleukin-1 receptor agonist, chemokine (C-X-C motif) ligand 6 [CXCL6], CXCL13, CXCL16, chemokine (C-C motif) ligand 8 [CCL8], CCL20, CCL23, CCL26, D-dimer, high mobility group protein-1 (HMG-1 ), tumor necrosis factor-oc, A-type natriuretic protein, B-type natriuretic protein, C-type natriuretic protein, C-reactive protein, caspase-3, calcitonin, procalcitonin₃- 6, soluble DPP-IV, soluble FAS ligand, creatinine kinase-BB, vascular endothelial growth factor, myeloperoxidase, and soluble intercellular adhesion molecule-1 . Early diagnosis of sepsis or in vitro prediction of a likelihood of an onset of sepsis are not mentioned.

A similar list of protein biomarkers is disclosed in US 2008/0050829 A1 for determining the status of sepsis in an individual using mass spectrometry.

Finally, not pre-published EP 09167018.2 filed on 31 . July 2009 by the present applicant with the EPO, having the title "Method of Diagnosing Organ Failure" describes a metabolomic method for predicting the likelihood of onset of an inflammation associated organ failure and/or sepsis associated organ failure from a biological sample of a mammalian subject in vitro. As biomarkers EP 09167018.2 uses a number of compounds such as amino acids, amino acid dimers, phenylthio carbamyl amino acids; carboxylic acids; ceramides with an N- acyl residue having from 1 to 30 carbon atoms in the acyl residue and having from 0 to 5 double bonds and from 0 to 5 hydroxy groups; carnitine and acylcarnitines having from 1 to 20 carbon atoms in the acyl residue; phospholipides; phosphatidylcholines having a total of 1 to 50 carbon atoms in the acyl residues; sphingolipids; prostaglandines; putrescine; oxysterols; biogenic amines and bile acids.

Thus, starting from the closest prior art of Moyer, E. et al., "Multiple Systems Organ Failure: VI. Death Predictors in the Trauma-Septic State - The Most Critical Determinants", The Journal of Trauma (1981 ) 21 , 862 - 869 using plasma amino acids, urea, lactate, ornithine, gucagon, glucose, alpha-aminobutyrate as deaths predictors in patients in a trauma-septic state, it was the problem of the present invention, to provide a reliable and easy to handle alternative method of early sepsis diagnosis with a high accuracy of prediction and clinical usefulness.

1. Engel, C; Brunkhorst, F.M.; Bone, H.G.; Brunkhorst, R.; Gerlach, H.; Grond, S.; Gruendling, M.; Huhle, G.; Jaschinski, U.; John, S.; Mayer, K.; Oppert, M.; Olthoff, D.; Quintel, M.; Ragaller, M.; Rossaint, R.; Stuber, F.; Weiler, N.; Welte, T.; Bogatsch, H.; Hartog, C.; Loeffler, M.; Reinhart, K. Epidemiology of sepsis in Germany: results from a national prospective multicenter study. Intensive Care Med, 2007, 33: 606-18.

2. Martin, G.S.; Mannino, D.M.; Eaton, S.; Moss, M. The epidemiology of sepsis in the United States from 1979 through 2000. N Engl J Med, 2003, 348: 1546-54.

3. Angus, D.C.; Linde-Zwirble, W.T.; Lidicker, J.; Clermont, G.; Carcillo, J.; Pinsky, M.R.

Epidemiology of severe sepsis in the United States: analysis of incidence, outcome, and associated costs of care. Crit Care Med, 2001, 29: 1303-10.

4. Angus, D.C.; Pereira, C.A.; Silva, E. Epidemiology of severe sepsis around the world. Endocr Metab Immune Disord Drug Targets, 2006, 6: 207-12.

5. Dellinger, R.P.; Levy, M.M.; Carlet, J.M.; Bion, J.; Parker, M.M.; Jaeschke, R.; Reinhart, K.;

Angus, D.C.; Brun-Buisson, C.; Beale, R.; Calandra, T.; Dhainaut, J.F.; Gerlach, H.; Harvey, M.; Marini, J.J.; Marshall, J.; Ranieri, M.; Ramsay, G.; Sevransky, J.; Thompson, B.T.; Townsend, S.; Vender, J.S.; Zimmerman, J.L.; Vincent, J.L. Surviving Sepsis Campaign: international guidelines for management of severe sepsis and septic shock: 2008. Crit Care Med, 2008, 36: 296-327. 6. Dellinger, R.P.; Carlet, J.M.; Masur, H.; Gerlach, H.; Calandra, T.; Cohen, J.; Gea-Banacloche, J.; Keh, D.; Marshall, J.C.; Parker, M.M.; Ramsay, G.; Zimmerman, J.L.; Vincent, J.L.; Levy, M.M. Surviving Sepsis Campaign guidelines for management of severe sepsis and septic shock. Crit Care Med, 2004, 32: 858-73.

7. American College of Chest Physicians/Society of Critical Care Medicine Consensus Conference: definitions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis. Crit Care Med, 1992, 20: 864-74.

8. Bone, R.C.; Balk, R.A.; Cerra, F.B.; Dellinger, R.P.; Fein, A.M.; Knaus, W.A.; Schein, R.M.;

Sibbald, W.J. Definitions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis. The ACCP/SCCM Consensus Conference Committee. American College of Chest Physicians/Society of Critical Care Medicine. Chest, 1992, 101: 1644-55.

9. Adib-Conquy, M.; Cavaillon, J.M. Stress molecules in sepsis and systemic inflammatory response syndrome. FEBS Lett, 2007, 581: 3723-33.

10. Brun-Buisson, C. The epidemiology of the systemic inflammatory response. Intensive Care Med, 2000, 26 Suppl 1: S64-74.

11. Marshall, J.C.; Reinhart, K. Biomarkers of sepsis. Crit Care Med, 2009, 37: 2290-8.

12. Marshall, J.C. Biomarkers of sepsis. Curr Infect Dis Rep, 2006, 8: 351-7.

13. Levi, M. Activated protein C in sepsis: a critical review. Curr Opin Hematol, 2008, 15: 481-6.

14. Bernard, G.R.; Vincent, J.L.; Laterre, P.F.; LaRosa, S.P.; Dhainaut, J.F.; Lopez-Rodriguez, A.;

Steingrub, J.S.; Garber, G.E.; Helterbrand, J.D.; Ely, E.W.; Fisher, C.J., Jr. Efficacy and safety of recombinant human activated protein C for severe sepsis. N 'Engl J Med, 2001, 344: 699-709.

15. Schuetz, P.; Christ-Crain, M.; Muller, B. Biomarkers to improve diagnostic and prognostic accuracy in systemic infections. Curr Opin Crit Care, 2007, 13: 578-85.

The invention provides the principle and the method for the generation of novel diagnostic tools to diagnose sepsis with superior sensitivities and specificities to address these problems.

Data integration of various "omics" data, e.g. to identify possible alterations of protein concentrations from altered RNA transcripts is an issue familiar to systems biology and to persons skilled in the arts for years.

Despite of that, the statistical combination of biomarker sets from different types of biomolecules, independent of data integration and biochemical interpretation to combined diagnostic signatures (combining several types of biomolecules) on a statistical basis applying various classification methods as described here is not obvious, unknown to persons skilled in the art, and has not been described in the literature. It clearly is distinct to approaches utilizing an integrative multi-dimensional analysis and combining e.g. genomes, epigenomes and transcriptomes (see SIGMA2: A system for the integrative genomic multi-dimensional analysis of cancer genomes, epigenomes, and transcriptomes, Raj Chari et al. BMC Bioinformatics 2008, 9:422) which attempt to analyse biological relationships between different omics data by various means.

Essentially, the method according to the present invention combines statistically significant biomolecule parameters of at least two different types of biomolecules on a statistical basis, entirely irrespective of known or unknown biological relationship of any kind, links or apparent biological plausibility to afford a combined biomarker composed of several types of biomolecules. The patient cases underlying the invention demonstrate that a diagnostic method and disease state specific classifier composed of at least two of the biomolecule types and those combined biomolecules of at least two distinct types describing the respective state of cells, a tissue, an organ or an organisms best among a collective of measured molecules, is superior to a composition of molecules or markers and their delineated molecular signatures. It is further superior to classifiers of biomolecules of just one type of biomolecules and as demonstrated here yields higher sensitivities and specificities in diagnostic applications. In that, the present invention goes far beyond the current state of the art and provides a method for generating diagnostic molecular signatures affording higher sensitivities and specificities and decreased false discovery rates compared to methods available so far.

We here show that a combination of two or more endogenous biomolecules out of two or more different types of biomolecules and selected by statistical means provides diagnostic tools with superior sensitivities and specificities in sepsis.

