Quantitative analysis of short double-stranded cell-free DNA in liquid biopsy

Abstract

The invention relates to quantitative analysis of short double-stranded cell-free DNA in liquid biopsy. The present invention relates to a method for identifying a cfDNA-derived biomarker of a target physiological condition, disease or condition, comprising: a) isolating cfDNA from a sample containing cfDNA of a subject known to be at the target physiological condition or having the target disease or condition; b) selecting a short double-stranded cfDNA fragment; c) determining a nucleic acid sequence of the short double-stranded cfDNA fragment; d) comparing the nucleic acid sequence of the short double-stranded cfDNA fragment with a reference; and; e) identifying, based on said comparison in step d), at least one cfDNA-derived biomarker associated with a target physiological condition, disease or condition, Furthermore, the invention relates to a method of assessing a disease, a method of determining a therapeutic effect, the use of a cfDNA-derived biomarker, and a device for detecting a cfDNA-derived biomarker.

Description

[0001] The present invention relates to a method for identifying cfDNA-derived biomarkers for a target physiological condition, disease or condition, comprising: a) isolating cfDNA from a sample containing cfDNA of a subject known to be in a target physiological condition or suffering from a target disease or condition; b) selecting short double-stranded cfDNA fragments; c) determining the nucleic acid sequence of the short double-stranded cfDNA fragments; d) comparing the nucleic acid sequence of the short double-stranded cfDNA fragments with a reference; and; e) based on the comparison in step d), identifying at least one cfDNA-derived biomarker associated with the target physiological condition, disease or condition. In addition, the present invention relates to a method for assessing a target physiological condition, disease or condition in a sample of a subject suspected of suffering from a target physiological condition, disease or condition, comprising a) isolating cfDNA from a sample containing cfDNA of the subject; b) selecting short double-stranded cfDNA fragments; c) determining the nucleic acid sequence of the short double-stranded cfDNA fragments; and d) determining the presence, absence or abundance of a cfDNA-derived biomarker associated with at least one disease or condition associated with the target physiological condition, disease or condition in the determined nucleic acid sequence of the short double-stranded cfDNA fragments. In addition, the present invention contemplates a method for determining the efficacy of treating a target physiological condition, disease or condition in a sample of a subject known to be in a target physiological condition or suffering from a target disease or condition, comprising: a) isolating cfDNA from a sample containing cfDNA of the subject; b) selecting short double-stranded cfDNA fragments; c) determining the nucleic acid sequence of the short double-stranded cfDNA fragments; d) determining the presence, absence or abundance of at least one cfDNA-derived biomarker associated with the target physiological condition, disease or condition in the determined nucleic acid sequences of the short double-stranded cfDNA fragments; e) repeating steps a) to d); and f) determining the efficacy of the treatment based on the determination over time. In addition, the present invention relates to the use of cfDNA-derived biomarkers in a sample in evaluating a subject suspected of having a disease or condition, and a device for detecting at least one cfDNA-derived biomarker in a liquid biopsy sample.

[0002] As a minimally invasive diagnostic procedure, liquid biopsy is based on various types of biomarkers, including circulating extracellular nucleic acids, such as cell-free DNA (cfDNA), extracellular vesicles, or circulating tumor cells. Cell-free DNA can be released into the blood by apoptosis, necrosis, or active secretion from almost all cell types and tissues, however, the largest proportion comes from the hematopoietic system (Sun et al., 2015). In addition to the physiological release of normal cell turnover, cancer cells or microbial pathogens are known to release their DNA into the blood circulation (Heitzer et al., 2020). The released genomic DNA is then degraded by DNA digesting enzymes (nucleases) to produce fragments with a size mainly ranging from 147 to 167 base pairs (bp), corresponding to one nucleosome (Lo et al., 2021). By high-throughput sequencing of cfDNA fragments, nucleosome positioning can be inferred with base pair resolution (Snyder et al., 2016). The exact position of nucleosomes and chromatin structure plays a key role in regulating gene expression by providing DNA paths for the transcriptional machinery. The dense packaging of DNA limits the access of the molecular transcription machinery, while the lack of an open structure of nucleosomes makes the DNA more accessible to key regulators of transcription, such as transcription factors, enhancers or repressors. Unfortunately, transcription factor footprints, i.e. the measurement of genome-wide binding of regulatory DNA-binding proteins, cannot yet be directly analyzed by liquid biopsy methods. To date, the diagnosis of various tumors and inflammatory diseases requires tissue biopsies. Due to their invasive nature, these are associated with risks for the patient and do not allow for repeated and / or regular and therefore dynamic analysis of clinical indications or physiological conditions.

[0003] Recently, Snyder et al. and Burnham et al. used ultra-deep sequencing of total cell-free DNA to identify a small fraction of double-stranded cfDNA that was significantly shorter than normal cfDNA, ranging between 35 and 80 nucleotides. They proposed that this short cfDNA might be protected from DNA-binding factors and thus might represent direct transcription factor binding events. However, ultra-deep sequencing of total cfDNA has shown limitations, as it may not allow for efficient and comprehensive analysis of genome-wide DNA binding events in liquid biopsies, which could also be used for routine diagnostic purposes at a reasonable cost.

[0004] Therefore, there is a need in the art to provide new and reliable means and methods for minimally invasive diagnosis. In particular, there is a need to provide means and methods that at least partially avoid and overcome the above-mentioned shortcomings of the prior art.

[0005] This need is met by the method, use and device having the features of the independent claims. Preferred embodiments are listed in the dependent claims which can be realized individually or in any arbitrary combination.

[0006] The inventors have developed a liquid biopsy sample such as plasma short double-stranded cfDNA fragments, especially those whose length distribution is shown as 20 to 60bp short double-stranded cfDNA fragments. The inventors have proved that this so-called DNA footprint fragment accumulates on open chromatin and gene regulatory elements. Based on this DNA footprint, the inventors identified the genomic regions and occupied transcription factor motifs that are differentially enriched between colorectal cancer and pancreatic cancer and between sepsis patients and clinical controls to facilitate the diagnosis and differentiation of various diseases. The method can also be applied to identify the physiological condition of the subject. Thus a new biomarker category is established, that is, quantitative analysis of short double-stranded cell-free DNA (footprint DNA) by liquid biopsy (liquid footprint).

[0007] Thus, the present invention relates to a method for identifying cfDNA-derived biomarkers for a target physiological condition, disease or disorder, comprising:

[0008] a) isolating cfDNA from a sample containing cfDNA from a subject known to be in a target physiological condition or to have a target disease or condition;

[0009] b) selecting short double-stranded cfDNA fragments;

[0010] c) determining the nucleic acid sequence of the short double-stranded cfDNA fragments;

[0011] d) comparing the nucleic acid sequence of the short double-stranded cfDNA fragments to a reference; and;

[0012] e) identifying, based on said comparing in step d), at least one cfDNA-derived biomarker associated with a target physiological condition, disease, or disorder.

[0013] The method is advantageous because it provides a minimally invasive yet reliable and effective means for adequately diagnosing and differentiating between various diseases and physiological conditions.The interactions of DNA binding proteins such as transcription factors and their important role in gene regulation can be analyzed in vivo.

