Method for rapid identification of genetic material
A method and kit for isolating and enriching microbial cell-free DNA fragments using a tailored buffer system addresses the challenges of low sensitivity in bloodstream infection diagnostics, enabling rapid and accurate pathogen identification and effective treatment.
Patent Information
- Application Number
- PCT/EP2026/063898
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2025-05-12
- Filing Date
- 2026-05-12
- Publication Date
- 2026-11-19
AI Technical Summary
Current methods for identifying pathogens in bloodstream infections, such as blood culturing, have long turnaround times and low sensitivity, leading to delayed and ineffective antimicrobial treatments, which contribute to increased mortality and antimicrobial resistance, while existing DNA sequencing technologies struggle with the low abundance and differentiation of microbial cell-free DNA in clinical samples.
A method and kit for isolating and enriching microbial cell-free DNA fragments using a composition with a chaotropic solute, non-ionic surfactant, and DNA binding matrix, preferentially binding DNA fragments larger than 75 nucleotides, followed by sequencing to identify pathogens and suggest targeted antimicrobial treatments.
Facilitates rapid and accurate identification of pathogens, improving treatment efficacy and reducing mortality by enhancing the signal-to-noise ratio of microbial DNA, allowing for personalized and timely antimicrobial therapy.
Smart Images

Figure EP2026063898_19112026_PF_FP_ABST
Abstract
Description
[0001] METHOD FOR RAPID IDENTIFICATION OF GENETIC MATERIAL
[0002] FIELD OF THE INVENTION
[0003] The present disclosure relates to the identification of an infectious microbial organism(s) and pathogens responsible for an infection based on the microbial cell-free DNA present in the sample. The disclosure also relates to methods of treating patients with antimicrobials known to be effective towards the identified microbial organism(s) in order to improve the clinical progression and outcome.
[0004] BACKGROUND OF THE INVENTION
[0005] Bloodstream infections (BSIs) are a major cause of mortality, often leading to sepsis or septic shock. Rapid initiation of effective antimicrobial therapy is essential for survival; however, the current gold standard for identifying pathogens in bloodstream infections, blood culturing, has limitations with long turnaround time and poor sensitivity. This delay in refining empirical broad-spectrum antimicrobial treatments contributes to increased mortality and the development of antimicrobial resistance.
[0006] Bloodstream infections (BSIs) are frequent and associated with high mortality due to the development of sepsis. The Global Burden of Disease study found that there were 50 million sepsis cases and 11 million sepsis-related deaths in 2017, constituting nearly 20% of all global deaths. To treat sepsis, the current international guideline recommends that broadspectrum empiric antimicrobial treatment should be administered ideally within 1 hour from its recognition. However, in 20% of cases, the empiric antimicrobial treatment does not cover the causative pathogen, leading to higher mortality. To facilitate targeted and effective antimicrobial treatment to improve sepsis survival, an accurate and rapid identification of the causative pathogen. However, gold-standard blood culturing fails to diagnose over 50% of patients with clinical sepsis, possibly due to prior antimicrobial treatment or a microbial load below the limit of detection. Furthermore, blood culturing has a turnaround time to pathogen identification of approximately 2 days, which contrasts the urgency to provide effective antimicrobial treatment. In addition to the urgency on a patient level, there is an increased focus on improving diagnostics to facilitate effective antimicrobial stewardship programs.
[0007] This has prompted the development of novel approaches to BSI diagnostics, including multiplex PCR assays, molecular target detection with magnetic resonance, and metagenomic next-generation sequencing (mNGS) of microbial cell-free DNA (cfDNA) in plasma. Of these, mNGS has gained particular attention due to its high sensitivity and hypothesis-free
[0008] approach. While cfDNA mNGS has an increased number of relevant detections of pathogens compared to the standard blood culturing, the method has generally been applied on DNA sequencing platforms with long turnaround times that compromise the urgency of diagnosing BSIs.
[0009] A major challenge in doing so is to "find the needle-in-the-haystack" as microbial DNA constitutes a small fraction of the total DNA in a clinical sample, and generally the load of microbial DNA compared to human DNA in the cell-free fraction constitutes a low percentage of the total DNA, giving rise to the issue that complete genomes are generally not constructible, making sequence-based analysis of e.g., microbial short cell-free DNA molecules (cfDNA) challenging.
[0010] Most of the human cell-free DNA has very defined lengths at 170 nucleotides, 340 nucleotides, 510 nucleotides etc. corresponding to the length of DNA wrapped around 1, 2, and 3 histone octamers, respectively, which protects the DNA from nuclease degradation. In contrast, microbial cell-free DNA (at least viral and bacterial) is short and fragmented, with mean lengths at 50-80 nucleotides.
[0011] Usually, DNA isolation is ignorant of the differences in the chemical nature of the mammalian cfDNA compared to microbial cfDNA, and therefore cfDNA extractions performed with the aim of analysing microbial cfDNA is often hampered by the massive excess of mammalian cfDNA in the sample.
[0012] It is therefore of great interest to explore methods that facilitate a more robust isolation of the smaller DNA fragments, predominantly consisting of microbial cfDNA, such that the downstream analysis of the microbial cfDNA may be simplified and more cost-effective.
[0013] Following isolation and sequencing independent classifications of DNA sequences by comparing them to a database of complete genomes has proved difficult because some DNA sequences are redundant between microbial genomes, which in some cases results in falsepositive identifications of microbial species. Accordingly, there is a need for methods that enable the fast and reliable identification of the source(s) of infection, that can enable the proper treatment of the infections within a short timeframe.
[0014] SUMMARY OF THE INVENTION
[0015] In a first aspect the present disclosure relates to a composition that facilitates the preferential binding of long DNA fragments with an approximate size of 75 nucleotides ormore, such as 75 nucleotides or more, 80 nucleotides or more, 90 nucleotides or more, or such as 100 nucleotides, or more to a matrix.
[0016] In a further aspect the present disclosure relates to a method of validating the DNA binding properties of a DNA binding buffer, comprising
[0017] a. adding a pool of DNA fragments comprising DNA fragments of different sizes such as e.g., ranging from 25-500 bp such as e.g., a DNA ladder, in a concentration of about 0.075ng / pL to the DNA binding buffer in the presence of a DNA binding matrix to facilitate binding of DNA fragments,
[0018] b. isolating the bound DNA from the unbound DNA to obtain an isolated DNA fraction and an unbound DNA fraction, and
[0019] c. analyse the size distribution and concentration of the isolated DNA fraction and the unbound DNA fraction.
[0020] In a further aspect the present disclosure relates to a kit for isolating microbial cell-free DNA fragments, comprising
[0021] a. a vial comprising a composition comprising
[0022] i. a chaotropic solute,
[0023] ii. a non-ionic surfactant,
[0024] iii. optionally, a chelating agent and
[0025] iv. optionally, a buffering agent, and
[0026] b. optionally, one or more matrices capable of binding DNA, optionally contained within a vial, and
[0027] c. optionally, means for performing the microbial cell-free DNA isolation, d. optionally, a DNA ladder, and
[0028] e. instructions for preparing the composition according to the first aspect, f. Optionally, instructions for isolating microbial cfDNA fragments using said kit.
[0029] In a further aspect the present disclosure relates to a method for enriching microbial cfDNA from a sample comprising host and microbial cfDNA using a composition according to the kit of the disclosure.
[0030] In a further aspect the present disclosure relates to a method for treating a patient suffering from an infection, the method comprising a) isolation of microbial cfDNA using a composition or a kit according the disclosure from a patient sample, and b) identification of at least one pathogen based on the isolated microbial cfDNA, and c) suggesting a clinically relevant effective antimicrobial treatment regimen targeting said pathogen.
[0031] In a further aspect the present disclosure relates to a method for determining the identity of at least one pathogen responsible for an infection in a patient, comprising the following steps:
[0032] a) obtaining a sample from the patient,
[0033] b) optionally, extracting the total DNA from said sample,
[0034] c) enriching the microbial cfDNA molecule fraction in the extracted DNA using a composition or a kit according to the disclosure,
[0035] d) optionally, amplifying the enriched microbial cfDNA molecules,
[0036] e) sequencing the enriched microbial cfDNA molecules,
[0037] f) establishing a dataset containing features of the microbial cfDNA molecules, and
[0038] g) determining at least one source of the microbial cfDNA from the sample as the identity of the pathogen(s) responsible for the infection.
[0039] In a further aspect the present disclosure relates to a method for treating a patient suffering from an infection, the method comprising a) identifying at least one pathogen according to the method for determining the identity of at least one pathogen responsible for an infection in a patient disclosed herein, and b) suggesting a clinically relevant effective antimicrobial treatment regimen targeting said pathogen.
[0040] In a further aspect the present disclosure relates to a method for excluding microbial infection(s) as the cause of a pathological states of a subject suspected of suffering from a microbial infection, the method comprising subjecting a sample from the subject to an analysis method according to the method for determining the identity of at least one pathogen responsible for an infection in a patient disclosed herein, and concluding based on an absence of any identified microbe(s) that the patient is infection-free.
[0041] BRIEF DESCRIPTION OF THE FIGURES
[0042] Figure 1. Flowchart of patient inclusion and selection for analysis.
[0043] Figure 2. (a) Flowchart of the sample handling workflow for blood cultures and mNGS samples (created with biorender.com). (b) A timeline for BSI diagnostics with Nanopore mNGS, Illumina mNGS, and blood culturing.
[0044] Figure 3. (a) Pie chart summarizing findings from blood culture and mNGS. (b) Total pathogen DNA burden for patients (measured in genome equivalents per microliter) positive by mNGS in BSI-confirmed and BSI-suspected and compared only for E. coli where identified, (c) Number of pathogens identified with mNGS per sample, (d) Plasma cfDNA concentrations across the different patient subgroups and blood donor samples.
[0045] Figure 4. Clinical characteristics of patients who were analyzed and excluded, a) CRP at blood culture for the patient subgroups BSI-confirmed, BSI-suspected, BSI-absent, and BSI-unlikely groups, b) qSOFA score at blood culture for the four groups, c) Systolic blood pressure at blood culture for the four groups, d) Temperature at blood culture for the four groups.
[0046] Figure 5. Conceptual illustration of the size dependent DNA isolation protocol performed in two steps 1st step: bind the "long" human DNA for removal and keep the supernatant with shorter fragments; 2nd step: bind all DNA in the supernatant from the 1st step to remove other impurities facilitating NGS library preparation.
[0047] Figure 6. Tapestation electropherogram for various cleanNGS solution ratios to sample (3-6x excess). Grey square denotes the desired 50 nucleotides DNA fragment peak
[0048] Figure 7. Tapestation electropherogram for various cleanNGS solution ratios to sample (3-6x excess) with various volumes of isopropanol. Grey square denotes the desired 50 nucleotides DNA fragment peak.
[0049] Figure 8. TapeStation electropherogram for various bead sizes A) 400 nm, B) 400+700nm and C) 700 nm, with long-fragment binding (LFB) buffer (3M GTC, about 5%v / v Tween20,about 12mM Tris, about 5mM EDTA 10 + Proteinase K +20 %v / v isopropanol ) eluted material shown in grey lines and short fragment binding (SFB) buffer (3M GTC, about 4%v / v Tween20, about 10mM Tris, about 3mM EDTA + 30%v / v isopropanol) eluted material shown in black lines with grey shading. DNA input was HyperLadder 25bp as 3 ng / pL (750ng). Grey dotted square denotes the 50 nucleotides DNA fragment peaks.
[0050] Figure 9. Tapestation electropherogram for various isopropanol concentrations A) 0 v / v% isopropanol B) 10%v / v and C) 20%v / v, in long-fragment binding (LFB) buffer (3M GTC, about 5%v / v Tween20, about 12mM Tris, about 5mM EDTA). 400+700 nm silica coated superparamagnetic beads. Grey dotted square denotes the 50 nucleotides DNA fragment peaks.
[0051] Figure 10. Tapestation electropherogram for various isopropanol concentrations in LFB A) 10%v / v isopropanol in LFB, B) 15%v / v and C) 20%v / v, with long-fragment binding (LFB) buffer (3M GTC, about 5%v / v Tween20, about 12mM Tris, about 5mM EDTA 10 +X %v / v isopropanol) and short fragment binding (SFB) buffer (3M GTC, about 4%v / v Tween20, about 10mM Tris, about 3mM EDTA + 30%v / v isopropanol) as replicates 1 (grey fully drawn lines) and 2 (black lines with shading). DNA input was HyperLadder 25bp as 3 ng / pL (750ng). 400+700 nm silica coated superparamagnetic beads. Grey dotted square denotes the 50 nucleotides DNA fragment peaks.
[0052] Figure 11. Tapestation electropherogram for various isopropanol concentrations A) 20v / v% and B) 22.5%v / v in long-fragment binding (LFB) buffer (3M GTC, about 5%v / v Tween20, about 12mM Tris, about 5mM EDTA) and short fragment binding (SFB) buffer (3M GTC, about 4%v / v Tween20, about 10mM Tris, about 3mM EDTA + 30%v / v isopropanol) as replicates 1 and 2. DNA input was HyperLadder 25bp as 3 ng / pL (750ng). 700 nm silica coated superparamagnetic beads. Grey dotted square denotes the desired 50 nucleotides DNA fragment peaks with grey fully drawn lines eluate from LFB and black lines with shading for SFB eluate.
[0053] Figure 12. Tapestation electropherogram for various isopropanol concentrations A) 25 v / v% and B) 27.5%v / v in long-fragment binding (LFB) buffer (3M GTC, about 5%v / v Tween20, about 12mM Tris, about 5mM EDTA) and short fragment binding (SFB) buffer (3M GTC, about 4%v / v Tween20, about lOmM Tris, about 3mM EDTA + 30%v / v isopropanol) as replicates 1 and 2. DNA input was HyperLadder 25bp as 3 ng / pL (750ng). 700 nm silica coated superparamagnetic beads. Grey dotted square denotes the desired 50 nucleotides DNAfragment peaks with grey fully drawn lines eluate from LFB and black lines with shading for SFB eluate.
[0054] Figure 13. TapeStation electropherogram for various DNA input A) 3 ng / pL (750ng), B) 0,2 ng / pL (50ng) and C) 0,0156 ng / pL (3.9ng) in human plasma using long-fragment binding (LFB) buffer (3M GTC, about 5%v / v Tween20, about 12mM Tris, about 5mM EDTA and 22.5 %v / v isopropanol) and short fragment binding (SFB) buffer (3M GTC, about 4%v / v Tween20, about 10mM Tris, about 3mM EDTA + 30%v / v isopropanol) as replicates 1 and 2. 700 nm silica coated superparamagnetic beads. Grey dotted square denotes the desired 50 nucleotides DNA fragment peaks with grey fully drawn lines eluate from LFB and black lines with shading for SFB eluate.
[0055] Figure 14. TapeStation electropherogram for various DNA Inputs A) 3 ng / pL (750ng), B) 0,2 ng / pL (50ng) and C) 0,0156 ng / pL (3.9ng) in human plasma using long-fragment binding (LFB) buffer (3M GTC, about 5%v / v Tween20, about 12mM Tris, about 5mM EDTA, and 0 %v / v isopropanol) and short fragment binding (SFB) buffer (3M GTC, about 4%v / v Tween20, about 10mM Tris, about 3mM EDTA + 30%v / v isopropanol). 700 nm silica coated superparamagnetic beads. Grey dotted square denotes the desired 50 nucleotides DNA fragment peaks with grey fully drawn lines eluate from LFB and black lines with shading for SFB eluate.
[0056] Figure 15. TapeStation electropherogram for various DNA Inputs A) 0.2 ng / pL (50ng), B) 0,075 ng / pL (18,75ng) and C) 0,025 ng / pL (6,25ng) in human plasma using long-fragment binding (LFB) buffer (3M GTC, about 5%v / v Tween20, about 12mM Tris, about 5mM EDTA, and 10 %v / v isopropanol) and short fragment binding (SFB) buffer (3M GTC, about 4%v / v Tween20, about lOmM Tris, about 3mM EDTA, 30%v / v isopropanol). 700 nm silica coated superparamagnetic beads. Grey dotted square denotes the desired 50 nucleotides DNA fragment peaks with grey fully drawn lines eluate from LFB and black lines with shading for SFB eluate.
[0057] Figure 16. TapeStation electropherogram for various DNA Inputs A) 0.7 ng / pL (14ng), B) 0,25 ng / pL (5ng) and C) 0,075 ng / pL (1.5ng) in human plasma using a long-fragment binding (LFB) buffer comprising 3M GTC, about 5%v / v Tween20, about 12mM Tris, about 5mM EDTA, 12.5 %v / v isopropanol. 700 nm silica coated superparamagnetic beads. Greydotted square denotes the desired 50 nucleotides DNA fragment peaks with grey fully drawn lines eluate from LFB and black lines with shading for SFB eluate.
[0058] Figure 17. Tapestation electropherogram for DNA in human plasma (total initial DNA concentration of 75 ng / m) using long-fragment binding (LFB) buffer (3M GTC, about 5%v / v Tween20, about 12mM Tris, about 5mM EDTA, 10, 12.5 or 15%v / v isopropanol). Grey dotted square denotes the desired 50 nucleotides DNA fragment peaks. Black lines indicate enriched fraction and grey line indicates the isolated long fragments.
[0059] Figure 18. Tapestation electropherogram for DNA in human plasma (total initial DNA concentration of 75 ng / m) using a long-fragment binding (LFB) buffer comprising 3M GTC, about 5%v / v Tween20, about 12mM Tris, about 5mM EDTA, and 15%v / v, 17.5 %v / v, or 20%v / v isopropanol. Grey dotted square denotes the desired 50 nucleotides DNA fragment peaks. Black lines indicate enriched fraction and grey line indicates the isolated long fragments.
[0060] Figure 19. Tapestation electropherogram for DNA in human plasma (total initial DNA concentration of 75 ng / m) using a long-fragment binding (LFB) buffer comprising 0.5, 1.0, 2.0 or 3.0M GTC, about 5%v / v Tween20, about 12mM Tris, about 5mM EDTA and 15%v / v isopropanol. Grey dotted square denotes the desired 50 nucleotides DNA fragment peaks. Black lines indicate enriched fraction and grey line indicates the isolated long fragments.
[0061] Figure 20. Tapestation electropherogram for DNA in human plasma (total initial DNA concentration of 75 ng / m) using a long-fragment binding (LFB) buffer comprising about 5%v / v Tween20, about 12mM Tris, about 5mM EDTA, 15%v / v isopropanol and 3.0, 4.0 or 4.2M GHCI, or 3.0M GTC. Grey dotted square denotes the desired 50 nucleotides DNA fragment peaks. Black lines indicate enriched fraction and grey line indicates the isolated long fragments.
[0062] Figure 21. Tapestation electropherogram for DNA in human plasma using a long-fragment binding (LFB) buffer comprising about 12mM Tris, about 5mM EDTA, 15%v / v isopropanol, 3.0 M GTC, and 0.4 %v / v, 4 %v / v or 8%v / v Tween20. Grey dotted square denotes the desired50 nucleotides DNA fragment peaks. Black lines indicate enriched fraction, and grey line indicates the isolated long fragments.
[0063] Figure 22. CfDNA fragment size distributions in clinical samples, top panel is control samples, middle panel is donor samples and lower panel is patient samples. Left panels - Read length kernel density estimation distribution plots. Right - cumulative distributions of kernel density estimation at 1 base pair increment.
[0064] Figure 23. TapeStation electropherogram for the ladder DNA (HyperLadder 25bp) in human plasma before DNA enrichment and isolation.
[0065] DETAILED DISCLOSURE
[0066] The present disclosure provides means for determining the identity of microbes(s) in a sample based on the cfDNA from said microbes isolated from said sample. The means include buffers systems optimized for use in the separation of microbial and non-microbial DNA based on size, where the buffer systems are tailored to facilitate the binding of DNA fragments based on their size, with a resolution of down to about 25 bp, meaning that the buffers used in the initial isolation can be used to separate the vast majority of mammalian cfDNA from the microbial cfDNA before sequencing and further analysis. The means also includes kits, quality control methods and algorithms which taken together improves the overall analysis flow of samples comprising mixed cfDNA fragments, thereby improving the initial separation, the downstream computation analysis, the speed of the analysis and the analysis depth, and further provide an exhaustive analysis of the microbial cfDNA and thereby the microbial presence in the sample.
[0067] A method for determining the identity of at least one pathogen responsible for an infection in a patient
[0068] The present disclosure relates to a method for determining the identity of at least one pathogen responsible for an infection in a patient. The method comprises a number of steps, which when combined may be used in a clinical setting to provide a general practitioner, physician, clinician or laboratory technician with valuable insights into the status of the patients and provide guidance as to which treatment the individual patients should be subjected to, thereby providing a personalized treatment.
[0069] In some infections the clinical response to suspected severe infections, leading to pathological states such as sepsis or meningitis, are often subject to treatment with very high levels of broad-spectrum antimicrobials even before the source(s) of the infection, i.e., the pathogen(s), is identified. A source of an infection is commonly referred to herein as a pathogen.
[0070] One of the reasons for subjecting patients to such harsh treatments is that it often takes days from obtaining a sample from a patient to positive identification, due to the need for elaborate analyses of the samples, often including attempted culturing of the sample to establish the source or sources of the infection, i.e., the pathogen or pathogens responsible for the infection.
[0071] As such the sample may originate from any source, preferably the sample is obtained from a host e.g., of mammalian origin, but as such there is no restriction on the origin of the sample. Preferably, the sample is a human sample, and more preferably the sample is selected from a tissue sample, a blood sample, a bronchial lavage fluid sample, an exhaled breath sample, serum sample, a plasma sample, a faecal sample, a cerebrospinal fluid sample, a BAL sample and a urine sample. In embodiments, the sample is a cell free sample or alternatively, the sample is a cell containing sample, wherein the host cells are sorted from the sample before initiation of the method described herein. In embodiments, the patient is a human patient. The sample may be obtained from a swab, such as a buccal swab, skin swab, cheek swab, nasopharyngeal swab, throat swab, or vaginal swab. Accordingly, the methods presented herein finds application in several fields related to analysis of genetic material, where different sources of the DNA are to be expected and where the target DNA is expected to be severely degraded or fragmented.
[0072] The present disclosure relates primarily to methods for identifying the pathogenic organism responsible for an infection in a host, but the methods disclosed herein may also find other uses within the metagenomic area. Examples of alternative uses is in identification of organisms responsible for an infestation. As such the methods presented herein may also be used to identify sources of contamination in samples such as e.g., food or pharmaceutical process contaminants. The methods may also be used to identify source(s) of DNA in e.g., environmental samples such as e.g., dirt samples or water samples.
[0073] The present disclosure provides a culturing free method for establishing the source or sources of an infection, in a hypothesis free manner, which can identify the source(s) of an infection based on the presence of genetic material from the source(s), in particular cell free DNA circulating in the plasma.
[0074] As such the method may comprise removing host DNA from the sample, followed by an enrichment of the microbial DNA fraction which is subsequently sequenced and analysed in order to determine the source of the microbial DNA to provide an identification of the pathogen responsible for the infection.
[0075] One of the common issues with analyses targeting microbial cfDNA is that the presence of the microbial cfDNA is often at least 1000-fold lower compared to the host cellular and non-cellular DNA i.e., cfDNA. Therefore, methods for addressing specifically the microbial cfDNA is often addressed with a needle-in-a-haystack analogy, since the presence of the microbial cfDNA is so low compared to the host DNA. Therefore, there is a need for improving the ratio of microbial cfDNA to host DNA, which improves the signal-to-noise ratio, with respect to the microbial cfDNA allowing faster diagnostic of the infection pathogen at lower cost.
[0076] The present disclosure therefore addresses methods of enriching and isolating the microbial DNA in a sample using a novel buffer system which allows for the enrichment of microbial DNA before the sequencing reaction, thereby reducing the amount of host DNA subjected to the sequencing reaction, thereby improving the overall signal-to-noise ratio with respect to the microbial cfDNA.
[0077] Once obtained, the sequences of the microbial cfDNA needs to be analysed to establish which fragments are from a credible source, leading to an infection, and which are simply contaminants from sample handling. Additionally, since some genes overlap between different microbial species and genera, the sequences need to be analysed in a way which allows for identification of the DNA sequences which hold the information about the true source of the infection, and which sequences only provides redundant information, that does not enable identification of the microbial source at a species or genus level.
[0078] Accordingly, once sequenced, the obtained data is to be provided as a dataset which enables the further bioinformatic analysis of the obtained sequences.
[0079] Based on such a dataset containing particular features about the individual sequences obtained from the sequencing reaction, the analysis provided herein enables the determination of the source or sources of an infection.
[0080] Accordingly, the methods provided herein for determining the source of an infection may be used for treatment purposes, where the determination of the source(s) of the infection may be used to recommend the specific treatment regimen for the patient, which is clinically relevant to ensure that the proposed treatment an effect against the particular infection. Preferably, the specific treatment regimen has been clinically relevant and / or verified to havean effect towards the specific microbe(s) identified. Also included in the method is also situations where the treatment has not yet been verified, but here clinicians use the method to test the efficiency of a given treatment by monitoring the development of the infection over time.
[0081] Further, the data obtained from the analysis may also be sued to construct a more detailed infection database, which can be used to monitor the spread of infections throughout populations, and to monitor the spread of antibiotic resistance based on the genetic information provided about the infection source.
[0082] Also, the information provided from the clinics may be used to monitor the efficiency of treatment towards specific microbes with a very high level of detail, since any infection may be monitored over time to establish the response of particular microbes towards specific antimicrobials.
[0083] In that way the present methods may also be used to address the development of an infection based on changes in the cfDNA profile between a number of samples.
[0084] Such information about the response of a microbe towards specific treatment regimens may be collected in a database to be used by clinicians in the process of suggesting the optimal treatment regimen for a given infection, based on its specific microbial cfDNA profile.
[0085] Accordingly, the present disclosure relates to a method for determining the identity of at least one pathogen responsible for an infection in a patient, comprising the following steps:
[0086] obtaining a sample from the patient,
[0087] optionally, extracting the total DNA from said sample,
[0088] optionally, enriching the microbial cfDNA molecule fraction in the extracted DNA, optionally, amplifying the enriched microbial cfDNA molecules,
[0089] sequencing the optionally enriched microbial cfDNA molecules,
[0090] establishing a dataset containing features of the microbial cfDNA molecules, and determining at least one source of the microbial cfDNA from the sample as the identity of the pathogen(s) responsible for the infection.
[0091] The disclosure further relates to a method for treating a patient suffering from an infection, which utilizes the determination of the infection source as disclosed herein, followed by a suggestion of a treatment option, preferably with a clinically relevant and / or verified effective antimicrobial targeting said infection source.
[0092] As is shown in example 1 and 8, the methods provided herein are exceptionally suited to providing guidance to clinicians regarding the source of the infection thereby aiding them in suggesting a suitable treatment for the infection.
[0093] Enriching the microbial cfDNA molecule fraction in the extracted DNA
[0094] Once a sample is obtained from a subject, such as e.g., a human, the method may comprise an extraction step, where DNA is extracted from the sample. Depending on the need and the specific DNA to be extracted, different DNA fractions may be extracted from samples in various ways, known to the skilled person. In example cellular DNA extraction involves breaking open cells (lysis) to release DNA, removing proteins and other contaminants, and then isolating the DNA, often using alcohol precipitation. The process typically includes mixing the sample with a lysis buffer, enzymatic treatment, centrifugation, and alcohol addition to visualize and collect the DNA. cfDNA on the other hand may be extracted from e.g., blood plasma firstly by separating blood cells from the remaining sample, minimizing the contamination from genomic DNA, followed by DNA extraction or direct capturing of the cfDNA on a matrix, followed by a washing and elution. Following the extraction, both the cellular DNA and cfDNA, whichever required, may be analysed by sequencing and bioinformatics or PCR based methods to determine the source of the DNA.
[0095] Most of the human cell-free DNA has lengths around 170 nucleotides and 340 nucleotides, corresponding to the length of DNA wrapped around 1, and 2 histone octamers, respectively, which protects the DNA from nuclease degradation. In contrast, microbial cell-free DNA (at least viral and bacterial) is short and fragmented, with mean lengths at 50-80 nucleotides (see Fig 22).
[0096] The inventors found that the purification of microbial cfDNA may be optimized by finetuning the buffer compositions used in the isolation, such that a two-step isolation is possible, which first separates a portion of the mammalian cfDNA from the sample, thereby enriching the remaining DNA in the sample including the microbial cfDNA fraction, which can then be isolated from the sample and purified for optimal downstream processing e.g. NGS preparation etc. Thereby the herein provided compositions and methods provide optimized means for isolation and preparation of microbial cfDNA from complex samples.
[0097] In embodiments, the purification is performed using a method for enriching microbial cfDNA in a sample, wherein the sample comprises a mix of microbial cfDNA and mammalian cfDNA, the method comprising a) mixing a sample with a first aqueous solution to obtain a composition that facilitates the preferential binding of long DNA fragments, preferably with anapproximate size larger than 75 nucleotides or more, such as fragments larger than 75, 80, 85, 90, 95, 100, 125, 150, or more to a matrix, b) contacting the mixed composition with a first matrix capable of binding DNA, thereby preferentially binding the longer cfDNA to the first matrix thereby increasing the ratio in unbound microbial cfDNA to unbound mammalian cfDNA in the enriched sample, and c) clearing the bound mammalian cfDNA on the matrix from the sample, to obtain a sample enriched in microbial cfDNA fragments.
[0098] As used herein "a composition that facilitates the preferential binding of long DNA fragments with an approximate size larger than 75 nucleotides or more, such as 75 nucleotides or more, 80 nucleotides or more, 90 nucleotides or more, or such as 100 nucleotides, or more to a matrix" means a composition which facilitates binding of larger DNA fragments to a matrix, such as those described in the Examples provided herein.
[0099] As is evident from the examples, a way to establish that a composition facilitate preferential binding of long DNA fragments with an approximate size of 75 nucleotides or more, may involve the use of an analysis comprising a) adding a pool of DNA fragments comprising DNA fragments of different sizes in a concentration of about 0.075ng / µl, e.g., ranging from 25-500 bp such as e.g., a DNA ladder to the DNA binding buffer in the presence of a DNA binding matrix to facilitate binding of DNA fragments, b) isolating the bound DNA from the unbound DNA to obtain an isolated DNA fraction and an unbound DNA fraction, and c) analyse the size distribution and concentration of the isolated DNA fraction and the unbound DNA fraction. The composition is considered to bind preferentially to the larger cfDNA fragments when the analysis shows that
[0100] 1. the concentration of the 100bp peak in the unbound fraction has a concentration of less than 25% of the 50bp peak,
[0101] 2. the concentration of the 50bp peak in the isolated DNA fraction is less than 7.5% of the concentration of the 150bp peak,
[0102] 3. the concentration of DNA in the unbound DNA fraction when measured e.g., using Qubit, is higher than 0.03 ng / pL, and
[0103] 4. the total recovery of DNA in the two fractions is higher than 70%.
[0104] It is evident from the above that the pool of DNA fragments of different sizes comprises fragments of at least 50bp, 100 bp and 150 bp, in order for the analysis to be applicable. Often the pool of DNA fragments will be a DNA ladder, which comprises fragments ranging from about 25 bp to about 500 bp in length, in an amount sufficient to enable the analysis, when performed using appropriate means, such as e.g., Tapestation or LabChip as shown in the Examples provided herein.