The above problem is solved by a method in accordance with claim 1 and a Kit according to claim 14.

In particular, the present invention provides a method for in vitro diagnosing sepsis or subtypes thereof, selected from the group consisting of: sepsis, severe sepsis, SIRS (Systemic Inflammatory Response Syndrome), septic shock, and sepsis related multiorgan failure.

More specifically, the present invention relates to: An in vitro method for predicting a likelihood of an onset of a sepsis in at least one biological sample of at least one tissue of a mammalian subject comprising the steps of: a) detecting at least two distinct types of biomolecules comprising a plurality of nucleic acids and a plurality of endogenous target metabolites, which metabolites have a molecular mass less than 1500 Da; b) measuring at least one parameter selected from the group consisting of nucleic acid expression level; presence or absence, level, amount, concentration of each individual biomolecule of each type in said sample and storing the obtained measured values as raw data in a database; c) mathematically preprocessing said raw data in order to reduce technical errors being inherent to the measuring procedures used in step b); d) selecting at least one suitable classifying algorithm from the group consisting of logistic regression, (diagonal) linear or quadratic discriminant analysis (LDA, QDA, DLDA, DQDA), perceptron, shrunken centroids regularized discriminant analysis (RDA), random forests (RF), neural networks (NN), Bayesian networks, hidden Markov models, support vector machines (SVM), generalized partial least squares (GPLS), partitioning around medoids (PAM), self organizing maps (SOM), recursive partitioning and regression trees, K- nearest neighbor classifiers (K-NN), fuzzy classifiers, bagging, boosting, and na^'ive Bayes; and applying said selected classifier algorithm to said preprocessed data of step c); e) said classifier algorithms of step d) being trained on at least one training data set containing preprocessed data from subjects being divided into classes according to their pathophysiological condition with respect to sepsis and/or their prognostic or responder conditions, in order to select a classifier function to map said preprocessed data to said conditions; f) applying said trained classifier algorithms of step e) to a preprocessed data set of a subject with unknown pathophysiological condition with respect to sepsis and/or their prognostic or responder conditions, and using the trained classifier algorithms to obtain a score for predicting the class label of said data set in order to calculate prognostic or responder likelihoods.

In accordance with the present invention, said endogenous target metabolites are selected from the group consisting of:

Bile acids, oxysterols; amino acids, phenylthio carbamyl amino acids (PTC-amino acids), dimethylarginine; carboxylic acids; Ceramides with an N-acyl residue having from 2 to 30 Carbon atoms in the acyl residue and having from 0 to 5 double bonds and having from 0 to 5 hydroxy groups; carnitine; acylcarnitines having from 1 to 20 carbon atoms in the acyl residue; acylcarnitines having from 3 to 20 carbon atoms in the acyl residue and having 1 to 4 double bonds in the acyl residue; acylcarnitines having from 1 to 20 carbon atoms in the acyl residue and having from 1 to 3 OH-groups in the acyl residue; acylcarnitines having from 3 to 20 carbon atoms in the acyl residue with 1 to 4 double bonds and 1 to 3 OH-groups in the acyl residue; phospholipides; phosphatidylcholines (diacylphosphatidylcholines) having a total of from 1 to 50 carbon atoms in the acyl residues; phosphatidylcholines having a total from 3 to 50 carbon atoms in the acyl residues and having a total of 1 to 8 double bonds in the acyl residues; sphingolipids, in particular sphingomyelines having a total number of carbon atoms in the acyl chains from 10 to 30; sphingomyelines having a total number of carbon atoms in the acyl chains from 10 to 30 and 1 to 5 double bonds; hydroxysphinogomyelines having a total number of carbon atoms in the acyl residues from 10 to 30; hydroxysphingoyelines having a total number of carbon atoms in the acyl residues from 10 to 30 and 1 to 5 double bonds; prostaglandines, namely 6-keto-prostaglandin F1 alpha, prostaglandin D2, thromboxane B2; putrescine; biogenic amines, namely histamine, serotonine, palmitoyl ethanolamine.

In a preferred embodiment, one type of distinct biomolecules are nucleic acids, preferably microRNAs and/or its DNAs and the other type of distinct biomolecules are bile acids as endogenous target metabolites. Preferably, the tissue is selected from the group consisting of blood, plasma, serum and other body fluids, cerebrospinal fluids, bone tissue, bone marrow tissue, muscular tissue, glandular tissue, brain tissue, nerve tissue, mucous tissue, connective tissue, and skin tissue and/or said sample is a biopsy sample and/or said mammalian subject includes humans; and/or further characterized in that the score is combined with a standard lab parameters commonly used in clinical chemistry, such as serum and/or plasma levels of low molecular weight biochemical compounds, enzymes, enzymatic activities, cell surface receptors and/or cell counts, in particular red and/or white cell counts, platelet counts, for predicting the class label of said data set in order to calculate prognostic or responder likelihoods.

It is preferred, that said step of mathematically preprocessing of said raw data obtained in step b) is carried out by a statistical method selected from the group consisting of: in case of raw data obtained by optical spectroscopy (UV, visible, IR, Fluorescence): background correction and/or normalization; in case of raw data obtained from metabolomics obtained by mass spectrometry or by 2D gel electrophoresis: smoothing, baseline correction, peak picking, optionally, additional further data transformation such as taking the logarithm in order to carry out a stabilization of the variances; in case of raw data obtained from transcriptomics: Summarizing single pixel to a single intensity signal; background correction; summarizing of multiple probe signals to a single expression value, in particular perfect match/mismatch probes; normalization;

A further preferred embodiment of the invention is a method in that after preprocessing step c) a further step of feature selection is inserted, in order to find a lower dimensional subset of features with the highest discriminatory power between classes; and said feature selection is carried out by a filter and/or a wrapper approach; wherein said filter approach includes rankers and/or feature subset evaluation methods.

Furthermore, in accordance with the invention, a pathophysiological condition corresponds to the label "diseased" and said physiological condition corresponds to the label "healthy" or said pathophysiological condition corresponds to different labels of "grades of a disease", "subtypes of a disease", different values of a "score for a defined disease"; said prognostic condition corresponds to a label "good", "medium", "poor", or "therapeutically responding" or "therapeutically non-responding" or "therapeutically poor responding".

Typically, high-throughput mass spectrometry data will be said metabolic data.

A preferred method according to the present invention is one in which said mammalian subject is a human being, said biological sample is blood and/or blood cells and/or bone marrow; wherein said target metabolites are bile acids which are selected from the group consisting of:

Compound Short

name

Taurocholic Acid TCA

Glycocholic Acid GCA

Glycochenodeoxycholic

GCDCA

Acid

Taurochenodeoxycholic

TCDCA

Acid

Taurolithococholic Acid TLCA

Glycolithocholic Acid GLCA

Taurolithocholic Acid

TLCAS

sulfate

Glycolithocholic Acid

GLCAS

sulfate

Taurodeoxycholic Acid TDCA

Glycodeoxycholic Acid GDCA

Cholic Acid CA

Chenodeoxycholic Acid CDCA

Ursodeoxycholic Acid UDCA

Deoxycholic Acid DCA

Tauroursodeoxycholic

TUDCA

Acid

Glycoursodeoxycholic

GUDCA

Acid

Lithocholic Acid LCA and said nucleic acid is microRNA, which is selected from the group consisting of SEQ-ID No. 9 to SEQ-ID No. 16;

wherein microRNA expression levels and serum bile acid concentration are used as said parameters of step b);

wherein raw data of microRNA expression are preprocessed using the generalized logarithm as variance-stabilizing normalization and summarizing the normalized multiple probe signals (technical replicates) to a single expression value, using the median;

wherein raw data of bile acids are preprocessed using the logarithm as variance- stabilizing normalization; wherein random forests are selected as suitable feature selection and classifying algorithm, the training of the classifying algorithm including preprocessed and filtered microRNA expression data, is carried out with a .632 bootstrap-validation;

applying said trained random forests classifier to said preprocessed microRNA expression data sets to a subject under suspicion of having sepsis or sepsis related disease, and using the trained classifiers to diagnose sepsis.

In a preferred embodiment of the invention the method is further characterized in that the following DNA probes for targeting said microRNA are used: Seq-ID No. 1 to Seq-ID No. 8; and/or

the following microRNA-target sequences are used: Seq-ID Nos. 9 to 16.

Typically, the metabolites in the samples are analyzed by liquid chromatography and mass spectrometry, wherein the quantification of the measured metabolite concentrations in said biological tissue sample is calibrated by reference to internal standards.