[0014] In general, the terms used herein should be given the common and customary meanings of ordinary technicians in the field, unless otherwise specified, without limitation to special or customized meanings. As described below, the terms "having", "including" or "comprising" or any arbitrary grammatical variants thereof are used in a non-exclusive manner. Therefore, these terms can refer to the situation where there are no other features in the entity described herein in addition to the features introduced by these terms, or the situation where there are one or more other features. For example, "A has B", "A includes B" and "A contains B" can all refer to the situation where there are no other elements in A except B (i.e., A is only and exclusively composed of B), or it can also refer to the situation where there are one or more other elements in entity A except B, such as element C, element C and D, or even other elements. In addition, as understood by those skilled in the art, the expression "including one" preferably means "including one or more", that is, equivalent to "including at least one". Therefore, unless otherwise stated, a statement related to one of a plurality of items is preferably related to at least one such item, more preferably to a plurality of items therein; therefore, for example, identifying a "cell" involves identifying at least one cell, preferably involves identifying a plurality of cells. The conventional meaning of the term "plurality" as used herein refers to a quantity more than one, i.e. at least two. Therefore, a plurality may preferably be a quantity of two or more, three or more, five or more, ten or more, 25 or more, 50 or more, 100 or more, 1000 or more. Unless otherwise stated, the terms "plurality" and "many" may be used interchangeably herein.

[0015] In addition, as described below, the terms "preferably", "more preferably" and "most preferably", "particularly", "more particularly", "specifically", "more specifically", "typically", "more typically" or similar terms are used in conjunction with optional features without limiting further possibilities. Therefore, the features introduced by these terms are optional features and are not intended to limit the scope of the claims in any way. Those skilled in the art will recognize that the present invention can be implemented by using alternative features. Similarly, features introduced by "in an embodiment" or similar expressions are intended to be optional features, without any limitation on other embodiments of the invention, without any limitation on the scope of the invention, and without any limitation on the possibility of combining features introduced in this manner with other optional or non-optional features of the invention.

[0016] The method described below is preferably an in vitro method. In principle, the method steps can be performed in any arbitrary order that the skilled person considers suitable, but preferably in the order indicated; in addition, one or more, preferably all of the steps can be assisted or performed by automated equipment. In addition, in addition to the steps explicitly mentioned herein, the method may also include other steps.

[0017] As used herein, if not otherwise specified, the term "about" refers to an indication value of technical precision generally accepted in the relevant field, preferably to an indication value of ±20%, more preferably ±10%, and most preferably ±5%. In addition, the term "substantially" means that there is no deviation that affects the indicated result or use, that is, the potential deviation will not cause the indicated result to deviate by more than ±20%, more preferably ±10%, and most preferably ±5%. Therefore, "substantially consisting of..." means including the specified ingredients, but does not include other ingredients except for materials present as impurities, materials unavoidable in the process of providing ingredients, and ingredients added for purposes other than the technical effects of the present invention. For example, the composition defined by the phrase "substantially consisting of..." includes any known acceptable additives, excipients, diluents, buffers, carriers, etc. Preferably, the composition consisting essentially of a group of ingredients will contain less than 5% by weight, more preferably less than 3% by weight, even more preferably less than 1% by weight, and most preferably less than 0.1% by weight of one or more non-specified ingredients.

[0018] The term "cfDNA" used herein refers to cell-free DNA, also referred to as circulating free DNA. The term generally includes degraded DNA fragments released into body fluids, typically nuclear or mitochondrial DNA fragments. cfDNA can be double-stranded (ds) or single-stranded (ss). According to the present invention, ds cfDNA is generally selected. Cell-free DNA can be released into blood or other body fluids by apoptosis, necrosis, or active secretion of almost all cell types and tissues, with the largest proportion generally coming from the hematopoietic system. Typically, cfDNA can be grouped into conventional cfDNA fragments with a typical size of about 100 to 500bp or even longer and short cfDNA fragments with a fragment size of less than 100bp. According to the present invention, short cfDNA fragments are of particular interest, typically short double-stranded cfDNA fragments are fragments with a size of less than 100bp, more typically with a size of between 10 and 80bp, and even more typically with a double-stranded DNA fragment between 20 and 60bp.

[0019] The short ds cfDNA according to the present invention may also be referred to as "footprint DNA". The inventors believe that "footprint DNA" is not bound by theory and reflects DNA sites protected by any type of protein-DNA interaction, such as transcription factor (TF) binding sites, binding sites of structural proteins, binding sites of regulatory proteins, transcription start sites or gene promoter sequences, CpG islands, preferably extensions of TF binding sites. A typical feature of short ds cfDNA is an increased average GC content. More typically, the GC content is increased by at least 5%, at least 7%, at least 10%, at least 12%, at least 15%, at least 17% compared to the average GC content of the genomic DNA of the reference genome. The GC content of an appropriate reference genome is known to those skilled in the art. In addition, those skilled in the art know how to select an appropriate reference genome. Typically, an appropriate reference genome is a genome of the same species as the subject or a species closely related thereto. Specifically, according to the present invention, the sample is a sample of human body fluids, and the reference genome is a human genome.

[0020] The term "fragment" of a biomacromolecule (preferably DNA) is used in a broad sense herein, and relates to any sub-portion of a corresponding biomacromolecule comprising a specified sequence, structure and / or function. Thus, the term includes sub-portions produced by actual breaks of biomacromolecules, particularly DNA molecules, and more particularly genomes, and also includes sub-portions derived from corresponding biomacromolecules in an abstract manner (e.g., on a computer). Thus, the term "subsequence" of a biological sequence (preferably a nucleic acid sequence) relates to any sub-portion of a corresponding biological sequence, which comprises a specified sequence, and, if specified, preferably has a specified function. Fragments according to the present invention particularly relate to cfDNA fragments as defined elsewhere herein.

[0021] The term "nucleic acid" as used herein refers to a linear or circular nucleic acid molecule, particularly a linear nucleic acid molecule. The term includes single-stranded and partially or completely double-stranded nucleic acid molecules. Preferably, the nucleic acid molecule is a DNA molecule, which may also be referred to as "DNA", more preferably a dsDNA molecule. In addition, chemically modified nucleic acid molecules are also included, including naturally occurring modified nucleic acid molecules, such as glycosylated or methylated nucleic acid molecules, or artificially modified derivatives, such as biotinylated p nucleic acid molecules, locked nucleic acids, peptide nucleic acids, etc. Typically, the DNA molecules according to the present invention are naturally occurring nucleic acid molecules, more typically cfDNA fragments as described elsewhere herein. Nucleic acid molecules comprising a variety of nucleic acids (e.g., at least ten nucleic acids) are also referred to as polynucleotides.