[0105] The parameters for establishing preferential binding may e.g., be adjusted depending on the samples and the specific needs. Thus, in embodiments, the 100bp peak may in for some application be at the most 50% concentration than that of 50bp peak in the enriched sample, such as at the most 10%, 20%, 25%, 30%, 35%, 40%, 45% or at the most 50% of the concentration of the 50bp peak in the enriched sample. In other embodiments, the 50bp peak at the most 25 % concentration than that of 150bp peak in the depleted sample, such as at the most 5%, 7.5%, 10%, 12.5%, 15%, 17.5%, 20%, 22.5% or such as at the most 25 % concentration than that of 150bp peak in the depleted sample. In embodiments, the total recovery of DNA in both the enriched sample and the isolated sample is at least 40%, such as at least 45%, 50%, 55%, 60%, 65%, 70%, 75%, 80%, 85% or such as least 90%.
[0106] As such the DNA fragments of different distinct sizes may be a DNA ladder. Such a ladder may comprise fragments with a size of 25bp, 50 bp, 75 bp, 100 bp, 125 bp, 150 bp, 175 bp, 200 bp, 250 bp, 300 bp, 400 bp and 500 bp. An example of such a ladder is the HyperLadder 25 bp.
[0107] Generally, the concentration of the different fractions may be evaluated in way known to the skilled person, and the skilled person is well equipped to analyse the size and amount of a given DNA population.
[0108] In the present disclosure the inventors have used an Agilent Tapestation to evaluate the concentration of the different DNA populations with varying sizes. The total concentration of DNA in the sample may be measured in a plethora of ways known to the skilled person. IN the present disclosure the inventors measured the total DNA concentration using Qubit Fluorometric Quantification.
[0109] As used herein "the depleted sample" refers to the sample of the initially isolated DNA following the first binding to the matrix, i.e., preferably containing the longer DNA fragments. As used herein "the enriched sample" refers to the sample where the initially bound DNA has been removed from following the first binding to the matrix, i.e., the enriched sample predominantly comprises the shorter DNA fragments.
[0110] Methods for establishing such parameters are provided in the examples and exemplary buffer composition which show preferential binding to DNA fragments with a size of more than 75 bp are also provided in the examples. As mentioned, the cfDNA may be extracted and / or enriched by binding of the cfDNA to matrix, preferably a solid support produced in a material, often silica based, which enables binding or absorption of cfDNA molecules under certain conditions.
[0111] As such, silica beads i.e., silica particles are commonly used to purify DNA from plasma due to their strong affinity for nucleic acids in the presence of a chaotropic solute such as e.g., guanidine thiocyanate, enabling efficient binding and wash steps. Silica beads are ideal for isolating DNA from crude samples such as e.g., plasma. In contrast, carboxylated magnetic particles / beads, are in some instances more suited for purification of extracted DNA, using solid-phase reversible immobilization (SPRI) mechanisms. Carboxylated particles / beads generally bind DNA via electrostatic interactions in the presence of polymers such as e.g., polyethylene glycol (PEG) and salt such as e.g., NaCI.
[0112] Accordingly, where carboxylated magnetic beads are used, it may be preferable to exchange the chaotropic solute with a polymer such as e.g., PEG.
[0113] By changing the conditions, the binding specificity of the matrix may be altered such that it preferentially binds longer DNA fragments, such as mammalian cellular or cell free DNA, thereby enabling an initial extraction and removal of mammalian DNA from the sample, followed by specific enrichment or purification of shorter DNA fragments, such as microbial cfDNA fragments, which are often shorter (50-80 nucleotides) compared to mammalian DNA fragments. Most of the human cell-free DNA has lengths around 170bp and 340bp corresponding to the length of DNA wrapped around 1 and 2 octamer histones, respectively, which protects the DNA from nuclease degradation.
[0114] From clinical samples the inventors found that the microbial cfDNA from the patients populated a smaller size distribution compared to the mammalian cfDNA (cf. Figure 22), and that an improved analysis could be obtained when using a cutoff at about 75bp, about 80 bp, about 85 bp, about 90, about 95 or about 100 bp.
[0115] Accordingly, the preferred cfDNA fragments are those which are smaller than about 75bp and less, since this cutoff provides an improved signal to noise ratio (cf. Figure 22). This in turn simplifies the downstream analysis by providing less unwanted mammalian sequences to be analysed, which means less computational requirements of the algorithm as less DNA fragments needs to be analysed. Furthermore, enriching the cfDNA for relevant microbial DNA enables better coverage of microbial genomes and / or plasmids allowing analysis of specific strains and genes e.g. genes responsible for antibiotic resistance. As can be seen form the examples, the binding buffers have been optimized to catch this specific selection, such that the predominant fraction enriched in the sample processing steps is the shorter DNA fragments, which are often also the microbial cfDNA fragments. Albeit the size cutoff is generally 75 bp, it is to be understood that the specific cutoff may be varied to the needs of the analysis, hence, the cutoff may be varied from 50 bp to 100 bp or more depending on the needs of the downstream analysis. Hence if it is evaluated that a shorter or longer cutoff ispreferred the methods can be adapted thereto, and the specific cutoff may therefore be varied depending on the required noise level intended. In the present disclosure the cutoff was set to 75bp as this provided a good selection of the mammalian and microbial cfDNA fragments, without losing too big a fraction of the microbial cfDNA, while also minimizing the amount of mammalian cfDNA.
[0116] The inventors found that in order to facilitate such separation, the DNA binding buffers had to be optimized such that they could facilitate a preferential binding of cfDNA fragments which were preferably smaller than 75bp, meaning that the DNA bound to the matrix was predominantly those fragments with a size predominantly larger than 75bp or more.
[0117] As can be seen from figure 6-21, different parameters may be used to obtain sufficient separation of the longer DNA fragments i.e., with a size of more than 75 bp from the shorter DNA fragments, i.e., those with an approximate size of less than 75 bp. The examples provided herein were prepared using a DNA ladder, where the DNA concentration and populations could be determined precisely before and after the analysis. This allowed the inventors to prepare binding buffers which preferentially bound DNA fragments larger than 75 bp, while not binding high amounts of lower bp fragments, in that way enriching the samples for the smaller fragments, by removal of the larger fragments.
[0118] While it is to be understood that the buffers preferentially bind fragments larger than 75bp, the buffer systems are not exact, meaning that the buffers predominantly facilitate the binding of cfDNA fragments that are larger than the intended cutoff. That the buffers preferentially bind larger cfDNA fragments, i.e., fragments larger than 75 bp, herein means that upon analysis of the distribution of DNA fragments in a sample, it will be evident that the population with a size of below 75bp is enriched in the sample following isolation of the bound DNA fragments compared to the untreated sample. Examples 2-6 show how such buffers may be optimized using DNA ladders comprising populations of DNA of known sizes and concentrations, but the examples also provide a method for assessing the buffer quality. What can be seen in Figures 17-21 is that the DNA fragments of the ladder with a size of 50 bp and 75 bp may be separated from the remaining DNA fragments of the ladder using tailored binding buffers optimized to bind the larger DNA fragments. This is especially seen from the figures when looking at the lower bp fraction, which are clearly enriched once optimal conditions are used.
[0119] Accordingly, the enrichment step of the method often comprises removing mammalian DNA molecules from the sample and purifying the microbial cfDNA molecules in the remaining sample.
[0120] The extraction of mammalian DNA from the sample comprises the use of a first DNA binding matrix, often a silica based solid support, and a first composition i.e., a mammalian DNA extraction composition that facilitate binding of mammalian DNA to the first matrix.
[0121] Preferably, the first composition has a pH of 7.5-8.5 and comprise a chaotropic solute, and an organic solvent in a concentration of 5-30%v / v, such as 5-28%v / v, preferably 10-18%v / v, more preferably 12.5-20% / v / v, more preferably 12%-17%v / v. In some embodiments the first composition comprises about 15%v / v isopropanol. In preferred embodiments the first composition comprises between 11.5-17.5%v / v of isopropanol. The binding of the larger fragments to the first matrix was also found to be concentration dependent.
[0122] Therefore, it may in some cases be preferred that the first composition comprises 5-20%v / v of the organic solvent, such as 5-18%v / v, preferably 10-17%v / v, 11.5-17.5%v / v or about 15%v / v. In other cases, it is preferred that the first composition comprises 15-30%v / v of the organic solvent. Most often it is preferable that the first composition comprise about 15%v / v isopropanol. In other instances, it is preferable that the first composition comprise 10-18%v / v isopropanol. It was found that an isopropanol level of about 15%v / v generally resulted in the most preferable isolation profile, but that other components of the composition also impacted the outcome of the separation, and especially the chaotropic solute and the non-ionic surfactant along with the specific beads was found to be critical for the separation cf. figs. 6-21.
[0123] As mentioned earlier, to obtain a sufficient isolation of the 50bp and 75 bp population in the DNA ladder it was needed to optimize the buffer compositions used both in the initial removal of longer DNA fragments, and in the subsequent isolation of the short DNA fragments, which predominantly comprises the microbial cfDNA fragments.
[0124] The subsequent purification of the microbial cfDNA fragments may comprise the use of a second DNA binding matrix, which in some instance is of the same composition as the first matrix, and a second composition i.e., a microbial cfDNA purification composition that facilitate binding of microbial cfDNA to the second matrix. Since the role of the second composition is to bind remaining DNA in the sample, i.e., not only specifically short fragments, the level of the organic solvent may be higher or lower than the indicated intervals, since the role of the composition is to bind remaining DNA in the sample, i.e., both long and short fragments. Generally, the shorter fragments will require a higher concentration of the organic solvent to binding to the matrix.
[0125] Accordingly, in some embodiments, the second binding buffer may be a standard DNA binding buffer, which binds all residual DNA in the sample following the first binding step. Themethod may therefore comprise contacting the sample with a first binding buffer as described herein, followed by removal of the bound DNA fragments from the solution, thereby obtaining a sample enriched in the desired cfDNA fragments with a size of 75 bp or less. As such what is intended is that the microbial cfDNA has a predominant size of 75 bp or less, meaning that the major fraction of the cfDNA in the enriched sample has a size of 75 bp or less, i.e., the sample may contain enriched smaller amounts of DNA fragments which have a size that is larger than 75 bp, as long as the majority of the DNA fragments in the enriched sample has a size which is smaller than 75 bp.
[0126] Herein "base-pair", "bp" and "nucleotides" is used interchangeably.
[0127] If further enrichment is required, it may be preferable to use a specific short fragment binding buffer as described herein.
[0128] In embodiments, the first and / or second matrix is a solid support with a size of 400-700nm, preferably a silica-based particle with a size of about 700nm. Preferably, the solid support is a magnetic solid support.
[0129] In embodiments, the first and / or second composition comprises guanidine thiocyanate as the chaotropic solute in an amount of about 3.0M. It is to be understood that the chaotropic solute can be any chaotropic solute which has the same effect as guanidine thiocyanate in a concentration of 3.0M. It is well known that different the chaotropic solutes are effective in different amounts. In example, it is to be expected that e.g., urea will require a higher concentration in order to provide the same effects as e.g., 3.0M guanidine thiocyanate, since its chaotropic activity is lower than guanidine thiocyanate. On the other hand, an example like Sodium dodecyl sulphate (SDS) has a higher chaotropic potential than both guanidine thiocyanate and urea, and it is therefore to be anticipated that the amount of SDS needed in the first and second composition would be much lower than the amount of either urea or guanidine thiocyanate. In embodiments, the chaotropic solute is guanidine thiocyanate in a concentration in the range of 2-3M.
[0130] As such the amount of the chaotropic solute should be sufficient to destabilize the hydrogen bonding, and it is well known to the skilled person that the optimal range for each of the different chaotropic solutes to obtain this effect varies. It is for instance commonly known that more urea is generally needed compared to guanidine thiocyanate to obtain the same effect.
[0131] It was found that the isolation of DNA fragments was highly dependent on the amount of organic solvent in the compositions. A lower concentration of the organic solvent entailsbinding of preferentially larger DNA fragments (>75 nucleotides) to the solid support, where a higher concentration of the organic solvent could facilitate binding of shorter cfDNA fragments (<75 nucleotides). Accordingly, by having an initial low or intermediate level of the organic solvent, it is possible to selectively bind the longer DNA fragments to the matrix, while the shorter fragments remain in solution. In the second step the higher concentration of the organic solvent will then facilitate the binding of the remaining DNA fragments in the solution, i.e., both the short and long fragments. It will also be evident to the skilled person that there are suitable alternatives to isopropanol which will have different effective concentrations and means for identifying such alternatives are well known to the skilled person. The inventors found that the amount of the organic solvent could be varied with the total amount of DNA in the sample, such that for higher concentrations of total DNA, more of the organic solvent could be added, to maintain the preferential binding of the mammalian cfDNA. As such where lower amounts of total DNA in the sample was measured, a level of about 10%v / v organic solvent was found to be optimal for binding of the mammalian cfDNA, while a higher level of total DNA required a higher concentration of organic solvent, where preferential binding, was seen for mammalian cfDNA up to 20%v / v, or in some cases up to 30% v / v of the organic solvent, without binding high amounts of microbial cfDNA to the matrix at the same time. As such different organic solvents may be used, but it is preferable that the organic solvent is isopropanol. In cases where the DNA concentration in the sample exceeds 500 pg, up to 30%v / v organic solvent in the composition for binding the larger DNA fragments may be used.
[0132] Examples of organic solvents are e.g., alcohol, such as methanol, ethanol, isopropanol, 1-propanol, n-butanol, and isopropanol, a hydrophilic aromatic solvent, such as phenol, o-cresol, o-chlorophenol and 2,6-dimethylphenol, a polar aprotic solvent such as dimethyl sulfoxide (DMSO), acetonitrile, dimethylformamide (DMF) and N-Methyl-2-pyrrolidone (NMP), an amide-based solvent such as formamide, N, N-Dimethylacetamide (DMAc), N, N-Dimethylpropionamide and methyl formamide, a halogenated organic solvent such as chloroform, dichloromethane (DCM), trichloro ethanol and carbon tetrachloride.
[0133] Since the samples from which the cfDNA is to be enriched often contain a plethora of different constituents which has an influence on the successfulness of an extraction of DNA from the sample, it is preferable that the composition further comprises agents which enhances the extraction efficiency from the sample. Examples of such agents are e.g., surfactants, buffers and chelating agents. The effects of each of these agents are well known to the skilled person.
[0134] As such it is preferable that the first and / or second composition further comprise an aqueous solvent, a non-ionic surfactant, a buffering agent and a chelating agent / divalent ionic captureagent. Preferably, the aqueous solvent is water, the non-ionic surfactant is Tween20, the buffering agent is tris and the chelating is EDTA.
[0135] The non-ionic surfactant, which generally acts to solubilize cell membranes and denature proteins, aiding in cell lysis and releasing DNA. Non-ionic surfactants also help remove lipids and membrane debris during DNA purification, improving yield and purity, while also reducing the clotting tendency of proteins and lipids in the sample. Examples of the non-ionic surfactant are Tween20, Tween80 and Triton X-100.
[0136] The chelating agent generally has the function of reducing the function of degradation enzymes, such as DNAses, by chelating the ions, e.g., Ca2+ or Mg2+ which is often required for their function. Examples of chelating agents that can chelate metal ions are e.g., EDTA, EGTA, DTPA, NTA, Citrate, Phytic acid and CDTA.
[0137] To maintain the pH of the compositions it is preferrable that the compositions comprise a buffering agent or buffering system. Examples of suitable buffering agents include HEPES, Tris, TAPS, Bicine, MOPS, AMPD, CHES, TES. Examples of buffering systems are e.g., phosphate buffer such as a sodium phosphate buffers or a potassium phosphate buffer. Preferably the buffer is a Tris or phosphate buffer.
[0138] A composition that facilitates binding of cell-free DNA to a matrix
[0139] As mentioned, the present disclosure provides compositions that facilitates the preferential binding of mammalian cell-free DNA or general DNA binding to suitable matrices. Since the mammalian cell-free DNA fragments are longer fragments, compared to e.g., bacterial fragments, the composition should facilitate the preferential binding of long DNA fragments with an approximate size of 75 nucleotides or more to a matrix, such as long DNA fragments with an approximate size of 100 nucleotides or more.
[0140] The inventors realized that by alternating the amount of some of the components in the compositions, the binding of mammalian cfDNA fragments, i.e., fragments of more than 75 nucleotides in a sample towards the suitable matrix could be manipulated. The specific buffer composition allows for isolation of mammalian cfDNA fragments from a sample, either as an end product, or in order to remove mammalian cfDNA from the sample, in order to purify microbial cfDNA.
[0141] Accordingly, the present disclosure provides primarily two distinct compositions, compositions that facilitates the preferential binding of the longer, often mammalian, cell-free DNA
[0142] fragments (cfDNA) to a matrix. The disclosure also provides compositions that facilitates general DNA binding which entails binding of the shorter, often microbial, cfDNA fragments to a matrix.
[0143] As used herein, "longer DNA fragments" is to be understood as DNA fragments with a size larger than 75-100 nucleotides, or larger than 100 nucleotides, 125 nucleotides, or larger than 150 nucleotides.
[0144] As used herein, "shorter DNA fragments" is to be understood as DNA fragments shorter than 75-100 nucleotides, preferably with a size between 10-75 nucleotides or 10-100 nucleotides. In some instances, the shorter DNA fragments may have a size of 10-80 nucleotides. In some instances, the shorter DNA fragments may have a mean size of less than about 75 nucleotides.
[0145] The enrichment compositions generally comprise an aqueous solvent, a chaotropic solute, a non-ionic surfactant, a buffering agent, a chelating agent, and a chaotropic solvent.
[0146] Compared to microbial cfDNA, mammalian cell-free DNA is often longer and has a different chemical composition, in that the mammalian cell-free DNA is often bound to histones.
[0147] Thereby, the inventors realized that the buffer compositions used in the isolation process can be tailored to enable the preferential binding of the larger DNA fragments, i.e., predominantly the mammalian DNA, or buffer compositions which are tailored to bind to all DNA in the sample, including the smaller microbial cfDNA fragments.
[0148] As used herein, a matrix is to be understood as a material which is capable of binding DNA under the conditions imposed by the composition disclosed herein. As such, the material of the matrix is non-limited and may be either solid, such as defined particles or semi-solid, such as gels or the like.
[0149] In embodiments, the matrix is a solid support. In other embodiments, the matrix is a semisolid support. In embodiments, the matrix is a silica based solid support or a carboxylic acid functionalized solid support. An example is silica based solid support which may be purchased in the desired sizes.
[0150] In embodiments, the matrix is silica based, or silica coated, such as e.g., silica beads or silica coated magnetic beads. In other embodiments the matrix is beads coated with carboxyl groups. In other embodiments the beads are coated with DNA oligos.
[0151] Accordingly, mixing of DNA with either of the compositions disclosed herein and a suitable matrix for binding DNA, enables the preferential binding of either mammalian cell-free DNA or the more general binding of DNA to the matrix, thereby enabling a two-step purification, where firstly, the mammalian cfDNA is extracted followed by an extraction of remaining DNA including microbial cfDNA.
[0152] Generally, the compositions are of chaotropic nature. With respect to the present disclosure, the function of the chaotropic aspect of the composition is to disrupt hydrogen bonding and denature proteins, nuclease inactivation, and enabling efficient DNA binding to the matrices, in particular, a first step of binding the larger mammalian DNA fragments, and a second step which comprises binding of remaining DNA including the shorter microbial cfDNA.
[0153] As such a chaotropic composition is an aqueous solution that contains chaotropic agents / solutes i.e., chemicals that disrupt the ordered structure of water and destabilize hydrogen bonding networks in biomolecules. Examples of chaotropic solutes are chaotropic salts such as e.g., guanidine thiocyanate and chaotropic nitrogen compounds such as e.g., urea.
[0154] In embodiments, the chaotropic solute is selected from a chaotropic salt, such as guanidinium chloride, guanidine thiocyanate, guanidium hydrochloride, lithium perchlorate, lithium acetate, magnesium chloride, sodium dodecyl sulphate, and a chaotropic nitrogen compound, such as urea and thiourea. Another example of a chaotropic solute is Sodium dodecyl sulfate (SDS), which is also a detergent.
[0155] In preferred embodiments, the chaotropic solute is selected from Guanidinium isothiocyanate (GTC), Guanidinium hydrochloride (GuHCI), Urea and Sodium perchlorate. Preferably, the chaotropic solute is GTC.
[0156] As such the amount of the chaotropic solute should be sufficient to destabilize the hydrogen bonding, and it is well known to the skilled person that the optimal range for each of the different chaotropic solutes to obtain this effect varies. It is for instance commonly known the more urea is generally needed compared to GuHCI to obtain the same effect.
[0157] As such the concentration of the chaotropic solute may be in the range 1-5 M, such as 1.0 M, 1.2 M, 1.5 M, 2.0 M, 2.5 M, 3.0 M, 3.5 M, 4.0 M, 4.5 M, 5.0 M, or such as 1.0-2.0 M, 1.5-2.5 M, 2.0-3.0 M, 2.5-3.5 M, 3.0-4.0 M or 3.5-5.0 M. Preferably, the concentration of the chaotropic solute is about 3.0M. Preferably, the composition which preferentially binds the larger DNA fragments comprise 2.0-3.0M of the chaotropic solute, and the chaotropic solute is preferably GTC. In embodiments, the composition comprises about 3.0M GTC.
[0158] The composition further comprises a non-ionic surfactant, which generally acts to solubilize and denature proteins. Non-ionic surfactants also help remove lipids and membrane debris during DNA purification, improving yield and purity, while also reducing the clotting tendency of proteins and lipids in the sample.
[0159] Non-limiting examples of non-ionic surfactant are Tween 20, Tween 80, Triton X-100, Triton X-114, Brij 35, Brij 58, NP-40, Span 20, Span 80, Pluronic F68, Pluronic F127, Lubrol PX and Nonidet P-40. In embodiments, the non-ionic surfactant is Tween20, Tween80 or Triton X-100. Preferably, the non-ionic surfactant is Tween20.
[0160] In compositions that enables the preferential binding of mammalian cfDNA, it may be preferable that the concentration of the non-ionic surfactant is in the range 5-8%v / v, such as 5%v / v, 5.5%v / v, 6%v / v, 6.5%v / v, 7%v / v, 7.5%v / v or 8%v / v, or such as 5-6%v / v, 5.5-6.5%v / v, 6-7%v / v, 6.5-7.5%v / v or 7-8%v / v.
[0161] In compositions that enables the preferential binding of longer DNA fragments, i.e., predominantly mammalian cfDNA, it may be preferable that the concentration of Tween20 is in the range l-8%v / v, such as l%v / v, 2%v / v, 3%v / v, 4%v%v, 5%v / v, 5.5%v / v, 6%v / v, 6.5%v / v, 7%v / v, 7.5%v / v or 8%v / v, or such as 5-6%v / v, 5.5-6.5%v / v, 6-7%v / v, 6.5-7.5%v / v or 7-8%v / v. In embodiments the Tween20 concentration is 3.5-5.5%v / v.
[0162] Preferably, the concentration of Tween20 is 4%v / v.
[0163] In compositions that provides general DNA binding, it may be preferable that the concentration of Tween20 is in the range l-4.5%v / v such as l%v / v, 1.5%v / v, 2%v / v, 2.5%v / v, 3%v / v, 3.5%v / v, 4%v / v, or 4.5%v / v, or such as l-2%v / v, 1.5-2.5%v / v, 2-3%v / v, 2.5-3.5%v / v, 3-4%v / v, 3.5-4.5%v / v, l-3%v / v, or 2-4.5%v / v.
[0164] The compositions further comprise a chelating agent, which has the function of reducing the function of degradation enzymes, such as DNAses, by chelating the ions, e.g., Ca2+ or Mg2+ which is often required for their function.
[0165] Examples of chelating agents that can chelate metal ions are e.g., EDTA, EGTA, DTPA, NTA, Citrate, Phytic acid and CDTA.
[0166] In preferred embodiments, the chelating agent is EDTA.
[0167] The concentration of the chelating agent may be varied such that it sufficiently chelates the metal ions and inhibits the DNA degradation enzymes.
[0168] The inventors found that the optimal concentration of the chelating agent for the binding of mammalian cfDNA was about 6mM. In some embodiments, the concentration of the chelating agent is in the range 4-7 mM such as 4 mM, 4.5 mM, 5 mM, 5.5 mM, 6 mM, 6.5 mM or 7 mM, or such as 4-5 mM, 4.5-5.5 mM, 5-6 mM, 5.5-6.5 mM, 6-7 mM, 4-6 mM, or 5-7 mM.
[0169] The inventors found that the optimal concentration of the chelating agent for the general binding of DNA was about 9.5 mM. In other embodiments, the concentration of the chelating agent is in the range 8-12 mM, such as 8 mM, 8.5 mM, 9 mM, 9.5 mM, 10 mM, 10.5 mM, 11 mM, 11.5 mM, or 12 mM, or such as 8-9 mM, 9-10 mM, 10-11 mM, 11-12 mM, 8-10 mM, 9-11 mM, or 10-12 mM.
[0170] In some embodiments, the concentration of the chelating agent is about 6 mM. In other embodiments, the concentration of the chelating agent is about 9.5 mM. In preferred embodiments, the chelating agent is EDTA and is present in 5-20 mM.
[0171] The amount of the organic solvent in the compositions was found to be essential for the preferential binding of mammalian cfDNA or simply larger DNA fragments population above the 75 bp cutoff, and by tailoring the amount of the organic solvent, the larger predominantly mammalian cfDNA could be isolated from a sample, leaving the smaller largely microbial cfDNA fragments in the solution, thereby enriching the sample for microbial cfDNA.
[0172] Accordingly, the compositions comprise an organic solvent, which has the function of reducing the solubility of DNA in solution, thereby enhancing the binding to the solid support. Since the solubility of the DNA is largely dependent on its size and other characteristics, the organic solvent may be used at different concentrations to preferably bind different DNA fractions. Preferably, the organic solvent is selected from organic solvent with these properties, such as e.g., selected from the group consisting of isopropanol, ethanol, DMSO, phenol, chloroform and formamide.
[0173] The inventors also found that the amount of the organic solvent could be varied with the total amount of DNA in the sample, such that for higher concentrations of total DNA, more of the organic solvent could be added, to maintain the preferential binding of the longer mammalian cfDNA. As such where lower amounts of total DNA in the sample was measured, a level of about 10%v / v organic solvent was found to be optimal for binding of the mammalian cfDNA, while a higher level of total DNA required a higher concentration of organic solvent, where preferential binding, was seen for the larger and predominantly mammalian cfDNA up to 20%v / v organic solvent, without binding high amounts of smaller and predominantly microbial cfDNA to the matrix at the same time.
[0174] Examples of organic solvents are e.g., alcohol, such as methanol, ethanol, isopropanol, 1-propanol, n-butanol, and isopropanol, a hydrophilic aromatic solvent, such as phenol, o-cresol, o-chlorophenol and 2,6-dimethylphenol, a polar aprotic solvent such as dimethyl sulfoxide (DMSO), acetonitrile, dimethylformamide (DMF) and N-Methyl-2-pyrrolidone (NMP), an amide-based solvent such as formamide, N, N-Dimethylacetamide (DMAc), N, N-Dimethylpropionamide and methylformamide, a halogenated organic solvent such as chloroform, dichloromethane (DCM), trichloroethanol and carbon tetrachloride.
[0175] In embodiments, the organic solvent is an alcohol. In embodiments, the organic solvent is an alcohol selected from the group consisting of methanol, ethanol, 1-propanol, n-butanol, and isopropanol. Preferably the organic solvent is isopropanol, and the preferred concentration of isopropanol is 10-18%v / v, such as e.g., 10%v / v, ll%v / v, 12%v / v, 13%v / v, 14%v / v, 15%v / v 16%v / v, 17%v / v or 18 %v / v, more preferably 12-17%v / v in the composition which preferentially binds the larger DNA fragments.
[0176] Generally, the binding of DNA to the solid matrix was found to be dependent on the amount of isopropanol used, the amount of chaotropic solute and the amount of non-ionic surfactant as can be seen from figures 6-21 and example 2-6. From the examples it was found that an increase in isopropanol from 0-35%v / v was found to correlate with DNA binding in a size dependent manner. At lower concentrations of isopropanol only larger DNA fragments were bound to the matrix, as could be seen by subsequent size analysis. Increasing the isopropanol then resulted in a shift in the size of the DNA fragments bound to the matrix, and at 30%v / v isopropanol all DNA in the sample was bound to the matrix. This prompted the investigation by the inventors to find the optimum of conditions which allowed for isolation of the smaller cfDNA fragments. Since the increase in isopropanol concentration increased general DNA binding, a high initial isopropanol concentration in the initial separation step would result in complete binding of all DNA fragments in the sample, meaning that both short and long fragments would be isolated i.e., both mammalian and microbial. To avoid this, the inventors split is the DNA isolation into two steps, a first step where a buffer is used which binds the larger DNA fragments to a solid matrix that can then be removed from the sample, and a second step where the remaining DNA fraction can then be bound to a solid matrix and isolated for further analysis and sequencing.
[0177] Accordingly, the first buffer composition used generally differs from the second buffer composition in that the first buffer is optimized to bind the larger DNA fragments, preferably those with a size of 75 bp or larger, which requires more specific buffer tailoring than the second buffer which generally acts to bind all remaining DNA.
[0178] Exemplary concentrations and suggested composition compounds are provided in table 1, which also proposes a number of potential alternatives.
[0179] Table 1 Overview of exemplary components and alternatives in relation to the long fragment binding composition
[0180]
[0181]
[0182] concentration. At high concentration of DNA, a high level e.g., 20-30%v / v of isopropanol is preferable whereas for lower DNA concentrations 10-12.5%v / v of isopropanol is more suitable.
[0183] It was found by the inventors, as shown in Example 3, that the isolation of the longer DNA fragments was optimal at an isopropanol level of about 15 %v / v, as this provided the desired cut-off of about 75 bp, as can be seen in figure 17 and 18.
[0184] In compositions for preferential binding of mammalian cfDNA, or large cfDNA fragments it may be optimal that the concentration of the organic solvent is in the range 5-20%v / v such as 5%v / v, 6%v / v, 7%v / v, 8%v / v, 9%v / v, 10%v / v, ll%v / v, 12%v / v, 13%v / v, 14%v / v, 15%v / v, or such as 5-7%v / v, 7-9%v / v, 9-ll%v / v, ll-13%v / v, 13-15%v / v, 15-17%v / v, 17-19%v / v, 10-20%v / v, 5-10%v / v, 10-15%v / v, or 15-20%v / v. In some cases where the initial DNA concentration in the sample is high, such as e.g., above 300 ng / sample, such as e.g., 750 ng / sample, then a higher initial concentration of the organic solvent may be needed to ensure proper binding of the longer DNA fragments. In such cases the compositions for preferential binding of mammalian cfDNA may comprise the organic solvent in a range of 20-28%v / v such as e.g., 20 %v / v, 21 %v / v, 22 %v / v, 23 %v / v, 24 %v / v, 25 %v / v, 26 %v / v, 27 %v / v or 28 %v / v.