Preferably, the microRNA expression data are obtained by quantitative real time PCR (q-RT- PCR) or by hybridization assays. Preferred metabolites for carrying out the method of the present invention, comprise compounds shown in the following table:

Compound Short name

Taurocholic Acid TCA

Glycocholic Acid GCA

Glycochenodeoxycholic Acid GCDCA

Taurochenodeoxycholic Acid TCDCA

Taurolithococholic Acid TLCA

Chenodeoxycholic Acid CDCA

Deoxycholic Acid DCA

Finally, the present invention relates to Kit for carrying out a method in accordance with the present invention, in a biological sample, comprising:

a) detection agents for the detection of at least two distinct types of biomolecules comprising a plurality of nucleic acids and a plurality of endogenous target metabolites;

b) positive and/or negative controls; and

c) classification software for classification of the results achieved with said detection agents.

The dependant claims are preferred embodiments of the present invention.

The present invention provides a solution to the problem described above, and generally relates to the use of "omics" data comprising, but not limited to nucleic acids expression data, and metabolomics data, statistical learning respectively machine learning for identification of molecular signatures and biomarkers. It comprises analysing of the aforementioned biomolecules via known methods and optimal composed marker sets are extracted by statistical methods and data classification methods.

The values of the individual markers of the different species of biomolecules thus are measured, compared to references, standards or controls and data processed to classifiers indicating diseased states etc. with superior sensitivities and specificities compared to procedures and biomarker confined to one type of biomolecules. A method for the selection and combination of biomarkers and molecular signatures of biomolecules in particular utilizing nucleic acids and metabolites in combination with the biomolecules obtained from body liquids or tissue, identified by use of statistical methods and classifiers derived from the data of these groups of molecules for use in diagnosis and early diagnosis, for patient stratification, therapy selection, therapy monitoring and theragnostics in sepsis is described.

Definitions

For the purpose of the present invention, the term "distinct types of biomolecules" means biological macromolecules such as nucleic acids, proteins, or polysaccharides, on the one hand, and on the other hand, smaller non-polymeric chemical entities such as endogenous naturally occurring metabolites, having a molecular mass less than 1500 Da.

Score: The term "score" might encompass a number score, e.g. a scale from 1 to 10 in which 1 represents the lowest probability or likelihood to develop a sepsis, and 10 represents the highest probability. However, the skilled person will also understand a colour score, e.g. green, yellow, red, in which "green" might represent the lowest probability or likelihood to develop a sepsis, and red the highest.

As used herein, the term "gene expression" refers to the process of converting genetic information encoded in a gene into ribonucleic acid, RNA (e.g., mRNA, rRNA, tRNA, or snRNA) through "transcription" of the gene (i.e., via the enzymatic action of an RNA polymerase), and for protein encoding genes, into protein through "translation" of mRNA. Gene expression can be regulated at many stages in the process. "Up-regulation" or "activation" refers to regulation that increases the production of gene expression products (i.e., RNA or protein), while "down-regulation" or "repression" refers to regulation that decrease production

Polynucleotide: A nucleic acid polymer, having more than 2 bases. "Peptides" are short heteropolymers formed from the linking, in a defined order, of a-amino acids. The link between one amino acid residue and the next is known as an amide bond or a peptide bond.

Proteins are polypeptide molecules (or consist of multiple polypeptide subunits). The distinction is that peptides are short and polypeptides/proteins are long. There are several different conventions to determine these, all of which have caveats and nuances.

Sepsis by definition comprises systemic inflammatory response syndrome due to an infection with pathogens.

Systemic inflammatory response syndrome (SIRS) is considered to be present when two or more of the following clinical findings are present:

Body temperature >38°C or <36°C;

Heart rate >90 min^"1 ;

Hyperventilation evidenced by a respiratory rate of >20 min^"1 or a PaC0₂ of <32 mm Hg; and White blood cell count of >12,000 cells μΙ_^"1 or <4,000 μΙ_^"1

Sepsis includes a systemic inflammatory response syndrome (SIRS) together with an infection.

Sepsis (commonly called a "blood stream infection) denotes the presence of bacteria (bacteremia) or other infectious organisms or their toxins in the blood (septicemia) or in other tissue of the body and the immune response of the host. Sepsis is currently thought to result from the interaction between the host response and the presence of micro-organisms and/or their toxins within the body. The observed host responses include the release of pro and antiinflammatory immune mediators as well as components of the coagulation system. Sepsis thus comprises a systemic response to infection, defined as hypothermia or hyperthermia, tachycardia, tachypnea, a clinically evident focus of infection or positive blood cultures, one or more end organs with either dysfunction or inadequate perfusion, cerebral dysfunction, hypoxaemia, increased plasma lactate or unexplained metabolic acidosis, and oliguria.

While sepsis usually is related to infection, it can also be associated with non-infectious insults due to trauma, burns, and pancreatitis. It is one of the most common causes of adult respiratory distress syndrome.

A precise definition of the term sepsis has been introduced and adopted by the ACCP/SCCM Consensus Conference Committee (1992) in the section: Definition for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis published inCrit Care Med. 20(6):864-874. International Sepsis Definitions Conference, attempted to improve the above definition with the aim of increasing the accuracy of the diagnosis of sepsis in 2001 by Levy M, Fink M, Mitchell P, Marshall JC, Abraham E, et al. for the International Sepsis Definitions Conference. 2001 SCCM/ESICM/ACCP/ATS/SIS. The conference statement suggested that although the SIRS concept was valid, in the future if supported by further epidemiologic data, it may be possible to use purely biochemical and/or immunological data, rather than clinical criteria to identify the inflammatory response. It also defined infection as a pathologic process induced by a micro-organism, and that sepsis should be defined as a patient with documented or suspected 'infection' exhibiting some of the following variables (Levy M, Fink M, Mitchell P, Marshall JC, Abraham E, et al. for the International Sepsis Definitions Conference. 2001 SCCM/ESICM/ACCP/ATS/SIS):

1. General variables

o Fever (core temperature >38.3 °C)

o Hypothermia (core temperature <36°C)

o Heart rate >90 min-1 or >2 SD above the normal value for age

o Tachypnea

o Altered mental status

o Significant oedema or positive fluid balance (>20 ml_/kg over 24 hrs) o Hyperglycemia (plasma glucose >7.7 mmol/L) in the absence of diabetes

2. Inflammatory variables

° Leukocytosis - WBC count >12,000 μί^"1

° Leukopaenia - WBC count <4000 μί^"1 o Normal WBC count with >10% immature forms

o Plasma C-reactive protein >2 SD above the normal value

o Plasma procalcitonin >2 SD above the normal value

3. Hemodynamic variables

o Arterial hypotension (SBP <90 mmHg, MAP <70 mmHg, or an SBP decrease

>40 mmHg in adults)

o SvO2a >70%

° Cardiac index > 3.5 Lmin^"1 M^"2

4. Organ dysfunction variables

o Arterial hypoxemia (PaO₂/FIO₂ <300)

o Acute oliguria (urine output <0.5 mLkg-1 hr^"1 for at least 2hrs)

o Creatinine increase >0.5 mg/dL

o Coagulation abnormalities (INR >1 .5 or aPTT >60 sees)

o Ileus (absent bowel sounds)

o Thrombocytopenia (platelet count <100μΙ_)

o Hyperbilirubinemia (plasma total bilirubin>4 mg/dL or 70 mmol/L)

5. Tissue perfusion variables

o Hyperlactatemia (>1 mmol/L)

o Decreased capillary refill or mottling

(WBC: white blood cell; SBP: systolic blood pressure; MAP: mean arterial blood pressure; SvO2: mixed venous oxygen saturation; INR: international normalized ratio; aPTT: activated partial thromboplastin time; tachycardia (may be absent in hypothermic patients), and at least one of the following indications of altered organ function: altered mental status, hypoxemia, increased serum lactate level, or bounding pulses.

The definition of severe sepsis remained unchanged and refers to sepsis complicated by organ dysfunction. Organ dysfunction is defined using Multiple Organ Dysfunction score (Marshall JC, Cook DJ, Christou NV, et al., "Multiple organ dysfunction score: A reliable descriptor of a complex clinical outcome", Crit Care Med 1995; 23: 1638-1652 or the definitions used for the Sequential Organ Failure Assessment (SOFA) score (Ferreira FL, Bota DP, Brass A, et al., "Serial evaluation of the SOFA score to predict outcome in critically ill patients", JAMA 2002; 286: 1754-1758. Septic shock in adults refers to a state of acute circulatory failure characterized by persistent arterial hypotension unexplained by other causes. Hypotension is defined by a systolic arterial pressure below 90 mm Hg, a MAP <70 mmHg, or a reduction in systolic blood pressure of >40 mm Hg from baseline, despite adequate volume resuscitation, in the absence of other causes for hypotension.