[0022] Unless otherwise specifically stated, the specific polynucleotides mentioned herein preferably include polynucleotide variants. The term "polynucleotide variant" used herein relates to a variant of a polynucleotide described herein, comprising a nucleic acid sequence characterized in that the sequence can be derived from the above-mentioned specific nucleic acid sequence by at least one nucleotide substitution, addition and / or deletion, wherein the polynucleotide variant should have the function and / or activity of a specific polynucleotide. Therefore, the variant of the marker sequence can be, for example, a sequence of an ortholog, a paralog or another homolog of a specific polynucleotide. Also preferably, the polynucleotide variant is or is derived from a non-naturally occurring allele of a specific polynucleotide. Polynucleotide variants also include polynucleotides comprising a nucleic acid sequence that can hybridize with the above-mentioned specific polynucleotide, preferably under stringent hybridization conditions. Technicians are aware of these stringent conditions, such as can be found in Current Protocols in Molecular Biology, John Wiley & Sons, NY (1989), 6.3.1-6.3.6. The skilled person knows how to determine the required hybridization conditions by referring to textbooks such as the above textbooks or the following textbooks: Sambrook et al., "Molecular Cloning", Cold Spring Harbor Laboratory, 1989; Hames and Higgins (Ed.) 1985, "Nucleic Acids Hybridization: A Practical Approach", IRL Press of Oxford University Press, Oxford; Brown (Ed.) 1991, "Essential Molecular Biology: A Practical Approach", IRL Press of Oxford University Press, Oxford. As used herein, "specific hybridization" preferably refers to hybridization under stringent conditions, more preferably refers to hybridization to a target sequence at least 2 times, preferably at least 5 times, more preferably at least 10 times, and even more preferably at least 100 times compared to non-target sequences. Therefore, a primer or probe that specifically hybridizes to a target sequence provides specific amplification and / or a specific signal. Other polynucleotide variants include polynucleotides comprising such a nucleic acid sequence, and the nucleic acid sequence has at least 70%, preferably at least 80%, more preferably at least 90%, even more preferably at least 95%, still more preferably at least 98%, and most preferably at least 99% identity with the specifically specified nucleic acid sequence. Preferably, the identity percentage value is calculated over the entire nucleic acid sequence region, preferably as described elsewhere herein. The polynucleotides of the present invention either consist of the above-mentioned nucleic acid sequence, or consist essentially of the above-mentioned nucleic acid sequence, or comprise the above-mentioned nucleic acid sequence. Therefore, they may also contain other nucleic acid sequences.

[0023] The degree of identity (e.g., expressed as "% identity") between two biological sequences (preferably DNA, RNA or amino acid sequences) can be determined by algorithms well known in the art. Preferably, the degree of identity is determined by comparing two optimally aligned sequences over a comparison window, wherein the sequence fragments in the comparison window may include additions or deletions (e.g., gaps or overhangs) compared to the comparison sequence used for optimal alignment. The percentage is calculated by determining (preferably over the entire length of the sequence, usually a polynucleotide sequence) the number of positions where the same residue appears in the two sequences to produce the number of matching positions, dividing the number of matching positions by the total number of positions in the comparison window, and multiplying the result by 100 to produce a percentage of sequence identity. The optimal alignment of sequences for comparison can be performed by the local homology algorithm of Smith and Waterman (1981), the homology alignment algorithm of Needleman and Wunsch (1970), the similarity search method of Pearson and Lipman (1988), computer implementations of these algorithms (such as BLAST, GAP, BESTFIT, PASTA or TFASTA) or visual inspection. In view of having determined that two sequences are compared, preferably GAP and BESTFIT are used to determine their optimal comparison, thereby determining the degree of identity. Preferably, a gap weight default value of 5.00 and a gap weight length default value of 0.30 are used. More preferably, the local alignment search basic tool (BLAST) is implemented to be used together with the default parameter values ​​for comparison. In the context of biological sequences involved herein, the term "substantially identical" represents that the identity % value is at least 80%, preferably at least 90%, more preferably at least 98%, and most preferably at least 99%. As will be understood, the term substantially identical includes 100% identity. The above description is applicable to the term "substantially complementary" after necessary modifications.

[0024] The term "sample" as used herein refers to any composition of a substance known or suspected to contain cfDNA, preferably known or suspected to contain short ds cfDNA to be detected. The sample can be any state of the substance. Therefore, preferably, the sample is a liquid, semi-solid, solid or gaseous sample. Preferably, the sample is a liquid or semi-solid sample, preferably an aqueous solution, which may optionally contain other solvents or other biological materials. Typically, the sample is derived from a body fluid or a body fluid sample, typically a body fluid of a vertebrate species. Generally, all types of vertebrate species are envisioned, however, in particular, the sample may involve a body fluid of a chicken, goose, duck, goat, sheep, cattle, pig, horse, dog, cat, hamster, rat, mouse, hamster or guinea pig or a human subject. Preferably, the body fluid is from a mammalian species. More preferably, the body fluid is human. According to the present invention, the sample is typically a body fluid sample, preferably selected from the group consisting of blood, plasma, serum, tears, urine, lymph, cerebrospinal fluid, bile, feces, sweat, amniotic fluid, synovial fluid and saliva, and more preferably the sample is a plasma sample.

[0025] Sample can be obtained by well-known technology, and it preferably includes scraping, swab or biopsy.Such sample can be obtained by using brush, (cotton) swab, tongue depressor, flushing / washing device, punch biopsy device, cavity puncture device such as needle or lancet or other surgical instruments.Usually, sample can be obtained from cell culture supernatant, body fluid or tissue or organ by separation techniques such as filtration, centrifugation or cell sorting.Sample can be blood as described herein or sample derived from blood, tear, urine, lymph, cerebrospinal fluid, bile, feces, sweat, amniotic fluid, synovial fluid or saliva sample.Sample according to the present invention can be generally referred to as liquid biopsy sample.“Liquid biopsy sample” generally refers to a body fluid sample that can be used for diagnostic purposes, which can replace the common biopsy sample usually obtained by invasive technology.The method according to the present invention may include the step of obtaining a sample from a subject.However, preferably, the method according to the present invention is an ex vivo method performed with a separated sample.

[0026] The use of a sample as described herein in the method according to the invention is advantageous because it can generally be obtained by non-invasive or minimally invasive techniques. It thus provides a minimally invasive but reliable diagnostic method for a large number of diseases and conditions that currently require large-scale invasive diagnostic procedures based on biopsies. It is also suitable for non-invasive prenatal diagnosis methods (NIPD).

[0027] The method according to the present invention includes a step of isolating cfDNA from a subject sample (e.g., a body fluid sample). It should be understood that in order to implement the method of the present invention, the sample may be further processed. The method according to the present invention is preferably an ex vivo method.

[0028] cfDNA and methods for its isolation are known in the art, for example from Aucamp et al. (2018) Biol. Rev. 93, 1649-1683, and further defined elsewhere herein.

[0029] In addition, according to the present invention, cfDNA can be enriched, for example, by precipitation and / or by binding to solid surface binding anionic compounds, such as magnetic beads that bind DNA. Also preferably, cells or other particulate matter can be removed by centrifugation and / or filtration. In addition, subfractions of cfDNA in the sample can be separated into cfDNA, for example, by size classification, separation or selection. Preferably, the above-mentioned enrichment and / or sub-classification is non-sequence specific. As described herein, all types of pre-treated samples are included in the term sample.

[0030] The term "biomarker" as used herein generally refers to an indicator of the severity or presence of a physiological or pathological condition (such as a disease or condition) of an organism. In particular, a "biomarker" according to the present invention is a nucleic acid molecule or a combination of nucleic acid molecules that makes it possible to identify a target physiological condition, disease or condition. The identification generally depends on the presence, absence or abundance of the biomarker. Therefore, the presence, absence or abundance of a nucleic acid molecule or a combination of nucleic acid molecules is generally specific to a target physiological condition, disease or condition. Therefore, however, the abundance of a biomarker may indicate a target physiological condition, disease or condition. The feature of a biomarker according to the present invention is typically a nucleic acid sequence, more typically a genomic sequence, such as a genomic sequence of a reference genome. In particular, the feature of a biomarker according to the present invention may be a specific target sequence; typically a sequence defined by a core target sequence and its upstream and downstream flanking sequences. According to the present invention, a biomarker is a biomarker of cfDNA origin associated with a target physiological condition, disease or condition.