[0185] In compositions for preferential binding of mammalian cfDNA it may be optimal that the concentration of the isopropanol is in the range 5-20%v / v such as 5%v / v, 6%v / v, 7%v / v, 8%v / v, 9%v / v, 10%v / v, ll%v / v, 12%v / v, 13%v / v, 14%v / v, 15%v / v, or such as 5-7%v / v, 7-9%v / v, 9-ll%v / v, ll-13%v / v, 13-15%v / v, 15-17%v / v, 17-19%v / v, 10-20%v / v, 5-10%v / v, 10-15%v / v, or 15-20%v / v. In some cases where the initial DNA concentration in the sample is high, such as e.g., above 300 ng / sample, such as e.g., 750 ng / sample, then a higher initial concentration of the isopropanol may be needed to ensure proper binding of the longer DNA fragments. In such cases the compositions for preferential binding of mammaliancfDNA may comprise the organic solvent in a range of 20-28%v / v such as e.g., 20 %v / v, 21 %v / v, 22 %v / v, 23 %v / v, 24 %v / v, 25 %v / v, 26 %v / v, 27 %v / v or 28 %v / v.
[0186] In compositions for binding most or all DNA fragments it may be optimal that the concentration of the organic solvent is in the range 20-45%v / v such as about 20%v / v, 22%v / v, 24%v / v, 26%v / v, 28%v / v, 30%v / v, 32%v / v, 34%v / v, 35%v / v, 36%v / v, 37%v / v, 38%v / v, 39%v / v, 40%v / v, 41%v / v, 42%v / v, 43%v / v, 44%v / v, or such as about 45%v / v, or such as 20-25%v / v, 25-30%v / v, 30-35%v / v, 35-40%v / v, 40-45%v / v, 22-28%v / v, 26-32%v / v, 20-30%v / v, 25-35%v / v or such as 35-45%v / v.
[0187] In compositions or binding most or all DNA fragments it may be optimal that the concentration of the isopropanol is in the range 20-45%v / v such as about 20%v / v, 22%v / v, 24%v / v, 26%v / v, 28%v / v, 30%v / v, 32%v / v, 34%v / v, 35%v / v, 36%v / v, 37%v / v, 38%v / v, 39%v / v, 40%v / v, 41%v / v, 42%v / v, 43%v / v, 44%v / v, or such as about 45%v / v, or such as 20-25%v / v, 25-30%v / v, 30-35%v / v, 35-40%v / v, 40-45%v / v, 22-28%v / v, 26-32%v / v, 20-30%v / v, 25-35%v / v or such as 35-45%v / v.
[0188] In embodiments, the organic solvent is a hydrophilic aromatic solvent. In embodiments, the organic solvent is a hydrophilic aromatic solvent selected from the group consisting of phenol, o-cresol, m-cresol, o-chlorophenol and 2,6-dimethylphenol.
[0189] In embodiments, the organic solvent is a polar aprotic solvent. In embodiments, the organic solvent is a polar aprotic solvent selected from the group consisting of dimethyl sulfoxide (DMSO), acetonitrile, dimethylformamide (DMF) and N-Methyl-2-pyrrolidone (NMP).
[0190] In embodiments, the organic solvent is an amide-based solvent. In embodiments, the organic solvent is an amide-based solvent selected from the group consisting of formamide, N, N-Dimethylacetamide (DMAc), N, N-Dimethylpropionamide and methylformamide.
[0191] In embodiments, the organic solvent is a halogenated organic solvent. In embodiments, the organic solvent is a halogenated organic solvent selected from the group consisting of formamide chloroform, dichloromethane (DCM), trichloroethanol and carbon tetrachloride.
[0192] To maintain the pH of the compositions it is preferrable that the compositions comprise a buffering agent or buffering system. Examples of suitable buffering agents include HEPES, Tris, TAPS, Bicine, MOPS, AMPD, CHES, TES. Examples of buffering systems are e.g., phosphate buffer such as a sodium phosphate buffers or a potassium phosphate buffer.
[0193] Preferably the buffer is a Tris or phosphate buffer.
[0194] In embodiments, the concentration of the buffering agent is in the range 5-20 mM such as 5 mM, 6 mM, 7 mM, 8 mM, 9 mM, 10 mM, 11 mM, 12 mM, 13 mM, 14 mM, 15 mM, 16 mM, 17 mM, 18 mM, 19 mM, or 20 mM, or such as 5-10 mM, 10-15 mM, 15-20 mM, 6-12 mM, 8-16 mM, 5-15 mM or 10-20 mM. In other embodiments, the concentration of the buffering agent is below 100 mM, such as in the range of 10-100 mM. For the compositions of the present disclosure the inventors found an optimal Tris concentration to be 14-15 mM.
[0195] In some embodiments, the buffer agent used has a buffering capacity in the range of pH 6 to pH 8.5. Preferably the buffer is active at pH 7.5 to pH 8.5.
[0196] In embodiments, the compositions have a pH of 5.0 to 9.5. In other embodiments, the compositions have a pH of 7.0 to 8.5. In preferred embodiments, the compositions have a pH of about pH 8.0. In other embodiments, the compositions have a pH of 5.0 to 8.5. In other embodiments, the compositions have a pH of 5.5-8.0.
[0197] In embodiments the composition that facilitates or binding most or all DNA fragments to a matrix comprises
[0198] a chaotropic solute, preferably guanidine thiocyanate, at a concentration of about 3.0 M,
[0199] a non-ionic surfactant, preferably Tween20 at a concentration of about 3.7- 4.0%v / v,
[0200] a chelating agent, preferably EDTA, at a concentration of about 9.5 mM and
[0201] an organic solvent at a concentration of about 30%v / v.
[0202] In embodiments, the composition that facilitates or binding most or all DNA fragments to a matrix comprises a non-ionic surfactant in a concentration of about 2-5%v / v, EDTA in a concentration of 8-12 mM, and isopropanol in a concentration of 25-45%v / v and the pH of the composition is in the range 5.0-8.5.
[0203] In embodiments, the composition that facilitates binding of most or all DNA fragments to a matrix comprises Tris in a concentration from 5-20 mM, a non-ionic surfactant selected from Tween 20, Tween 80, Triton X-100, Triton X-114, Brij 35, Brij 58, NP-40, Span 20, Span 80, Pluronic F68, Pluronic F127, Lubrol PX and Nonidet P-40, and a chaotropic solute is selectedfrom solute is selected from a chaotropic salt a chaotropic nitrogen compound, preferably guanidine thiocyanate (GTC).
[0204] In embodiments a composition that facilitates preferential binding of mammalian cfDNA fragments to a matrix comprises
[0205] a chaotropic solute, preferably guanidine thiocyanate, at a concentration of about 3.0 M,
[0206] a non-ionic surfactant, preferably Tween20 at a concentration of about 6.5%v / v,
[0207] a chelating agent, preferably EDTA, at a concentration of about 6.1 mM and an organic solvent at a concentration of about 10%v / v.
[0208] In embodiments, the composition that facilitates preferential binding of mammalian cfDNA fragments to a matrix comprises a non-ionic surfactant in a concentration of about 5-10%v / v, EDTA in a concentration of 4-7 mM, and isopropanol in a concentration of 7-13 %v / v and the pH of the composition is in the range 5.0-8.5, preferably 7.5-8.5.
[0209] In embodiments, the composition that facilitates preferential binding of mammalian cfDNA fragments to a matrix comprises Tris in a concentration from 5-20 mM, a non-ionic surfactant selected from Tween 20, Tween 80, Triton X-100, Triton X-114, Brij 35, Brij 58, NP-40, Span 20, Span 80, Pluronic F68, Pluronic F127, Lubrol PX and Nonidet P-40, and a chaotropic solute is selected from a chaotropic salt, or a chaotropic nitrogen compound, preferably the chaotropic solute is guanidine thiocyanate (GTC).
[0210] Exemplary concentrations and suggested composition compounds are provided in table 2, which also proposes a number of potential alternatives.
[0211] Table 2 Overview of exemplary components and alternatives in relation to the short fragment binding composition
[0212]
[0213]
[0214] A kit comprising the compositions for isolation of small DNA fragments
[0215] The present disclosure also relates to a kit which is suitable for use in isolating microbial cfDNA fragments from a sample.
[0216] In embodiments such a kit comprises
[0217] a. a vial comprising a composition comprising
[0218] i. a chaotropic solute,
[0219] ii. a non-ionic surfactant,
[0220] Hi. optionally, a chelating agent and
[0221] iv. optionally, a buffering agent, and
[0222] b. optionally, one or more matrices capable of binding DNA, optionally contained within a vial, and
[0223] c. optionally, means for performing the microbial cell-free DNA isolation, d. optionally, a DNA ladder, and
[0224] e. instructions for preparing the composition for binding long DNA fragments according to the present disclosure, and
[0225] f. Optionally, instructions for isolating microbial cfDNA fragments using said kit.
[0226] The kit may also comprise one or more vials comprising wash buffer(s), and further one or more vials comprising elution buffer(s).
[0227] A washing buffer may comprise 15-40%v / v isopropanol, 1-8M or l-10%v / v of the chaotropic solute, a chelating agent in an amount of 1-10 mM, and a non-ionic surfactant in an amount of 0.1-10%v / v. A further washing buffer may comprise buffered ethanol, such as a composition comprising 1-lOOmM Tris, pH 8.0, l-5mM EDTA and 80%v / v ethanol, or 100% ethanol. The kit may further comprise a vial comprising an elution buffer, and the elution buffer is generally a TE buffer, a TAE buffer or a TBE buffer. In embodiments the first washing buffer comprises 3 M GTC, 30 v / v% isopropanol, 5% Tween20, 50 mM Bis-Tris pH 6, and 4 mM EDTA. In embossments, the elution buffer comprises 10 mM Tris-HCI, 0.1 mM EDTA.
[0228] Accordingly, the present disclosure provides a kit for use in the enrichment of predominantly microbial cfDNA fragments, preferably with a predominant size of 75 bp or less.
[0229] Generally, isopropanol is provided in a separate vial or provided by the user, such that the first vial provided in the kit provides the remaining constituents of the composition, and the person using the kit will mix a portion of the composition of the vial with a portion of the sample and a portion of isopropanol to reach the final composition, which is capable of binding the intended DNA fragments.
[0230] In example, in a final volume of lOOOpI with a target concentration of isopropanol of 10%v / v, about 250pl of a sample may be mixed with 750 pl of the composition in the first vial. In such a case the concentration of isopropanol in the first vial of the kit would be about 13%v / v. Likewise, in cases where the final concentration of isopropanol is 20%, and 500pl of thesample is mixed with 500pl of the composition of the first vial, the concentration of isopropanol in the first vial of the kit would have to be about 40%v / v. Such calculations for the dilutions are well known to the skilled person. As such what is meant by "mixing with a sample in a suitable amount" is that the amount of constituents in the vials on the kit are dependent on the intended mixing ratios between sample and the composition in the vial, in some cases, the total volume is lOOOul, and the sample volume is about 250pl while the volume to be used from the vial is about 750ul, thereby the concentration in the vial needs to be about 33% higher than the final concentration. Where the sample and the composition in the vial are mixed in a 1:1 ratio, the concentration in the vial of the constituents needs to be twice the final concentration. As such the sample and the composition of the first vial may be mixed in any practical ratio, such as e.g., 1:1, 2:1, 3:1, 1:2, 1:3, 1:5, 1:7, 1:10, 1:20, 1:50 or 1:100.
[0231] Where the exact amount of a constituent is not provided with the kit any volumes of the first and / or further vial(s) to be mixed with the sample may be provided in the instructions for use.
[0232] In some embodiments, the a vial of the kit comprises isopropanol in an amount of 10-90 %v / v. In another embodiment, a second vial of the kit comprises isopropanol in an amount of 25-90 %v / v.
[0233] In some embodiments, the kit further comprises one or more vials comprising a wash buffer as disclosed herein, preferably at least two different washing buffers as disclosed herein, more preferably the first and second washing buffer as disclosed herein. In further embodiments, the kit also comprises one or more vials comprising an elution buffer as described herein.
[0234] In some embodiments, the one or more matrices capable of binding DNA is provided with the kit. In other embodiments, the one or more matrices is capable of binding DNA is provided separately from the kit.
[0235] The matrices, such as silica beads, may be premixed with the vial(s) or it may be added to the vial(s) / sample upon performing a method as disclosed herein.
[0236] In embodiments, the kit comprises the vials for the first and second washing buffer. In further embodiments, the first washing buffer comprises 15-40%v / v isopropanol, 1-8M or 1-10%v / v of the chaotropic solute, a chelating agent in an amount of 1-10 mM, and a non-ionic surfactant in an amount of 0.1-10%v / v. In additional embodiments, the second washingbuffer comprises buffered ethanol, such as a composition comprising 1-lOOmM Tris, pH 8.0, l-5mM EDTA and 80%v / v Ethanol.
[0237] In embodiments, the kit comprises the vial comprising an elution buffer. In further embodiments, the elution buffer is a TE buffer, a TAE buffer or a TBE buffer.
[0238] A method for enriching microbial cfDNA in a sample, wherein the sample comprises a mix of microbial cfDNA and mammalian cfDNA
[0239] The present disclosure also relates to a method of enriching microbial cfDNA from a sample comprising a mix of microbial cfDNA and mammalian cfDNA. The method comprises the use of the compositions of the present disclosure, wherein a first composition is optimized as described herein to preferentially bind and isolate mammalian cfDNA, and a second composition which is optimized to preferentially bind microbial cfDNA is optimally used in a second binding and isolation step.
[0240] The concentration of the constituents in the composition are to be understood such that it is the final concentration of constituents present during the binding between the DNA and the matrix. Accordingly, the specific concentration of constituents in the compositions which are prepared is often higher than the final binding concentration. As such the compositions as disclosed herein are often prepare from a stock solution e.g., an aqueous solution, which is to be mixed with the sample in order to reach the target concentration of the different constituents disclosed herein. Accordingly, the stock solution may vary depending on the mixing ratio between sample and stock solution. In some instances, a 1000 pL reaction volume is desired this may be obtained by mixing any amount of a stock solution with any amount of sample, optionally also with addition of a predefined amount of the matrix. In examples, 1-800 pL sample may be mixed with 999-200 pL of the stock solution, such as e.g., 10 pL of sample mixed with 990 pL of stock solution, 50 pL with 950 pL, 100 pL with 900 pL, 200 pL with 800 pL, 400 pL with 600 pL, 600 pL with 400 pL, or 800 pL with 200 pL of stock solution. Additionally, samples may be pooled before or after addition of the stock solution, thereby changing the total volume, without changing the concentration. In other instances, a first mixing is performed, followed by a second mixing or pooling step. Any such different methods of arriving at the final conditions set by the composition disclosed here encompassed herein and are within the capabilities of the skilled person to derive.
[0241] In other examples, the sample is mixed with the stock solution in a ratio of a 1:1. In such a case the concentration of the constituents in the stock solution needs to be twice the final concentration i.e., twice the concentration of the first composition for binding longer DNAfragments. As such the sample and stock solution may be mixed in any practical ratio, such as e.g., 1:1, 2:1, 3:1, 1:2, 1:3, 1:5, 1:7, 1:10, 1:20, 1:50 or 1:100. In several cases the solubility and other practical parameters set the boundaries for the concentration of the constituents in the stock solution, such boundaries are easily derived from the constituents and means to evaluate the maximal practical concentration of each of the components may be calculated without undue burden.
[0242] As such the method for enriching the short cfDNA fragments, predominantly comprising the microbial cfDNA, may comprise the steps of
[0243] a) mixing a sample with a first aqueous solution or stock solution to obtain a composition that facilitates the preferential binding of longer (>75 nucleotides) mammalian cell-free DNA fragments (cfDNA) to a matrix,
[0244] b) contacting the sample with a first matrix capable of binding DNA, thereby binding longer DNA fragments i.e., predominantly mammalian cfDNA, to the first matrix, thereby increasing the ratio in unbound microbial cfDNA to unbound mammalian cfDNA in the sample, and
[0245] c) clearing the matrix bound mammalian cfDNA from the sample, to obtain a sample enriched in short DNA fragmentation i.e., predominantly microbial cfDNA fragments.
[0246] Figures 7-21 show how the two different compositions may be used to enrich the short DNA fragments.
[0247] As such the first aqueous solution that is to be mixed with the sample is generally a more concentrated version of the final composition that facilitates the preferential binding of mammalian cell-free DNA fragments (cfDNA) to a matrix. In example, about 250 pL sample may be mixed with about 750 pL of the aqueous solution. In such an instance the first aqueous solution is diluted about 33%, wherefore the first aqueous solution in this example should have a concentration of the constituents which is 33% higher than the final composition, such calculations are well known by the skilled person. Accordingly, depending on the mixing ratio the first aqueous solution may have different concentrations. As such the skilled person will know how to calculate the exact amount of each of the components to be added to a given amount of the sample to reach the different target concentrations as provided by the compositions described herein.
[0248] As such any volume, also referred to as reaction volume, of the sample and the first aqueous solution may be mixed, as long as the final mixture has the intended features i.e., that the final mixture facilitates the preferential binding of longer (>75 nucleotides) mammalian cell-free DNA fragments (cfDNA) to a matrix. In embodiments, the reaction volume is 100-100.000 pl, such as about 10 pl, 20 pl, 50 pl, 100 pl, 150 pl, 200 pl, 300 pl, 500 pl, 750 pl,1.000 pl, 1.250 pl, 1.500 pl, 3.000 pl, 5.000 pl, 7.000 pl, 10.000 pl, 25.000 pl, 50.000 pl, 70.000 pl, or such as about 100.000 pl. In some embodiments, the volume of the sample is 100-10. OOOul, such as about 10 pl, 20 pl, 50 pl, 100 pl, 150 pl, 200 pl, 300 pl, 500 pl, 750 pl, 1.000 pl, 1.250 pl, 1.500 pl, 3.000 pl, 5.000 pl, 7.000 pl, or such as about 10.000 pl. In some embodiments, the volume of the first aqueous solution is 10-10.000 pl, such as about 10 pl, 20 pl, 50 pl, 100 pl, 150 pl, 200 pl, 300 pl, 500 pl, 750 pl, 1.000 pl, 1.250 pl, 1.500 pl, 3.000 pl, 5.000 pl, 7.000 pl, or such as about 10.000 pl. In some embodiments, the volume of the first matrix capable of binding DNA is 1-10.000 pl, such as about 1 pl, 2 pl, 3 pl, 5 pl, 7 pl, 10 pl, 20 pl, 50 pl, 100 pl, 150 pl, 200 pl, 300 pl, 500 pl, 750 pl, 1.000 pl, 1.250 pl, 1.500 pl, 3.000 pl, 5.000 pl, 7.000 pl, or such as about 10.000 pl.
[0249] Once steps a)-c) have been completed the sample is principally enriched in microbial cfDNA, since the buffer composition have ensured that it is preferentially the mammalian cfDNA which is bound to the matrix. It is preferable that the microbial cfDNA fragments content in the sample is significantly enriched following step c).
[0250] Examples of such enrichment may be seen e.g., in figure 6-21 and 23 and Examples 3 and 8.
[0251] The separation or clearing performed in step c) can principally be conducted in a plethora of ways known to the skilled person, depending on the matrix used. In example, beads or particles, such as silica particles bound with the longer, often mammalian, cfDNA may be separated from the solution e.g., by centrifugation, sedimentation or column separation. Alternatively, where e.g., silica coated magnetic particles are used, the particles bound with the mammalian cfDNA may be isolated using magnetic separation.
[0252] Once the majority of the mammalian cfDNA is cleared from the sample, the microbial cfDNA may be further enriched or purified via a number of further steps e.g., using the composition which is optimized to promote general DNA binding including the shorter (<75 nucleotides) microbial cfDNA to the matrix.
[0253] Accordingly, in embodiments, the method further comprises the additional steps:
[0254] d) contacting the sample enriched for microbial cfDNA fragments from step c) with a second aqueous solution to obtain a composition that promote general DNA binding, including microbial cfDNA, to a matrix as disclosed herein, e) contacting the sample with a second matrix capable of binding DNA thereby enriching for microbial cfDNA on the second solid or semi-solid support, f) isolating the matrix preferentially bound with microbial cell-free DNA from remaining sample,
[0255] g) optionally, one or more washing steps using one or more washing buffers, andh) optionally, eluting the microbial cfDNA from the second matrix to obtain a sample further enriched in microbial cfDNA fragments.
[0256] In some embodiments, the volume of the second aqueous solution is 10-10.000 pl, such as about 10 pl, 20 pl, 50 pl, 100 pl, 150 pl, 200 pl, 300 pl, 500 pl, 750 pl, 1.000 pl, 1.250 pl,
[0257] 1.500 pl, 3.000 pl, 5.000 pl, 7.000 pl, or such as about 10.000 pl. In some embodiments, the added volume of the second matrix is 1-10.000 pl, such as about 1 pl, 2 pl, 3 pl, 5 pl, 7 pl, 10 pl, 20 pl, 50 pl, 100 pl, 150 pl, 200 pl, 300 pl, 500 pl, 750 pl, 1.000 pl, 1.250 pl, 1.500 pl,
[0258] 3.000 pl, 5.000 pl, 7.000 pl, or such as about 10.000 pl.
[0259] As such, isolation and / or clearing may be carried out, in a plethora of ways known to the skilled person, including, centrifugation, sedimentation, column separation and magnetic separation.
[0260] Further distinct compositions include those applicable for washing of the DNA bound matrices, i.e., the mammalian cfDNA and / or microbial cfDNA bound matrices, thereby removing unwanted loosely associated DNA fragments and other biological contaminants such as proteins, lipids, sugars etc. are also described herein.
[0261] The method may also comprise a series of washing steps, which serves to remove contaminants bound to the solid support following binding of the short fragments i.e., microbial cfDNA to the solid support. Generally, by washing is meant that the solid support bound with the DNA is isolated from the solution e.g., centrifugation, whereafter the supernatant is discarded, and the beads are suspended in a washing buffer.
[0262] In embodiments, the methods comprise at least two washing steps. In embodiments, the first wash is performed using a first washing composition which may be identical to the composition that facilitates preferential binding of short DNA fragments, preferably with an approximate size of 75 nucleotides or less to a matrix as described herein, preferably, with a pH of 5.5-8.0. In further embodiments, a second washing step is performed using a second washing composition comprising buffered ethanol, e.g., comprising 50 mM Tris pH 8.0, 2mM EDTA and 80%v / v Ethanol. In further embodiments, a third washing step is performed using a composition that comprises pure ethanol. In some embodiments, the washing steps are performed once each. In other embodiments, washing steps are each performed more than one time. In some embodiments, the first washing step is repeated multiple times. In other embodiments, the second washing step is repeated multiple times. In further embodiments, the third washing step is repeated multiple times. Preferably the washing buffer used in the terminal step before elution is compatible with the intended NGS platform.
[0263] Elution from the matrix may be done by means suitable therefore and means to elute DNA from specific matrices are readily available to the skilled person, such as e.g., by exposing the DNA bound matrix to salt gradients and / or pH gradients. In some cases, elution may be used to prepare the isolated DNA for sequencing or other analysis methods. In other cases, the DNA is analysed directly while bound to the matrix.
[0264] Elution is usually performed using an elution buffer such as e.g., a Tris / EDTA buffer (TE buffer), a Tris / acetate / EDTA (TAE buffer) buffer or a Tris / borate / EDTA buffer (TBE buffer). The compositions of commonly used DNA elution buffers are well known to the skilled person.
[0265] It is preferable that the amount of microbial cfDNA fragment in the sample is enriched following step g) compared to the crude sample before isolation of the mammalian cfDNA. Examples of such enrichments may e.g., be seen from Example 2, 3 and 8, and figures 6-21. The enrichment may be addressed e.g., by comparing the electropherograms of the crude sample to that of the enriched sample e.g., using an Agilent Tapestation System to measure the electropherograms or by comparing the ratio or counts between sequenced microbial and host DNA (DNA reads) of crude samples to that of enriched samples as outlined in example 8.
[0266] It is to be understood that the sample may be any sample that is suspected of containing both mammalian and microbial cfDNA. As such the sample may be a mammalian sample or a sample associated with mammalian sources. Preferably, the sample is a sample from a human subject.
[0267] As such the type of sample is not limited, and may principally be any sample containing DNA, it is however preferable that the sample is selected from a cell-free fluid sample, preferably, a serum sample, a plasma sample, a cerebrospinal fluid sample, a BAL sample, exhaled condensate sample and a urine sample. Alternatively, the sample may be a cell containing sample, which is pretreated before incubation with the compositions of the present disclosure, such that once the cfDNA is to be isolated the sample is preferably, cell-free.
[0268] The origin the microbial cfDNA is as such not limited but should preferably have a chemical composition which allows for the isolation according to the methods disclosed herein. As such the microbial cell-free DNA fragments may be of bacterial, viral, parasite or fungi species origin.
[0269] When performing the method, the matrices used in the clearing or isolation steps may either be the same matrix in both steps, or it may be different matrices in the different steps.
[0270] For instance, the matrix used to bind mammalian cfDNA may be an optimized matrix, which has been optimized to bind preferentially to mammalian cfDNA.
[0271] On the other hand, the matrix used to bind microbial cfDNA may be an optimized matrix, which has been optimized to bind preferentially to microbial cfDNA.
[0272] Accordingly, in embodiments, the first and second matrices are of the same composition.
[0273] In other embodiments, the first and second matrices are of a different composition.
[0274] As mentioned herein the first and / or second matrix may be a solid support or semi-solid support.
[0275] A method for microbial cfDNA sequence analysis
[0276] Following enrichment and sequencing the obtained sequencing data is to be analysed to determine the source of the infection.
[0277] Once complete, the sequencing provides a raw list of DNA reads that have been identified in the sample.
[0278] To determine the source of the DNA and to provide a meaningful insight into the source of the infection, a series of steps are required to analyse the sequencing data.
[0279] Current methodologies for identification of sources of an infection, such as e.g. PCR based diagnostic methods, are often limited in the number of species which can be detected in a single sample, since, in the case of PCR, the method can only address the species for which the PCR primers are specific. On the other hand, sequencing-based methods can detect very small amounts of DNA and is therefore exceptionally exposed to contaminants, leading to a high false-positive rate. Additionally, methods that rely on the culturing of the infectious material is in some cases tedious due to the problem that not all microbes will grow in standard media and conditions. Additionally, low microbial loads in the sample or fastidious pathogens may show little or no growth where the conditions for the microbe are suboptimal. Further, administration of antimicrobials prior to sampling may reduce growth of the microorganism which may lead to false-negative results. Further, the culturing can take several days from sample to test result, which in some subjects leads to a worsening in the infection, and in severe cases the infection develops from mild infection to sepsis in the patient, which may ultimately lead to death of the patient if adequate treatment is notprovided in time. It is therefore essential that new methodologies are provided, which can provide an unbiased and faster diagnosis for patients suffering from infections, such that the correct treatment for the infection can be initiated as soon as possible.
[0280] Accordingly, there is an urgent need for better and faster diagnostic methods which can establish the source of infections without the bias of PCR- or current sequencing-based methods and which at the same time does not require tedious sample handling.
[0281] The present disclosure provides herein methods for identifying the source or sources of an infection in a fast and unbiased manner, where multiple species can be detected from a single sample, and wherein the false-positive rate is vastly improved over other sequencingbased methods.
[0282] It is to be understood that the bioinformatic analysis methods provided are not solely applicable to host samples, and identification of a source of infection, but may also be applicable to other areas where identification of DNA from (complex) samples may require substantial analysis. As such the methods may also be applicable for identification of a source of infestation. A source of infestation is not limited to infectious agents in animal hosts: the source may be responsible for infestation of food products, building materials such as e.g., like wood, textiles or other biobased materials. Also, in agriculture or industrial fermentation, undesired microbial activity may cause loss of yields, and a swift identification of the infesting agent may be of high relevance. The method of the first aspect of the invention can in practice be applied to any such situation and provide a rapid determination of the causative microbial agent, which in turn allow for rapid intervention.
[0283] A common problem with establishment of a source of an infection from a (patient) sample is often, that the amount of microbial short cell-free DNA molecules (cfDNA) in a sample is very limited compared to the overall cfDNA content in the sample, thereby making it much more difficult to obtain reliable sequencing data from the samples, and secondly that contaminants and common genes make accurate prediction more difficult, often leading to false negative or false positive identifications. The method for identifying the microbial sources is also referred to herein as the " SeeQFilter tool", " SeeQFilter method" or " SeeQFilter analysis".
[0284] To overcome these problems, the inventors have found that, in addition to the sample enrichment steps provided herein, the false negative and false positive identifications may be reduced in the down-stream sequence analysis by performing a number of sequence filtering steps followed by statistical analyses, which accounts for the distribution of microbial cfDNA across the respective genomes, the number of reads, and the length of the cfDNA or the mean read length obtained from sequencing.
[0285] A further benefit of the present invention is that the method has a fast throughput compared to the classical cultivation-based methods. The method herein provided allows for analysis of a sample from a patient, without having to perform additional cultivation steps before isolation and sequencing of the cultivated species, which can take several days and is prone to false negative bias, since the specific media used to cultivate the microbes might not be suitable for cultivation of the specific species, which might have a very narrow window of cultivation such as e.g., virus, overall leading to a high number of false negatives from cultivation samples.
[0286] The present disclosure therefore relates to a method for identifying the microbial source(s) of an infection in a host as the microbial source(s) of a plurality of short cell-free DNA molecules (cfDNA) in a sample; in more simple terms, the present invention facilitates that the most likely microbial species causing an infection are identified from the cfDNA profile found in the infected individual. In embodiments, the method is a computer implemented method. In some aspects the present disclosure also relates to a computer or computer system, which implements the method of the first aspect of the invention.
[0287] It will be evident to the skilled person that particular methods of the present disclosure may also be used to analyse non-cfDNA fragments, i.e., DNA extracted from cells. Such uses are also encompassed by the present disclosure.
[0288] In embodiments, the sample comprises subsets of cfDNA sequences originating from different sources and the method is configured to identify the source of each of the subset of cfDNA sequences. In some embodiments, the cfDNA sequences originating from a plethora of clinically relevant and clinically irrelevant sources and the method is configured to identify the source of each of the subset of cfDNA sequences.
[0289] As such clinically relevant sources may include pathogenic or opportunistic microbes, such as pathogenic bacteria or yeast. As such clinically irrelevant sources may include contaminant microbes or mammalian sources.
[0290] Such different sources may be divided in to clinically relevant and clinically irrelevant source. Such division may be based on the species, such as e.g., mammalian vs. microbial, but also based on the size of the cfDNA fragments, as discussed below, since microbial cfDNA and mammalian cfDNA generally exhibits different features.
[0291] Accordingly, the method may provide both the microbial source(s) of an infection in a host and any microbial source(s) of contaminants, if present. Also, where no source of infection can be provided the method may still be used to identify any microbial contaminants in thesample, such as e.g., those datasets which fails the statistical test(s). Accordingly, if no microbial source(s) of an infection is identified in the method, but a source of at least one microbial contaminant is identified in step d) of the method, then the method should identify the at least one microbial contaminant as such.