Metabolite: as used here, the term "metabolite" denotes endogenous organic compounds of a cell, an organism, a tissue or being present in body liquids and in extracts obtained from the aforementioned sources with a molecular weight typically below 1500 Dalton. Typical examples of metabolites are carbohydrates, lipids, phospholipids, sphingolipids and sphingophospholipids, amino acids, cholesterol, steroid hormones and oxidized sterols and other compounds such as collected in the Human Metabolite database (http://www.hmdb.ca/) and other databases and literature. This includes any substance produced by metabolism or by a metabolic process and any substance involved in metabolism.

"Metabolomics" as understood within the scope of the present invention designates the comprehensive quantitative measurement of several (2-thousands) metabolites by, but not limited to, methods such as mass spectroscopy, coupling of liquid chromatography, gas chromatography and other separation methods chromatography with mass spectroscopy.

"Oligonucelotide arrays "or" oligonucleotide chips" or "gene chips": relates to a "microarray", also referred to as a "chip", "biochip", or "biological chip", is an array of regions having a suitable density of discrete regions, e. g., of at least 100/cm², and preferably at least about 1000/cm². The regions in a microarray have dimensions, e. g. diameters, preferably in the range of between about 10-250 μππ, and are separated from other regions in the array by the same distance. Commonly used formats include products from Agilent, Affymetrix, lllumina as well as spotted fabricated arrays where oligonucleotides and cDNAs are deposited on solid surfaces by means of a dispenser or manually.

It is clear to a person skilled in the art that nucleic acids, proteins and peptides as well as metabolites can be quantified by a variety of methods including the above mentioned array systems as well as, but not limited to: quantitative sequencing, quantitative polymerase chain reaction and quantitative reverse transcription polymerase chain reaction (qPCR and RT- PCR), immunoassays, protein arrays utilizing antibodies, mass spectrometry.

"microRNAs" (miRNAs) are small RNAs of 19 to 25 nucleotides that are negative regulators of gene expression.

Under different species or types or classes of biomolecules in this context is understood: RNA, microRNA, proteins and peptides of various lengths as well as metabolites.

A biomarker in this context is a characteristic, comprising data of at least two biomolecules of at least two different types (RNA, microRNA, proteins and peptides, metabolites) that is measured and evaluated as an indicator of biologic processes, pathogenic processes, or responses to an therapeutic intervention. A combined biomarker as used here may be selected from at least two of the following types of biomolecules: sense and antisense nucleic acids, messenger RNA, small RNA i.e. siRNA and microRNA, polypeptides, proteins including antibodies, small endogenous molecules and metabolites.

Data classification is the categorization of data for its most effective and efficient use. Classifiers are typically deterministic functions that map a multi-dimensional vector of biological measurements to a binary (or n-ary) outcome variable that encodes the absence or existence of a clinically-relevant class, phenotype, distinct physiological state or distinct state of disease. To achieve this various classification methods such as, but not limited to, logistic regression, (diagonal) linear or quadratic discriminant analysis (LDA, QDA, DLDA, DQDA), perceptron, shrunken centroids regularized discriminant analysis (RDA), random forests (RF), neural networks (NN), Bayesian networks, hidden Markov models, support vector machines (SVM), generalized partial least squares (GPLS), partitioning around medoids (PAM), self organizing maps (SOM), recursive partitioning and regression trees, K-nearest neighbor classifiers (K-NN), fuzzy classifiers, bagging, boosting, and na^'ive Bayes and many more can be used. The term "binding", "to bind", "binds", "bound" or any derivation thereof refers to any stable, rather than transient, chemical bond between two or more molecules, including, but not limited to, covalent bonding, ionic bonding, and hydrogen bonding. Thus, this term also encompasses hybridization between two nucleic acid molecules among other types of chemical bonding between two or more molecules.

DETAILED DESCRIPTION OF THE INVENTION

In the method of the present invention, biomarker data and classifier obtained by combination of different types of biomolecules comprising: nucleic acids and metabolites, identified according to the invention afford a description of a physiological state and can be used as a superior tool for diagnosing complex diseases.

The discrimination of pathological samples or tissues from healthy specimens requires a combination of data of at least two distinct types of biomolecules, a measurement of their parameters and a statistical processing and classifier generation.

As mentioned above a biological link between molecules combined in a biomarker by means of classification is entirely irrelevant to the outcome and selection of the issues and can not be necessarily explained by biological models.

The method according to the present invention comprises essentially the following 7 steps as table 1 :

Step 1 :

Biological sample obtained

Step 2:

different species of biomolecules are detected

Step 3:

Measurement of raw data and deposit in data base

Step 4:

Preprocessing of raw data from data base

Step 5:

Select suitable classifier algorithm

Step 6:

Train classifier based on data of a biomarker composed of

at least two types of biomolecules

Step 7:

Use of the classifier to assess physiological state, as diagnostic tool to indicate sepsis or as a prognostic tool

Table 1: Schematic diagram of proposed method. The detail of each step is described as follows:

1. In the stepl , a biological sample of at least one tissue of a mammalian subject is obtained.

The biological sample can be obtained from any mammalian subject including human, that has any risk of developing sepsis in the sense of the present invention.

Samples to be used in the invention can be obtained in any manner known to a skilled artisan. The invention is not limited to just sample believed to be altered (with regard to concentrations of biomolecules such as nucleic acids and metabolites) due to sepsis. Instead, samples can be derived from any part of the subject containing at least some tissue, cells or body liquids such as blood. Examples of the samples include blood, plasma, serum and other body fluids, cerebrospinal fluids, bone tissue, bone marrow tissue, muscular tissue, glandular tissue, brain tissue, nerve tissue, mucous tissue, connective tissue, and skin tissue. The sample may be a biopsy sample

2. In the step 2, different species of biomolecules comprising a plurality of nucleic acids and a plurality of endogenous target metabolites are detected. In other words, at least two nucleic acids and at least two metabolites are detected. Preferably, 2 to 16 nucleic acids and 2 to 17 metabolites will be used. However, e.g. also 2 to 8 nucleic acids, or less, and e.g. 2 to 8, or less metabolites may be used in accordance with the present invention.

The nucleic acids and the metabolites may be detected from a same biological sample (tissue) or from different same biological samples (tissues).

As used herein, the term "endogenous target metabolite" refers to a compound/metabolite that is differentially present or differentially concentrated in septic organisms compared to non-septic organisms. For example, sepsis predictive metabolites are present in septic tissues but not in non-septic tissues.

The examples of the nucleic acids include RNAs and their DNA counterparts, mRNAs and their DNA counterparts, non-coding RNAs and their DNA counterparts, microRNAs and their DNA counterparts. The nucleic acids are preferably microRNAs and/or their DNA counterparts, more preferably microRNAs.

Preferably the plurality of microRNAs comprise 2, 3, 4, 5, 6, 7 or all of the microRNAs shown in the following table: Seq-ID microRNA ID Target sequence