[0031] The term "cfDNA-derived biomarker" generally refers to a nucleic acid molecule or a combination of nucleic acid molecules identified by the method of the present invention. In particular, "cfDNA-derived biomarkers" represent a typical assembled nucleic acid sequence of one or more genomic positions of a reference genome. cfDNA-derived biomarkers generally correspond to gene regulatory regions in the reference genome, such as transcription start sites, transcription factor binding sites, gene promoter regions, CpG islands, and binding sites for structural proteins, more typically transcription start sites, transcription factor binding sites, and gene promoter regions. The length of the nucleic acid sequence generally corresponds to a single short cfDNA fragment, which can be up to several hundred bases in length. The nucleic acid sequence is generally a continuous sequence of the reference genome. If multiple genomic positions are involved, cfDNA-derived biomarkers may also be referred to as abundance matrices or biomarker features. More typically, cfDNA-derived biomarkers are characterized by target sequences of at least 20bp to 500bp or even more on the reference genome. The sequence and size of cfDNA-derived biomarkers vary specifically according to the target physiological condition, disease, or condition. In particular, the cfDNA-derived biomarkers according to the present invention may be characterized by a target sequence defined by a core target sequence of at least 20 bp on a reference genome and its upstream and downstream flanking sequences, starting from 5 bp on each side, up to 200 bp. The preferred size of the cfDNA-derived biomarkers found in the present invention is between 30 bp and 300 bp.

[0032] A cfDNA-derived biomarker can generally be characterized by its presence or abundance in a sample of a subject known to be in a target physiological condition or suffering from a target disease or condition, or can be characterized by its absence in the sample. Typically, a cfDNA-derived biomarker according to the present invention corresponds to a differentially enriched region identified by comparing short cfDNA fragments of a sample to a reference, particularly as described in further detail elsewhere herein.

[0033] Using appropriate software tools known in the art, for example, using the implemented edgeR method using the R package DEBrowser (v1.2.0), differential enrichment regions (DERs) can be identified, which are typically between nucleic acid sequences determined from a sample and a reference. Typically, differential enrichment analysis is performed on the determined nucleic acid sequences of short ds cfDNA fragments, more typically on the raw read counts of the common peaks defined elsewhere herein; particularly in the concentration of common peaks. More typically, the nucleic acid sequence of a sample is subjected to differential enrichment analysis compared to a reference, particularly compared to the nucleic acid sequence of a reference sample. Differential enrichment can typically be determined by comparing the nucleic acid sequence of a sample with the nucleic acid sequence of a reference. When the difference is statistically significant, the nucleic acid sequence is generally considered to be differentially enriched between a sample and a reference. The difference may be in the presence, absence or abundance of the determined nucleic acid sequence being compared. Typically, the sequence is a mapped sequence, more typically a common peak of a mapped sequence as described elsewhere herein.

[0034] Those skilled in the art can generally use various well-known statistical evaluation tools, such as determination of confidence intervals, determination of p-values, Student's t-test, Mann-Whitney test, etc., to determine whether a value is statistically significant without further effort. For details, see Dowdy and Wearden, Statistics for Research, John Wiley & Sons, New York 1983. Preferred confidence intervals are at least 50%, at least 60%, at least 70%, at least 80%, at least 90%, at least 95%. In embodiments, the p-value is 0.5, 0.1, 0.05 or 0.01.

[0035] Preferably, the nucleic acid sequence is considered to be differentially enriched between the sample and the reference with an adjusted p-value of less than 0.5, and / or a fold change of ≤-1.5 or ≥ 1.5. More typically, the p-value is less than 0.2, even more typically the p-value is less than 0.1, even more typically the p-value is less than 0.1 and / or the fold change is ≤-2 or ≥2.

[0036] Preferably, highly significant DERs relate to the 100 most significant DERs identified by comparing the short cfDNA fragments of the sample to a reference, even more typically the 80 most significant, 50 most significant, 30, 20, 10 or 5 most significant DERs. Preferably, a combination of combined DERs is used as a cfDNA-derived biomarker for assessing a target physiological condition, disease or disorder, as explained elsewhere herein. In complex physiological or pathological conditions, the use of a combination of DERs is advantageous because it is believed that it can achieve a particularly highly reproducible and reliable diagnostic tool.

[0037] The phrase "associated with a target physiological condition, disease or disorder" refers to a biomarker that is indicative of the target physiological condition, disease or disorder. It is noted that a biomarker identified in a sample from a pregnant subject may be indicative of a target physiological condition, disease or disorder in a fetus.

[0038] In particular, the presence, absence or abundance of a particular biomarker or combination of biomarkers in a sample indicates a target physiological condition or a target disease or condition. Biomarkers can generally be determined by any method known to those skilled in the art. Specifically, biomarkers derived from cfDNA can be identified and / or their presence, absence or abundance determined by the methods according to the present invention as outlined elsewhere herein.

[0039] In the following examples, preferred biomarkers characterized by DER are described with reference to the human genome, see Tables 4-7.

[0040] The term "target physiological condition" as used herein relates to any type of physiological condition. Typically, the target physiological condition may relate to pregnancy, the subject's developmental stage (including fetal developmental stage) and aging-related conditions, training status, aerobic and anaerobic conditions of the body, hormonal conditions (including menstrual cycle and insulin status), and other conditions that are generally non-pathological and therefore generally do not require medical treatment. In particular, the target physiological condition is related to pregnancy or developmental stage, and more specifically, the term "target physiological condition" may relate to the subject's developmental stage, health, growth or nutritional status, and even more specifically, a fetus in a pregnant subject.

[0041] Typically, the target disease or condition is cancer or an infectious disease. The term "cancer" as used herein relates to a disease in animals, including humans, characterized by the uncontrolled growth of a group of somatic cells ("cancer cells"). This uncontrolled growth may be accompanied by invasion and destruction of surrounding tissues (infiltration), and may cause cancer cells to spread to other parts of the body (metastasis). Preferably, the term cancer also includes cancer recurrence (recurrence). Preferably, the cancer is selected from the list consisting of: acute myeloid leukemia (AML), acute lymphocytic leukemia, adrenocortical carcinoma, AIDS-related lymphoma, anal cancer, appendix cancer, astrocytoma, atypical teratoma, basal cell carcinoma, bile duct cancer, bladder cancer, bone cancer, brain stem glioma, breast cancer, Burkitt's lymphoma, carcinoid tumor, cerebellar astrocytoma, cervical cancer, chordoma, chronic lymphocytic leukemia, chronic myeloid leukemia, colon cancer , colorectal cancer, craniopharyngioma, endometrial cancer, ependymoblastoma, ependymoma, esophageal cancer, extracranial germ cell tumor, extragonadal germ cell tumor, extrahepatic bile duct cancer, fibrosarcoma, gallbladder cancer, gastric cancer, gastrointestinal stromal tumor, gestational trophoblastic tumor, hairy cell leukemia, head and neck cancer, hepatocellular carcinoma, Hodgkin lymphoma, hypopharyngeal cancer, hypothalamic and visual pathway glioma, intraocular melanoma, Kaposi's sarcoma, laryngeal cancer, medulloblastoma, medullary epithelioma, melanoma Pigmentoma, Merkel cell carcinoma, mesothelioma, oral cancer, multiple endocrine neoplasms, multiple myeloma, mycosis fungoides, nasal and paranasal sinus cancer, nasopharyngeal cancer, neuroblastoma, non-Hodgkin lymphoma, non-small cell lung cancer, oral cancer, oropharyngeal cancer, osteosarcoma, ovarian cancer, ovarian epithelial cancer, ovarian germ cell tumor, ovarian low malignant potential tumor, pancreatic cancer, pancreatic ductal adenocarcinoma, papilloma, sinus and nasal cancer, parathyroid cancer, penile cancer, nasopharyngeal cancer, chromaffin More preferably, the cancer is selected from the list consisting of pancreatic ductal adenocarcinoma (PDAC), colorectal cancer (CRC), non-small cell lung cancer (NSCLC), bone cancer; even more preferably selected from pancreatic ductal adenocarcinoma (PDAC) and colorectal cancer (CRC).