[0292] In a first step the method may comprise obtaining at least one cfDNA sequencing dataset comprising; (i) a plurality of cfDNA sequences of the cfDNA molecules, where each of the cfDNA sequences aligns to the genome of a microbial species, (ii) the nucleotide length of each of the cfDNA molecules in the plurality of cfDNA molecules, (iii) the genomic sequence of the microbial species, (iv) the genomic coordinates(s) of each alignment, (v) the number of reads of each alignment and (vi) the total number of cfDNA sequencing reads in the dataset.
[0293] In some embodiments, the dataset comprises none of features (i)- (vi). In some embodiments, the dataset comprises feature (vi). In some embodiments, the dataset comprises feature (v). In some embodiments, the dataset comprises features (v) and (vi). In some embodiments, the dataset comprises feature (ii). In some embodiments, the dataset comprises features (ii) and (vi). In some embodiments, the dataset comprises features (ii) and (v). In some embodiments, the dataset comprises features (ii), (v), and (vi). In some embodiments, the dataset comprises features (i), (iii), and (iv). In some embodiments, the dataset comprises features (i), (iii), (iv), and (vi). In some embodiments, the dataset comprises features (i), (iii), (iv), and (v). In some embodiments, the dataset comprises features (i), (iii), (iv), (v), and (vi). In some embodiments, the dataset comprises features
[0294] (i), (iii), (iv), and (ii). In some embodiments, the dataset comprises features (i), (iii), (iv),
[0295] (ii), and (vi). In some embodiments, the dataset comprises features (i), (iii), (iv), (ii), and (v). In some embodiments, the dataset comprises features (i), (ii), (iii), (iv), (v), and (vi). In preferred embodiments, the dataset comprises at least features (i), (iii) and (iv), and optionally also features (ii), (v), and (vi).
[0296] Accordingly, in embodiments, establishing a dataset containing relevant information about the identified cfDNA sequences comprise
[0297] optional, performing an alignment of each of the cfDNA sequences in the list to host genomic sequences,
[0298] removing from the list cfDNA sequences that align with the host genome, thereby establishing a filtered list,
[0299] performing an alignment of the cfDNA sequences of the filtered list to a plurality of microbial genome sequences,
[0300] grouping the cfDNA sequences in the filtered list that align with the microbial genome sequence(s) into microbial species and / or genus specific genomic set(s)
[0301] applying a last common ancestor approach to remove cfDNA sequences where the alignment of the sequence to the microbial species and / or genus specific genomic set(s) is ambiguous, thereby establishing a microbe specific cfDNA molecule list.
[0302] In the context of the present disclosure, "ambiguous" is used to indicate an outcome of a sequence alignment algorithm wherein the scoring function identifies two or more candidate alignment locations that are within a predefined threshold of each other, such that the algorithm cannot unambiguously resolve a preferred alignment. Accordingly, an ambiguous alignment may refer to an alignment scenario in which the resolution or discriminatory power of the alignment process is insufficient to uniquely map the query sequence to a specific locus in the database, due to similarity across multiple loci.
[0303] Upon performing the above steps, a dataset of the microbial cfDNA molecules may be provided which may comprise all or some of the following features:
[0304] the cfDNA sequences of the microbial cfDNA,
[0305] the nucleotide length of each of the microbial cfDNA molecules in said dataset, the number of reads of the microbial cfDNAs in said dataset,
[0306] the genomic position(s) in the microbial genome of each alignment,
[0307] the genomic sequence(s) of the microbial species and / or genus to which the specific microbial cfDNA align and
[0308] the total number of microbial cfDNA reads in the dataset.
[0309] The provided dataset may then be used in an analysis to determine at least one source of the microbial cfDNA from the sample as the identity of the pathogen(s) responsible for the infection. Alternatively, the provided dataset may be used in an analysis to determine an absence of microbial cfDNA in the sample as described herein.
[0310] The inventors also realised that upon degradation of microbes by the immune system, the microbial cfDNA fragments are degraded into fragments which are more or less evenly distributed throughout the genome of the microorganism, due to the stochastic nature of DNA degradation by the mammalian nucleases.
[0311] The analysis may hence comprise one or more of the following steps:
[0312] performing for each cfDNA sequencing dataset(s),
[0313] o a statistical analysis of the uniformity of the distribution of cfDNA alignments across the genome of the microbial species, and / or
[0314] o a statistical analysis of the distribution of the nucleotide length of the cfDNA molecules in the dataset, and / or
[0315] o an informatic analysis of the 5'- and 3'- sequences of the cfDNA molecules in the dataset,
[0316] determining the pathogen(s) responsible for the infection as the microbe(s) for which:
[0317] o the distribution of cfDNA alignments across the genome of the microbial species does not differ significantly from a Poisson distribution and / or o the identified length of the cfDNA molecules can be statistically sampled from a normal distribution defined by a cfDNA dataset from known infectious host samples and / or
[0318] o the 5'- or 3'- termini of cfDNA molecules are host-specific enzymatic degradation termini sequences.
[0319] In some embodiments, the analysis comprises the following steps:
[0320] performing for each cfDNA sequencing dataset(s),
[0321] o a statistical analysis of the uniformity of the distribution of cfDNA alignments across the genome of the microbial species, and / or
[0322] o a statistical analysis of the distribution of the nucleotide length of the cfDNA molecules in the dataset, and / or
[0323] o an informatic analysis of the 5'- and 3'- sequences of the cfDNA molecules in the dataset,
[0324] determining the pathogen(s) responsible for the infection as the microbe(s) for which:
[0325] o the distribution of cfDNA alignments across the genome of the microbial species does not differ significantly from a Poisson distribution and / or o the identified length of the cfDNA molecules can be statistically sampled from a normal distribution defined by a cfDNA dataset from known infectious host samples and / or
[0326] o the 5'- or 3'- termini of cfDNA molecules are host-specific enzymatic degradation termini sequences.
[0327] In some embodiments, the analysis comprises the following steps:
[0328] performing for each cfDNA sequencing dataset(s),
[0329] o a statistical analysis of the uniformity of the distribution of cfDNA alignments across the genome of the microbial species, and / or
[0330] o a statistical analysis of the distribution of the nucleotide length of the cfDNA molecules in the dataset,
[0331] determining the pathogen(s) responsible for the infection as the microbe(s) for which:
[0332] o the distribution of cfDNA alignments across the genome of the microbial species does not differ significantly from a Poisson distribution and / or o the identified length of the cfDNA molecules can be statistically sampled from a normal distribution defined by a cfDNA dataset from known infectious host samples.
[0333] In some embodiments, the analysis comprises one of the analysis steps. In other embodiments, the analysis comprises two or three of the analysis steps.
[0334] In embodiments, the analysis comprises a statistical analysis of the uniformity of the distribution of cfDNA alignments across the genome of the microbial species. In other embodiments, the analysis comprises a statistical analysis of the distribution of the nucleotide length of the cfDNA molecules in the dataset. In other embodiments, the analysis comprises an informatic analysis of the 5'- and 3'- sequences of the cfDNA molecules in the dataset.
[0335] In embodiments, the analysis comprises a statistical analysis of the uniformity of the distribution of cfDNA alignments across the genome, including plasmids etc., of the microbial species and a statistical analysis of the distribution of the nucleotide length of the cfDNA molecules in the dataset.
[0336] In embodiments, the analysis comprises a statistical analysis of the uniformity of the distribution of cfDNA alignments across the genome of the microbial species and an informatic analysis of the 5'- and 3'- sequences of the cfDNA molecules in the dataset.
[0337] In other embodiments, the analysis comprises a statistical analysis of the distribution of the nucleotide length of the cfDNA molecules in the dataset and an informatic analysis of the 5'-and 3'- sequences of the cfDNA molecules in the dataset.
[0338] Once performed, the analysis generally provides a determination of the source of the infection or causative pathogen.
[0339] cfDNA size and genome distribution
[0340] The size of the cfDNA from microbial sources, such as bacteria and virus, differs from the eukaryotic e.g., mammalian cfDNA, due to the histone winding of eukaryotic DNA, which defines the length of the mammalian cfDNA fragments. Since microbial DNA of bacterial orviral origin does not have the same structural definition, cfDNA from the microbes typically has a shorter length compared to mammalian cfDNA. In example, human cell-free DNA has lengths at 170 and 340 nucleotides corresponding to the length of DNA wrapped around 1 or 2 histones, respectively, which protects the DNA from nuclease degradation. In contrast, most microbial cell-free DNA is short and fragmented, with lengths at 50-80 nucleotides.
[0341] Accordingly, in some instances eukaryotic cfDNA in the sample may be filtered out based on the sequencing data, simply by adding a read length filter, filtering out reads of more than e.g., 300 nucleotides or more than 250 nucleotides, thereby reducing the computational requirement for alignments towards the genomes as specified herein. Since mammalian cfDNA has discrete preferred sizes due to the winding around the histones, the cfDNA from a sample may also be sorted based on a statistical test of the overall size distributions in the sample, in example, it may be expected that a sample comprising cfDNA will give rise to a multimodal size distribution once analysed, with several distinct peaks, and several overlapping normal distributions. Accordingly, cfDNA of mammalian origin and cfDNA of microbial origin can be separated based on a statistical test of size distribution. As mentioned herein the larger DNA fragments are predominantly removed prior to the sequencing analysis, as this improves the speed and cost of the algorithm significantly.
[0342] The size distribution of cfDNA fragments in a sample may be evaluated by several methods known to the skilled person. The herein provided methods rely on statistical analysis of the size distribution, but other alternative methods for analysing distributions are well known to the skilled person.
[0343] In embodiments, the size distribution of cfDNA fragments from a given species is evaluated by fitting the size distribution data to a normal distribution or a log-normal distribution, preferably a normal distribution.
[0344] As the inventors herein realised that upon degradation of microbes by the immune system, the microbial cfDNA fragments are degraded into fragments with particular (small) sizes, which are specific to the taxonomical species of the microbe, the inventors realised that this could be used to improve the false-positive rate of detection, such that less false-positive identification are made.
[0345] To be more specific, the inventors realised that the size of the cfDNA, which results from degradation by the immune system of microbial cells, falls within a normal distribution, while cfDNA which results from contaminants in the sample preparation, often has a more heterogenous size distribution.
[0346] Therefore, the realization by the inventors that this difference could be quantified upon analysing the read length of the sequencing hits from a sample, allowed for an improved filtering of the data, which provides a more robust identification of the actual source of an infection, that have a reduced false-positive rate, compared to current state-of-the-art methods.
[0347] Additionally, the inventors realised that for true positives, there exists a general uniform distribution of cfDNA throughout the genome of the microbial species, due to the unspecific degradation microbial DNA by the mammalian nucleases. This unspecific cleavage of the microbial DNA leads to an approximately uniform distribution of cfDNA alignments to the genome of the microbes for true-positives while for contaminants and false-positives, a more heterogenous distribution is seen.
[0348] Accordingly, by performing a statistical analysis of the distribution of the cfDNA alignments across the genome of the initial hit i.e., a species of a microbe, the inventors were able to show that for true-positives the distribution generally followed a Poisson distribution, while false-positives tended to have a more uneven cfDNA distribution throughout the genome of the species.
[0349] This enables a more robust identification of the source of infections in patients suffering from a microbial infection, which ultimately can lead to a more robust diagnosis and treatment of the patient. Genomes as defined herein includes microbial genomes such as viral, bacterial, fungal, parasites and potential specie specific plasmids if applicable.
[0350] Accordingly, in embodiments, the method further comprises a step of performing a statistical analysis of the uniformity of the distribution of cfDNA alignments across the genome of the microbial species, and / or a statistical analysis of the distribution of the nucleotide length of the cfDNA molecules in the dataset, for each cfDNA sequencing dataset. In some embodiments, the method comprises a step of performing a statistical analysis of the uniformity of the distribution of cfDNA alignments across the genome of the microbial species for each cfDNA sequencing dataset. In other embodiments, the method comprises a step of performing a statistical analysis of the distribution of the nucleotide length of the cfDNA molecules in the dataset for each cfDNA sequencing dataset. In further embodiments, the method comprises a step of performing a statistical analysis of the uniformity of the distribution of cfDNA alignments across the genome of the microbial species and a statistical analysis of the distribution of the nucleotide length of the cfDNA molecules in the dataset for each cfDNA sequencing dataset.
[0351] Following the statistical analysis, the method generally comprises identifying the microbial source of the infection as the microbe(s) for which the distribution of cfDNA alignments across the genome of the microbial species is uniform. The uniformity may be addressed in several ways, such as e.g., deviation from the relevant distributions, such as e.g., the Poisson distribution.
[0352] In embodiments, the method may comprise identifying the microbial source of the infection as the microbe(s) for which the distribution of cfDNA alignments across the genome of the microbial species does not deviate significantly from a Poisson distribution and / or the identified length of the cfDNA molecules can be statistically sampled from a normal distribution defined by a cfDNA dataset from known infectious host samples distribution. Accordingly, in embodiments, the method comprises identifying the microbial source of the infection as the microbe(s) for which the distribution of cfDNA alignments across the genome of the microbial species does not deviate significantly from a Poisson distribution.
[0353] In other embodiments, the method comprises identifying the microbial source of the infection as the microbe(s) for which the identified length of the cfDNA molecules can be statistically sampled from a normal distribution or log-normal distribution, preferably normal distribution, defined by a cfDNA dataset from known infectious host samples. Accordingly, the length of the cfDNA molecules can be subjected to a normality test, where they are tested against a normal distribution defined by a known set of species specific cfDNA molecules. Such test may be performed for the microbial cfDNA or the mammalian cfDNA or both. In examples, a sample may comprise a plurality of cfDNAs originating from a range of microbes. For some organisms the expected distribution of cfDNA fragments is known, and will when analysed follow a specific normal distribution, with a species-specific shape of the distribution.
[0354] Accordingly, a set of cfDNAs aligning to a particular microbial species may fall outside the expected normal distribution for the identified species, potentially suggesting that the assignment of the cfDNA to that species is faulty, and that the specific cfDNA molecule is wrongfully assigned to a species with which genomic sequence it aligns. Accordingly, in some cases, the microbial cfDNA will align to multiple species, and there may be a need to sample the length of the cfDNA(s) against multiple species to find the species that has the most likely match for the specific cfDNA, thereby enhancing the likelihood of a true-positive species assignment of the individual cfDNA molecules.
[0355] In embodiments, the method comprises identifying the microbial source of the infection as the microbe(s) for which the distribution of cfDNA alignments across the genome of the microbial species does not deviate significantly from a Poisson distribution and the length of the cfDNA molecules in the dataset does not deviate significantly from a normal- or lognormal distribution.
[0356] Deviation from a Poisson distribution of cfDNA alignments across the genome of the microbial species may be tested in a plethora of ways, known to the skilled person such as a chi-square goodness-of-fit test, a Kolmogorov-Smirnov, a Cramer-von Mises test or an Anderson-Darling test adapted for Poisson distributions.
[0357] Generally, to perform the analysis for the Poisson distribution of the alignments, the genome(s) is / are segmented, and the Poisson distribution is used to address the uniformity of the genomic coverage of the sequencing data. In embodiments, the genome of the microbial species is segmented into at least two segments, and the number of reads of the microbe specific cfDNAs in each segment is counted (k). In additional embodiments, the statistical analysis of the uniformity of the distribution of cfDNA alignments across the genome of the microbial species is based on the counted alignments in each segment. As such the genome may be segmented into any number of segments, depending on the alignments in the dataset. Generally, segmentation of the genome should be performed after the alignments, such that the segments are constructed to have the minimal effect on the number of reads being split between individual segments. In embodiments, the genome is segmented into at least 10 segments, such as at least 20, 30, 40, 50, 60 70, 80, 90, 100, 110, 120, 130, 140, 150 segments or such as at least 160 segments. In other embodiments, the genome segmented into segments comprises 1 kbp-500 kbp sections, such as 10-100 kbp sections, such as 20-80 kbp sections, such as 30-70 kbp sections, or such as 40-80 kbp sections. In some embodiments, the genome is segmented into segments defied by gene coding and / or non-coding regions. In embodiments, the gene coding regions comprises one or more gene clusters.
[0358] Poisson distribution tests
[0359] Accordingly, the data used to analyse the distribution of the alignments across the genome relies on the cfDNA sequence as obtained from sequencing of the cfDNA in the sample, and the genomic sequence of at least one microbial species, often retrieved from a database.
[0360] In some embodiments, the statistical test comprises the chi-square goodness-of-fit test applied to observed frequencies and expected Poisson frequencies. In some embodiments, the statistical test comprises the Kolmogorov-Smirnov test with a Poisson distribution as the reference. In some embodiments, the statistical test comprises a bootstrap-based goodness-of-fit test comparing empirical counts to a simulated Poisson distribution. Additional statistical analyses are known to the skilled person, as may e.g., include, but is not limited to regression-based tests, probability plot correlation coefficient (PPCC) tests, rootograms, or nonparametric tests.
[0361] cfDNA fragment size distribution tests
[0362] As disclosed herein, where the cfDNA originates from a microbe, the specific microbe often gives rise to a specific degradation pattern, meaning that the cfDNA fragments from that microbe often has a defined length of the cfDNA fragments and a set of specific cfDNA fragments. While the latter may be used in a sequence-based analysis as described above, the length of the cfDNA fragments, originating from the host often follows a normal distribution defined by the specific host. As such, the lengths of the cfDNA fragments in a sample is often a mixture of the host defined distribution and the distribution(s) of the cfDNA originating from any other source e.g., microbes. The normal distribution of the cfDNA may be obtained from a database containing cfDNA fragments of particular microbes, where each microbe has a specific distribution of sizes. Alternatively, a database of cfDNA fragments from different microbes is prepared in order to obtain the distribution of cfDNA fragments from that microbe. The cfDNA fragments in a sample may then be sampled against the normal distribution of the database provided distributions to aid in the identification of the microbe. For instance, due to genome overlaps between microbes there may be situations where an alignment identifies a source, but the cfDNA fragment pattern / distribution does not match that of the microbe, as provided by the database. In such a situation, further sequence analysis may be needed to establish the true-positive source of the infection.
[0363] In an example, the microbe is first identified by alignment of the cfDNA fragment towards the genome of the microbe, and the known cfDNA fragment distribution of that microbe is different from that of the cfDNA fragment distribution in the sample. In such a case it is likely that the microbe is misidentified, and further sequence alignments towards different species may be required to arrive at the true source of the cfDNA. Such analysis may rely on a full analysis of the normal distribution of the cfDNA fragments in the sample, or it may be based on sampling of the cfDNA fragments in the sample.
[0364] In embodiments, the length of the cfDNA molecules originating from the sample are statistically sampled from a normal- or log-normal distribution defined by a dataset of known cfDNA molecules from microbes. Such sampling may be used to identify misidentified microbes.
[0365] Statistical analysis of the sampling of a normal distribution may be tested in a plethora of ways, known to the skilled person such as a t-test, an Anova test, a Shapiro-Wilk test, a Kolmogorov-Smirnov test, an Anderson-Darling test etc.
[0366] In some embodiments, the statistical test for normal or log-normal distribution of lengths of cfDNA comprises a Shapiro-Wilk test. In some embodiments, the statistical test for normal orlog-normal distribution of lengths of cfDNA comprises a Kolmogorov-Smirnov test using a normal distribution as the reference.
[0367] Other alternative statistical tests will be evident to the skilled person, such as but not limited to an Anderson-Darling test, a Lilliefors test, an D'Agostino-Pearson omnibus test, a Jarque-Bera test, tests based on skewness and kurtosis, or bootstrap-based goodness-of-fit normality tests. In some embodiments, the statistical test for normal or log-normal distribution of lengths of cfDNA comprises a Bayesian model comparison between a Gaussian and an alternative non-parametric distribution. In some embodiments, the statistical test for normal or log-normal distribution of lengths of cfDNA comprises a measure of deviation from linearity in a normal probability plot.
[0368] The statistical test for normal or log-normal distribution of lengths of cfDNA may also include the use of a machine learning classifier trained to detect non-normality in distributions.
[0369] Terminal sequence identity and non-host mediated DNA degradation
[0370] It is well known that foreign DNA is processed and degraded by specific enzymes such as e.g., nucleases, responsible for the degradation of foreign DNA in e.g., the blood of mammals. Examples of sequence specific nucleases which degradation products i.e., DNA fragments may be identified by their 5'-end or 3'-end sequence are e.g., DFFB (DNA fragmentation factor subunit beta), DNASE1 (Deoxyribonuclease 1) and DNASE1L3 (Deoxyribonuclease I Like 3). Examples of nucleases are e.g., described in Han et al 2021 (Han, Diana S. C. et al. " The Nexus of cfDNA and Nuclease Biology", Trends in Genetics, Volume 37, Issue 8, 758 - 770), Wang et al 2023 (Wang, G., et al., " Fragment Ends of Circulating Microbial DNA as Signatures for Pathogen Detection in Sepsis", Clinical Chemistry, Volume 69, Issue 2, February 2023, Pages 189-201) and Zhou et al 2022 (Zhou., Q, " Epigenetic analysis of cell-free DNA by fragmentomic profiling", Proc. Natl. Acad. Sci.
[0371] U. S. A. 119 (44) e2209852119), all incorporated herein in their entirety. Accordingly, upon degradation in a host, the cfDNA will have a host-specific enzymatic degradation pattern, which entails that the 5'- or 3'- termini of cfDNA molecules primarily presents host-specific enzymatic degradation termini sequences.
[0372] The inventors herein realised that upon degradation of DNA by nucleases in the host, the DNA will be cut in distinctive positions, since the nucleases of the host often has a particular consensus sequence as to which they are specific, to have preference to.
[0373] In bacteria, such nucleases are e.g., the restriction enzymes commonly used in cloning protocols, to generate a series of DNA fragments with defined 5'-ends and 3'-ends, to be ligated following the degradation with the restriction enzymes. Much in the same mannermammals have similar enzymes, nucleases, which degrade both native and foreign DNA into fragments. The mechanism behind the DNA degradation is for the cell to avoid immunological responses towards the longer DNA fragments, which could otherwise have critical consequences for the host.
[0374] Generally, the difference between the restriction enzymes of bacteria and the mammalian nucleases, is generally that nucleases are as such less sequence specific, and more specific towards structural elements in the DNA, such as hairpins, or are recruited to specific sites by other proteins rather than binding strictly to precise DNA sequences. Other means of recognition may e.g., be unmethylated CpG motifs, lack of histone association and abnormal or uncommon mammalian DNA structures e.g., long dsDNA, hairpins, loop etc.). Examples of mammalian nucleases that are involved in degradation of foreign DNA are e.g., TREX1, EndoG and DNase II.
[0375] cfDNA fragments originating from host mediated DNA degradation may be identified from the terminal sequences of the cfDNA, as compared to DNA which is degraded e.g., by exposure to UV-A and UV-B light, hydrolysis, oxidative damage, radiation, heat exposure, and exposure to pH extremes such as e.g., pH below 2 or pH above 12. Often such physical / chemical degradation of DNA is of a more random nature than host mediated DNA degradation. Accordingly, cfDNA which originates from host mediated degradation and cfDNA which originates from physical / chemical degradation may be differentiated, and the cfDNA fragments resulting from physical / chemical degradation, such as e.g., samples handling, may be filtered out once identified as such.
[0376] The terminals which may be recognised as originating from degradation by the host are normally between 2 nucleotides and 10 nucleotides long.
[0377] Accordingly, by performing a bioinformatic analysis of the 5'- and / or 3'-end sequences of the of the cfDNA molecules may be used to establish and identify source microbial cfDNA originating from the sample as such, and not from contaminant sources.
[0378] In some embodiments, the termini sequence of cfDNA molecules primarily display hostspecific enzymatic degradation termini sequences. In additional embodiments, the cfDNA predominantly have a host-specific enzymatic degradation pattern.
[0379] Last common ancestor analysis
[0380] Generally, a Last Common Ancestor (LCA) analysis is a method used to determine the most recent shared ancestor of a group of organisms, genes, or sequences. It's often applied in phylogenetics, taxonomy, and metagenomics to help understand evolutionary relationships.
[0381] Generally, an LCA analysis is based on a series of alignments, where multiple matches are found for one or more of the alignments (taxonomic hits). The LCA analysis is then performed to establish the most specific taxonomic level common to all hits.
[0382] An LCA analysis often follows a series of analysis steps relying on the taxonomy of the hits, and may include,
[0383] one or more filtering steps e.g., comprising
[0384] o a first round of filtering of the hits such that only the top N hits or those above a specific score threshold is kept in the list, such a score may be based on e.g., % identity or E-values,
[0385] o a second round of filtering, e.g., to remove low-complexity hits or ambiguous taxa,
[0386] retrieval of the taxonomic lineages for each remaining hit, such as e.g., species, genus, family, order etc.,
[0387] Traverse the taxonomic lineages and find the lowest (most specific) level that is common to all valid hits and
[0388] Identify the lowest common taxonomic lineage as the last common ancestor, such as e.g., the read belongs to the Bacteroidetes phylum.
[0389] The benefit performing an LCA analysis is that it avoids overclassification of the hits, where there are multiple likely outcomes of the alignments. In example, a dataset of cfDNA sequences may be obtained from a sample from a host and alignment shows that the sequences match more than one known species, such as e.g., Escherichia coii, Shigella flexneri and Salmonella enterica. In such a situation, all three might be equally likely when analysed solely based on their alignment. Accordingly, an LCA analysis will in such a case result in the family Enterobacteriaceae, which is the last common ancestor, accordingly, since the alignment as such does not provide the specific species solely based on the identified sequence.
[0390] Accordingly, combining the other analyses disclosed herein with an LCA analysis allows for an additional step of control for overclassification, thereby strengthening the identification of true positives, by reducing overclassification of the identified hits.
[0391] Accordingly, following the LCA analysis, the cfDNA sequences where the alignment of the sequence to the microbial species and / or genus specific genomic set(s) is ambiguous may be removed from the list, since these are not likely to have been correctly assigned following the initial alignment.
[0392] Accordingly, in embodiments, the preparation of the cfDNA dataset as herein described method may comprise the steps of
[0393] aligning the sequences of the filtered list with a plurality of microbial genome sequences,
[0394] grouping all the microbial sequences aligning to the cfDNA sequences in the filtered list on a genus and / or species level,
[0395] performing a last common ancestor analysis, by identifying all near-best alignments of the cfDNA sequences in the groups and
[0396] removing from the list the cfDNA sequences where the alignment of the sequence to the microbial species and / or genus specific genomic set(s) is ambiguous, thereby establishing a microbe specific cfDNA list.
[0397] In some embodiments, if less than half of near-best alignments, as assessed by one or more of the criteria outlined above, of a read map is to the same genus, the classification of this cfDNA sequence may be disregarded to avoid misclassifications due to shared genomic elements and wrongly segmented genome contigs. Accordingly, in some embodiments, only assignment in which more than half of the read map is assigned to the same genus, are considered reliable.
[0398] A method comprising one or more analyses
[0399] The method of the present disclosure may comprise one or more of the above-mentioned analyses, in order to identify the source of the infection. The inventors found that the optimal results were achieved upon combination of the different analyses, but each of the analysis steps in the method may be used to identify the source microbes and / or contaminant microbes in the sample.
[0400] The method may further comprise a step of identifying contaminants in the sample, originating e.g., from the sample preparation procedures. Accordingly, the method may further comprise a step of identifying microbial contaminant source(s). In embodiments, microbial contaminant source(s) are identified as the microbes for which the distribution of cfDNA alignments across the genome of the microbial species deviates significantly from a Poisson distribution and / or for which the length of the cfDNA molecules in the dataset may not be statistically sampled from a normal- or log-normal distribution defined by a dataset of known cfDNA molecules from microbes and / or for which the 5'- or 3'end terminals of the cfDNA molecules in the dataset suggests that the cfDNA fragments are not a result of host mediated DNA degradation.
[0401] In embodiments, contaminants in the sample are identified where the length of the cfDNA molecules in the dataset are incompatible with statistical sampling from a normal distribution or log-normal distribution defined by a dataset of known cfDNA molecules from microbes. Inembodiments, contaminants in the sample are identified where the length of the cfDNA molecules in the dataset does not allow for statistical sampling from a normal distribution or log-normal distribution defined by a dataset of known cfDNA molecules from microbes. In embodiments, contaminants in the sample are identified where the length of the cfDNA molecules in the dataset does not enable a statistical sampling from a normal distribution or log-normal distribution defined by a dataset of known cfDNA molecules from microbes. In embodiments, contaminants in the sample are identified where the length of the cfDNA molecules in the dataset cannot be statically sampled from a normal distribution or lognormal distribution defined by a dataset of known cfDNA molecules from microbes.
[0402] In embodiments, microbial contaminant source(s) are identified as the microbes for which the distribution of cfDNA alignments across the genome of the microbial species deviates significantly from a Poisson distribution and for which the length of the cfDNA molecules in the dataset may not be statistically sampled from a normal- or log-normal distribution defined by a dataset of known cfDNA molecules from microbes.
[0403] In a preferred embodiment, the present invention relates to a method for identifying the microbial source(s) of an infection in a host as the microbial source(s) of a plurality of short cell-free DNA molecules (cfDNA) in a sample, the method comprising
[0404] a) obtaining at least one cfDNA sequencing dataset comprising
[0405] i. a plurality of cfDNA sequences of microbe specific cfDNA,
[0406] II. the nucleotide length of each of the cfDNA molecules in the plurality of cfDNA molecules,
[0407] ill. the genomic sequence of the microbial species,
[0408] iv. the genomic position(s) of each alignment,
[0409] v. the number of reads of each alignment and
[0410] vi. the total number of cfDNA sequencing reads in the dataset,
[0411] b) performing for each cfDNA sequencing dataset
[0412] i. a statistical analysis of the uniformity of the distribution of cfDNA alignments across the genome of the microbial species, and
[0413] II. a statistical analysis of the distribution of the nucleotide length of the cfDNA molecules in the dataset,
[0414] c) identifying the microbial source of the infection as the microbe(s) for which;
[0415] i. the distribution of cfDNA alignments across the genome of the microbial species does not differ significantly from a Poisson distribution and
[0416] II. the identified length of the cfDNA molecules can be statistically sampled from a normal distribution defined by a cfDNA dataset from known infectious host samples,
[0417] d) identifying microbial contaminant source(s) as the microbes for which the distribution of cfDNA alignments across the genome of the microbial species deviates significantly from a Poisson distribution and for which the distribution of lengths of the cfDNA molecules are incompatible with statistical sampling from a normal distribution or log-normal distribution defined by a cfDNA dataset from known infectious host samples and the termini sequence of cfDNA molecules are non-host-specific enzymatic degradation termini sequences.
[0418] In a preferred embodiment, the present disclosure relates to a method for identifying the microbial source(s) of an infection in a host as the microbial source(s) of a plurality of short cell-free DNA molecules (cfDNA) in a sample, the method comprising
[0419] a) obtaining at least one cfDNA sequencing dataset comprising
[0420] i. a plurality of cfDNA sequences of microbe specific cfDNA,
[0421] II. the nucleotide length of each of the cfDNA molecules in the plurality of cfDNA molecules,
[0422] ill. the genomic sequence of the microbial species,
[0423] iv. the genomic position(s) of each alignment,
[0424] v. the number of reads of each alignment and
[0425] vi. the total number of cfDNA sequencing reads in the dataset, b) performing for each cfDNA sequencing dataset
[0426] i. a statistical analysis of the uniformity of the distribution of cfDNA alignments across the genome of the microbial species, II. a statistical analysis of the distribution of the nucleotide length of the cfDNA molecules in the dataset,
[0427] c) identifying the microbial source of the infection as the microbe(s) for which;
[0428] i. the distribution of cfDNA alignments across the genome of the microbial species does not differ significantly from a Poisson distribution and
[0429] II. the identified length of the cfDNA molecules can be statistically sampled from a normal distribution defined by a cfDNA dataset from known infectious host samples.