9 hsa-miR-19a GCAGUCCUCUGUUAGUUUUGCAUAGUUGCACUACAAGAAGAAUGUA

GUUGUGCAAAUCUAUGCAAAACUGAUGGUGGCCUGC

10 hsa-miR-449b UGACCUGAAUCAGGUAGGCAGUGUAUUGUUAGCUGGCUGCUUGGG

UCAAGUCAGCAGCCACAACUACCCUGCCACUUGCUUCUGGAUAAAU UCUUCU

1 1 hsa-miR-380 AAGAUGGUUGACCAUAGAACAUGCGCUAUCUCUGUGUCGUAUGUAA

UAUGGUCCACAUCUU

12 hsa-miR-485 ACUUGGAGAGAGGCUGGCCGUGAUGAAUUCGAUUCAUCAAAGCGA

GUCAUACACGGCUCUCCUCUCUUUUAGU

13 hsa-miR-297 UGUAUGUAUGUGUGCAUGUGCAUGUAUGUGUAUAUACAUAUAUAU

GUAUUAUGUACUCAUAUAUCA

14 hsa-miR-200a CCGGGCCCCUGUGAGCAUCUUACCGGACAGUGCUGGAUUUCCCAG

CUUGACUCUAACACUGUCUGGUAACGAUGUUCAAAGGUGACCCGC

15 hsa-miR-205 AAAGAUCCUCAGACAAUCCAUGUGCUUCUCUUGUCCUUCAUUCCAC

CGGAGUCUGUCUCAUACCCAACCAGAUUUCAGUGGAGUGAAGUUC

AGGAGGCAUGGAGCUGACA

16 hsa-miR-509-1 CAUGCUGUGUGUGGUACCCUACUGCAGACAGUGGCAAUCAUGUAU

AAUUAAAAAUGAUUGGUACGUCUGUGGGUAGAGUACUGCAUGACAC

AUG

The endogenous target metabolites have a molecular mass less than 1500 Da. Examples of such endogenous target metabolites include Bile acids, oxysterols; amino acids, phenylthio carbamoyl amino acids (PTC-amino acids), dimethylarginine; carboxylic acids; Ceramides with an N-acyl residue having from 2 to 30 Carbon atoms in the acyl residue and having from 0 to 5 double bonds and having from 0 to 5 hydroxy groups; carnitine; acylcarnitines having from 1 to 20 carbon atoms in the acyl residue; acylcarnitines having from 3 to 20 carbon atoms in the acyl residue and having 1 to 4 double bonds in the acyl residue; acylcarnitines having from 1 to 20 carbon atoms in the acyl residue and having from 1 to 3 OH-groups in the acyl residue; acylcarnitines having from 3 to 20 carbon atoms in the acyl residue with 1 to 4 double bonds and 1 to 3 OH-groups in the acyl residue; phospholipides; phosphatidylcholines (diacylphosphatidylcholines) having a total of from 1 to 50 carbon atoms in the acyl residues; phosphatidylcholines having a total from 3 to 50 carbon atoms in the acyl residues and having a total of 1 to 8 double bonds in the acyl residues; sphingolipids, in particular sphingomyelines having a total number of carbon atoms in the acyl chains from 10 to 30; sphingomyelines having a total number of carbon atoms in the acyl chains from 10 to 30 and 1 to 5 double bonds; hydroxysphinogomyelines having a total number of carbon atoms in the acyl residues from 10 to 30; hydroxysphingoyelines having a total number of carbon atoms in the acyl residues from 10 to 30 and 1 to 5 double bonds; prostaglandines, namely 6-keto- prostaglandin F1 alpha, prostaglandin D2, thromboxane B2; putrescine; biogenic amines, namely histamine, serotonine, palmitoyi ethanolamine. The preferable endogenous target metabolites are bile acids.

Preferably the bile acids comprise following compounds shown in the following table:

Compound Short

name

Taurocholic Acid TCA

Glycocholic Acid GCA

Glycochenodeoxycholic Acid GCDCA

Taurochenodeoxycholic Acid TCDCA

Taurolithococholic Acid TLCA

Glycolithocholic Acid GLCA

Taurolithocholic Acid sulfate TLCAS

Glycolithocholic Acid sulfate GLCAS

Taurodeoxycholic Acid TDCA

Glycodeoxycholic Acid GDCA

Cholic Acid CA

Chenodeoxycholic Acid CDCA

Ursodeoxycholic Acid UDCA

Deoxycholic Acid DCA

Tauroursodeoxycholic Acid TUDCA

Glycoursodeoxycholic Acid GUDCA

Lithocholic Acid LCA

More preferably the bile acids comprise 2 3, 4, 5, 6 or all of the compounds shown in the following table:

Compound Short name

Taurocholic Acid TCA

Glycocholic Acid GCA

Glycochenodeoxycholic Acid GCDCA

Taurochenodeoxycholic Acid TCDCA

Taurolithococholic Acid TLCA

Chenodeoxycholic Acid CDCA

Deoxycholic Acid DCA 3. In the step3, at least one parameter of each individual biomolecule of the sample is measured and the obtained measured values is stored as raw data in a database.

The examples of the parameter include, nucleic acid expression level; presence or absence, level, amount, concentration. Measuring the parameter can be performed in any manner known by one skilled in the art.

The parameter of nucleic acids can be measured for example by microarrays, transcriptomics quantitative reverse transcriptase polymerase chain reaction, PCR or quantitation and relative quantitation applying sequencing or second generation sequencing.

Another example of a method of quantifying nucleic acids is as follows: hybridizing at least a portion of the nucleic acids with a fluorescent nucleic acid, and reacting the hybridized nucleic acids with a fluorescent reagent, wherein the hybridized nucleic acids emit a fluorescent light. Another method of quantifying the amount of the nucleic acids in a sample is by hybridizing at least a portion the nucleic acids to a radio-labeled complementary nucleic acid. In instances when a nucleic acid capable of hybridizing to the RNA or microRNA is used in the measuring step, in case of the microRNA the nucleic acid is at least 5 nucleotides, at least 10 nucleotides, at least 15 nucleotides, at least 20 nucleotides, at least 25 nucleotides, at least 30 nucleotides or at least 40 nucleotides; and may be no longer than 25 nucleotides, no longer than 35 nucleotides; no longer than 50 nucleotides; no longer than 75 nucleotides, no longer than 100 nucleotides or no longer than 125 nucleotides in length. The nucleic acid is any nucleic acid having at least 80% homology, 85% homology, 90% homology, 95% homology or 100% homology with any of the complementary sequences for the microRNAs.

The parameter of metabolite can be measured for example by mass spectrometry (MS), 2D gel electrophoresis or other methods. Mass spectrometry is used in three main formats:, gas chromatography-mass spectrometry(GC-MS), flow-injection analysis-tandem mass spectrometry, (FIA-MS/MS), and HPLC-tandem mass spectrometry

(HPLC-MS/MS). Each of these formats can be utilized with varying degrees of mass resolution, for example, from unit mass on quadrupole instruments to higher resolution techniques such as Fourier transform mass spectroscopy (FT-ICR-MS) and Orbitrap instruments. The mass spectrometer can be employed as a mass analyzer alone, in which case the ion abundance is plotted against them/z value, or as a tandem mass analyzer, where, for example, in the acquisition of a product-ion spectrum the first analyzer (MS1 ) operates to select ions of a given m/z value and these ions are then fragmented, usually by collision with an inert gas in a collision-induced dissociation (CID) process within a collision cell. The second analyzer (MS2) is then used to measure the m/z value of the resultant fragment ions to give a MS/MS or MS2 spectrum. This process can be extended further by utilizing a third analyzer (MS3), where a fragment ion selected by MS2 is dissociated and its MS/MS/MS or MS3 spectrum recorded. Mass analyzers can be arranged in series in space, such as on ion-beam instruments (for example, tandem quadrupole, Q-TOF instruments), or in time, such as with ion traps. Ion traps are particularly adept at recording MS3 and further MSn spectra. Beam instruments offer other advantages such as SRM "scans", where MS1 is "parked" on an m/z value of interest and MS2 on the m/z value(s) of a known fragment ion. No scanning occurs here, and thus ion transmission to the detector is maximized. Thus, when a component with the m/z value defined by MS1 elutes from the column and fragments to give ions defined by MS2, a signal is recorded. Many SRMs can be programmed to occur in a given "scan", in which case the analyzers jump from one SRM to another. The result is a multiple reaction monitoring (MRM) "scan". In a chromatographic analysis, the SRMs can be timeprogrammed so as to coincide with the approximate time of elution of the components of interest, thereby avoiding "scanning" redundant m/z values during analyte elution and maximixing the sensitivity. Similar "scans" can be recorded on MS-only instruments, where MS1 is "parked" on them/z value of interest and selected ion recording or selected ion monitoring (SIM) chromatograms generated.

Two other important scan modes utilized on ion-beam instruments are precursor ion scans and neutral loss scans. In precursor ion scans, MS2 is set to transmit the m/z value of a known fragment, and MS1 is scanned to identify ions dissociating to give this fragment. In the neutral loss scan, MS1 and MS2 are scanned in parallel and offset by a set m/z value which corresponds to the loss of a defined neutral fragment. The quantification of the measured metabolite concentrations in the sample preferably is calibrated by reference to internal (added to the sample before extraction) or external standards (added to the sample after extraction) so that individual variance between the preparation of the samples and, more importantly, matrix effects and other inferences in the sample are minimized.

The parameter of the biomolecules can be measured by reacting the biomolecules with a reagent capable of binding to at least a portion of the biomolecules to forms a measurable product or complex. The measurable complex is measured according to methods known to a skilled artisan.

The reagent capable of binding to the biomolecules can be any compound known to a skilled artisan as being able to bind to the nucleic acids and metabolites, in a manner that enables one to detect the presence and the amount of the molecule. An example of a compound capable of binding to the biomolecules is a nucleic acid capable of hybridizing or an aptamer. The nucleic acid has at least 5 nucleotides, at least 10 nucleotides, at least 15 nucleotides, at least 20 nucleotides, at least 25 nucleotides, at least 30 nucleotides, at least 40 nucleotides or at least 50 nucleotides; and no longer than 200 nucleotides in length. The nucleic acid is any nucleic acid having at least 80% homology, 85% homology, 90% homology, 95% homology or 100% homology with a sequence complementary to a RNA or microRNA or an aptamer capable of binding RNA, microRNA, and metabolite. One specific example of a nucleic acid capable of binding to RNA or microRNA is a nucleic acid primer for use in a reverse transcriptase polymerase chain reaction.