[0042] The term "infectious disease" or "infection" generally refers to various infections with pathogenic microorganisms in a subject, including fungal, parasitic, viral or bacterial infections. According to the present invention, the target disease or condition is preferably an infection caused by a pathogenic microorganism. Preferably, the infection causes an inflammatory disease, i.e., a health state or pathophysiological condition in which the body of the infected subject produces an immune response to the pathogenic microorganism. The infection may be a local infection, such as an infection at the site of infection, may be tissue or organ specific, such as a viral or bacterial infection of a target organ, such as hepatitis, pancreatitis or encephalitis, or may be systemic, affecting the whole body of the subject, such as Lyme disease or sepsis. In addition, the target disease or condition may be a chronic inflammatory disease, including chronic inflammatory bowel disease (IBD), such as ulcerative colitis or Crohn's disease; an autoimmune disease, such as multiple sclerosis and rheumatism; or another chronic disease or condition, such as irritable bowel syndrome, metabolic disorders, diabetes, genetic diseases and developmental abnormalities.

[0043] According to the present invention, the target disease or condition can be particularly selected from the group consisting of: cancer, including pancreatic ductal adenocarcinoma (PDAC), colorectal cancer (CRC), non-small cell lung cancer (NSCLC), bone cancer; fungal, parasitic, viral or bacterial infection, including sepsis, pancreatitis, Lyme disease, encephalitis; chronic inflammatory diseases, including chronic inflammatory bowel disease (IBD), such as ulcerative colitis or Crohn's disease; autoimmune diseases, including multiple sclerosis (MS), rheumatism; irritable bowel syndrome; metabolic disorders; genetic diseases or developmental abnormalities, especially genetic diseases or developmental abnormalities of the fetus, including trisomy.

[0044] Preferably, the disease or disorder is cancer or an infectious disease, more preferably selected from pancreatic ductal adenocarcinoma (PDAC), colorectal cancer (CRC) and sepsis.

[0045] The term "subject" as used herein relates to a vertebrate, preferably livestock or companion animals, such as chickens, geese, ducks, goats, sheep, cattle, pigs, horses, dogs, cats, hamsters, rats, mice, hamsters or guinea pigs. More preferably, the subject is a mammal. Even more preferably, the subject is a human. Typically, the subject is known or suspected of having a disease or being in a physiological condition as described above, more typically, the subject is known or suspected of having PDAC, CRC or sepsis. The method according to the present invention can be applied to non-invasive or minimally invasive prenatal diagnostic methods. In this case, the term "subject" includes a pregnant woman with a fetus.

[0046] The method according to the present invention includes a step of separating cfDNA from a sample containing cfDNA of a subject known to be in a target physiological condition or suffering from a target disease or condition. Methods for separating cfDNA are known in the art, such as from Aucamp et al. (2018) Biol. Rev. 93, 1649-1683. These methods may generally rely on centrifugal force or vacuum extraction based on silica membranes. cfDNA separation kits are available from Qiagen or Thermo Fisher Scientific in Hilden, Germany. Typically, separating cfDNA from a sample includes separating cellular components from the sample, more typically followed by protease treatment to remove proteins from the remaining liquid sample. Then, cfDNA fragments can be separated from the liquid sample by binding to magnetic beads. According to the present invention, the separation of cfDNA preferably refers to the separation of total cfDNA. In this context, the term "total cfDNA" more preferably refers to cfDNA fragments of all sizes separated in the method according to the present invention, typically referring to the total amount of cfDNA of double-stranded cfDNA separated before any size selection step. The step of isolating cfDNA from a sample may also be referred to as preparing cfDNA from the sample, thereby producing a cfDNA preparation that can be further processed.

[0047] The method generally further comprises the step of generating a double-stranded cfDNA library, preferably a sequencing library of the double-stranded cfDNA.

[0048] This can usually be achieved by preparing a double-stranded DNA specific library; thus double-stranded cfDNA fragments are usually selected. Other methods known in the art for selecting ds cfDNA, such as non-denaturing polyacrylamide gel electrophoresis separation methods, can be used. The library is preferably a next-generation sequencing library of ds cfDNA. Library generation procedures are known in the art. Kits for these procedures are commercially available, such as NEXTFLEX Cell-Free DNA Sequencing Kit (V2) (NOVA-5150-02). According to the present invention, ds DNA specific primers can be used to generate sequencing libraries. The size of the primers can usually exceed the size of short cfDNA, such as 30 to 100bp, typically between 50 and 70bp. The method according to the present invention may also include obtaining a ds cfDNA library.

[0049] In addition, the method according to the present invention includes a step of selecting short double-stranded cfDNA fragments from a cfDNA preparation. Preferably, the method includes a step of size selection of short ds cfDNA fragments using any method known in the art. Specifically, the selection of short double-stranded cfDNA fragments includes gel electrophoresis size separation of double-stranded cfDNA, gradient gel electrophoresis, PCR-based methods, hybridization and capture, and / or selection using binding protein cross-linking and capture, more specifically, gel electrophoresis size separation of double-stranded cfDNA sequencing libraries.

[0050] More specifically, gel electrophoresis size separation, gradient gel electrophoresis or PCR-based methods can be used to select fragments up to 100 bp; more typically, fragments between 20 bp and 60 bp, and even more specifically, gel electrophoresis size separation or gradient gel electrophoresis. Even more typically, size selection of cfDNA, in particular size selection of cfDNA libraries, can be performed using an appropriate preparative gel electrophoresis instrument, particularly with a pulsed field, such as a BluePippin instrument (Sage Science). Selection of short dscfDNA fragments can be performed on an agarose gel, such as 3% agarose. Other methods including size-specific solid phase extraction, size-dependent chromatographic separation, and other methods known in the art are also possible. Alternatively, selective enrichment can be performed based on known sequences, including sequence-guided selection, such as hybridization and capture methods. In addition, after size selection, a quality check can usually be performed to ensure normal size distribution. Size selection can include a re-amplification step for samples of the selected size. In addition, if fragments outside the selected size are detectable, the size selection can be repeated. Additionally, the quality of size selection can be assessed using appropriate means in the art, such as the Fragment Analyzer High Sensitivity DNA Kit (Agilent).