[0430] In embodiments, the method further comprises a step d) comprising identifying microbial contaminant source(s) as the microbes for which the distribution of cfDNA alignments across the genome of the microbial species deviates significantly from a Poisson distribution and for which the distribution of lengths of the cfDNA molecules are incompatible with statistical sampling from a normal distribution or log-normal distribution defined by a cfDNA dataset from known infectious host samples.
[0431] In some embodiments the bioinformatic analysis is performed using the of the 5'-sequence of the of the cfDNA molecules. In other embodiments, the bioinformatic analysis is performed on the of the 3'-sequence of the of the cfDNA molecules. In other embodiments, the bioinformatic analysis is performed on the of the 5'-end and 3'-end sequence of the of the cfDNA molecules. In additional embodiments, the bioinformatic analysis is performed on the at least 3 most terminal residues, such as the 3, 4, 5, 6, 7, 8, 9, 10, 15 or such as the 20 most terminal residues of either or both terminals of the cfDNA sequence.
[0432] In some embodiments, the disclosure relates to a method for identifying the microbial source(s) of an infection in a host as the microbial source(s) of a plurality of short cell-free DNA molecules (cfDNA) in a sample, the method comprising
[0433] a) obtaining at least one cfDNA sequencing dataset comprising
[0434] i. a plurality of cfDNA sequences of the cfDNA molecules, where each of the cfDNA sequences aligns to the genome of a microbial species, II. the genomic sequence of the microbial species,
[0435] ill. the genomic position(s) of each alignment,
[0436] b) performing a statistical analysis of the uniformity of the distribution of cfDNA alignments across the genome of the microbial species for each cfDNA sequencing dataset and
[0437] c) identifying the microbial source of the infection as the microbe(s) for which the distribution of cfDNA alignments across the genome of the microbial species does not deviate significantly from a Poisson distribution.
[0438] In embodiments, the method further comprises a step d) comprising identifying microbial contaminant source(s) as the microbes for which the distribution of cfDNA alignments across the genome of the microbial species deviates significantly from a Poisson distribution.
[0439] In some embodiments, the disclosure relates to a method for identifying the microbial source(s) of an infection in a host as the microbial source(s) of a plurality of short cell-free DNA molecules (cfDNA) in a sample, the method comprising
[0440] a) obtaining at least one cfDNA sequencing dataset comprising the nucleotide length of each of the cfDNA molecules in the plurality of cfDNA molecules,
[0441] b) performing a statistical analysis of the distribution of the nucleotide length of the cfDNA molecules in the dataset for each cfDNA sequencing dataset and c) identifying the microbial source of the infection as the microbe(s) for which the length of the cfDNA molecules in the dataset can be statistically sampled from a normal- or log-normal distribution defined by a dataset of known cfDNA molecules from microbes.
[0442] In embodiments, the method further comprises a step d) comprising identifying microbial contaminant source(s) as the microbes for which the length of the cfDNA molecules in the dataset may not be statistically sampled from a normal- or log-normal distribution defined by a dataset of known cfDNA molecules from microbes.
[0443] In some embodiments, the disclosure relates to a method for identifying the microbial source(s) of an infection in a host as the microbial source(s) of a plurality of short cell-free DNA molecules (cfDNA) in a sample, the method comprising
[0444] a) obtaining at least one cfDNA sequencing dataset comprising a plurality of cfDNA sequences of the cfDNA molecules, where each of the cfDNA sequences aligns to the genome of a microbial species,
[0445] b) performing a bioinformatic analysis of the 5'- and / or 3'-end sequences of the of the cfDNA molecules for each cfDNA sequencing dataset and
[0446] c) identifying the microbial source of the infection as the microbe(s) for which the 5'end and 3'end terminals of the cfDNA originating from the sample are defined by a host enzymatic degradation pattern.
[0447] In embodiments, the method further comprises a step d) comprising identifying microbial contaminant source(s) as the microbes for which the 5'- or 3'end terminals of the cfDNA molecules in the dataset are primarily described by non-enzymatic degradation processes or non-host specific enzymatic activity.
[0448] In some embodiments, the disclosure relates to a method for identifying the microbial source(s) of an infection in a host as the microbial source(s) of a plurality of short cell-free DNA molecules (cfDNA) in a sample, the method comprising
[0449] a) obtaining at least one cfDNA sequencing dataset comprising
[0450] i) a plurality of cfDNA sequences of the cfDNA molecules, where each of the cfDNA sequences aligns to the genome of a microbial species, ii) the nucleotide length of each of the cfDNA molecules in the plurality of cfDNA molecules,
[0451] iii) the genomic sequence of the microbial species and iv) the genomic position(s) of each alignment,
[0452] b) performing for each cfDNA sequencing dataset
[0453] i) a statistical analysis of the uniformity of the distribution of cfDNA alignments across the genome of the microbial species and ii) a statistical analysis of the distribution of the nucleotide length of the cfDNA molecules in the dataset,
[0454] c) identifying the microbial source of the infection as the microbe(s) for which;
[0455] i) the distribution of cfDNA alignments across the genome of the microbial species does not deviate significantly from a Poisson distribution and
[0456] ii) the length of the cfDNA molecules in the dataset can be statistically sampled from a normal- or log-normal distribution defined by a dataset of known cfDNA molecules from microbes.
[0457] In embodiments, the method further comprises a step d) comprising identifying microbial contaminant source(s) as the microbes for which
[0458] i) the distribution of cfDNA alignments across the genome of the microbial species deviates significantly from a Poisson distribution and
[0459] ii) the length of the cfDNA molecules in the dataset may not be statistically sampled from a normal- or log-normal distribution defined by a dataset of known cfDNA molecules from microbes.
[0460] In some embodiments, the disclosure relates to a method for identifying the microbial source(s) of an infection in a host as the microbial source(s) of a plurality of short cell-free DNA molecules (cfDNA) in a sample, the method comprising
[0461] a) obtaining at least one cfDNA sequencing dataset comprising
[0462] i) a plurality of cfDNA sequences of the cfDNA molecules, where each of the cfDNA sequences aligns to the genome of a microbial species,
[0463] ii) the genomic sequence of the microbial species, iii) the genomic position(s) of each alignment,
[0464] b) performing for each cfDNA sequencing dataset
[0465] i) a statistical analysis of the uniformity of the distribution of cfDNA alignments across the genome of the microbial species, and c) identifying the microbial source of the infection as the microbe(s) for which;
[0466] i) the distribution of cfDNA alignments across the genome of the microbial species does not deviate significantly from a Poisson distribution.
[0467] In embodiments, the method further comprises a step d) comprising identifying microbial contaminant source(s) as the microbes for which
[0468] i) the distribution of cfDNA alignments across the genome of the microbial species deviates significantly from a Poisson distribution.
[0469] In some embodiments, the disclosure relates to a method for identifying the microbial source(s) of an infection in a host as the microbial source(s) of a plurality of short cell-free DNA molecules (cfDNA) in a sample, the method comprising
[0470] a) obtaining at least one cfDNA sequencing dataset comprising
[0471] i) a plurality of cfDNA sequences of the cfDNA molecules, where each of the cfDNA sequences aligns to the genome of a microbial species and ii) the nucleotide length of each of the cfDNA molecules in the plurality of cfDNA molecules,
[0472] b) performing for each cfDNA sequencing dataset
[0473] i) a statistical analysis of the distribution of the nucleotide length of the cfDNA molecules in the dataset, and
[0474] c) identifying the microbial source of the infection as the microbe(s) for which; i) the length of the cfDNA molecules in the dataset can be statistically sampled from a normal- or log-normal distribution defined by a dataset of known cfDNA molecules from microbes.
[0475] In embodiments, the method further comprises a step d) comprising identifying microbial contaminant source(s) as the microbes for which
[0476] i) the length of the cfDNA molecules in the dataset may not be statistically sampled from a normal- or log-normal distribution defined by a dataset of known cfDNA molecules from microbes.
[0477] Number of reads and false positives
[0478] Since sequencing-based methods are highly sensitive, very small amounts of DNA may give rise to a read. Accordingly, besides the above-mentioned tests for significance etc., the number of reads can also be used as a qualifier for inclusion of data in the dataset, in order to avoid that species with a very low number of cfDNA sequence reads are given the same importance as highly abundant data. The lower limit may be static or dynamic depending on the number of total reads in the sample. One way to filter out false-positives may be by filtering the hits such that each of the cfDNA sequencing datasets comprises minimum number of total sequencing reads accredited to that source, such as e.g., less than 10 sequencing reads. Alternatively, the filtering is done by combining the statistical tests provided herein with a more static qualifier such as a given number of reads or a lowest % of total reads. Another alternative is to use a false discovery rate analysis, such as e.g., a Benjamini-Hochberg (BH) procedure (see e.g., Benjamini, Yoav, and Yosef Hochberg.
[0479] " Controlling the False Discovery Rate: A Practical and Powerful Approach to Multiple Testing." Journal of the Royal Statistical Society. Series B (Methodological), vol. 57, no. 1, 1995, pp. 289-300. JSTOR').
[0480] A common issue with identification the source of short DNA fragments, such as the microbial cfDNA fragments discussed herein is often a low signal-to-noise ratio, due to the low abundance of the microbial cfDNA molecules compared to remaining DNA in the sample. Thesignal-to-noise ratio is influenced by several factors, both relating to the sample, but also relating to the sample handling and equipment used. To reduce the noise in the sample, originating from sample handling, difference in equipment etc., the method may comprise a step of background subtraction. As an example, the background subtraction may be done by preparing and sequencing a series of samples preferably from health donors, in order to obtain a location specific and / or temporal background, including any potential location and / or temporal specific contaminants. For instance, samples may be obtained from e.g., 5 healthy donors followed by preparation of the sample, sequencing and analysis as disclosed herein, whereafter the results are used to establish an average background for that specific site. Preferably, the background samples are prepared, sequenced and analysed in parallel to the samples suspected to contain the target pathogens. Accordingly, in embodiments, the method disclosed herein comprises a step of establishing set of location and / or temporal specific background contaminant sequences to be subtracted from the cfDNA sequencing dataset. As used herein "location" refer to the physical location at which the sample is obtained, preparation of the sample is performed or where the sequencing and analysis of the sample is performed. As used herein "temporal" refer to the time of the analysis, where e.g., a temporal specific background refers to a background sample which is run before or after the patient sample, or which is run in parallel with the patient sample, or within a specific timeframe from the analysis of the patient sample. This site may include a laboratory, clinical facility, point-of-care location, or any setting equipped to conduct molecular or genetic diagnostic procedures, including sample extraction, sample preparation, amplification, sequencing or analysis. The site of analysis may be remote from the site of sample collection, and the background samples may be used to account for such location differences.
[0481] As the method identifies a source of an infection, the method may also comprise providing at least one suggestion for a treatment to which a subject is eligible based on the identity / identities of the microbial source(s) of the infection. The method may also provide more than one suggestion for treatment. This will often be the case in polymicrobial infections, where there is not a single treatment option covering treatment of all the sources of the infection.
[0482] cfDNA sequencing dataset
[0483] Depending on the specific analysis, the dataset used for the method may comprise a number of elements, as specified herein.
[0484] Before performing the method for identification of the source(s) of the infection, the dataset is generated by a number of steps, which filters the sequencing hits such that the analysis isperformed on only the relevant cfDNA alignments. By performing a pre-filtering step of all the sequencing data, the memory requirement for the analysis may be reduced.
[0485] As such the dataset may comprise the combination of features (i)-(vi) as discussed in the above sections.
[0486] Accordingly, the cfDNA sequencing dataset may be prepared, generated or obtained by providing a list comprising a plurality of cfDNA sequences,
[0487] aligning each one of the cfDNA sequences in the list to host genomic sequences, removing from the list cfDNA sequences that align to the host genome thereby establishing a filtered list,
[0488] aligning the sequences of the filtered list with a plurality of microbial genome sequences,
[0489] grouping all the microbial sequences aligning to the cfDNA sequences in the filtered list on a genus an / or species level,
[0490] performing a last common ancestor analysis, by identifying all near-best alignments of the cfDNA sequences in the groups,
[0491] removing from the list the cfDNA sequences where the alignment of the sequence to the microbial species and / or genus specific genomic set(s) is ambiguous, thereby establishing a microbe specific cfDNA molecule list, and
[0492] based on the microbe specific cfDNA list, provide at least one cfDNA sequencing dataset comprising
[0493] cfDNA sequences of the microbe specific cfDNA list,
[0494] the nucleotide length of each of the cfDNA molecules in said list, the number of reads of the microbe specific cfDNAs in said list, the genomic position(s) in the microbial genome of each alignment, the full genomic sequence(s) of microbial species to which the cfDNA sequences preferentially aligns to and
[0495] the total number of cfDNA reads in the dataset.
[0496] It is to be understood that the dataset may comprise all of the above mentioned parts, or only a sub-set of the above mentioned parts, depending on the analysis to be performed in the method as disclosed herein, such as e.g., where the method compares the use of alignments, the dataset should at least comprise the sequence of the cfDNA to be aligned and hit(s) to which it aligns. Such logics may naturally be applied for selection of the relevant requirements to the dataset, and hence sub-selection from the preparation of the dataset may be performed to obtain only the required data for the analysis.
[0497] As used herein "align to" refers to the arrangement of sequence elements by application of a recognized computational method, such as local or global alignment algorithms, whereby regions of significant similarity are identified and mapped. Such an alignment accommodates the introduction of insertions, deletions, or substitutions as necessary to optimize the correspondence between the sequences. As such, sequences are considered to align significantly when they exhibit a Levenshtein distance of less than 20% relative to the length of the compared sequence e.g., the cfDNA. Additionally, the similarity between the sequence should be of sufficient magnitude to enable reliable discrimination of a sequence originating from a particular strain, individual, or variant within a species from sequences derived from organisms of different species, preferably independent of an absolute percentage identity.
[0498] As such "aligning to" may also comprise that an cfDNA has a sequence which is identical or substantially identical to a microbial or human DNA sequence.
[0499] In the context of the present disclosure, "ambiguous" is used to indicate an outcome of a sequence alignment algorithm wherein the scoring function identifies two or more candidate alignment locations that are within a predefined threshold of each other, such that the algorithm cannot unambiguously resolve a preferred alignment. Accordingly, an ambiguous alignment may refer to an alignment scenario in which the resolution or discriminatory power of the alignment process is insufficient to uniquely map the query sequence to a specific locus in the database, due to similarity across multiple loci.
[0500] In embodiments, cfDNA sequences that align to the human genome with more than 50% of the total cfDNA sequence are removed from the list to establish the filtered list.
[0501] In embodiments, the sequence of a strand of an cfDNA is identical to or complementary to a microbial DNA sequence.
[0502] In embodiments the human sequences are removed where the sequence of a strand of an cfDNA is identical to or complementary to a human DNA sequence.
[0503] Accordingly, the selection may provide at least one cfDNA sequencing dataset comprising cfDNA sequences of the microbe specific cfDNA list,
[0504] the nucleotide length of each of the cfDNA molecules in said list, the number of reads of the microbe specific cfDNAs in said list, the genomic position(s) in the microbial genome of each alignment, the genomic sequence(s) of a microbial species to which the cfDNA sequences align and
[0505] the total number of cfDNA reads in the dataset.
[0506] In other embodiments, a cfDNA sequencing dataset is obtainable by
[0507] providing a list comprising a plurality of cfDNA sequences,
[0508] aligning each one of the cfDNA sequences in the list to human genomic sequences,
[0509] removing from the list cfDNA sequences that align to the human genome thereby establishing a filtered list,
[0510] aligning the sequences of the filtered list with a plurality of microbial genome sequences,
[0511] grouping all the microbial sequences aligning to the cfDNA sequences in the filtered list on a genus an / or species level,
[0512] performing a last common ancestor analysis, by identifying all near-best alignments of the cfDNA sequences in the groups,
[0513] removing from the list the cfDNA sequences where the alignment of the sequence to the microbial species and / or genus specific genomic set(s) is ambiguous, thereby establishing a microbe specific cfDNA list, and
[0514] based on the microbe specific cfDNA list, provide at least one cfDNA sequencing dataset comprising
[0515] (i) cfDNA sequences of the microbe specific cfDNA list,
[0516] (ii) the nucleotide length of each of the cfDNA molecules in said list,
[0517] (iii) the number of reads of the microbe specific cfDNAs in said list,
[0518] (iv) the genomic position(s) in the microbial genome of each alignment,
[0519] (v) the genomic sequence(s) of a microbial species to which the cfDNA sequences align and
[0520] (vi) the total number of cfDNA reads in the dataset.
[0521] In embodiments, a single dataset is provided. In other embodiments, a plurality of datasets is provided, such as e.g., 2, 5, 10, 20, 100, or 1000 datasets or between 1-1000 datasets.
[0522] In addition, the present disclosure also relates to a computer or computer system, which implements the method disclosed herein.
[0523] Method of treatment
[0524] As mentioned above, the method of determining the identity of at least one pathogen responsible for an infection in a patient as described herein may be used in a clinical settingto provide critical perspectives on the status of the patients and may be used to provide guidance as to which treatment the individual patients should be subjected to, thereby providing a personalized treatment. As such the compositions, kits and methods of analysis provided herein may all be used in different steps of the method of treatment provided herein.
[0525] Accordingly, the present disclosure also relates to a method for treating a patient suffering from an infection, the method comprising identifying at least one pathogen according to the method disclosed herein and based on that identification suggesting a clinically relevant and / or verified effective antimicrobial treatment regimen targeting said pathogen.
[0526] Examples of suggested empirical antimicrobial treatment of bloodstream infections are e.g., described in Timsit et al., 2014, included herein by reference (TimsitJF, Soubirou JF, Voiriot G, Chemam S, Neuville M, Mourvillier B, Sonneville R, Mariotte E, Bouadma L, Wolff M.
[0527] Treatment of bloodstream infections in ICUs. BMC Infect Dis. 2014 Nov 28; 14:489.)
[0528] As such the infection may be a bacterial, fungal, viral, or parasitic infection. The pathogen may therefore be a bacterium, a fungus, a parasite or a virus. The method is especially suited for identification of viral and bacterial sources, since the cfDNA originating from these are often shorter than the cfDNA fragments from eukaryotic sources, wherefore the methods provided herein are especially optimized for and suited to identify the viral and bacterial cfDNA in the samples.
[0529] In some cases the infection has resulted in a systemic or localized pathological state in the patient, such as e.g., abscess, cellulitis, sepsis, bacteraemia, viremia, fungemia, toxaemia, tuberculosis, chronic osteomyelitis, meningitis, encephalitis or brain abscess.
[0530] In embodiments, more than one source of infection i.e., pathogen is identified. In other embodiments, a plurality of sources of infection i.e., pathogens are identified. In additional embodiments, the infection is a polymicrobial infection.
[0531] A method for excluding microbial infectionfs) as the cause of a pathological states of a subject
[0532] The present disclosure also relates to methods for clearing patients from a suspicion of infection, i.e., a tool for clinicians to be used where the aetiology of the disease means that the patient is suspected of having one or more infection, such as e.g., in sepsis patients. In some cases, where the cause of the disease is unknown, but where it is suspected that thedisease is due to a microbial infection, it can be very beneficial to be able to rule out infections as source of the disease, such that the treatment can be adapted to better accommodate the needs of the patient. Examples of non-infection caused diseases which may have the same clinical presentation as sepsis is e.g., systemic inflammatory response syndrome (SIRS), but other examples are other diseases leading to inflammation, such as e.g., pancreatitis, trauma, burns, ischemia-reperfusion injury, drug reactions, haematological malignancies, and autoimmune disorders. In general, the diagnosis of any disease or patient state where the abnormal release of damage associated molecular patterns (DAMPs) results in enhanced inflammation will benefit from exclusion of the potential for a microbial infection, such that the treatment can be aimed at the actual underlying disease. Such exclusion has historically been shown to be beneficial for patient outcome when using the inflammation related markers procalcitonin and C-reactive protein, but with limitations on sensitive.
[0533] Accordingly, the present disclosure also relates to a method for excluding microbial infection(s) as the cause of a pathological states of a subject suspected of suffering from a microbial infection, the method comprising subjecting a sample from the subject to an analysis method according to the method for determining the identity of at least one pathogen responsible for an infection in a patient disclosed herein, and concluding, based on an absence of any identified microbe(s) that the patient is infection-free.
[0534] In preferred embodiments, the subject is a mammal, more preferably the subject is human patient, and most preferable the subject is a human patient suspected of suffering from a microbial infection.
[0535] As such the pathological state may be selected from abscess, cellulitis, sepsis, bacteraemia, viremia, fungemia, toxaemia, tuberculosis, chronic osteomyelitis, meningitis, encephalitis, brain abscess, pancreatitis, trauma, burns, ischemia-reperfusion injury, drug reactions and autoimmune disorders, preferably the pathological state is sepsis. In some aspects, the patient suffers from toxaemia, chronic osteomyelitis such as e.g., non-bacterial osteomyelitis, meningitis such as e.g., non-infectious meningitis, or encephalitis such as e.g., autoimmune encephalitis. In some cases, the patient suffers from a disease of unknown aetiology. In some cases the pathological states is an enhanced inflammation state caused by abnormal release of damage associated molecular patterns (DAMPs).
[0536] SPECIFIC EMBODIMENTS
[0537] 1. A method for determining the identity of at least one pathogen responsible for an infection in a patient, comprising the following steps:
[0538] a) obtaining a sample from the patient,
[0539] b) optionally, extracting the total DNA from said sample,
[0540] c) optionally, enriching the microbial cfDNA molecule fraction in the extracted DNA,
[0541] d) optionally, amplifying the enriched microbial cfDNA molecules,
[0542] e) sequencing the enriched microbial cfDNA molecules,
[0543] f) establishing a dataset containing features of the microbial cfDNA molecules, and
[0544] g) determining at least one source of the microbial cfDNA from the sample as the identity of the pathogen(s) responsible for the infection.
[0545] The method according to item 1, wherein the microbial cfDNA molecule fraction in the extracted DNA is enriched by,
[0546] isolating the majority of DNA molecules with a size of 75 bp or more from the sample and
[0547] purifying the DNA molecules in the remaining sample, thereby enriching the microbial cfDNA.
[0548] The method according to item 2, wherein
[0549] the extraction of mammalian DNA from the sample comprises the use of
[0550] o a first DNA binding matrix and
[0551] o a first composition as defined in any one of items 68-89.
[0552] The method according to item 2 or 3, wherein
[0553] the purification of the microbial cfDNA fragments comprises the use of
[0554] o a second DNA binding matrix and
[0555] o a second composition as defined in any one of items 90-97.
[0556] The method according to item 3 or 4, wherein the first and / or second matrix is a support defined in any one of items 104-110.
[0557] The method according to any one of items 3-5 wherein first composition comprises about 15%v / v isopropanol, about 3M GTC, 4%v / v Tween 20.
[0558] The method according to any one of the preceding items, wherein step f) comprises steps i.-vii. of item 30.
[0559] The method according to any of the preceding items, wherein step g) comprises the use of a method as defined in any one of items 30-52.
[0560] 9. The method according to any of the preceding items, wherein step g) comprises gl) performing for each cfDNA sequencing dataset(s) of step f),
[0561] i. a statistical analysis of the uniformity of the distribution of cfDNA alignments across the genome of the microbial species, and / or
[0562] II. a statistical analysis of the distribution of the nucleotide length of the cfDNA molecules in the dataset, and / or
[0563] g2) determining the pathogen(s) responsible for the infection as the microbe(s) for which:
[0564] ill. the distribution of cfDNA alignments across the genome of the microbial species does not differ significantly from a Poisson distribution and / or
[0565] iv. the identified length of the cfDNA molecules can be statistically sampled from a normal distribution defined by a cfDNA dataset from known infectious host samples.
[0566] 10. The method according to anyone of the preceding items, wherein the method comprises a quantification of the amount of microbial cfDNA molecules in the sample.
[0567] 11. The method according to any one of the preceding claims, wherein the sample is obtained from a mammalian subject.
[0568] 12. The method according to any one of the preceding claims, wherein the sample is obtained from a human subject.
[0569] 13. The method according to any one of the preceding claims, wherein the sample is obtained from a human subject, suspected of suffering from a microbial infection.
[0570] 14. The method according to any one of the preceding claims, wherein the sample is obtained from a human subject that has been clinically validated to suffer from a microbial infection of unknown aetiology.
[0571] 15. The method according to any one of the preceding items, wherein the sample is tissue, blood, bronchial lavage fluid, exhaled breath, serum, plasma, faecal, cerebrospinal fluid, BAL, or urine.
[0572] 16. A method for treating a patient suffering from an infection, the method comprising identifying at least one pathogen according to the method of any one of items 1- 15, and
[0573] suggesting a clinically relevant and / or verified effective antimicrobial treatment regimen targeting said pathogen.
[0574] 17. The method according to any one of items 1-15, or the method according to item 16, wherein the infection is a bacterial, fungal, viral, or parasitic infection.
[0575] 18. The method according to any one of items 1-15, or the method according to item 16 or 17, wherein the infection has resulted in a systemic or localized pathological state in the patient.
[0576] 19. The method according to item 18, wherein the pathological state is selected from abscess, cellulitis, sepsis, bacteraemia, viremia, fungemia, toxaemia, tuberculosis, chronic osteomyelitis, meningitis, encephalitis and brain abscess.
[0577] 20. The method according to any one of items 1-15, or the method according to item 16- 19, wherein the pathogen is a bacterium, a fungus, a parasite or a virus.
[0578] 21. The method according to any one of items 1-15, or the method according to any one of items 16-20, wherein the patient is a human patient.
[0579] 22. The method according to any one of items 1-15, or the method according to any one of items 16-21, wherein more than one pathogen is identified.
[0580] 23. The method according to any one of items 1-15, or the method according to any one of items 16-22, wherein a plurality of pathogens are identified.
[0581] 24. A method for excluding microbial infection(s) as the cause of a pathological states of a subject suspected of suffering from a microbial infection, the method comprising subjecting a sample from the subject to an analysis method according to the method of any one of items 1-15, and concluding, based on an absence of any identified microbe(s) that the patient is infection-free.
[0582] 25. The method according to item 24, wherein the pathological state is selected from abscess, cellulitis, sepsis, bacteraemia, viremia, fungemia, toxaemia, tuberculosis, chronic osteomyelitis, meningitis, encephalitis and brain abscess.
[0583] 26. The method according to item 24 or 25, wherein the subject is a mammal.
[0584] 27. The method according to item 24 or 25, wherein the subject is a human patient suspected of suffering from a microbial infection.
[0585] 28. The method according to any one of items 24-27, wherein the pathological state is sepsis.
[0586] 29. The method according to any one of items 24-27, wherein the pathological state is an enhanced inflammation state caused by abnormal release of damage associated molecular patterns (DAMPS).
[0587] 30. A method for identifying the microbial source(s) of an infection in a host as the microbial source(s) of a plurality of short cell-free DNA molecules (cfDNA) in a sample, the method comprising
[0588] a) obtaining at least one cfDNA sequencing dataset comprising
[0589] i. a plurality of cfDNA sequences of microbe specific cfDNA,
[0590] II. the nucleotide length of each of the cfDNA molecules in the plurality of cfDNA molecules,
[0591] ill. the genomic sequence of the microbial species,
[0592] iv. the genomic position(s) of each alignment,
[0593] v. the number of reads of each alignment and
[0594] vi. the total number of cfDNA sequencing reads in the dataset, b) performing for each cfDNA sequencing dataset
[0595] i. a statistical analysis of the uniformity of the distribution of cfDNA alignments across the genome of the microbial species, and / or
[0596] II. a statistical analysis of the distribution of the nucleotide length of the cfDNA molecules in the dataset, and / or
[0597] ill. an informatic analysis of the 5'- and / or 3'- sequences of the cfDNA molecules in the dataset
[0598] c) identifying the microbial source of the infection as the microbe(s) for which;
[0599] i. the distribution of cfDNA alignments across the genome of the microbial species does not differ significantly from a Poisson distribution and / or
[0600] II. the identified length of the cfDNA molecules can be statistically sampled from a normal distribution defined by a cfDNA dataset from known infectious host samples and / or
[0601] ill. the 5'- and / or 3'- termini of cfDNA molecules are host-specific enzymatic degradation termini sequences, and optionally,
[0602] d) identifying microbial contaminant source(s) as the microbes for which the distribution of cfDNA alignments across the genome of the microbial speciesdeviates significantly from a Poisson distribution and / or for which the distribution of lengths of the cfDNA molecules are incompatible with statistical sampling from a normal distribution or log-normal distribution defined by a cfDNA dataset from known infectious host samples and / or the termini sequence of cfDNA molecules are non-host-specific enzymatic degradation termini sequences.
[0603] The method according to item 30, wherein a cfDNA sequencing dataset is obtainable by performing the following steps:
[0604] i. providing a list comprising a plurality of cfDNA sequences,
[0605] II. aligning each one of the cfDNA sequences in the list to host genomic sequences,
[0606] ill. removing from the list cfDNA sequences that align to the host genome thereby establishing a filtered list,
[0607] iv. aligning the sequences of the filtered list with a plurality of microbial genome sequences,
[0608] v. grouping all the cfDNA sequences in the filtered list that align to microbial sequences in a microbial species / genus specific genomic set of cfDNA sequences,
[0609] vi. applying a last common ancestor approach to remove cfDNA sequences where the alignment of the sequence to the microbial species and / or genus specific genomic set(s) is ambiguous, thereby establishing a microbe specific cfDNA sequence list,
[0610] vii. based on the microbe specific cfDNA sequences in the microbe specific cfDNA sequence list, provide the at least one cfDNA sequencing dataset comprising
[0611] 1) the cfDNA sequences of the microbe specific cfDNA,
[0612] 2) the nucleotide length of each of the cfDNA molecules in said dataset,
[0613] 3) the number of reads of the microbe specific cfDNAs in said dataset,
[0614] 4) the genomic position(s) in the microbial genome of each alignment with a cfDNA in said dataset,
[0615] 5) the full genomic sequence(s) of the microbial species to which the cfDNA derived microbial genomic compositions predominantly align and
[0616] 6) the total number of cfDNA reads in the dataset.
[0617] The method according to item 30 or 31, wherein method comprises step d).
[0618] 33. The method according to any one of items 30-32, wherein the method is a computer implemented method.
[0619] 34. The method according to any one of items 30-33, wherein the normal- or log-normal distribution is statistically tested using a Kolmogorov-Smirnov Test, a Shapiro-Wilk Test and / or an Anderson-Darling Test.
[0620] 35. The method according to any one of items 30-34, wherein the Poisson distribution is tested using a Chi-Square goodness-of-fit test.
[0621] 36. The method according to any one of items 30-35, wherein the significance level (a) of the statistical analysis is o=0.05 or less, preferably o = 0.01 or less.
[0622] 37. The method according to any one of items 30-36, wherein the genome of the microbial species is segmented into at least two segments, and the number of reads of the microbe specific cfDNAs in each segment is counted (k).