4. In the step4, the raw data from the database are mathematically preprocessed in order to reduce technical errors.

In case the raw data is obtained by optical spectroscopy (UV, visible, IR, Fluorescence), preprocessing of the data is typically carried out by background correction and/or normalization. The skilled person is aware of a number of suitable known background correction and normalization strategies; a comparative survey in case of Affymetrix data is given in L.M. Cope et al., Bioinformatics 2004, 20(3), 323-331 or R.A. Irizarry et al.,Bioinformatics 2006, 22(7), 789-794, respectively.

In case the raw data is obtained by mass spectrometry or by 2D gel electrophoresis, preprocessing of the data is typically carried out by smoothing, baseline correction, peak picking, optionally, additional further data transformation such as taking the logarithm in order to carry out a stabilization of the variances.

In case the raw data is obtained by transcriptomics, preprocessing of the data is typically carried out by summarizing single pixel to a single intensity signal; background correction; summarizing of multiple probe signals to a single expression value, in particular perfect match/mismatch probes; normalization.

Depending on the data at hand, it may also consist of some variance stabilizing transformation or transformation to normality as for instance taking the logarithm or using Box-Cox power transformations [Box, G. E. P. and Cox, D. R. An analysis of transformations (with discussion). Journal of the Royal Statistical Society B 1964, 26, 21 1 -252].

Often also scaling e.g. by standard deviation or median absolute deviation (MAD) might be used to transform the raw data. However, this step is not necessary for all kind of data, respectively all kind of further statistical analyses and hence may also be omitted.

5. In the step 5, at least one suitable classifying algorithm is selected and the selected classifier algorithm is applied to the preprocessed data.

Due to the high dimensionality of the data, most classification algorithms cannot be directly applied. One reason is the so-called curse of dimensionality: With increasing dimensionality the distances among the instances assimilate. Noisy and irrelevant features further contribute to this effect, making it difficult for the classification algorithm to establish decision boundaries. Further reasons why classification algorithms are not applicable on the full dimensional space are performance limitations. Ultimately, feature transformation techniques are applied before classification. Furthermore, also for the task of identifying unknown marker candidates, the use of traditional methods is limited due to the high dimensionality of the data.

To identify diseased subjects with the highest possible sensitivity and specificity is the main goal in diagnostic development. For this purpose, a large number of classification algorithms can be chosen e.g. logistic regression, (diagonal) linear or quadratic discriminant analysis (LDA, QDA, DLDA, DQDA), shrunken centroids regularized discriminant analysis (RDA), random forests (RF), neural networks (NN), support vector machines (SVM), generalized partial least squares (GPLS), partitioning around medoids (PAM), self organizing maps (SOM), recursive partitioning and regression trees, K-nearest neighbor classifiers (K-NN), bagging, boosting, na^'ive Bayes and many more can be applied to develop new marker candidates. These algorithms are trained on a training data set which contains instances labeled according to classes, e.g. healthy and diseased, and then tested on a test data set which includes novel instances not used for the training. Finally, the classifier will be used to predict the class label of novel unlabeled instance [T. M. Mitchell. Machine Learning. Mc Graw-Hill, 1997].

Then, the measured values of nucleic acids and metabolites detected in the sample is compared to either a standard amount of the respective biomolecule measured in a normal cell or tissue or a reference amount of the respective biomolecule stored in a database. If the amount of the biomolecules of interest in the sample is different to the amount of the biomolecules determined in the standard or control sample, the differential concentration data are processed and used for classifier generation.

6. In the step 6, the classifier algorithms of step 5 is trained on at least one training data set containing preprocessed data from subjects being divided into classes according to their pathophysiological condition with respect to sepsis and/or their prognostic or responder conditions, in order to select a classifier function to map the preprocessed data to the conditions. Classifiers are typically deterministic functions that map a multi-dimensional vector of biological measurements to a binary (or n-ary) outcome variable that encodes the absence or existence of a clinically-relevant class, phenotype or distinct state of disease. The process of building or learning a classifier involves two steps: (1 ) selection of a family functions that can approximate the systems response, and (2) using a finite sample of observations (training data) to select a function from the family of functions that best approximates the system's response by minimizing the discrepancy or expected loss between the system's response and the function predictions at any given point.

Depending on the chosen feature selection strategy, the combination of the different data can take place before or after feature selection. The combined data is then used as input to train and validate the classifier. However, it is also possible to train several different classifiers for the different data separately and then combine the classifiers to the predictive signature. As the data types may be very different from qualitative/categorical to quantitative/numerical, not all classifiers may work for such multilevel data; e.g., some classifiers accept only quantitative data. Hence, depending on the data types one has to choose a class of functions for classification which has an appropriate domain.

7. In the step 7, the trained classifier algorithms of step 6 is applied to a preprocessed data set of a subject with unknown pathophysiological condition with respect to sepsis to predict the class label of said data set in order to calculate prognostic or responder likelihoods.

The trained classifier utilizes data from at least two groups of biomolecules of the aforementioned types and afford a value or a score. This score is assigned to an altered physiological state of plasma, tissue or an organ with a computed probability and can indicate a diseased state, a state due to intervention (e.g. therapeutic intervention by treatment, surgery or pharmacotherapy) or an intoxication with some probability. This score can be used as a diagnostic tool to indicate that the subject or the organism is diagnosed as diseased, to indicate intoxication as having sepsis. The pathophysiological condition corresponds to the label "diseased" and said physiological condition corresponds to the label "healthy" or said pathophysiological condition corresponds to different labels of "grades of a disease", "subtypes of a disease", different values of a "score for a defined disease"; said prognostic condition corresponds to a label "good", "medium", "poor", or "therapeutically responding" or "therapeutically non-responding" or "therapeutically poor responding".

The score and time-dependent changes of the score can be used to assess the success of a treatment or the success of a drug administered to the subject or the organism or assess the individual response of a subject or an organism to the treatment or to make a prognosis of the future course of the physiological state or the disease and the outcome. The prognoses are relative to a subject without the disease or the intoxication having normal levels or average values of the score or classifier composed of at least two biomolecules of two distinct types

Furthermore, the score can be combined with the standard lab parameters commonly used in clinical chemistry, such as serum and/or plasma levels of low molecular weight biochemical compounds, enzymes, enzymatic activities, cell surface receptors and/or cell counts, in particular red and/or white cell counts, platelet counts to predict the class label of said data set in order to calculate prognostic or responder likelihoods

The method according to the present invention may optionally comprises the feature (variable, measurement) selection step. However, it is recommended if the number of features is larger than the number of samples. Feature selection methods try to find the subset of features with the highest discriminatory power.

Numerous feature selection strategies for classification have been proposed, for a comprehensive survey see e.g. [M. A. Hall and G. Holmes. IEEE Transactions on Knowledge and Data Engineering, 15(6): 1437-1447, 2003.]. Following a common characterization, it is distinguished between filter and wrapper approaches. Filter approaches use an evaluation criterion to judge the discriminating power of the features. Among the filter approaches, it can further be distinguished between rankers and feature subset evaluation methods. Rankers evaluate each feature independently regarding its usefulness for classification. As a result, a ranked list is returned to the user. Rankers are very efficient, but interactions and correlations between the features are neglected. Feature subset evaluation methods judge the usefulness of subsets of the features. The information of interactions between the features is in principle preserved, but the search space expands to the size of 0(2<d>). For high-dimensional data, only very simple and efficient search strategies, e.g. forward selection algorithms, can be applied because of the performance limitations.

The wrapper attribute selection method uses a classifier to evaluate attribute subsets. Cross- validation is used to estimate the accuracy of the classifier on novel unclassified objects. For each examined attribute subset, the classification accuracy is determined. Adapted to the special characteristics of the classifier, in most cases wrapper approaches identify attribute subsets with higher classification accuracies than filter approaches, cf. Pochet, N., De Smet, F., Suykens, J. A., and De Moor, B.L., Systematic benchmarking of microarray data classification: assessing the role of non-linearity and dimensionality reduction. Bioinformatics, 20(17):3185-95 (2004).

As the attribute subset evaluation methods, wrapper approaches can be used with an arbitrary search strategy. Among all feature selection methods, wrappers are the most computational expensive ones, due to the use of a learning algorithm for each examined feature subset.