[0051] Preferably, the selection of short ds cfDNA fragments results in an enrichment of short ds cfDNA fragments by at least 2-fold, at least 4-fold, at least 10-fold, at least 50-fold, at least 75-fold, or at least 100-fold compared to an unenriched total cfDNA preparation. Enrichment generally refers to an enrichment of at least two-fold or more defined by the ratio of short cfDNA to conventional cfDNA; more preferably, short cfDNA refers to cfDNA fragments having a length of at most 100 bp or less. The enrichment of short cfDNA can be determined by quantifying the amount of total cfDNA and the amount of short cfDNA before and after enrichment. For example, in a typical ds cfDNA preparation of a healthy subject, the amount of short cfDNA fragments is typically in the range of 0.2%-3% of the total cfDNA (Hudecova et al., 2022). After size selection as described herein, the amount of short cfDNA fragments increases significantly, typically up to 5% to 99% of the total cfDNA in the preparation after size selection. More typically, size selection results in an amount of short cfDNA fragments in the preparation that is at least 50% of the total cfDNA after size selection; even more typically, at least 60%, at least 70%, at least 80%, at least 90%, at least 95%, at least 96%, at least 97%, at least 98% or at least 99% of the total cfDNA in the preparation after size selection.

[0052] Size selection in the methods according to the present invention is advantageous because it enables enrichment of short cfDNA fragments and allows downstream analysis of short cfDNA fragments with high accuracy, but does not require ultra-deep sequencing or high coverage, i.e., 10-fold or more.

[0053] The method according to the present invention includes a step of determining the nucleic acid sequence of the short double-stranded cfDNA fragment. The term generally refers to the identification of nucleic acid molecules of all sizes, particularly the nucleic acid sequence of cfDNA fragments. Devices and methods for determining nucleic acid sequences are known in the art. Those skilled in the art can select appropriate sequencing methods. The determination of nucleic acid sequences includes qualitative determination, i.e., identification of nucleic acid sequences, and quantitative or semi-quantitative determination, i.e., determining the amount or relative amount of nucleic acid sequences.

[0054] According to the present invention, nucleotide sequence is typically determined using next generation sequencing and / or deep sequencing known in the art, preferably using Illumina sequencing.More typically, sequence is determined with a coverage rate equivalent to 10 times of reference genome at most, and even more typically with a coverage rate equivalent to 5 times, 2 times, 1 times, 0.5 times, 0.2 times or 0.1 times of reference genome at most.According to the present invention, the coverage rate of conventional application is equivalent to 1 times of reference genome at most.For example, in the embodiment provided below, only about 30,000,000 100bp / readings provide enough sequencing depths for samples derived from human DNA, i.e., human genome is a reference genome.The method for calculating coverage, such as Lander / Waterman equation, is known in the art.In the case of sequence-directed selection, coverage refers to the equivalence with selected target region.

[0055] Preferably, the determined nucleic acid sequence of the short ds cfDNA fragments comprises a protein binding footprint DNA sequence, the protein binding footprint DNA sequence comprising a protein-DNA interaction site, such as a transcription factor (TF) binding site, a structural protein binding site, a regulatory protein binding site, a transcription start site or a gene promoter sequence, a CpG island, preferably a TF binding site. Without wishing to be bound by theory, the inventors believe that interaction or binding with proteins can prevent further fragmentation and / or complete degradation of short dscfDNA fragments.

[0056] The term "reference" used in accordance with the present invention relates to a reference sample or any amount (reference amount) or value (reference value) that can be compared with the determined nucleic acid sequence of the short double-stranded cfDNA fragments to identify a cfDNA-derived biomarker of a target physiological condition, disease or condition, or to assess a target physiological condition, disease or condition in a sample of a subject suspected of having the target physiological condition, disease or condition. Thus, a reference sample or reference amount or reference value can be obtained from a healthy subject.

[0057] Preferably, the reference according to the present invention comprises one or more short double-stranded cfDNA fragments, which (i) are derived from a body fluid sample of a healthy subject or from a body fluid sample of a subject who is not in the target physiological condition or does not suffer from the target disease or condition, or (ii) are known short double-stranded cfDNA fragments.

[0058] The "reference sample" according to the present invention is preferably a preparation of short double-stranded cfDNA fragments from a sample of a healthy subject of the same species, or a solution of synthetically produced short ds cfDNA fragments. In particular, the reference sample can be a sample of a body fluid or tissue of a healthy subject of the same species. The term "solution of synthetically produced short ds cfDNA fragments" covers any type of aqueous solution to which known short ds cfDNA fragments are added and / or an aqueous solution known to contain a predetermined amount of known short ds cfDNA fragments. According to the present invention, the reference sample is particularly preferably a sample from a healthy subject, more specifically, the reference sample is a sample of a healthy subject who has recently undergone surgery, i.e., a subject in a postoperative condition, such as a subject who has recently undergone surgery; wherein "recently" generally refers to a few minutes to a few hours after surgery to 2 days after surgery.

[0059] In the method for identifying a biomarker derived from cfDNA according to the present invention, the reference is preferably a reference sample. A "healthy subject" is preferably a subject who is apparently healthy and / or known not to be in the target condition and / or not suffering from the target disease or condition. In particular, a healthy subject does not suffer from the target disease or condition. The term "healthy subject" according to the present invention preferably includes a postoperative subject and a subject suffering from any disease or condition other than the target disease or condition including cancer.

[0060] According to the present invention, the term "reference amount" or "reference value" may typically refer to an amount or value indicating the abundance of at least one short ds cfDNA fragment (preferably a specific cfDNA fragment). More typically, the term "reference amount" or "reference value" may indicate the value of the average amount of the abundance of any short ds cfDNA fragment. In particular, in the method for assessing a target physiological condition, disease or condition in a subject sample according to the present invention, the reference may refer to a "reference amount" or "reference value".

[0061] According to the present invention, the term "compared with a reference" typically refers to any comparison of a sample with a reference of a biomarker of a cfDNA source suitable for identifying a target physiological condition, disease or condition. More typically, any known statistical method for inferring discrimination ability or mean / median difference or distribution difference can be applied to the comparison. In these statistical methods, for example, generalized linear models are used, and differential analysis is performed using the importance of features in likelihood ratio tests, t tests, Wilcoxon rank sum tests, Anova, Kolmogorov Smirnov tests, Wald tests, receiver operating characteristics, effect sizes, mutual information, and machine learning models (such as random forests). More typically, "compared with a reference" refers to differential enrichment analysis as described elsewhere herein.

[0062] According to the present invention, preferably, at least one short double-stranded cfDNA fragment or multiple short double-stranded cfDNA fragments that are not present in the reference short double-stranded cfDNA fragments or that are different in abundance from the reference short double-stranded cfDNA fragments are identified as at least one cfDNA-derived biomarker associated with the target physiological condition, disease or condition. According to the present invention, the phrase "different from the abundance of the short ds cfDNA fragments of the reference" includes the absence of short ds cfDNA fragments in the sample. Preferably, in order to identify cfDNA-derived biomarkers associated with the target physiological condition, disease or condition, the method of the present invention is repeated for many samples of different subjects known to be in the same target physiological condition or known to suffer from the same target disease or condition, preferably for many biological replicates. More preferably, the repetitions can be performed in parallel, for example, multiple samples are processed simultaneously.