[0623] 38. The method according to item 37, wherein the statistical analysis of the uniformity of the distribution of cfDNA alignments across the genome of the microbial species is based on the counted alignments in each segment.
[0624] 39. The method according to item 37 or 38, wherein the genome is segmented into at least 10 segments, such as at least 20, 30, 40, 50, 60 70, 80, 90, 100, 110, 120, 130, 140, 150 segments or such as at least 160 segments.
[0625] 40. The method according to any one of items 37-39, wherein the genome segmented into segments comprises 1 kbp-500 kbp sections, such as 10-100 kbp sections, such as 20-80 kbp sections, such as 30-70 kbp sections, or such as 40-80 kbp sections.
[0626] 41. The method according to any one of items 37-40, wherein the genome is segmented into segments defined by gene coding sections.
[0627] 42. The method according to any one of item 41, wherein the gene coding section comprises one or more gene clusters.
[0628] 43. The method according to any one of items 30-42, wherein each cfDNA sequencing dataset comprises at least 100 total sequencing reads.
[0629] 44. The method according to any one of items 30-43, wherein the sample is obtained from any mammalian host, preferably a human.
[0630] 45. The method according to any one of items 30-44, wherein the cfDNA sequences are obtained from a sample, such as e.g., a tissue, blood, bronchial lavage fluid, exhaled breath, serum, plasma, faecal, or urine sample.
[0631] 46. The method according to item 45, wherein the sample comprises subsets of cfDNA sequences originating from a plurality of sources that may be clinically relevant or clinically irrelevant, and the method is configured to identify the source of each of the subset of cfDNA sequences.
[0632] 47. The method according to any one of items 30-46, wherein the method comprises providing at least one suggestion for a treatment to which a subject is eligible based on the identity / identities of the microbial source(s) of the infection.
[0633] 48. The method according to any one of items 30-47, wherein the method provides the microbial source(s) of an infection and microbial source(s), of microbial contaminants, if any.
[0634] 49. The method according to any one of items 30-48, wherein if no microbial source(s) of an infection is identified in step c), and a source of at least one microbial contaminant is identified in step d), then the method should identify at least one microbial contaminant as such.
[0635] 50. The method according to any one of items 30-49, insofar as items 32-49 depend on item 31, wherein the sequence of a strand of a cfDNA is identical to or complementary to a microbial DNA sequence.
[0636] 51. The method according to any one of items 30-50, wherein, if no microbial source(s) of an infection is identified, which is / are not defined as contaminants, the safe discharge of the patient is supported.
[0637] 52. The method according to any one of items 30-51, wherein the method comprises a step of establishing set of location specific background contaminant sequences to be subtracted from the cfDNA sequencing dataset.
[0638] 53. A method of treating an infection in a subject in the need thereof, comprising
[0639] a) identifying the source of an infection as the likely source of the microbial cfDNA as defined in any of the preceding items, and
[0640] b) administering to the subject an effective amount of an anti-microbial compound with a verified activity against the identified microbe.
[0641] 54. The method according to item 53, wherein the microbe is viral, bacterial, fungal or parasitic.
[0642] 55. The method according to any one of items 53 or 54, wherein more than one source of infection is identified.
[0643] 56. The method according to any one of items 53 or 54, wherein a plurality of sources of infection are identified.
[0644] 57. A method according to any one of items 30-52 for use in a diagnostic process for infectious diseases.
[0645] 58. A method for excluding microbial infection(s) as the cause of a pathological states of a subject suspected of suffering from a microbial infection, the method comprising subjecting a sample from the subject to an analysis method according to the method of any one of items 30-52, and concluding, based on an absence of any identified microbe(s), which is / are not defined as contaminants, that the patient is infection- free.
[0646] 59. The method according to item 58, wherein the pathological state is selected from abscess, cellulitis, sepsis, bacteraemia, viremia, fungemia, toxaemia, tuberculosis, chronic osteomyelitis, meningitis, encephalitis and brain abscess.
[0647] 60. The method according to item 58 or 59, wherein the subject is a mammal.
[0648] 61. The method according to item 58 or 59, wherein the subject is a human patient suspected of suffering from a microbial infection.
[0649] 62. The method according to any one of items 58-51, wherein the pathological state is sepsis.
[0650] The method according to any one of items 58-62, wherein the pathological state is an enhanced inflammation state caused by abnormal release of damage associated molecular patterns (DAMPs).
[0651] A composition that facilitates the preferential binding of long DNA fragments with an approximate size of 75 nucleotides or more, such as 75 nucleotides or more, 80 nucleotides or more, 90 nucleotides or more, or such as 100 nucleotides, or more to a matrix.
[0652] The composition according to item 64, wherein the composition is considered to facilitate preferential binding of long DNA fragments with an approximate size of 75 nucleotides or more, when an analysis comprising
[0653] a) adding a pool of DNA fragments comprising DNA fragments of different sizes in a concentration of about 0.075ng / µl, e.g., ranging from 25-500 bp such as e.g., a DNA ladder to the DNA binding buffer in the presence of a DNA binding matrix to facilitate binding of DNA fragments,
[0654] b) isolating the bound DNA from the unbound DNA to obtain an isolated DNA fraction and an unbound DNA fraction, and
[0655] c) analyse the size distribution and concentration of the isolated DNA fraction and the unbound DNA fraction,
[0656] shows that
[0657] i. the concentration of the 100bp peak in the unbound fraction has a concentration of less than 25% of the 50bp peak,
[0658] II. the concentration of the 50bp peak in the isolated DNA fraction is less than 7.5% of the concentration of the 150bp peak,
[0659] ill. the concentration of DNA in the unbound DNA fraction when measured e.g., using Qubit, is higher than 0.03 ng / pL, and
[0660] iv. the total recovery of DNA in the two fractions is higher than 70%.
[0661] The composition according to item 65, wherein DNA fragments of different distinct sizes is a DNA ladder, preferably a ladder comprising fragments with a size of at least 50 bp, 100 bp and 150 bp, preferably comprising fragments with a size of 25bp 50 bp, 75 bp, 100 bp, 125 bp, 150 bp, 175 bp, 200 bp, 250 bp, 300 bp, 400 bp and 500 bp, such as e.g., a HyperLadder 25 bp.
[0662] The composition according to item 65 or 66, wherein the analysis of the size distribution and concentration of the isolated DNA fraction and the unbound DNA fraction is performed using Tapestation or LabChip.
[0663] The composition according to item 65 or 66, wherein the composition comprises a chaotropic solute and an organic solvent.
[0664] The composition according to any one items 65-68, wherein the composition comprises
[0665] a) an organic solvent,
[0666] b) a chaotropic solute,
[0667] c) a non-ionic surfactant,
[0668] d) optionally, a chelating agent and
[0669] e) optionally, a buffering agent, and
[0670] optionally, wherein the pH of the composition is in the range 5.0-8.5, preferably 7.5-8.5.
[0671] The composition according to any one of items 64-69, wherein the matrix is a solid or semi-solid support.
[0672] The composition according to any one of items 64-70, wherein the composition comprises isopropanol in an amount of 5-20 %v / v as an organic solvent and 0.001-8.0M of a chaotropic solute.
[0673] The composition according to any one of items 64-71, wherein the pH is in the range 5.0-8.5, preferably 7.5-8.5.
[0674] The composition according to any one of items 64-71, wherein the composition further comprises
[0675] a non-ionic surfactant in a concentration of 0.1-10%v / v,
[0676] a chelating agent in a concentration of 1-10 mM,
[0677] a buffering agent in a concentration of 1-100 mM and wherein the pH of the composition is in the range 5.0-8.5, preferably 7.5-8.5.
[0678] The composition according to any one of items 69-73, wherein the chaotropic solute is selected from a chaotropic salt, such as guanidinium chloride, guanidine thiocyanate, guanidium hydrochloride, lithium perchlorate, lithium acetate, magnesium chloride, sodium dodecyl sulphate, and a chaotropic nitrogen compound, such as urea and thiourea.
[0679] The composition according to any one of items 69-74, wherein the chaotropic solute is guanidine thiocyanate (GTC).
[0680] 76. The composition according to any one of items 69-75, wherein the non-ionic surfactant is selected from Tween 20, Tween 80, Triton X-100, Triton X-114, Brij 35, Brij 58, NP-40, Span 20, Span 80, Pluronic F68, Pluronic F127, Lubrol PX and Nonidet P-40.
[0681] 77. The composition according to any one of items 71-76, wherein the non-ionic surfactant is Tween 20.
[0682] 78. The composition according to any one of items 69-77, wherein the buffering agent is selected from HEPES, Tris, TAPS, Bicine, MOPS, AMPD, CHES, TES and phosphate buffering systems such as a NaH2PO4 and Na2HPO4 buffer system.
[0683] 79. The composition according to any one of items 69-78, wherein the buffering agent is Tris.
[0684] 80. The composition according to any one of items 69-79, wherein the chelating agent is selected from EDTA, EGTA, DTPA, NTA, Citrate, Phytic acid and CDTA.
[0685] 81. The composition according to any one of items 69-80, wherein the chelating is EDTA.
[0686] 82. The composition according to any one of items 69-70, or 72-81, wherein the organic solvent is selected from an alcohol, such as methanol, ethanol, isopropanol, 1- propanol, n-butanol, and isopropanol, a hydrophilic aromatic solvent, such as phenol, o-cresol, o-chlorophenol and 2,6-dimethylphenol, a polar aprotic solvent such as dimethyl sulfoxide (DMSO), acetonitrile, dimethylformamide (DMF) and N-Methyl-2- pyrrolidone (NMP), an amide-based solvent such as formamide, N, N- Dimethylacetamide (DMAc), N, N-Dimethylpropionamide and methylformamide, a halogenated organic solvent such as chloroform, dichloromethane (DCM), trichloroethanol and carbon tetrachloride.
[0687] 83. The composition according to any one of items 69-82, wherein the concentration of the chaotropic solute is in the range 0.01-8.0 M, such as 0.01 M, 0.02 M, 0.05 M, 0.07 M, 0.1 M, 0.2 M, 0.3 M, 0.5 M, 0.7 M, 1.0 M, 1.2 M, 1.5 M, 2.0 M, 2.5 M, 3.0 M, 3.5 M, 4.0 M, 4.5 M, 5.0 M, 5.5 M, 6.0 M, 6.5 M, 7.0 M, 7.5 M, or 8.0M, or such as 0.01-0.1 M, 0.05-0.5 M, 0.1-1.0 M, 1.0-2.0 M, 1.5-2.5 M, 2.0-3.0 M, 2.5-3.5 M,
[0688] 3.0-4.0 M, 3.5-5.0 M, 4.0-5.5 M, 4.5-6.0 M, 5.0-6.5 M, 5.5-7.0 M, 6.0-7.5 M or 6.5- 8.0 M.
[0689] 84. The composition according to any one of items 69-83, wherein the concentration of the non-ionic surfactant is in the range 0.1-10%v / v, such as 5%v / v, 5.5%v / v, 6%v / v, 6.5%v / v, 7%v / v, 7.5%v / v or 8%v / v, or such as 5-6%v / v, 5.5-6.5%v / v, 6- 7%v / v, 6.5-7.5%v / v or 7-8%v / v.
[0690] 85. The composition according to any one of items 69-85, wherein the concentration of the buffering agent is in the range 1-100 mM such as 5 mM, 6 mM, 7 mM, 8 mM, 9 mM, 10 mM, 11 mM, 12 mM, 13 mM, 14 mM, 15 mM, 16 mM, 17 mM, 18 mM, 19 mM, or 20 mM, or such as 5-10 mM, 10-15 mM, 15-20 mM, 6-12 mM, 8-16 mM, 5- 15 mM or 10-20 mM.
[0691] 86. The composition according to any one of items 69-86, where the concentration of the chelating agent is in the range 1-10 mM such as 4 mM, 4.5 mM, 5 mM, 5.5 mM, 6 mM, 6.5 mM or 7 mM, or such as 4-5 mM, 4.5-5.5 mM, 5-6 mM, 5.5-6.5 mM, 6-7 mM, 4-6 mM, or 5-7 mM.
[0692] 87. The composition according to any one of items 69-87, wherein the concentration of the organic solvent is in the range 5-20%v / v such as 5%v / v, 6%v / v, 7%v / v, 8%v / v, 9%v / v, 10%v / v, ll%v / v, 12%v / v, 13%v / v, 14%v / v, 15%v / v, 16 %v / v, 17 %v / v, 18 %v / v, 19 %v / v, or 20 %v / v or such as 5-7%v / v, 7-9%v / v, 9-ll%v / v, ll-13%v / v, 13-15%v / v, 15-17% / v / v, 17-20%v / v, 5-10%v / v, 10-15%v / v or 15-20 %v / v.
[0693] 88. The composition according to any one of items 69-88, wherein
[0694] the concentration of the chaotropic solute is about 3.0 M,
[0695] the concentration of the non-ionic surfactant is about 6.5%v / v, the concentration of the chelating agent is about 6.1 mM and the concentration of the organic solvent is 10-20%v / v.
[0696] 89. A composition that facilitates binding of short DNA fragments with an approximate size of 75 nucleotides or less to a matrix.
[0697] 90. The composition according to item 89, wherein the matrix is a solid or semi-solid support.
[0698] 91. The composition according to item 89 or 90, wherein the composition comprises a chaotropic solute and an organic solvent.
[0699] 92. The composition according to any one of items 89-91, wherein the composition comprises an organic solvent with a chaotropic activity.
[0700] The composition according to any one of items 89-92, wherein the composition comprises isopropanol in an amount of 20-45%v / v as an organic solvent and 0.001-8.0M of a chaotropic solute.
[0701] The composition according to any one of items 89-93, wherein the composition comprises the constituents defined in any one of items 71-86.
[0702] The composition according to any one of items 89-94, wherein the concentration of the organic solvent is in the range 20-45%v / v such as 20%v / v, 22%v / v, 24%v / v, 26%v / v, 28%v / v, 30%v / v, 32%v / v, 34%v / v, 36%v / v, 38%v / v, 40%v / v, 42%v / v, 43%v / v, 44%v / v, or 45%v / v, or such as 20-25%v / v, 25-30%v / v, 30-35%v / v, 35-40%v / v, 40-45%v / v, 22-28%v / v, 26-32%v / v, 20-30%v / v, 25-35%v / v, 20-35%v / v, 30-40%v / v, or 35-45%v / v.
[0703] The composition according to any one of items 89-95, wherein
[0704] the concentration of the chaotropic solute is about 3.0 M,
[0705] the concentration of the non-ionic surfactant is about 3.7%v / v
[0706] the concentration of the chelating agent is about 9.5 mM and
[0707] the concentration of the organic solvent is about 30-45%v / v.
[0708] A method for enriching microbial cfDNA in a sample, wherein the sample comprises a mix of microbial cfDNA and mammalian cfDNA, the method comprising
[0709] a) mixing a sample with a first aqueous solution to obtain a composition as defined in any one of items 64-88,
[0710] b) contacting the mixed composition with a first matrix capable of binding DNA, thereby preferentially binding the mammalian cfDNA to the first matrix thereby increasing the ratio in unbound microbial cfDNA to unbound mammalian cfDNA in the enriched sample, and
[0711] c) clearing the bound mammalian cfDNA on the matrix from the sample, to obtain a sample enriched in microbial cfDNA fragments.
[0712] The method according to item 97, wherein the method further comprises the subsequent steps
[0713] d) contacting the sample enriched for microbial cfDNA fragments from step c) with a second aqueous solution to obtain a composition as defined in any one of items 89-96,
[0714] e) contacting the composition with a second matrix capable of binding DNA thereby enriching for microbial cfDNA on the second solid or semi-solid support,
[0715] f) isolating the matrix preferentially bound with microbial cell-free DNA from remaining sample, and
[0716] g) optionally, washing the isolated matrix using a washing buffer,
[0717] h) eluting the microbial cfDNA from the second solid or semi-solid support using an elution buffer to obtain a sample further enriched in microbial cfDNA fragments.
[0718] 99. The method according to item 97 or 98, wherein the microbial cfDNA fragments content in the sample is more than significantly enriched following step c) or following step h).
[0719] 100. The method according to any one of items 97-99, wherein the sample is obtained from a mammalian subject, preferably a human subject.
[0720] 101. The method according to any one of items 97-100, wherein the sample is selected from a cell-free fluid sample, preferably a blood sample, a serum sample, a plasma sample, BAL sample, exhaled condensate sample and a urine sample.
[0721] 102. The method according to any one of items 97-100, wherein the microbial cell- free DNA fragments are of bacterial, viral, parasite or fungi species origin.
[0722] 103. The method according to any one of items 97-102, wherein the first and second matrices are of the same composition.
[0723] 104. The method according to any one of items 97-103, wherein the first and second matrices are of a different composition.
[0724] 105. The method according to any one of items 97-104, wherein the second matrix is specific towards mammalian cfDNA fragments.
[0725] 106. The method according to any one of items 97-105, wherein the second matrix is specific towards microbial cfDNA fragments.
[0726] 107. The method according to any one of items 97-106, wherein the first and / or second matrix is a solid support or semi-solid support.
[0727] . The method according to any one of items 97-107, wherein the first and / or second matrix is a silica based solid support or a carboxylic acid functionalized solid support.
[0728] . The method according to any one of items 97-108, wherein the first and / or second matrix is a silica based solid support or a carboxylic acid functionalized solid support, with a size of 300-1000 pm, such as about 300 pm, 400 pm, 500 pm, 600 pm, 700 pm, 800 pm, 900 pm, or 1000 pm, or such as 300-500 pm, 400-700 pm, 500-800 pm, 600-900 pm or 700-1000 pm.
[0729] . The method according to any one of items 97-109, wherein the clearing in step c) is performed by sedimentation, centrifugation, magnetic separation and / or column separation.
[0730] . The method according to any one of items 97-110, wherein the clearing in step c) is performed by sedimentation, centrifugation, magnetic separation and / or column separation.
[0731] . A composition that facilitates the preferential binding of long DNA fragments with an approximate size of 75 nucleotides or more, such as 75 nucleotides or more, 80 nucleotides or more, 90 nucleotides or more, or such as 100 nucleotides, or more to a matrix.
[0732] . The composition according to item 112, wherein the composition is considered to facilitate preferential binding of long DNA fragments with an approximate size of 75 nucleotides or more, when an analysis comprising
[0733] a. adding a pool of DNA fragments comprising DNA fragments of different sizes in a concentration of about 0.075ng / µl, e.g., ranging from 25-500 bp such as e.g., a DNA ladder to the DNA binding buffer in the presence of a DNA binding matrix to facilitate binding of DNA fragments,
[0734] b. isolating the bound DNA from the unbound DNA to obtain an isolated DNA fraction and an unbound DNA fraction, and
[0735] c. analyse the size distribution and concentration of the isolated DNA fraction and the unbound DNA fraction,
[0736] shows that
[0737] i. the concentration of the 100bp peak in the unbound fraction has a concentration of less than 25% of the 50bp peak,
[0738] II. the concentration of the 50bp peak in the isolated DNA fraction is less than 7.5% of the concentration of the 150bp peak,
[0739] Hi. the concentration of DNA in the unbound DNA fraction when measured e.g., using Qubit, is higher than 0.03 ng / pL, and iv. the total recovery of DNA in the two fractions is higher than 70%.
[0740] 114. The composition according to item 113, wherein DNA fragments of different distinct sizes is a DNA ladder, preferably a ladder comprising fragments with a size of 25bp 50 bp, 75 bp, 100 bp, 125 bp, 150 bp, 175 bp, 200 bp, 250 bp, 300 bp, 400 bp and 500 bp, such as e.g., a HyperLadder 25 bp.
[0741] 115. The composition according to item 112 or 113, wherein the analysis of the size distribution and concentration of the isolated DNA fraction and the unbound DNA fraction is performed using Tapestation or LabChip.
[0742] 116. The composition according to any one of items 112-115, wherein the composition comprises
[0743] a. an organic solvent,
[0744] b. a chaotropic solute,
[0745] c. a non-ionic surfactant,
[0746] d. optionally, a chelating agent and
[0747] e. optionally, a buffering agent, and
[0748] optionally, wherein the pH of the composition is in the range 5.0-8.5, preferably 7.5-8.5.
[0749] 117. The composition according to any one of items 112-116, wherein the composition comprises
[0750] a) an organic solvent being isopropanol in an amount of 10-18 %v / v, preferably 12-17%v / v,
[0751] b) a chaotropic solute being guanidine thiocyanate (GTC),
[0752] c) a non-ionic surfactant being Tween20,
[0753] d) optionally, a chelating agent and
[0754] e) optionally, a buffering agent, and
[0755] optionally, wherein the pH of the composition is in the range 5.0-8.5, preferably 7.5- 8.5.
[0756] 118. The composition according to item 117, wherein GTC is present in an amount of 2-4M, preferably about 3M.
[0757] 119. The composition according to 117 or 118, wherein Tween20 is present in an amount of 0.1-10%v / v, preferably about 4%v / v.
[0758] 120. The composition according to any one of items 117-119, wherein the chelating agent is present in an amount of 1-10 mM.
[0759] 121. The composition according to any one of items 117-120 wherein the chelating agent is EDTA.
[0760] 122. The composition according to any one items 117-121, wherein the buffering agent is present in an amount of 1-100 mM.
[0761] 123. The composition according to any one of items 117-122, wherein the buffering agent is Tris.
[0762] 124. The composition according to any one of items 117-123, wherein the composition comprises about 15 %v / v isopropanol, about 3 M of guanidine thiocyanate, 4-6 mM EDTA, 10-20 mM Tris, and has a pH of 5.0-8.5, preferably 7.5- 8.5.
[0763] 125. A method of validating the DNA binding properties of a DNA binding buffer, comprising
[0764] a. adding a pool of DNA fragments comprising DNA fragments of different sizes such as e.g., ranging from 25-500 bp such as e.g., a DNA ladder, in a concentration of about 0.075ng / pL to the DNA binding buffer in the presence of a DNA binding matrix to facilitate binding of DNA fragments,
[0765] b. isolating the bound DNA from the unbound DNA to obtain an isolated DNA fraction and an unbound DNA fraction, and
[0766] c. analyse the size distribution and concentration of the isolated DNA fraction and the unbound DNA fraction.
[0767] 126. The method according to item 125, wherein method is for validating the binding properties of a DNA binding composition that facilitates the preferential binding of long DNA fragments with an approximate size of 75 nucleotides or more, wherein the composition is considered to facilitate the preferential binding of long DNA fragments with an approximate size of 75 nucleotides or more when,
[0768] i. the concentration of the 100bp peak in the unbound DNA fraction has a concentration of less than 25% of the 50bp peak,
[0769] II. the concentration of the 50bp peak in the isolated DNA fraction is less than 7.5% of the concentration of the 150bp peak, ill. the concentration of DNA in the unbound DNA fraction when measured e.g., using Qubit, is higher than 0.03 ng / pL, and
[0770] iv. the total recovery of DNA in the two fractions is higher than 70%.
[0771] . The method according to item 125 or 126, wherein DNA fragments of different distinct sizes is a DNA ladder, preferably a ladder comprising fragments with a size of 25bp 50 bp, 75 bp, 100 bp, 125 bp, 150 bp, 175 bp, 200 bp, 250 bp, 300 bp, 400 bp and 500 bp, such as e.g., a HyperLadder 25 bp.
[0772] . The method according to any one of items 125 to 127, wherein the analysis of the size distribution and concentration of the isolated DNA fraction and the unbound DNA fraction is performed using Tapestation or LabChip.
[0773] . The method according to any one of items 125-128, wherein the DNA binding buffer is a composition according to any one of items 112-124.
[0774] . A kit for isolating microbial cell-free DNA fragments, comprising
[0775] a. a vial comprising a composition comprising
[0776] i. a chaotropic solute,
[0777] ii. a non-ionic surfactant,
[0778] iii. optionally, a chelating agent and
[0779] iv. optionally, a buffering agent, and
[0780] b. optionally, one or more matrices capable of binding DNA, optionally contained within a vial, and
[0781] c. optionally, means for performing the microbial cell-free DNA isolation, d. optionally, a DNA ladder, and
[0782] e. instructions for preparing the composition according to items 112-124, f. Optionally, instructions for isolating microbial cfDNA fragments using said kit.
[0783] . The kit according to item 130, further comprising
[0784] g. one or more vials comprising wash buffer(s) and
[0785] h. one or more vials comprising elution buffer(s).
[0786] . The kit according to item 130 or 131, wherein the kit comprises a first and second washing buffer, wherein
[0787] i. the first washing buffer comprises 15-40%v / v isopropanol, 1-8M or l-10%v / v of the chaotropic solute, a chelating agent in an amount of 1-10 mM, and a non-ionic surfactant in an amount of 0.1-10%v / v and
[0788] j. the second washing buffer comprises buffered ethanol, such as a composition comprising 1-lOOmM Tris, pH 8.0, l-5mM EDTA and 80%v / v Ethanol.
[0789] 133. The kit according to any one of items 130-132, wherein the kit comprises the vial comprising the elution buffer, and the elution buffer is a TE buffer, a TAE buffer or a TBE buffer.
[0790] 134. The kit according to any one of items 130-133, for use in the enrichment of predominantly microbial cfDNA fragments, preferably with a predominant size of 75 bp or less.
[0791] 135. A method for enriching microbial cfDNA from a sample comprising host and microbial cfDNA using a composition according to any one of items 112-124.
[0792] 136. The method according to item 135, wherein the microbial cfDNA has a predominant size of 75 bp or less.
[0793] 137. A method for treating a patient suffering from an infection, the method comprising
[0794] a. Isolation of microbial cfDNA using a composition according to any one of items 112-124, or a kit according to any one of items 130-134 from a patient sample, and
[0795] b. Identification of at least one pathogen based on the isolated microbial cfDNA, and
[0796] c. suggesting a clinically relevant effective antimicrobial treatment regimen targeting said pathogen.
[0797] 138. A method for determining the identity of at least one pathogen responsible for an infection in a patient, comprising the following steps:
[0798] a) obtaining a sample from the patient,
[0799] b) optionally, extracting the total DNA from said sample,
[0800] c) enriching the microbial cfDNA molecule fraction in the extracted DNA using a composition according to any one of items 112-124 or a kit according to any one of items 130-134,
[0801] d) optionally, amplifying the enriched microbial cfDNA molecules,
[0802] e) sequencing the enriched microbial cfDNA molecules,
[0803] f) establishing a dataset containing features of the microbial cfDNA molecules, and
[0804] g) determining at least one source of the microbial cfDNA from the sample as the identity of the pathogen(s) responsible for the infection.
[0805] 139. The method according to item 138, wherein the microbial cfDNA molecule fraction in the extracted DNA is enriched by,
[0806] removing mammalian DNA molecules from the sample using a composition according to any one of items 112-124 or a kit according to any one of items 130- 134, and
[0807] purifying the microbial cfDNA molecules in the remaining sample.
[0808] 140. The method according to item 138 or 139, wherein step f) comprises providing from the sequencing in step e) a list of cfDNA sequences, Optionally, performing an alignment of each of the cfDNA sequences in the list to host genomic sequences and removing from the list cfDNA sequences that align to the host genome thereby establishing a filtered list,
[0809] performing an alignment of the cfDNA sequences of the filtered list to a plurality of microbial genome sequences,
[0810] grouping the cfDNA sequences in the filtered list that align with the microbial genome sequence(s) into microbial species and / or genus specific genomic set(s), applying a last common ancestor approach to remove cfDNA sequences where the alignment of the sequence to the microbial species and / or genus specific genomic set(s) is ambiguous, thereby establishing a microbe specific cfDNA molecule list, based on the microbe specific cfDNA molecules in the microbe specific cfDNA molecule list, provide a dataset of the microbial cfDNA molecules comprising v. the cfDNA sequences of the microbial cfDNA,
[0811] vi. the nucleotide length of each of the microbial cfDNA molecules in said dataset,
[0812] vii. optionally, the number of reads of the microbial cfDNAs in said dataset,
[0813] viii. the genomic position(s) in the microbial genome of each alignment, ix. the genomic sequence(s) of the microbial species and / or genus to which the specific microbial cfDNA align and
[0814] x. the total number of microbial cfDNA reads in the dataset.
[0815] 141. The method according to any one of items 138-140, wherein step g) comprises gl) performing for each cfDNA sequencing dataset(s) as defined in item 138, i. a statistical analysis of the uniformity of the distribution of cfDNA alignments across the genome of the microbial species, and / or II. a statistical analysis of the distribution of the nucleotide length of the cfDNA molecules in the dataset,
[0816] g2) determining the pathogen(s) responsible for the infection as the microbe(s) for which:
[0817] ill. the distribution of cfDNA alignments across the genome of the microbial species does not differ significantly from a uniformity test, such as a Poisson distribution, and / or
[0818] iv. the identified length of the cfDNA molecules can be statistically sampled from a normal distribution defined by a cfDNA dataset from known infectious host samples.
[0819] 142. The method according to any one of items 138-141, wherein the sample is tissue, blood, bronchial lavage fluid, exhaled breath, serum, plasma, faecal, cerebrospinal fluid, BAL, or urine.
[0820] 143. The method according to any one of items 138-141, wherein the sample is a human sample.
[0821] 144. A method for treating a patient suffering from an infection, the method comprising
[0822] a. identifying at least one pathogen according to the method of any one of items 138-143, and
[0823] b. suggesting a clinically relevant effective antimicrobial treatment regimen targeting said pathogen.
[0824] 145. A method for excluding microbial infection(s) as the cause of a pathological states of a subject suspected of suffering from a microbial infection, the method comprising subjecting a sample from the subject to an analysis method according to the method of any one of items 138-143, and concluding based on an absence of any identified microbe(s) that the patient is infection-free.
[0825] EXAMPLES
[0826] Example 1
[0827] The present example aims to show the clinical relevance of the methods disclosed herein for determining the source of an infection in an individual.
[0828] Methods
[0829] Study design
[0830] We performed a cross-sectional study in the emergency ward at Aalborg University Hospital, which serves as the reference hospital in the North Denmark Region (catchment population approximately 600,000 inhabitants). Patients were prospectively enrolled upon suspicion of BSI between 1 June and 31 December 2022. For each patient, a blood sample (9 mL K2EDTA) for mNGS was drawn from the same venipuncture as a blood culture for routine diagnostics. Following inclusion, patient metadata was collected from their electronic medical record, which included medical history and diagnosis, clinical microbiology results, biochemical blood parameters, and antimicrobial therapy. As a negative reference group, healthy blood donor samples were collected from the blood bank at Aalborg University Hospital from 12 donors and handled in parallel with the clinical samples.
[0831] Study population
[0832] Study inclusion criteria comprised that the patient was admitted to the emergency ward at Aalborg University Hospital upon suspicion of BSI as adjudicated by a requested blood culture. Patients were required to be alert, oriented, and able to read and understand participant information. Patients were excluded from the study if they were younger than 18 years old or were unable to provide written informed consent. Enrolment took place on selected weekdays from 8 a.m. to 4 p.m.