Compound Short Precursor Product ion DP CE

name ion (m/z) (V) (V)

(m/z)

Taurocholic Acid TCA 514 80 -300 -128

Glycocholic Acid GCA 464 74 -175 -74

Glycochenodeoxycholic

GCDCA 448 74 -170 -72

Acid

Taurochenodeoxycholic TCDCA 498 80 -300 -130 Acid

Taurolithococholic Acid TLCA 482 80 -265 -1 18

Glycolithocholic Acid GLCA 432 74 -185 -76

Taurolithocholic Acid

TLCAS 562 482

sulfate -145 -34

Glycolithocholic Acid

GLCAS 512 432 -70 -46

sulfate

Taurodeoxycholic Acid TDCA 498 80 -300 -1 30

Glycodeoxycholic Acid GDCA 448 74 -125 -78

Cholic Acid CA 407 407 -135 -1 8

Chenodeoxycholic Acid CDCA 391 391 -160 -21

Ursodeoxycholic Acid UDCA 391 391 -160 -21

Deoxycholic Acid DCA 391 391 -160 -21

Tauroursodeoxycholic 498

TUDCA 80

Acid -285 -1 30

Glycoursodeoxycholic 448

GUDCA 74

Acid -155 -80

Lithocholic Acid LCA 375 375 -155 -27

Table 2: Analyte-specific MS parameters.

Another aspect of the invention is a kit for diagnosing, or prognosticating sepsis. In other words, a kit for carrying out the method according to the present invention is provided. The kit of present invention comprises: a) detection agents for the detection of different species of biomolecules comprising a plurality of nucleic acids and a plurality of endogenous target metabolites; b) positive and/or negative controls; and c) classification software for classification of the results achieved with said detection agents.

Another embodiment of this aspect is a kit for prognosticating a sepsis, wherein the prognosis is an expected response by a subject to a treatment of the sepsis. In another embodiment of this aspect, the kit is for prognosticating a sepsis, wherein the prognosis is an expected survival of a subject with sepsis. The agents and sequences specified in the described examples are diagnostic for sepsis, or prognosticates the expected response or survival of the subject. The binding of the nucleic acid or aptamer or antibody to the nucleic acids and metabolites is diagnostic for sepsis, prognosticates an expected response to a treatment, or prognosticates an expected survival of a subject having sepsis. The present invention further provides a method for evaluating candidate therapeutic agents. The method can be applied to identify molecules that modulate the concentrations of nucleic acids and metabolites.

One aspect of the invention is a method of detecting for early diagnosing sepsis, prognosticating an expected response to a treatment, or prognosticating an expected survival.

In a specific embodiment, data obtained from two different types of biomolecules are used and processed together according to the method thus providing improved classification and diagnosis of complex diseases.

Detailed example: Method utilizing microRNA and metabolite data

Samples: 41 mixed sepsis vs. 27 controls

As an example, we use the microRNA and metabolites (bile acids) of lllumina HumanMI V2 BeadChip

Analysis:

For developing and validating a classifier based on these data we used random forest in combination with .632 bootstrap where each analysis step - including low level analysis - was repeated in each bootstrap step. Of course, we could also have used a split-sample or a k-fold cross-validation approach for validation. Moreover, we could have used a different class of functions for classification e.g. (diagonal) linear or quadratic discriminant analysis (LDA, QDA, DLDA, DQDA), shrunken centroids regularized discriminant analysis (RDA), random forests (RF), neural networks (NN), support vector machines (SVM), generalized partial least squares (GPLS), partitioning around medoids (PAM), self organizing maps (SOM), recursive partitioning and regression trees, K-nearest neighbor classifiers (K-NN), bagging, boosting, na^'ive Bayes and many more. In the case of microRNA data, the low level analysis consisted of the variance stabilizing transformation of Huber et al. (2002) [Huber W, von Heydebreck A, Sueltmann H, Poustka A, Vingron M. Variance Stabilization Applied to Microarray Data Calibration and to the Quantification of Differential Expression. Bioinformatics 2002, 18: 96-104] (often called normalization). There is a large number of alternative methods which could have been used. Several examples are given in L.M. Cope, Bioinformatics 2004, 20(3), 323-331 or R.A. Irizarry,. Bioinformatics 2006, 22(7), 789-794. In the case of bile acids the logarithm was applied as variance stabilizing transformation. In the case of random forest feature selection is canonically part of the classification algorithm. Again there are numerous other feature selection strategies we could have used, some examples are given in [M. A. Hall and G. Holmes. IEEE Transactions on Knowledge and Data Engineering, 15(6): 1437-1447, 2003.]. In the case of microRNA data we obtain the estimated errors given in Table 3.

Table 3: microRNA data, classification error via .632 bootstrap estimate (500 replications)

The estimated overall accuracy using bootstrap is 82.2%. Using only bile acids for classification we obtain the results given in Table 4 which corresponds to an overall accuracy of 84.6 %.

Table 4: bile acids data, classification error via .632 bootstrap estimate (500 replications) Now, if we combine the information of the microRNA probes with the information of the bile acids, we obtain the results given in Table 5. That is, the estimated overall accuracy using .632 bootstrap is 89.4 %. Hence, this combination increases the overall accuracy from 82.2% respectively, 84.6 % to 89.4 %.

Table 5: combination of microRNA and metabolite data, classification error via .632 bootstrap estimate (500 replications)

The microRNA probes selected during bootstrap validation are given in Table 6.

Table 6: microRNA probes selected during .632 bootstrap validation (1 column is SEQ-ID- No)

The results of the Sanger sequence search in accordance with Griffiths-Jones et al. 2008 [Griffiths-Jones S, Saini HK, van Dongen S, Enright AJ. miRBase: tools for microRNA genomics. NAR 2008 36(Database lssue):D154-D158] for known human microRNAs are given in Table 7. Seq-ID microRNA ID Target sequence

9 has-miR-19a GCAGUCCUCUGUUAGUUUUGCAUAGUUGCACUACAAGAAGAAUGUA

GUUGUGCAAAUCUAUGCAAAACUGAUGGUGGCCUGC

10 has-miR-449b UGACCUGAAUCAGGUAGGCAGUGUAUUGUUAGCUGGCUGCUUGGG

UCAAGUCAGCAGCCACAACUACCCUGCCACUUGCUUCUGGAUAAAU UCUUCU

1 1 has-miR-380 AAGAUGGUUGACCAUAGAACAUGCGCUAUCUCUGUGUCGUAUGUAA

UAUGGUCCACAUCUU

12 has-miR-485 ACUUGGAGAGAGGCUGGCCGUGAUGAAUUCGAUUCAUCAAAGCGA

GUCAUACACGGCUCUCCUCUCUUUUAGU

13 has-miR-297 UGUAUGUAUGUGUGCAUGUGCAUGUAUGUGUAUAUACAUAUAUAU

GUAUUAUGUACUCAUAUAUCA

14 has-miR-200a CCGGGCCCCUGUGAGCAUCUUACCGGACAGUGCUGGAUUUCCCAG

CUUGACUCUAACACUGUCUGGUAACGAUGUUCAAAGGUGACCCGC

15 has-miR-205 AAAGAUCCUCAGACAAUCCAUGUGCUUCUCUUGUCCUUCAUUCCAC

CGGAGUCUGUCUCAUACCCAACCAGAUUUCAGUGGAGUGAAGUUC

AGGAGGCAUGGAGCUGACA

16 has-miR-509-1 CAUGCUGUGUGUGGUACCCUACUGCAGACAGUGGCAAUCAUGUAU

AAUUAAAAAUGAUUGGUACGUCUGUGGGUAGAGUACUGCAUGACAC

AUG

Table 7: Results of the Sanger sequence search for known human microRNAs for microRNA probes selected during .632 bootstrap validation (1 column is SEQ-ID-No). The bile acids selected during bootstrap validation are given in Table 8.

Compound Short name

Taurocholic Acid TCA

Glycocholic Acid GCA

Glycochenodeoxycholic Acid GCDCA

Taurochenodeoxycholic Acid TCDCA

Taurolithococholic Acid TLCA

Chenodeoxycholic Acid CDCA

Deoxycholic Acid DCA

Table 8: bile acids selected during bootstrap validation

The example described above demonstrates that the method functions with a combination of two different types of biomolecules (nucleic acids and metabolites) having sepsis with a performance which is superior than that of a test or diagnostic or prognostic tool comprising a set of preselected biomolecules composed of just one type such as nucleic acids or metabolite solely. The above descriptions are illustrative and not restrictive. It is to be understood that this invention is not limited to particular methods, and experimental conditions described, as such methods and conditions may vary.

The sequence listing accompanying the present application comprising sequences with SEQ- IDs No 1 to 16 is part of the disclosure of the present invention.