[0063] The method also includes a step of mapping the determined nucleic acid sequence of the short double-stranded cfDNA fragment to a reference genome, preferably obtaining a mapped nucleic acid sequence, also referred to herein as a mapping sequence. Mapping the determined nucleic acid sequence to the reference genome can be used as a suitable tool for identifying the origin of cfDNA fragments, in other words, to identify the position of cfDNA fragments on the reference genome. The mapped sequence can generally be aligned with the gene regulatory region in the reference genome, such as the transcription start site, the transcription factor binding site, the gene promoter region, the CpG island and the binding site of the structural protein, more typically with the transcription start site, the transcription factor binding site and the gene promoter region, preferably corresponding. The device and method for mapping the nucleic acid sequence to the reference genome are known in the art, and are easily obtained by those skilled in the art, and the reference genome is a genome with a known assembly sequence, such as the human reference genome assembly GRCh37.

[0064] Preferably, the reference genome is a vertebrate genome, preferably a mammalian genome, more preferably a human, goat, sheep, cattle, pig, horse, dog, cat, hamster, rat, mouse, hamster or guinea pig genome, still more preferably a human genome. A reference genome is typically a genome with a known assembled sequence.

[0065] According to the present invention, the step of mapping the determined nucleic acid sequence can generally be performed using available software tools known to those skilled in the art, such as NextGenMap (v0.5.5), or any other short read alignment or mapping tool known in the art, including Bowtie, BWA, HISAT, STAR.

[0066] In addition, before mapping, the method may include at least one or more of the following steps: sequencing quality control, sequence adapter removal, removal of terminal poly G sequences, quality fine-tuning, removal of low-complexity sequencing reads. Typically, the mapped sequence is processed using an appropriate software tool, such as SAMtools by Heng Li, which is available at www.htslib.org. Processing of the mapped sequence may include at least one or more of the following steps: deduplication; filtering of blacklisted regions, MapQ filtering; deep tool downstream analysis, as known in the art.

[0067] Mapped sequences can generally serve as the basis for identifying common peaks. The term "common peak" as used herein generally refers to multiple mapping sequences within the same position of a reference genome identified in a sample of a subject known to be in a target physiological condition or suffering from a target disease or condition. Common peaks can generally be established in more than one sample of different subjects known to be in a target physiological condition or suffering from a target disease or condition. The common peak may be narrow, for example, between 20bp and 100bp, or it may be wide, for example, between 101bp and 500bp. Generally, a common peak is defined as a genomic position, particularly where a narrow peak is identified in at least three of four samples; more specifically, the sample refers to a sample containing cfDNA of a subject known to be in a target physiological condition or suffering from a target disease or condition. Common peaks can generally be derived from mapped sequences using appropriate bioinformatics tools (such as MACS2) known to those skilled in the art. Common peaks associated with target physiological conditions, diseases or conditions are preferably identified using appropriate bioinformatics tools (such as R as known in the art).

[0068] In the methods of identifying a biomarker according to the invention, steps a) to e) may generally be repeated on samples from different subjects known to be in the target physiological condition or suffering from the target disease or disorder; more typically, thereby validating the biomarker.

[0069] Preferably, according to the present invention, the comparison in step d) includes comparing the mapping sequence of the short double-stranded cfDNA fragments with a reference, in particular the mapping sequence of the short double-stranded cfDNA fragments of the reference sample. More preferably, the method further comprises the step of obtaining a common peak from the mapping sequence. Still more preferably, the method comprises comparing the common peak obtained from the mapping sequence of the short double-stranded cfDNA fragments of the sample with the common peak obtained from the mapping sequence of the short double-stranded cfDNA fragments of the reference. The comparison typically includes a differential enrichment analysis, and therefore more typically may lead to the identification of differentially enriched regions between the sample and the reference short ds cfDNA fragments. In other words, the comparison generally includes or consists of a differential enrichment analysis of the common peaks obtained from the mapping sequences of the short double-stranded cfDNA fragments of the sample and the reference. Preferably, the method for identifying a biomarker derived from cfDNA according to the present invention includes identifying a differentially enriched region (DER). According to the present invention, the method for identifying a biomarker derived from cfDNA comprises step d), comparing the nucleic acid sequence of the short double-stranded cfDNA fragments with a reference; typically identifying a differentially enriched region (DER) thereby.

[0070] Furthermore, in the method for identifying cfDNA-derived biomarkers according to the present invention, the cfDNA-derived biomarkers associated with at least one disease or condition are preferably based on a combination of more than one DER.

[0071] The above definitions apply mutatis mutandis to the following. The additional definitions and explanations given further below also apply mutatis mutandis to all embodiments described in this specification. In particular, unless otherwise stated, the definitions and explanations regarding the above method steps may apply to the following method steps.

[0072] In addition, the present invention relates to a method for assessing a target physiological condition, disease or condition in a sample of a subject suspected of being in a target condition or suffering from a target disease or condition, comprising: a) isolating cfDNA from a sample containing cfDNA of the subject; b) selecting short double-stranded cfDNA fragments; c) determining the nucleic acid sequences of the short double-stranded cfDNA fragments; and d) determining the presence, absence or abundance of at least one cfDNA-derived biomarker associated with the target physiological condition, disease or condition in the determined nucleic acid sequences of the short double-stranded cfDNA fragments.

[0073] As used herein, the term "determining the presence, absence or abundance of at least one cfDNA-derived biomarker" refers to each measure for determining the presence, absence or quantity of a cfDNA-derived biomarker in a sample. Therefore, the term generally includes qualitative determinations, such as determining the presence of a cfDNA-derived biomarker sequence above a predetermined detection limit, and quantitative or semi-quantitative determinations, i.e., determining the amount or relative amount of a cfDNA-derived biomarker. The terms "abundance" and "quantity" are used interchangeably herein. The determination is based on the determination of the nucleic acid sequence in step c). The determination preferably includes selecting a target sequence from the determined nucleic acid sequence and determining its abundance.

[0074] Preferably, in the method for assessing a target physiological condition, disease or disorder according to the present invention, at least one cfDNA-derived biomarker associated with the target physiological condition, disease and disorder is identified or is capable of being identified by the method for identifying biomarkers described elsewhere herein.

[0075] The methods for assessing a target physiological condition, disease, or disorder according to the present invention may also include comparing the determined abundance of at least one cfDNA-derived biomarker associated with the target physiological condition, disease, or disorder to a reference.

[0076] Typically, the methods for assessing a target physiological condition, disease or disorder according to the present invention further comprise the step of assessing the target condition, disease or disorder in a sample from a subject.

[0077] The term "assessment" as used herein refers to determining the target physiological condition of a subject, or diagnosing a target disease or condition. It will be understood by those skilled in the art that assessment is generally not intended to be correct for 100% of the subjects to be studied. However, the term requires that the assessment is correct for a statistically significant portion of the subjects (e.g., a cohort in a cohort study). Those skilled in the art can use various well-known statistical assessment tools, such as determination of confidence intervals, determination of p-values, Student's t-tests, Mann-Whitney tests, etc., to effortlessly determine whether a portion is statistically significant. For details, see Dowdy and Wearden, Statistics for Research, John Wiley & Sons, New York 1983. Preferred confidence intervals are at least 90%, at least 95%, at least 97%, at least 98%, or at least 99%. The p-value is preferably 0.1, 0.05, 0.01, 0.005, or 0.0001.