[0833] Patient subgrouping
[0834] Based on the medical records, patients were subdivided into the following groups as adjudicated by a clinical microbiologist: patients with a positive blood culture (BSI-confirmed); patients with a negative blood culture but a high suspicion of BSI (BSI-suspected); patients with a negative blood culture and no suspicion of an infectious disease (BSI-absent); and patients with a negative blood culture and low suspicion of a true BSI (BSI-unlikely). Patients were assigned based on all available microbiology test results (blood culture results, urine culture, respiratory samples, etc.) combined with a clinical evaluation of the patients including vital signs (e.g., temperature, pulse, blood pressure, respiration rate), biochemical parameters (e.g., CRP [C-reactive protein], white blood cell count), and imaging (e.g., chest X-ray, CT abdomen and chest, endoscopies).
[0835] Patients in groups BSI-confirmed, BSI-suspected, and BSI-absent were selected for analysis. Patients in the BSI-unlikely group were excluded from metagenomic sequencing analysis due to limitations in funding. Patients in the BSI-absent group were analysed to evaluate specificity (Fig. 1).
[0836] Clinical microbiology
[0837] All clinical microbiology results were subsequently obtained from patient medical records (wwBakt, Autonik AB, Sweden), and no follow-up analyses were conducted. Clinical microbiology analyses were conducted at the Department of Clinical Microbiology at Aalborg University Hospital. A standard blood culture using two BD BACTEC Plus Aerobic medium and one BD BACTEC Lytic Anaerobic medium glass culture vial were obtained from a peripheral site upon admission and incubated in the BACTEC FX Top instrument (Becton Dickinson AB, Stockholm, Sweden). Species identification was performed using conventional biochemical tests and matrix-assisted laser desorption ionization-time of flight (MALDI Biotyper 3.1, Bruker Daltonics Microflex LT, MBT 6903 MSP Library), with scores >2 for identification at the species level. However, Escherichia coli was identified based on the growth of red colonies on the CHROMagar Orientation medium and a positive indole test, according to the manufacturer's instructions.
[0838] Plasma preparation and DNA extraction
[0839] Plasma was prepared from study blood samples by centrifugation at 1,600 g at 4°C for 10 minutes within 1 hour from blood draw at the Department of Biochemistry at Aalborg University Hospital. The plasma supernatant was subsequently centrifuged again at 16,000 g at 4°C for 10 minutes and the new supernatant was obtained and stored at -80°C until analysis. At analysis, plasma samples were thawed, and single-stranded oligonucleotides with lengths from 50 to 200 bp and a distinct, conserved region for bioinformatic discrimination were added in a series of relevant concentrations to allow for absolute quantification of microbial DNA, in specific 16 SNSs as described herein were designed with 4 unique central regions (i) (SEQ ID NO: 1-4), and in 4 different total lengths for each unique central region (i). DNA was extracted from 2 mL plasma using the QIAmp MinElute ccfDNA kit (Qiagen) standard protocol with DNA elution in 30 pL. At least one process control was included in each batch, consisting of 2 mL nuclease-free water (Qiagen). DNA was quantified with the Qubit dsDNA HS Assay Kit on a Qubit 4 Fluorometer (Thermo Fisher). DNA length was assessed with the D1000 High Sensitivity ScreenTape Assay on a Tapestation 4150 (Agilent).
[0840] Library preparation and sequencing
[0841] Ten to 18 pL of extracted DNA was used as input to the SRSLY DNA NGS Library Preparation PicoPlus Kit (Claret Biosciences). The library preparation was carried out according to the SRSLY protocol for extreme short fragment retention. At least one process control was included in each batch, consisting of 18 pL nuclease-free water (Qiagen). Subsequently, samples were prepared for Oxford Nanopore sequencing with the ligation kit (Oxford Nanopore Technologies, SQK-LSK114), where four or five samples were multiplexed with a total of 200 fmol DNA as input. Libraries were sequenced on a PromethION 24 on R10.4.1flow cells and basecalled with the super-accurate model (Guppy 7.0.9) with a quality cut-off at a Phred score of 10 applied (Fig. 2).
[0842] Data analysis and bioinformatics
[0843] Basecalled reads were demultiplexed, trimmed for adapters, and checked for any remaining barcodes or adapters with Cutadapt (version 4.4). Reads were aligned to the human genome (consecutive rounds of alignment with minimap2 version 2.26, bwa version 0.7.17, and BLAST version 2.15.0). Reads with no significant alignments were aligned with Bowtie2 (-k 10 --mp 2,2 --score-min G,20,8) to a database consisting of Genome Taxonomy Database representative bacterial and archaeal genomes from release 207 and virus (complete genomes) and fungi (all genomes) from NCBI RefSeq release 215 as well as the reference NCBI Homo sapiens genome (T2T). All microbial reads were re-evaluated in a last-common-ancestor approach and only accepted if more than 50% of near-best alignments for a read (edit distance <10% of read length and aligned length >90% read length) matched the same genus. Species with a non-uniform genome coverage of reads were discarded.
[0844] Quantification of absolute load of microbial DNA
[0845] Reads from the single-stranded spike-ins were used to quantify DNA for all identified pathogens in genome equivalents per microliter (GPM). This was done by computing sizedependent transfer functions based on the read counts of SNSs, and using the median read length of the reads for a pathogen taxonomy the read count of that species was converted into an absolute count of molecules (million molecules per microliter [MPM]). This was normalised to genome equivalents per microliter (GPM) using the genome size of the reference genome for that taxonomy. The absolute load of microbial DNA in patient samples was compared to the 12 blood donor samples for each species and included only when the absolute load in a patient sample was five times that observed in the control sample with the highest load of that species.
[0846] Assessment of clinical relevance of mNGS results
[0847] To evaluate the relevance of mNGS findings, results were contextualized with the patient's other microbiological results and overall clinical picture. mNGS results were grouped on a patient level based on the clinical relevance of identifications as adjudicated by two clinical microbiologists. " Confirmed" was assigned only when at least one of the pathogens identified by mNGS was also identified by blood culture. " Probable" was assigned when at least one of the pathogens identified by mNGS was also identified in other samples from the suspected infection site or when the mNGS results aligned very well with the clinical picture (e.g., E. coll identified in a patient with suspected urinary tract infection [UTI]). " Possible" was assigned when at least one of the pathogens identified by mNGS was consistent with theclinical picture. " Unlikely" was assigned when the pathogens identified by mNGS were considered irrelevant in the context of the acute infection.
[0848] Potential clinical impact analysis
[0849] To evaluate the potential impact on patient antimicrobial treatment, the microbial identifications from mNGS were compared to results from routine diagnostics, medical record (suspected site of infection), and the antimicrobials administered prior to and during the relevant admission by two clinical microbiologists. As the mNGS analysis was conducted in batches following completed patient inclusion, the impact analysis was based on the turnaround time in a potential real-time setup, which was estimated to be around 6 hours. This aligns with previously reported turnaround times for Nanopore mNGS. Patient courses were evaluated for three different potential impact types. Firstly, a change from ineffective to effective antimicrobial treatment was assigned when mNGS results could have led to a correction from ineffective (either due to antimicrobial resistance or the antimicrobial not covering the pathogen) to effective antimicrobial treatment. This was adjudicated only when the first-line antimicrobial against the pathogen identified by mNGS would have been effective. Secondly, an addition of an extra antimicrobial for coverage was assigned when mNGS results could have led to the addition of one or more extra antimicrobials to cover all identified pathogens relevant to the acute infection. Thirdly, extended treatment was assigned when the finding of a pathogen in the blood with DNA sequencing that was already covered by the antimicrobial treatment could have led doctors to consider prolonging IV antimicrobial treatment and admission. Only the patients where pathogens from mNGS were adjudicated as relevant to the acute infection were considered in the impact analysis. In the assessment of impact, all clinical microbiology diagnostics were considered, and e.g., a change to effective treatment was only adjudicated if no relevant diagnostic answers were available at the time (6 hours post admission) to guide the treatment. Negative mNGS answers were not assessed for potential in guiding de-escalations.
[0850] Statistical analysis
[0851] Confidence intervals for sensitivity and specificity were calculated using the Clopper-Pearson exact method. If not specified otherwise, a two-sided Wilcoxon-Mann-Whitney test was used to compare distributions. In case of multiple comparisons, the P-values were adjusted with the Holm-Bonferroni method. The comparison of mNGS results to gold-standard blood culturing was conducted as follows: a true positive was adjudicated when at least one pathogen identified by mNGS was in accordance with blood culturing; a true negative was adjudicated when no pathogens were identified by either blood culture or mNGS; a false positive was adjudicated when blood culture was negative and mNGS was positive; a false negative was adjudicated when blood culture was positive and mNGS was negative for that pathogen. All statistical analyses were conducted in R (version 4.2.0).
[0852] Results
[0853] Patient inclusion and sample selection
[0854] During the study period, a total of 186 patients provided written informed consent. Following inclusion, 40 patients were selected for analysis after subdividing the patients into four groups: BSI-confirmed (n = 11), BSI-suspected (n = 25), BSI-absent (n = 4), and BSI-unlikely (n = 141), while five patients were excluded (Fig. 1). Only samples from patients in BSI-confirmed, BSI-suspected, and BSI-absent were subjected to mNGS analysis.
[0855] Patient characteristics
[0856] A subset of 40 patients (50% female) with a median age of 71 years (interquartile range [IQR]: 59-84) was analyzed using mNGS (Table 15).
[0857] The most common infection types were UTI (n = 12) and lower respiratory tract infection (LRTI, n = 10). Of the 12 patients admitted with a UTI, 5 were suspected to have pyelonephritis (Table 16, 17 and 18). More than half of the 40 patients had one or more comorbidities, including diabetes (n = 13), cancer (n = 7), and chronic obstructive pulmonary disease (n = 5). Emphasizing the success of our BSI categorization, we observed significantly higher CRP levels in the BSI-confirmed (n = 11, median: 216 mg / L, P = 0.047 and P = 0.002, Wilcoxon-Mann-Whitney) and BSI-suspected (n = 25, median: 204 mg / L, P = 0.035 and P < 0.0001, Wilcoxon-Mann-Whitney) groups, compared to BSI-absent (n = 4, median: 1.3 mg / L) and BSI-unlikely (n = 141, median: 67 mg / L) groups (Fig. 4).
[0858] Table 15 Patient characteristics of the three groups analysed with metagenomic DNA sequencing.
[0859]
[0860] ^Immunosuppressive therapy covers ATC codes L04 and H02.
[0861] *Oral antimicrobials administered at the general practitioner in relation to the relevant infection.cqSOFA, Quick Sequential Organ Failure Assessment.
[0862] d., Not relevant.
[0863] ePICC, peripherally inserted central catheter.
[0864] Table 16 - Results from conventional microbiological tests and mNGS for the BSI confirmed group
[0865]
[0866]
[0867] a., Not relevant.
[0868] faKSHV, Kaposi's sarcoma-associated herpesvirus.
[0869] CCMV, cytomegalovirus.
[0870] dRS virus, human respiratory syncytial virus.
[0871] eCould not be determined to species level with mNGS.
[0872] Table 17 - Results from conventional microbiological tests and mNGS for the BSI suspected group
[0873]
[0874]
[0875] a., Not relevant.
[0876] faKSHV, Kaposi's sarcoma-associated herpesvirus.
[0877] CCMV, cytomegalovirus.
[0878] dRS virus, human respiratory syncytial virus.
[0879] eCould not be determined to species level with mNGS.
[0880] Table 18 - Results from conventional microbiological tests and mNGS for the BSI-absent group
[0881]
[0882] Clinical microbiology
[0883] The 11 positive blood cultures from patients in the BSI-confirmed group identified E. coli n = 7), Staphylococcus epidermidis n = 2), Enterococcus faecalis n = 1), and Staphylococcusaureus (n = 1), reflecting that most patients in this group were admitted with urosepsis (Table 16). Of the 25 patients in the BSI-suspected group, 14 (56%) had positive cultures from urine (n = 10), sputum (n = 2), bronchoalveolar lavage (n = 1), or other sites (n = 3), while the remaining patients in this group had no positive microbiological tests to guide the antimicrobial treatment. Pathogens identified in other cultures for BSI-suspected patients were E. coll n = 7), Pseudomonas aeruginosa n = 2), and others (n = 9).
[0884] Metagenomic DNA sequencing analysis
[0885] From all 40 patients, plasma cfDNA was sequenced (Fig. 2a) and patient samples had a mean read depth of 10. IM reads (s.d. 5.6M) with a mean Phred quality of 19.0 (s.d. 0.7) and an average read length N50 of 176 bp (s.d. 11.1 bp). Patients in BSI-confirmed (n = 11, median: 10.1 ng / mL) had significantly higher levels of cfDNA in plasma compared to blood donor samples (n = 12, median: 2.3 ng / mL) (P = 0.00058, Wilcoxon-Mann-Whitney), which was also the case for patients in BSI-suspected (n = 25, median: 14.3 ng / mL) (P < 0.00001, Wilcoxon-Mann-Whitney) (Fig. 3d).
[0886] mNGS confirmed all positive blood culture findings, and three patients in this group had one or two additional pathogens identified (Table 16). The additional findings comprised two bacteria (Enterococcus faecium [pl39] and Citrobacter koseri [pl83]) and two viruses (human cytomegalovirus [human betaherpesvirus 5, CMV] [pl39] and Kaposi's sarcoma-associated herpesvirus (human gammaherpesvirus 8, KSHV) [pl04]). In the BSI-suspected group, one to six pathogens were identified in 12 of 25 patients (Fig. 3c). Eleven of the 12 patients with a positive mNGS result had one or more pathogens identified that were deemed relevant for the acute infection by a clinical microbiologist and assigned either " Probable" (n = 10) or " Possible" (n = 1) (Fig. 3a). This included common pathogens like E. coll, P. aeruginosa, and Streptococcus pneumoniae, but also pathogens like Serratia spp., Raoultella planticola, and Veillonella atypica. One patient with a positive mNGS result, interpreted as " Unlikely" to be relevant to the acute infection, had Clostridioides difficile identified (pl43). Interestingly, in 5 of the 12 patients in the BSI-suspected group with a pathogen identified by mNGS, another microbiological test result supported the finding (Table 16). All patients who received antimicrobials prior to admission (n = 7) were in the BSI-suspected group, of which three (p072, pl05, and pl40) were positive with mNGS (data not shown). No pathogens were identified in the BSI-absent group. Evaluated against blood culture findings, the sensitivity was 100% (confidence interval 71.5%-100.0%) and the specificity was 59% (confidence interval 38.9%-76.5%) at a patient level (Clopper-Pearson exact method). Since most pathogen identifications were deemed relevant to the acute infection, the low specificity is likely due to blood culturing being a poor reference method.
[0887] With the single-stranded DNA spike-ins, the load of pathogen DNA in plasma samples was quantified, and pathogens were reported in titers from 0.5 to 321 GPM, which is in accordance with previously reported pathogen DNA titers in BSIs. Comparing the load of E. collâ the most frequently identified pathogenâ across BSI-confirmed (n = 7, median: 41.3 GPM) and BSI-suspected (n = 6, median: 4.37 GPM), there was no significant difference P = 0.30, Wilcoxon-Mann-Whitney), which was also the case when comparing the total pathogen load in patients from BSI-confirmed (n = 11, median: 41.3 GPM) and BSI-suspected (n = 12, median: 4.71 GPM) with a positive mNGS test P = 0.10, Wilcoxon-Mann-Whitney) (Fig. 3b).
[0888] Case series
[0889] To contextualize the results of mNGS to the patient clinical presentation, we present a more detailed review of selected cases from BSI-confirmed and BSI-suspected groups, who were positive by mNGS. Cases that highlight possibilities and challenges of the method were selected once all results were available. From the BSI-confirmed group, three patients were selected for a more detailed analysis.
[0890] The first patient (pOOl), a 77-year-old male, was admitted with symptoms of septicemia with LRTI as the suspected primary focus and a CRP of 109 mg / L (normal reference range <8 mg / L). The patient was treated with empirical IV cefuroxime in combination with gentamicin. However, the focus was diagnosed after 48 hours as urosepsis with the identification of E. faecalis in blood and urine cultures, prompting a change to IV ampicillin. mNGS also identified E. faecalis at 1 GPM.
[0891] The second patient (pl04), a 66-year-old male with Kaposi's sarcoma and recurrent episodes of spontaneous bacterial peritonitis, secondary to liver cirrhosis, was admitted with acute abdominal pain and an elevated CRP of 41 mg / L, with no obvious signs of infection. Initial treatment included IV piperacillin-tazobactam and oral ciprofloxacin. After 48 hours, blood cultures taken from both a peripheral vein and a peripherally inserted central catheter (PICC) line were positive for S. epidermidis, with the PICC line identified as the suspected source. The PICC line had been in place for an extended period, though the exact date of insertion was unknown. This prompted a change in treatment to IV vancomycin, and the PICC line was removed. mNGS identified S. epidermidis (52.4 GPM) and KSHV (1 GPM).
[0892] The third patient (pl39), a 64-year-old male, was admitted with infection with unknown focus and an elevated CRP of 248 mg / L. The patient had a medical history of a renal transplant and received immunosuppressives. E. coll was cultured from blood taken at admission after 46 hours, and this was also identified by mNGS at 102 GPM. Additionally, mNGS identified Enterococcus faecium (0.7 GPM) and CMV (26.6 GPM). A PET-CT scan identified the infectious focus as multiple infected renal and liver cysts, and the patient wasadministered IV piperacillin-tazobactam for 31 days with gradual improvement. On day 11 of the hospitalization, a quantitative PCR for CMV was positive with 290 copies / mL.
[0893] From the BSI-suspected group, three patients were selected for a more detailed description.
[0894] The first patient (pl28), a 72-year-old female, was admitted upon suspicion of pneumonia with a CRP of 470 mg / L and a low systolic blood pressure of 99 mmHg. The patient was known with chronic obstructive pulmonary disease and autoimmune hepatitis. The patient was seen by a general physician prior to the admission concerning the LRTI, but it was unclear from the medical record if the patient was treated with antimicrobials before admission. All microbiology tests taken at admission (blood culture, urine culture, S. pneumoniae urine antigen test) were negative, while S. pneumoniae was identified by mNGS at 3.5 GPM. The patient was administered IV benzylpenicillin at admission with a good clinical response.
[0895] The second patient (pl73) was a 66-year-old female with a medical history of nephrolithiasis treated with double-J stent and numerous episodes of UTIs. She was admitted on suspicion of pyelonephritis presenting a CRP of 164 mg / L. Prior to admission, the patient was treated with ciprofloxacin, and a urine culture obtained 5 days before was positive
[0896] for Enterococcus spp. (1 x 10e4 CFU / mL). However, the blood and urine cultures taken at admission were negative and the patient was treated with IV vancomycin. mNGS detected several pathogens, with Klebsiella spp. showing the highest titer at 5.8 GPM. Three weeks prior to the relevant admission, the patient had severe sepsis requiring treatment in the intensive care unit, following the ureteral stone removal. Notably, Klebsiella pneumoniae had been cultured from both blood and urine during this admission, suggesting a likely relevance of the mNGS result.
[0897] The third patient (p072), a 64-year-old male, was admitted upon suspicion of pyelonephritis, but blood and urine cultures taken at admission were negative, likely due to the patient already receiving pivmecillinam ordered from his general physician or clinician. At admission, the treatment was escalated to IV ampicillin with good clinical response. mNGS was positive for E. coli (2.8 GPM).
[0898] Potential clinical impact
[0899] The mNGS results for the 22 patients, where the result was either confirmed by blood culture or deemed relevant by a clinical microbiologist, were analysed for potential clinical impact. In a real-time setting, the mNGS result could have affected the antimicrobial treatment in 13 cases (59%) subgrouping to five changes from ineffective to effective treatment, seven escalations from oral to IV, and one addition of an extra antimicrobial. The five cases where achange from ineffective to effective treatment could have been facilitated are particularly interesting. In patients from BSI-confirmed, this was achieved by identifying an unsuspected pathogen more rapidly than the blood culture (n = 4), while it in patients from BSI-suspected was achieved by identifying an unsuspected pathogen not detected by blood culture (n = 1).
[0900] We demonstrated that Nanopore mNGS can identify pathogens clinically relevant to the acute infection in twice as many patients as gold-standard blood culture and that the method detects a broad range of aetiologies. In summary, this study provides a proof-of-concept for the use of Nanopore cfDNA metagenomics as a diagnostic tool for patients with a BSI, and the study encourages the method to be evaluated in larger cohorts and subjected to comprehensive analytical validation.
[0901] Example 2 - Identification of optimal buffer compositions using defined DNA fragments
[0902] In this example, organic solvent isopropanol and chaotropic solute guanidine thiocyanate were tested at various concentrations in the long fragment binding buffer (LFB) buffer to preferentially increase the affinity of longer DNA fragments to a hydrophilic support, compared with shorter fragments that remain in solution, see e.g., Table 1. The LFB composition was further evaluated for the benefits of adding various concentrations of nonionic surfactants (e.g. Tween-20) to solubilize contaminating lipids, sugars and proteins, easing working routines. Finally, a buffer component was added to reduce chemical DNA degradation (e.g. Tris), and a divalent ion scavenger (e.g., EDTA) was added to reduce enzymatic degradation. A similar evaluation was performed for the short fragment binding buffer (SFB) focusing on binding of DNA fragments of 10-100 nucleotides by increasing the amount of chaotropic solute and organic solvent, see e.g., Table 2.
[0903] To identify the optimal buffer compositions in these two steps e.g. step 1 LFB and step 2 SFB a defined composition of DNA fragments was evaluated in the form of commercially available DNA ladders.
[0904] Methods
[0905] 1. Defining the optimal ratio of solid support, sample and binding buffers as a function of isopropanol
[0906] A DNA control was produced by mixing 38 pL of nuclease-free water with 12 pL of commercially available ladder (Promega, G4511), resulting in a concentration of 43.6 ng / pL DNA (input of 523 ng).
[0907] The samples listed in the Table 3 below were prepared by mixing x-fold 50 pL of cleanNGS (alternating layers of magnetite and negatively charged carboxylated (non-styrene) polymer coatings around a polystyrene core) bead solution (CleanNA, NL) with a total of 50 pL DNA control, and in some cases x-fold 50 pL 100%v / v isopropanol.
[0908] Table 3 - Investigated bead to sample ratios and isopropanol
[0909]
[0910] The samples were then incubated for 10 min at room temperature. Pelleted on a magnet, washed 2x with 80%v / v EtOH and eluted in 25 pL nuclease-free water
[0911] The effectiveness to specifically capture about 50 nucleotides DNA fragments (corresponding to the average of microbial cfDNA) was assessed by electropherogram (Tapestation, HS D1000 screen tape) as outlined in the results.
[0912] 2. Selection of optimal solid support composition in human plasma
[0913] Plasma for evaluation was prepared by centrifugation of whole blood (healthy donors) at 1,600 g (4°C) for 10 minutes within 1 hour from blood draw. The plasma supernatant was subsequently centrifuged again at 16,000 g (4°C) for 10 minutes and the new supernatant was flash-frozen at -80°C until use.
[0914] 245 pL freshly frozen plasma was carefully thawed and 5 pL HyperLadder 25bp (Cat#BIO-33057; Bioline Reagents Ltd, UK) was added in duplicate for each of the below mixtures.
[0915] The below mixtures in Table 4 were prepared to evaluate the influence of bead size (400 nm, 700 nm and 1:1 ratio mixture of 400 / 700 nm), keeping the surface area constant, on the size selection of a DNA ladder in plasma using silica coated superparamagnetic beads. Theevaluation was performed for both removal of long fragments (LFB buffer) and capturing of short fragments (SFB buffer).
[0916] Table 4 - Overview of LFB and SFB buffers for evaluating different bead sizes
[0917]
[0918] To each of the 250 pL ladder-spiked plasma samples, 737 pL LFB buffer stock solution was added to achieve the desired final concentrations in Table 4 adding 11 pL of Proteinase K (e.g., Thermo Scientific, EO0491) with incubation at 53°C for 1 H.
[0919] The 3 different bead mixtures (e.g. 400 nm, 700 nm and 400+700 nm) were added as 13.7 pL followed by vortexing and incubation at room-temperature for 10 min as two replicates. The magnetic beads were pelleted, and the supernatant transferred to a separate tube for the purification described below to obtain the shorter DNA fragments.
[0920] The bead pellets were washed with three different buffers, Wash #1 (3M GTC, about 5%v / v Tween20, about 40mM Tris, about 2mM EDTA and 30%v / v isopropanol), Wash #2 (about 40mM Tris, about 2mM EDTA and about 80%v / v EtOH) and 100%v / v EtOH followed by elution in standard TE-buffer to obtain DNA long fragments bound at this step, for later analysis by Tapestation.
[0921] The supernatants from the above capture had 600 pL of SFB buffer added, followed by 22 pL of the 3 different bead mixtures followed by vortexing and incubation at room temperature for 10 min. Washing of the beads, and elution of the final short-fragment DNA was performed as outlined above.
[0922] All samples were assessed by electropherogram (Tapestation, HS D1000 screen tape) as outlined in the results.
[0923] 3. Selection of optimal surfactants and their concentrations
[0924] To evaluate the optimal surfactants, a number of commercially available reagents were evaluated, including non-ionic detergents (Tween-20, Tween-80, NP-40, Triton X-100, Triton-X-114, Brji-35 and Brij-48) and zwitterionic detergents (CHAPS and CHAPSO) for their ability to solubilize biomaterials in the sample (proteins, lipids, sugars etc.) with the goal of enhanced processing capability.
[0925] Plasma for evaluation was prepared by centrifugation of whole blood (healthy donors) at 1,600 g (4°C) for 10 minutes within 1 hour from blood draw. The plasma supernatant was subsequently centrifuged again at 16,000 g (4°C) for 10 minutes and the new supernatant was flash-frozen at -80°C until use
[0926] The freshly frozen plasma was carefully thawed and 5 pL HyperLadder 25bp (Cat#BIO-33057; Bioline Reagents Ltd, UK) was added to 245 pL in duplicate for each of the below mixtures.
[0927] To each of the 250 ladder-spiked plasma samples, 737 pL LFB buffer stock solutions were added. 13.7 pL 700nm silica coated superparamagnetic beads e.g., was added followed by vortexing and incubation at room-temperature for 10 min as two replicates. The magnetic beads were pelleted, and the supernatant transferred to a separate tube for analysis on SDS-PAGE for protein removal detection, and the following purification to obtain the shorter DNA fragments.
[0928] The bead pellets were washed with three different buffers, Wash #1 (3M GTC, about 5%v / v Tween20, about 40mM Tris, about 2mM EDTA and 30%v / v isopropanol), Wash #2 (about 40mM Tris, about 2mM EDTA and 80%v / v EtOH) and 100%v / v EtOH followed by elution in standard TE-buffer to obtain fragments bound at this step, for later analysis by Tapestation.
[0929] The supernatants from the above capture had 600 pL of SFB buffer added to achieve the desired final concentrations of the evaluated surfactants at 0.5x, lx, 2x and lOx their critical micelle concentration (CMC in mM). Then 22 pL of the 700nm silica coated superparamagnetic beads were added followed by vortexing and incubation at room temperature for 10 min. The beads were pelleted, and the supernatant was saved for SDS-PAGE based analysis. Washing of the beads, and elution of the final short-fragment DNA was performed as outlined above.
[0930] Samples were evaluated by visual inspection in regard to the ability of the beads to form a pellet when exposed to a magnet combined with SDS-PAGE of extracted proteins.
[0931] 4. Selection of optimal buffer type, pH and stabilization agents
[0932] To evaluate the optimal buffer type, pH and stabilization agents, a subset of buffers known to be beneficial in the literature for DNA purification and storage were tested (nuclease free water, TRIS (pH 4,6,8), TRIS / EDTA (TE, pH 8), TRIS / Citrate (pH 8), NaCI / TRIS / EDTA (STE, pH 8), NaCI / TRIS / EDTA / TWEEN (STET, pH 8), potassium phosphate / DMSO / glycerol (KPDG, pH 8).
[0933] Plasma for evaluation was prepared by centrifugation of whole blood (healthy donors) at 1,600 g (4°C) for 10 minutes within 1 hour from blood draw. The plasma supernatant was subsequently centrifuged again at 16,000 g (4°C) for 10 minutes and the new supernatant was flash-frozen at -80°C until use
[0934] The freshly frozen plasma was carefully thawed, in some instances the plasma was diluted xlO in the ten prepared buffers and 5 pL HyperLadder 25bp (Cat#BIO-33057; Bioline Reagents Ltd, UK) was added to 245 pL of these dilution in duplicates. Similar samples were prepared, using 245 pL of the prepared buffers with no plasma, adding 5 pL of the Hyperladder. All samples were then incubated for 1 hr at room temperature.
[0935] To each of the 250 pL ladder-spiked plasma samples, 737 pL LFB buffer stock solution was added. 13.7 pL 700nm silica coated superparamagnetic beads were then added, followed by vortexing and incubation at room-temperature for 10 min. The magnetic beads were pelleted, and the supernatant transferred to a separate tube for the following purification to obtain shorter DNA fragments.
[0936] The bead pellets were washed with three different buffers, Wash #1 (3M GTC, about 5%v / v Tween20, about 40mM Tris, about 2mM EDTA, 30%v / v isopropanol), Wash #2 (about 40mM Tris, about 2mM EDTA, 80%v / v EtOH) and 100%v / v EtOH followed by elution in standard TE-buffer to obtain fragments bound at this step, for later analysis by Tapestation.
[0937] The supernatants from the above capture had 600 pL of SFB buffer added. Then 22 pL of the 700nm silica coated superparamagnetic beads was added followed by vortexing and incubation at room temperature for 10 min. Washing of the beads, and elution of the final short-fragment DNA was performed as outlined above.
[0938] 5. Selection of optimal isopropanol concentration for the LFB buffer as a function of DNA input
[0939] Plasma for evaluation was prepared by centrifugation of whole blood (healthy donors) at 1,600 g (4°C) for 10 minutes within 1 hour from blood draw. The plasma supernatant was subsequently centrifuged again at 16,000 g (4°C) for 10 minutes and the new supernatant was flash-frozen at -80°C until use.
[0940] To evaluate the effects of varying amounts of cfDNA in the plasma sample, 245 pL freshly frozen plasma mixed with HyperLadder 25bp for final concentrations of 0,017 ng / pL (3.9 ng), 0,025 (6,25ng), 0,075 (18,75ng), 0,2 ng / pL (50 ng), or 3 ng / pL (750 ng) (Cat#BIO-33057; Bioline Reagents Ltd, UK) in duplicate for each of the below mixtures.
[0941] The below LFB buffers were prepared to evaluate the influence of different %v / v isopropanol on the size selection of the DNA ladder in plasma using 700 nm or 400 nm + 700 nm mixed silica coated superparamagnetic beads. The evaluation was performed for removal of long fragments (LFB buffer) keeping the final binding conditions (SFB buffer) constant for the capturing of short fragments.
[0942] Table 5 - Overview of LFB and SFB buffers for evaluating different isopropanol %v / v and DNA load
[0943]
[0944] To each of the 250 pL ladder-spiked plasma samples, 737 pL LFB buffer stock solution was added to achieve the desired final concentrations in Table 5 with incubation at 53C for 1 H.
[0945] 13.7 pL silica coated superparamagnetic beads were added followed by vortexing and incubation at room-temperature for 10 min as two replicates. The magnetic beads werepelleted, and the supernatant transferred to a separate tube for the purification described below to obtain the shorter DNA fragments.