Claims

Claims

1 . An in vitro method for predicting a likelihood of an onset of a sepsis in at least one biological sample of at least one tissue of a mammalian subject comprising the steps of: a) detecting at least two distinct types of biomolecules comprising a plurality of nucleic acids and a plurality of endogenous target metabolites, which metabolites have a molecular mass less than 1500 Da; b) measuring at least one parameter selected from the group consisting of nucleic acid expression level; presence or absence, level, amount, concentration of each individual biomolecule of each type in said sample and storing the obtained measured values as raw data in a database; c) mathematically preprocessing said raw data in order to reduce technical errors being inherent to the measuring procedures used in step b); d) selecting at least one suitable classifying algorithm from the group consisting of logistic regression, (diagonal) linear or quadratic discriminant analysis (LDA, QDA, DLDA, DQDA), perceptron, shrunken centroids regularized discriminant analysis (RDA), random forests (RF), neural networks (NN), Bayesian networks, hidden Markov models, support vector machines (SVM), generalized partial least squares (GPLS), partitioning around medoids (PAM), self organizing maps (SOM), recursive partitioning and regression trees, K-nearest neighbor classifiers (K-NN), fuzzy classifiers, bagging, boosting, and na^'ive Bayes; and applying said selected classifier algorithm to said preprocessed data of step c); e) said classifier algorithms of step d) being trained on at least one training data set containing preprocessed data from subjects being divided into classes according to their pathophysiological condition with respect to sepsis and/or their prognostic or responder conditions, in order to select a classifier function to map said preprocessed data to said conditions; f) applying said trained classifier algorithms of step e) to a preprocessed data set of a subject with unknown pathophysiological condition with respect to sepsis and/or their prognostic or responder conditions, and using the trained classifier algorithms to obtain a score for predicting the class label of said data set in order to calculate prognostic or responder likelihoods.

2. A method according to claim 1 , characterized in that said endogenous target metabolites are selected from the group consisting of:

Bile acids, oxysterols; amino acids, phenylthio carbamoyl amino acids (PTC-amino acids), dimethylarginine; carboxylic acids; Ceramides with an N-acyl residue having from 2 to 30 Carbon atoms in the acyl residue and having from 0 to 5 double bonds and having from 0 to 5 hydroxy groups; carnitine; acylcarnitines having from 1 to 20 carbon atoms in the acyl residue; acylcarnitines having from 3 to 20 carbon atoms in the acyl residue and having 1 to 4 double bonds in the acyl residue; acylcarnitines having from 1 to 20 carbon atoms in the acyl residue and having from 1 to 3 OH-groups in the acyl residue; acylcarnitines having from 3 to 20 carbon atoms in the acyl residue with 1 to 4 double bonds and 1 to 3 OH-groups in the acyl residue; phospholipides; phosphatidylcholines (diacylphosphatidylcholines) having a total of from 1 to 50 carbon atoms in the acyl residues; phosphatidylcholines having a total from 3 to 50 carbon atoms in the acyl residues and having a total of 1 to 8 double bonds in the acyl residues; sphingolipids, in particular sphingomyelines having a total number of carbon atoms in the acyl chains from 10 to 30; sphingomyelines having a total number of carbon atoms in the acyl chains from 10 to 30 and 1 to 5 double bonds; hydroxysphinogomyelines having a total number of carbon atoms in the acyl residues from 10 to 30; hydroxysphingoyelines having a total number of carbon atoms in the acyl residues from 10 to 30 and 1 to 5 double bonds; prostaglandines, namely 6-keto- prostaglandin F1 alpha, prostaglandin D2, thromboxane B2; putrescine; biogenic amines, namely histamine, serotonine, palmitoyl ethanolamine.

3. A method according to claim 1 or 2, characterized in that one type of distinct biomolecules are nucleic acids, preferably microRNAs and/or its DNAs and the other type of distinct biomolecules are bile acids as endogenous target metabolites.

4. A method according to anyone of claims 1 through 3, characterized in that said tissue is selected from the group consisting of blood, plasma, serum and other body fluids, cerebrospinal fluids, bone tissue, bone marrow tissue, muscular tissue, glandular tissue, brain tissue, nerve tissue, mucous tissue, connective tissue, and skin tissue and/or said sample is a biopsy sample and/or said mammalian subject includes humans; and/or further characterized in that the score is combined with a standard lab parameters commonly used in clinical chemistry, such as serum and/or plasma levels of low molecular weight biochemical compounds, enzymes, enzymatic activities, cell surface receptors and/or cell counts, in particular red and/or white cell counts, platelet counts, for predicting the class label of said data set in order to calculate prognostic or responder likelihoods.

5. A method according to anyone of claims 1 to 4, characterized in that said step of mathematically preprocessing of said raw data obtained in step b) is carried out by a statistical method selected from the group consisting of: in case of raw data obtained by optical spectroscopy (UV, visible, IR, Fluorescence): background correction and/or normalization; in case of raw data obtained from metabolomics obtained by mass spectrometry or by 2D gel electrophoresis: smoothing, baseline correction, peak picking, optionally, additional further data transformation such as taking the logarithm in order to carry out a stabilization of the variances; in case of raw data obtained from transcriptomics: Summarizing single pixel to a single intensity signal; background correction; summarizing of multiple probe signals to a single expression value, in particular perfect match/mismatch probes; normalization; 6. Method according to anyone of claims 1 to 5, characterized in that after preprocessing step c) a further step of feature selection is inserted, in order to find a lower dimensional subset of features with the highest discriminatory power between classes; and said feature selection is carried out by a filter and/or a wrapper approach;

wherein said filter approach includes rankers and/or feature subset evaluation methods.

7. Method according to anyone of claims 1 to 6, characterized in that said pathophysiological condition corresponds to the label "diseased" and said physiological condition corresponds to the label "healthy" or said pathophysiological condition corresponds to different labels of "grades of a disease", "subtypes of a disease", different values of a "score for a defined disease"; said prognostic condition corresponds to a label "good", "medium", "poor", or "therapeutically responding" or "therapeutically non-responding" or "therapeutically poor responding".

8. Method according to anyone of claims 1 to 7, characterized in that said metabolic data is high-throughput mass spectrometry data.

9. Method according to anyone of claims 1 to 8, characterized in that said mammalian subject is a human being, said biological sample is blood and/or blood cells and/or bone marrow;

wherein said target metabolites are bile acids which are selected from the group consisting of: Compound Short

name

Taurocholic Acid TCA

Glycocholic Acid GCA

Glycochenodeoxycholic

GCDCA

Acid

Taurochenodeoxycholic

TCDCA

Acid

Taurolithococholic Acid TLCA

Glycolithocholic Acid GLCA

Taurolithocholic Acid

TLCAS

sulfate

Glycolithocholic Acid

GLCAS

sulfate

Taurodeoxycholic Acid TDCA

Glycodeoxycholic Acid GDCA

Cholic Acid CA

Chenodeoxycholic Acid CDCA

Ursodeoxycholic Acid UDCA

Deoxycholic Acid DCA

Tauroursodeoxycholic

TUDCA

Acid

Glycoursodeoxycholic

GUDCA

Acid

Lithocholic Acid LCA

and said nucleic acid is microRNA, which is selected from the group consisting of SEQ-ID No. 9 to SEQ-ID No. 16;

wherein microRNA expression levels and serum bile acid concentration are used as said parameters of step b);

wherein raw data of microRNA expression are preprocessed using the generalized logarithm as variance-stabilizing normalization and summarizing the normalized multiple probe signals (technical replicates) to a single expression value, using the median;

wherein raw data of bile acids are preprocessed using the logarithm as variance- stabilizing normalization; wherein random forests are selected as suitable feature selection and classifying algorithm, the training of the classifying algorithm including preprocessed and filtered microRNA expression data, is carried out with a .632 bootstrap-validation;

applying said trained random forests classifier to said preprocessed microRNA expression data sets to a subject under suspicion of having sepsis or sepsis related disease, and using the trained classifiers to diagnose sepsis.

10. Method according to claim 9, characterized in that the following DNA probes for targeting said microRNA are used: Seq-ID No. 1 to Seq-ID No. 8;

and/or

the following microRNA-target sequences are used: Seq-ID Nos. 9 to 16.

1 1 . Method according to anyone of claims 1 to 10, characterized in that the metabolites in the samples are analyzed by liquid chromatography and mass spectrometry, wherein the quantification of the measured metabolite concentrations in said biological tissue sample is calibrated by reference to internal standards.

12. Method according to anyone of claims 1 to 1 1 , characterized in that the microRNA expression data are obtained by quantitative real time PCR (q-RT-PCR) or by hybridization assays.

13. Method according to anyone of claims 1 to 12, characterized that said metabolites comprise compounds shown in the following table:

Compound Short name

Taurocholic Acid TCA

Glycocholic Acid GCA

Glycochenodeoxycholic Acid GCDCA

Taurochenodeoxycholic Acid TCDCA

Taurolithococholic Acid TLCA

Chenodeoxycholic Acid CDCA

Deoxycholic Acid DCA

14. Kit for carrying out a method in accordance with anyone of claims 1 to 14, in a biological sample, comprising: a) detection agents for the detection of at least two distinct types of biomolecules comprising a plurality of nucleic acids and a plurality of endogenous target metabolites; b) positive and/or negative controls; and

c) classification software for classification of the results achieved with said detection agents.