[0946] The bead pellets were washed with three different buffers, Wash #1 (3M GTC, about 5%v / v Tween20, about 40mM Tris, about 2mM EDTA, 30%v / v isopropanol ), Wash #2 (about 40mM Tris, about 2mM EDTA, 80%v / v EtOH) and 100%v / v EtOH followed by elution in standard TE-buffer to obtain DNA long fragments bound at this step, for later analysis by Tapestation.
[0947] The supernatants from the above capture had 600 pL of SFB buffer added, followed by 22 pL of silica coated superparamagnetic beads followed by vortexing and incubation at room temperature for 10 min. Washing of the beads, and elution of the final short-fragment DNA was performed as outlined above.
[0948] All samples were assessed by electropherogram (Tapestation, HS D1000 screen tape) as outlined in the results.
[0949] 6. Selection of optimal chaotropic salt and its concentration for the LFB buffer
[0950] To evaluate the optimal chaotropic salt and their relevant concentration, a number of commercially available reagents were evaluated (guanidine thiocyanate, guanidinium chloride, lithium perchlorate, urea, sodium dodecyl sulfate, sodium iodide) for their ability to selectively enhance DNA fragment binding depending on size.
[0951] Plasma for evaluation was prepared by centrifugation of whole blood (healthy donors) at 1,600 g (4°C) for 10 minutes within 1 hour from blood draw. The plasma supernatant was subsequently centrifuged again at 16,000 g (4°C) for 10 minutes and the new supernatant was flash-frozen at -80°C until use
[0952] The freshly frozen plasma was carefully thawed and HyperLadder 25bp (Cat#BIO-33057; Bioline Reagents Ltd, UK) was added to 245 pL in duplicate for each of the below mixtures for a final DNA input of 0.2 ng / pL (50ng).
[0953] To each of the 250 pL ladder-spiked plasma samples, 737 pL LFB buffer stock solution was added as prepared below (Table 6). 13.7 pL 700nm silica coated superparamagnetic beads were added followed by vortexing and incubation at room-temperature for 10 min as two replicates. The magnetic beads were pelleted, and the supernatant transferred to a separate tube for shorter DNA fragment capture using the below SFB buffer for all experiments.
[0954] Table 6 - Overview of LFB and SFB buffers for evaluating different isopropanol %v / v and DNA load
[0955]
[0956] The bead pellets were washed with three different buffers, Wash #1 (3M GTC, about 5%v / v Tween20, about 40mM Tris, about 2mM EDTA, 30%v / v isopropanol), Wash #2 (about 40mM Tris, about 2mM EDTA, 80%v / v EtOH) and 100%v / v EtOH followed by elution in standard TE-buffer to obtain fragments bound at this step, for later analysis by Tapestation.
[0957] The supernatants from the above capture had 600 pL of SFB buffer added. Then 22 pL of the 700nm silica coated superparamagnetic beads was added followed by vortexing and incubation at room temperature for 10 min. Washing of the beads, and elution of the final short-fragment DNA was performed as outlined above.
[0958] 7. Selection of optimal isopropanol concentration for the SFB buffer as a function of DNA input
[0959] Plasma for evaluation was prepared by centrifugation of whole blood (healthy donors) at 1,600 g (4°C) for 10 minutes within 1 hour from blood draw. The plasma supernatant was subsequently centrifuged again at 16,000 g (4°C) for 10 minutes and the new supernatant was flash-frozen at -80°C until use.
[0960] To evaluate the effects of varying amounts of cfDNA in the plasma sample, 245 pL freshly frozen plasma mixed with HyperLadder 25bp for final concentrations of 0,025 (6,25ng), 0,075 (18,75ng), 0,2 ng / pL (50 ng), (Cat#BIO-33057; Bioline Reagents Ltd, UK) in duplicate for each of the below mixtures.
[0961] The LFB buffer was kept constant for size selection of the DNA ladder directly in plasma using 700 nm or 400 nm + 700 nm mixed silica coated super-paramagnetic beads. A range of different %v / v isopropanol was then evaluated, to locate the optimal composition for retaining 50-100 nucleotides DNA fragments but removing smaller contaminants that could affect further processing of the sample.
[0962] Table 7 - Overview of LFB and SFB buffers for evaluating different isopropanol %v / v and DNA load
[0963]
[0964] To each of the 250 pL ladder-spiked plasma samples, 737 pL LFB buffer stock solution was added to achieve the desired final concentrations in Table 7 with incubation at 53C for 1 H.
[0965] 13.7 pL silica coated superparamagnetic beads were added followed by vortexing and incubation at room-temperature for 10 min as two replicates. The magnetic beads were pelleted, and the supernatant transferred to a separate tube for the purification described below to obtain the shorter DNA fragments.
[0966] The bead pellets were washed with three different buffers, Wash #1 (3M GTC, about 5%v / v Tween20, about 40mM Tris, about 2mM EDTA, 30%v / v isopropanol ), Wash #2 (about 40mMTris, about 2mM EDTA, 80%v / v EtOH) and 100%v / v EtOH followed by elution in standard TE-buffer to obtain long DNA fragments bound at this step, for later analysis by Tapestation.
[0967] The supernatants from the above capture had 600 pL of SFB buffer added with varying %v / visopropanol as listed, followed by 22 pL of silica coated superparamagnetic beads followed by vortexing and incubation at room temperature for 10 min. Washing of the beads, and elution of the final short-fragment DNA was performed as outlined above.
[0968] Results
[0969] 1. Defining the optimal ratio of solid support, sample and binding buffers as a function of isopropanol
[0970] The electropherograms from the performed Tapestation analysis can be found on Figure 2 for different ratios of sample and bead solution, while varied sample to bead ratios in combination with isopropanol can be found on Figure 3. From analyses of the area-under-the-curve, using the manufacturer provided analysis software (Agilent, US), it is evident that addition of isopropanol in any amount (20-43%v / v) is beneficial for capturing short DNA fragments <75bp, while the most optimal sample:beads ratio seems to be 3x excess e.g. 50pL sample to 150 pL bead solution with 150 pL of 100 v / v% isopropanol, resulting in a final isopropanol concentration of 43%v / v see Table 8.
[0971] Table 8 - Overview of area-under-the-curve (AUC) using manufacturer supplied software (Agilent) for a Tapestation for 50 nucleotides and 75bp fragments. Ladder direct load control is the true 50 nucleotides / 75bp ratio.
[0972]
[0973] 2. Selection of optimal solid support composition in human plasma
[0974] The comparison of 400 nm, 700 nm and mix 400 / 700 showed variable total recovery levels (75-81%, 82-88%, 88-92%), but in general better for 700nm beads. The general trend was better removal of long fragments (>150 nucleotides) for 700 nm beads (87-88%, 97-98%) when compared to 400 nm (57-90%, 58-96%) and the mixed (41-76%, 42-85%), see Figure 4 and Table 9. Recovery of short fragments (50 nucleotides and 75bp) was clearly superior for the 700 nm beads (81-87%, 83-87%) when compared to the smaller 400 nm beads (58-63%, 27-31%) and the mix (68-70%, 33-35%). All in all the 700 nm beads seem to be significantly superior to the 400 or 400 / 700 mix, even though the 400 sized bead has better binding capacity pr mg of bead.
[0975] Table 9 - Long fragments (>150 nucleotides) removal efficiency and capture efficiency of short fragments ( <100 nucleotides) for various bead sizes
[0976]
[0977] 3. Selection of optimal isopropanol concentration for the LFB buffer as a function of DNA input
[0978] Various isopropanol concentrations were evaluated as part of the LFB buffer (3M GTC, about 5%v / v Tween20, about 12mM Tris, about 5mM EDTA 10), with electropherograms for 0-27.5%v / v shown on Figure 5, Figure 6, Figure 7 and Figure 8. From these presented data it is evident that the binding of DNA fragments in the 50-200 nucleotides range is surprisingly sensitive to the isopropanol in the sample, with an optimum around 15-25%v / v, more specifically 20-22, 5%v / v slightly dependent on nature of the solid support (e.g. 700 nm size vs mixed 400+700 nm). At 22.5%v / v almost all fragments larger than 150 nucleotides are retained on 700 nm silica coated superparamagnetic beads, while <100 nucleotides mostly remain in the supernatant. This selection at 750 ng of input DNA is highly relevant for sequence independent enrichment of microbial cfDNA (50-80 nucleotides fragments) over human cfDNA (>150 nucleotide fragments). Capture of the short DNA fragments is clearly possible with the applied SFB buffer containing 30%v / v isopropanol.
[0979] To ensure the applicability of the binding buffers, DNA levels corresponding to the range expected from clinical human samples, where the subject is showing signs of sepsis was evaluated (Figure 9, Figure 10 and Figure 11). From these collected data it is evident that the size of fragments retained by the LFB buffer in combination with 700 nm silica coated superparamagnetic beads is highly dependent on the amount of DNA in the sample (4-750ng) due to enhanced binding of larger fragments over short fragments. Therefore, the amount of isopropanol (20-22.5%v / v) that is optimal for very high DNA levels (750 ng) potentially not reached in the clinical setting, results in problematic binding of the 50-100 nucleotides fragments in the LFB buffer to the silica beads at more physiological DNA levels. It was therefore evaluated to use 0%v / v and 10%v / v isopropanol, where it was evident that 0%v / v isopropanol allowed all fragments below 250-300 nucleotide fragments to remain in the supernatant during the long fragment binding step, reducing the enrichment level of the 50-80 nucleotide fragments significantly. The 10%v / v isopropanol, when applied on plasma spiked samples, evaluated in more physiological relevant range of DNA (6,25-50ng) showed good performance for all concentrations tested, enabling binding of all DNA above 100 nucleotides fragments to the beads in the long fragment binding step, while allowing the 50-80 nucleotides fragments to remain in solution for capture in the short fragment binding step. Based on these results, the most optimal isopropanol %v / v during long-fragment binding is in the 10-25%v / v range, depending on the input DNA amount, with 10%v / v being optimal when working in the physiological range of DNA (5-50ng).
[0980] It was further evaluated how a preceding DNA extraction step would affect the cut-off for size selection, by mixing the desired DNA concentration in water only. Here it was evident that a small size selection shift is observed so that a slightly higher isopropanol concentration of 12.5%v / v during long fragment binding is optimal for some applications (Figure 12).
[0981] Example 3 - Further optimization of isopropanol content
[0982] Methods
[0983] Samples were prepared as described in Example 2, with alterations to the binding buffer labelled long binding buffer, to evaluate the effect of changes to the isopropanol concentration on the recovery of the DNA fragments of different sizes.
[0984] Binding buffers were prepared with 10, 12.5, 15.0, 17.5 and 20 %v / v isopropanol in the long fragment binding buffer to evaluate the optimal conditions for removal of longer DNA fragments larger than 75 bp.
[0985] Samples were prepared using human plasma with a total DNA concentration (human plasma + ladder (HyperLadder 25bp from Meridian Biosciences Cat. No. BIO-33057)) of 75 ng / mL.
[0986] Remaining steps are carried out as described in Example 2 and the distribution of DNA sizes were evaluated using the Tapestation assay as also described in Example 2 and the results are presented in figures 16 and 17.
[0987] Results
[0988] To have an optimal DNA isolation of the shorter DNA fragments, it was decided that the isolation method should live up to a set of quality parameters, which defines when the sample isolation is sufficiently efficient in order to enable an optimal subsequent sequencing and bioinformatic analysis.
[0989] The four initial quality control check points where:
[0990] ⢠That the concentration of the DNA in the lOObp peak should not amount to more than 25% of the concentration of DNA in the 50bp peak in the enriched sample
[0991] ⢠The concentration of the 50 bp peak should not amount to more than 7.5% of the concentration of DNA in the 150bp peak in the depleted sample
[0992] ⢠The total DNA concentration in the enriched sample should surpass 0.03 ng / pL
[0993] ⢠Total recovery of input DNA should be at least 70% across the enriched and depleted fractions.
[0994] The results are shown in Figure 17, Figure 18 and table 10.
[0995] Table 10 - Evaluation of the optimal isopropanol concentration for binding and removal of longer DNA fragments
[0996]
[0997]
[0998] As can be seen from the table the optimal conditions were found to be the 15%v / v isopropanol, as other sample conditions showed either a too large proportion of the lOObp peak in the enriched sample, or too large proportion of the 50bp peak in the depleted sample.
[0999] Accordingly, the optimal isopropanol concentration was found to be 15%v / v, however it was seen that both 10%v / v and 12.5%v / v isopropanol allowed for sufficient recovery, concentration and absence of the smaller fractions in the depleted sample. However, the 10%v / v and 12.5%v / v isopropanol sample did not provide sufficient removal of the larger fragments as can be seen in Table 10 and Figure 17 and 18.
[1000] Example 4 - Further optimization of GTC content
[1001] Methods
[1002] Samples were prepared as described in Example 2, with alterations to the binding buffer labelled long binding buffer, to evaluate the effect of changes to the GTC concentration.
[1003] Binding buffers were prepared with 0.5, 1.0, 2.0 and 3.0M GTC in the long fragment binding buffer to evaluate the optimal conditions for removal of longer DNA fragments larger than 75 bp.
[1004] Samples were prepared using human plasma with a total DNA concentration (human plasma + ladder) of 75 ng / mL
[1005] Remaining steps are carried out as described in Example 2 and the distribution of DNA sizes were evaluated using the Tapestation assay as also described in Example 2 and the results are presented in Figure 19 and in table 11.
[1006] Results
[1007] Table 11 - Evaluation of the optimal GTC concentration for binding and removal of longer DNA fragments
[1008]
[1009] As can be seen from the table the optimal conditions were found to be the 3M GTC, as other sample conditions showed either a too large proportion of the lOObp peak in the enriched sample, and too low recovery (0.5M and 1.0M samples).
[1010] Accordingly, the optimal GTC concentration was found to 3.0M, however it was seen that 2.0M also allowed for sufficient recovery, concentration and absence of the smaller fractions in the depleted sample. However, the 2.0M GTC sample did not provide sufficient removal of the larger fragments as can be seen in Table 11 and Figure 19, according to the quality parameters set out in Example 3.
[1011] Example 5 - Test of alternative to GTC
[1012] Methods
[1013] The experiments in the examples were performed as described in Example 2, however substituting GTC with GHCI at the concentrations indicated in table 12.
[1014] Samples were prepared using human plasma with a total DNA concentration (human plasma + ladder) of 75 ng / mL
[1015] The distribution of DNA sizes were evaluated using the Tapestation assay as also described in Example 2 and the results are presented in Figure 20.
[1016] Results
[1017] An alternative to GTC was evaluated using different concentrations of GHCI as indicated in
[1018] Table 12.
[1019] Table 12 - Evaluation of the alternative chaotropic solute Guadinium hydrochloride (GHCI)
[1020]
[1021]
[1022] As can be seen in table 12 and Figure 20, GHCI buffers all resulted in a non-specific isolation of the smaller DNA fragments, with both low recovery and a high presence of larger fragments in the enriched sample, and the GHCI samples therefore did not meet the requirements set out by the quality parameters provided in Example 3.
[1023] Accordingly, the results suggests that GTC provides the optimal conditions for the isolation of smaller DNA fragments when compared to GHCI.
[1024] Example 6 - Further optimization of surfactant content
[1025] Methods
[1026] The experiments in the examples were performed as described in Example 2, however modifying the Tween20 concentrations (0.4, 4 or 8%v / v) to identify the optimal Tween20 concentration for removal of the longer DNA fragments from the sample.
[1027] The distribution of DNA sizes was evaluated using the Tapestation assay as also described in Example 2 and the results are presented in table 13 and Figure 21.
[1028] Results
[1029] Table 13 - Evaluation of the optimal surfactant concentration
[1030]
[1031] As can be seen in table 13, the 4%v / v Tween20 sample provided the most optimal conditions for isolation of the populations of smaller fragments of DNA, where the high surfactant sample (8% Tween20) resulted in a higher degree of the smaller fragments being in the depleted sample i.e., providing a lower recovery, while the low surfactant sample (0.4%v / v Tween20) resulted in both low recovery an unspecific recover as seen from the presence of the lOObp population in the enriched sample.
[1032] Example 7 - Distribution of cfDNA in clinical samples
[1033] Microbial cfDNA are degraded differently from human cfDNA, causing microbial cfDNA reads to be shorter.
[1034] Based on sequencing data from a diverse patient population the distribution of microbial cfDNA read lengths is compared to the lengths of all reads sequenced through shotgun metagenomics sequencing of cfDNA from plasma (cf. Fig. 22). The data was obtained from patients at the emergency care ward at Aalborg University Hospital including a diverse group of patients presenting with suspected infection and a wide variety of co-morbidities (e.g., diabetes, COPD, alcoholism, etc.). Patients were included on the ordering of standard infectious disease diagnostics (i.e. blood culture). 40 patients were recruited and analysed, along with 13 negative controls (water) and 14 non-infected donor samples (Table 14).
[1035]
[1036] All reads were processed using cutadapt to perform demultiplexing and enable processing of per-sample reads, as well as trimming and eliminating sequencing artefacts such as adapters and multiplex barcodes.
[1037] These trimmed reads represent the DNA that was actually present in the plasma sample, including both human and microbial DNA. Human DNA reads were eliminated using standard alignment tools against a human reference genome (T2T). Reads remaining after elimination of human reads are deemed microbial.
[1038] Read length distribution plots were generated by performing kernel density estimations per read length in 1 base pair increment for each sample individually, then visualizing the median density and 95% confidence interval per sample type (negative control, donor, and patient samples) for microbial reads and the total reads (cf. Fig. 22). Cumulative distributions are visualized as well, depicting the sum of the densities of reads at 1 base pair increments.
[1039] The data shows that the bacterial cfDNA fraction has an overall smaller fragment size compared to the overall cfDNA found in the sample. The data also shows that in the patient samples, the 75 bp cutoff improves data quality by reducing the noise from the non-bacterial cfDNA fragments in the sample.
[1040] Example 8 - Clinical validation of cfDNA enrichment
[1041] Aim
[1042] The aim of the present example is to evaluate the clinical application of using the long fragment binding buffer shown in the previous examples in the extraction of microbial cfDNA from clinical plasma samples.
[1043] Methods
[1044] Six clinical patient plasma samples from patients, one donor sample spiked with fragmented E. coli DNA and a water sample were treated either with a standard protocol for cfDNA extraction, where the cfDNA was extracted directly using conditions similar to the short fragment buffer to extract all cfDNA, or by a size selection protocol, obtaining the short cfDNA fraction directly from the plasma samples.
[1045] In the standard protocol a binding buffer comprising 30%v / v isopropanol and 3M GTC was added to each sample, followed by mixing with 700nm beads (Cytiva, 29357373). The mixed sample was spun down briefly, and the beads were pelleted on the magnet and the supernatant was discarded. Thereafter, washes of the pelleted beads were performed three times using a first wash buffer (3 M GTC, 30 v / v% isopropanol, 5% Tween20, 50 mM Bis-Tris pH 6, 4 mM EDTA), a second wash buffer (50 mM Tris-HCI, 0.5 mM EDTA, 80% EtOH), and finally using 100% Ethanol.
[1046] The washed samples were spun down and pelleted using a magnet. Once the ethanol was evaporated the bound DNA was eluted by resuspension in an elution buffer (10 mM Tris-HCI, 0.1 mM EDTA) followed by vortexing and removal of the beads by pelleting the beads using a magnet, and collection of the supernatant.
[1047] In the size selection protocol, the plasma samples were treated as in the standard protocol, only preceded by a step where the long DNA was bound using the long fragment DNA binding buffer as disclosed herein, comprising about 10%v / v isopropanol and 3M GTC.
[1048] Following the total extraction (standard protocol) or size specific extraction, the remaining DNA was sequenced, and the sequencing reads were analysed and aligned to the respective microbe.
[1049] The number of reads from the relevant pathogen per million human or total reads (Table 19) is calculated for each sample and normalized to the standard treatment.
[1050] Table 19
[1051]
[1052] Results
[1053] The size selection protocols provided a remarkable enrichment across all samples as is shown in figure 24 and 25, both when evaluated in relation to the number of human reads where the mean enrichment was 6.9x and when evaluated in relation to the number of total reads where the mean enrichment was 6. lx. This highlights that human DNA is successfully discarded in the size selection protocol without compromising the performance of the assay end to end.
[1054] Accordingly, the present example shows that the size selection enrichment enriches the pathogen cfDNA in clinical samples compared to standard protocols.
Claims
1. CLAIMS1. A composition that facilitates the preferential binding of long DNA fragments with an approximate size of 75 nucleotides or more, such as 75 nucleotides or more, 80 nucleotides or more, 90 nucleotides or more, or such as 100 nucleotides, or more to a matrix.
2. The composition according to claim 1, wherein the composition is considered to facilitate preferential binding of long DNA fragments with an approximate size of 75 nucleotides or more, when an analysis comprisinga. adding a pool of DNA fragments comprising DNA fragments of different sizes in a concentration of about 0.075ng / µl, e.g., ranging from 25-500 bp such as e.g., a DNA ladder to the DNA binding buffer in the presence of a DNA binding matrix to facilitate binding of DNA fragments,b. isolating the bound DNA from the unbound DNA to obtain an isolated DNA fraction and an unbound DNA fraction, andc. analyse the size distribution and concentration of the isolated DNA fraction and the unbound DNA fraction,shows thati. the concentration of the 100bp peak in the unbound fraction has a concentration of less than 25% of the 50bp peak,II. the concentration of the 50bp peak in the isolated DNA fraction is less than 7.5% of the concentration of the 150bp peak,ill. the concentration of DNA in the unbound DNA fraction when measured e.g., using Qubit, is higher than 0.03 ng / pL, and iv. the total recovery of DNA in the two fractions is higher than 70%.
3. The composition according to claim 2, wherein DNA fragments of different distinct sizes is a DNA ladder, preferably a ladder comprising fragments with a size of 25bp 50 bp, 75 bp, 100 bp, 125 bp, 150 bp, 175 bp, 200 bp, 250 bp, 300 bp, 400 bp and 500 bp, such as e.g., a HyperLadder 25 bp.
4. The composition according to claim 2 or 3, wherein the analysis of the size distribution and concentration of the isolated DNA fraction and the unbound DNA fraction is performed using Tapestation.
5. The composition according to any one of the preceding claims, wherein the composition comprisesa. an organic solvent,b. a chaotropic solute,c. a non-ionic surfactant,d. optionally, a chelating agent ande. optionally, a buffering agent, andoptionally, wherein the pH of the composition is in the range 5.0-8.5, preferably pH 7.5-8.5.
6. The composition according to any one of the preceding claims, wherein the composition comprisesa. an organic solvent being isopropanol in an amount of 10-18 %v / v, preferably 12-17%v / v,b. a chaotropic solute being guanidine thiocyanate (GTC),c. a non-ionic surfactant being Tween20,d. optionally, a chelating agent ande. optionally, a buffering agent, andoptionally, wherein the pH of the composition is in the range 5.0-8.5, preferably pH 7.5-8.5.
7. The composition according to claim 6, wherein GTC is present in an amount of 2-4M, preferably about 3M.
8. The composition according to claim 6 or 7, wherein Tween20 is present in an amount of 0.1-10%v / v, preferably about 4%v / v.
9. The composition according to any one of claims 6-8, wherein the chelating agent is present in an amount of 1-10 mM.
10. The composition according to any one of claims 6-9 wherein the chelating agent is EDTA.
11. The composition according to any one of claims 6-10, wherein the buffering agent is present in an amount of 1-100 mM.
12. The composition according to any one of claims 6-11, wherein the buffering agent is Tris.
13. The composition according to any one of claims 6-12, wherein the composition comprises about 15 %v / v isopropanol, about 3 M of guanidine thiocyanate, 4-6 mM EDTA, 10-20 mM Tris, and has a pH of 5.0-8.5, preferably 7.5-8.5.
14. A method of validating the DNA binding properties of a DNA binding buffer, comprisinga. adding a pool of DNA fragments comprising DNA fragments of different sizes such as e.g., ranging from 25-500 bp such as e.g., a DNA ladder, in a concentration of about 0.075ng / pL to the DNA binding buffer in the presence of a DNA binding matrix to facilitate binding of DNA fragments,b. isolating the bound DNA from the unbound DNA to obtain an isolated DNA fraction and an unbound DNA fraction, andc. analyse the size distribution and concentration of the isolated DNA fraction and the unbound DNA fraction.
15. The method according to claim 14, wherein the method is for validating the binding properties of a DNA binding composition that facilitates the preferential binding of long DNA fragments with an approximate size of 75 nucleotides or more, wherein the composition is considered to facilitate the preferential binding of long DNA fragments with an approximate size of 75 nucleotides or more when,i. the concentration of the 100bp peak in the unbound DNA fraction has a concentration of less than 25% of the 50bp peak,II. the concentration of the 50bp peak in the isolated DNA fraction is less than 7.5% of the concentration of the 150bp peak, ill. the concentration of DNA in the unbound DNA fraction when measured e.g., using Qubit, is higher than 0.03 ng / pL, andiv. the total recovery of DNA in the two fractions is higher than 70%.
16. The method according to claim 14 or 15, wherein DNA fragments of different distinct sizes is a DNA ladder, preferably a ladder comprising fragments with a size of at least 50 bp, 100 bp and 150 bp, preferably comprising fragments with a size of 25bp 50 bp, 75 bp, 100 bp, 125 bp, 150 bp, 175 bp, 200 bp, 250 bp, 300 bp, 400 bp and 500 bp, such as e.g., a HyperLadder 25 bp.
17. The method according to any one of claims 14 to 16, wherein the analysis of the size distribution and concentration of the isolated DNA fraction and the unbound DNA fraction is performed using Tapestation or LabChip.
18. The method according to any one of claims 14-17, wherein the DNA binding buffer is a composition according to any one of claims 1-13.
19. A kit for isolating microbial cell-free DNA fragments, comprisinga. a vial comprising a composition comprisingi. a chaotropic solute,ii. a non-ionic surfactant,iii. optionally, a chelating agent andiv. optionally, a buffering agent, andb. optionally, one or more matrices capable of binding DNA, optionally contained within a vial, andc. optionally, means for performing the microbial cell-free DNA isolation, d. optionally, a DNA ladder, ande. instructions for preparing the composition according to claims 1-13, f. Optionally, instructions for isolating microbial cfDNA fragments using said kit.
20. The kit according to claim 19, further comprisinga. one or more vials comprising wash buffer(s) andb. one or more vials comprising elution buffer(s).
21. The kit according to claim 19 or 20, wherein the kit comprises a first and second washing buffer, whereina. the first washing buffer comprises 15-40%v / v isopropanol, 1-8M or l-10%v / v of the chaotropic solute, a chelating agent in an amount of 1-10 mM, and a non-ionic surfactant in an amount of 0.1-10%v / v andb. the second washing buffer comprises buffered ethanol, such as a composition comprising 1-lOOmM Tris, pH 8.0, l-5mM EDTA and 80%v / v Ethanol.
22. The kit according to any one of claims 19-21, wherein the kit comprises the vial comprising the elution buffer, and the elution buffer is a TE buffer, a TAE buffer or a TBE buffer.
23. The kit according to any one of claims 15-22, for use in the enrichment of predominantly microbial cfDNA fragments, preferably with a predominant size of 75 bp or less.
24. A method for enriching microbial cfDNA from a sample comprising host and microbial cfDNA using a composition according to any one of claims 1-13.
25. The method according to claim 24, wherein the microbial cfDNA has a predominant size of 75 bp or less.
26. A method for treating a patient suffering from an infection, the method comprisinga. Isolation of microbial cfDNA using a composition according to any one of claims 1-13, or a kit according to any one of claims 19-23 from a patient sample, andb. Identification of at least one pathogen based on the isolated microbial cfDNA, andc. suggesting a clinically relevant effective antimicrobial treatment regimen targeting said pathogen.
27. A method for determining the identity of at least one pathogen responsible for an infection in a patient, comprising the following steps:a) obtaining a sample from the patient,b) optionally, extracting the total DNA from said sample,c) enriching the microbial cfDNA molecule fraction in the extracted DNA using a composition according to any one of claims 1-13 or a kit according to any one of claims 19-23,d) optionally, amplifying the enriched microbial cfDNA molecules,e) sequencing the enriched microbial cfDNA molecules,f) establishing a dataset containing features of the microbial cfDNA molecules, andg) determining at least one source of the microbial cfDNA from the sample as the identity of the pathogen(s) responsible for the infection.
28. The method according to claim 27, wherein the microbial cfDNA molecule fraction in the extracted DNA is enriched by,removing mammalian DNA molecules from the sample using a composition according to any one of claims 1-13 or a kit according to any one of claims 19-23, andpurifying the microbial cfDNA molecules in the remaining sample.
29. The method according to claim 27 or 28, wherein step f) comprisesproviding from the sequencing in step e) a list of cfDNA sequences, Optionally, performing an alignment of each of the cfDNA sequences in the list to host genomic sequences and removing from the list cfDNA sequences that align to the host genome thereby establishing a filtered list,performing an alignment of the cfDNA sequences of the filtered list to a plurality of microbial genome sequences,grouping the cfDNA sequences in the filtered list that align with the microbial genome sequence(s) into microbial species and / or genus specific genomic set(s),applying a last common ancestor approach to remove cfDNA sequences where the alignment of the sequence to the microbial species and / or genus specific genomic set(s) is ambiguous, thereby establishing a microbe specific cfDNA molecule list, based on the microbe specific cfDNA molecules in the microbe specific cfDNA molecule list, provide a dataset of the microbial cfDNA molecules comprising i. the cfDNA sequences of the microbial cfDNA,II. the nucleotide length of each of the microbial cfDNA molecules in said dataset,ill. optionally, the number of reads of the microbial cfDNAs in said dataset,iv. the genomic position(s) in the microbial genome of each alignment, v. the genomic sequence(s) of the microbial species and / or genus to which the specific microbial cfDNA align andvi. the total number of microbial cfDNA reads in the dataset.
30. The method according to any one of claims 27-29, wherein step g) comprises gl) performing for each cfDNA sequencing dataset(s) as defined in claim 29, i. a statistical analysis of the uniformity of the distribution of cfDNA alignments across the genome of the microbial species, and / orII. a statistical analysis of the distribution of the nucleotide length of the cfDNA molecules in the dataset,g2) determining the pathogen(s) responsible for the infection as the microbe(s) for which:ill. the distribution of cfDNA alignments across the genome of the microbial species does not differ significantly from a uniformity test, such as a Poisson distribution, and / oriv. the identified length of the cfDNA molecules can be statistically sampled from a normal distribution defined by a cfDNA dataset from known infectious host samples.
31. The method according to any one of claims 27-30, wherein the sample is tissue, blood, bronchial lavage fluid, exhaled breath, serum, plasma, cerebrospinal fluid, faecal, BAL, or urine.
32. The method according to any one of claims 27-31, wherein the sample is a human sample.
33. A method for treating a patient suffering from an infection, the method comprisinga. identifying at least one pathogen according to the method of any one of claims 27-32, andb. suggesting a clinically relevant effective antimicrobial treatment regimen targeting said pathogen.
34. A method for excluding microbial infection(s) as the cause of a pathological states of a subject suspected of suffering from a microbial infection, the method comprising subjecting a sample from the subject to an analysis method according to the method of any one of claims 27-32, and concluding based on an absence of any identified microbe(s) that the patient is infection-free.