[0078] More specifically, when assessing a target physiological condition, disease or disorder, the methods of the invention include identifying whether the subject is in the target physiological condition or has a target disease or disorder.

[0079] The present invention also contemplates a method for determining the efficacy of treating a target physiological condition, disease or condition in a sample from a subject known to be in a target physiological condition or suffering from a target disease or condition, comprising: a) isolating cfDNA from a sample containing cfDNA from the subject; b) selecting short double-stranded cfDNA fragments; c) determining the nucleic acid sequences of the short double-stranded cfDNA fragments; d) determining the presence, absence or abundance of at least one cfDNA-derived biomarker associated with the target physiological condition, disease or condition in the determined nucleic acid sequences of the short double-stranded cfDNA fragments; e) repeating steps a) to d); and f) determining the efficacy based on the determination over time.

[0080] As used herein, the step of "determining therapeutic efficacy based on determinations over time" refers to each and every measure of treatment success determined over time. Typically, the determination involves comparing samples after treatment of subjects known to be in the condition, or suffering from the disease or condition, with samples before or without treatment of subjects known to be in the condition, or suffering from the disease or condition. The term "after treatment" typically refers to an appropriate time point, or typically more than one time point, after the start of an appropriate treatment, at which the desired effect typically occurs.

[0081] The method may also include the step of determining a treatment response. In particular, a treatment success is determined if the subject's physiological condition or disease or condition improves after treatment, or a treatment failure is determined if the subject's physiological condition or disease or condition worsens after treatment, or if the physiological condition, disease or condition remains unchanged.

[0082] In the method for determining the therapeutic efficacy of a target physiological condition, disease or disorder, steps a) to d) are preferably repeated at appropriate intervals; more preferably, at least once a year to several times a week; even more preferably, at least once every four months to at least ten times a month.

[0083] In addition, the present invention relates to a device for detecting at least one cfDNA-derived biomarker in a liquid biopsy sample, the device comprising at least: 1) a sample collector and a selector for separating and selecting cfDNA; 2) a sequencer, preferably for high-throughput sequencing; and 3) an evaluation unit, which includes at least a processor. The sample collector and the selector are generally units of the device, suitable for receiving from a sample of a subject, preferably from a plurality of samples to typically obtain cfDNA. More typically, the sample collector and the selector can separate cfDNA from the sample by an automatic cfDNA preparation procedure, and can specifically select short dscfDNA. In addition, the sample collector and the selector can generally prepare cfDNA preparations for further automated processing in the sequencer. The sequencer can be any sequencer known in the art for short ds cfDNA sequencing, and can generally be a next-generation sequencing device known in the art. The evaluation unit includes at least a processor, typically a microprocessor, including tangibly embedded executable instructions for performing at least one step of the method described herein. Preferably, the evaluation unit also includes a storage medium. The sample collector and the selector, the sequencer, and the evaluation unit are generally assembled into a functional unit.

[0084] Preferably, the device is suitable for performing the methods according to the present invention, in particular methods for identifying cfDNA-derived biomarkers for a target physiological condition, disease or disorder, and / or methods according to the present invention for assessing a target physiological condition, disease and disorder, and / or methods described elsewhere herein for determining the efficacy of treating a target physiological condition, disease or disorder.

[0085] In addition, the present invention also contemplates cfDNA-derived biomarkers for diagnosing a disease or condition in a subject suspected of having a disease or condition. Preferably, the cfDNA-derived biomarker is obtained and / or obtainable by the method for identifying cfDNA-derived biomarkers disclosed herein.

[0086] In view of the above, the following implementation plans are specifically envisaged:

[0087] 1. A method for identifying a cfDNA-derived biomarker for a target physiological condition, disease, or disorder, comprising:

[0088] a) isolating cfDNA from a sample containing cfDNA from a subject known to be in a target physiological condition or to have a target disease or condition;

[0089] b) selecting short double-stranded cfDNA fragments;

[0090] c) determining the nucleic acid sequence of the short double-stranded cfDNA fragments;

[0091] d) comparing the nucleic acid sequence of the short double-stranded cfDNA fragments to a reference; and;

[0092] e) identifying, based on said comparing in step d), at least one cfDNA-derived biomarker associated with a target physiological condition, disease, or disorder.

[0093] 2. The method of embodiment 1, wherein the short double-stranded cfDNA fragments are double-stranded DNA fragments with a fragment size of less than 100 bp, preferably between 10 and 80 bp, and more preferably between 20 and 60 bp.

[0094] 3. The method of embodiment 1 or 2, wherein the sample is a body fluid sample, and the body fluid is preferably selected from the group consisting of blood, plasma, serum, tears, urine, lymph, cerebrospinal fluid, bile, feces, sweat, amniotic fluid, synovial fluid and saliva, and more preferably the sample is a plasma sample.

[0095] 4. The method of any one of embodiments 1 to 3, wherein the selection in step b) results in an enrichment of short ds cfDNA fragments by at least 2-fold, at least 4-fold, at least 10-fold, at least 50-fold, at least 75-fold or at least 100-fold, preferably, wherein enrichment refers to an enrichment of at least two times as defined by the ratio of short cfDNA to regular cfDNA; more preferably, wherein short cfDNA refers to cfDNA with a length of at most 100 bp or less.

[0096] 5. The method according to any one of embodiments 1 to 4, further comprising the step of generating a double-stranded cfDNA library, preferably a double-stranded cfDNA sequencing library.

[0097] 6. The method of any one of embodiments 1 to 5, wherein the selection of short double-stranded cfDNA fragments comprises gel electrophoresis size separation of double-stranded cfDNA, gradient gel electrophoresis, PCR-based methods, hybridization and capture, and / or selection using binding protein cross-linking and capture, preferably gel electrophoresis size separation of double-stranded cfDNA sequencing library.

[0098] 7. The method according to any one of embodiments 1 to 6, wherein the sequence is determined using next generation sequencing, preferably Illumina sequencing.

[0099] 8. The method according to any one of embodiments 1 to 7, wherein the determined nucleic acid sequence comprises a protein binding footprint DNA sequence, which comprises a protein-DNA interaction site, such as a transcription factor (TF) binding site, a binding site for a structural protein, a binding site for a regulatory protein, a transcription start site or a gene promoter region, a CpG island, preferably a TF binding site.

[0100] 9. The method of any one of embodiments 1 to 8, wherein steps a) to e) are typically repeated on samples from different subjects known to be in the target physiological condition or suffering from the target disease or disorder; more typically, thereby validating the biomarker.

[0101] 10. The method of embodiment 8, wherein the abundance and / or type of protein binding footprint DNA sequences in the determined nucleic acid sequences of the short double-stranded cfDNA fragments are different from that of the reference.

[0102] 10. The method of any one of embodiments 1 to 9, wherein the reference comprises one or more short double-stranded cfDNA fragments, which are (i) derived from a body fluid sample of a healthy subject or a body fluid sample of a subject who is not in the target physiological condition or does not suffer from the target disease or condition, or (ii) are known short double-stranded cfDNA fragments.

[0103] 11. The method of any one of embodiments 1 to 10, wherein at least one short double-stranded cfDNA fragment or multiple short double-stranded cfDNA fragments are identified as at least one cfDNA-derived biomarker associated with a target physiological condition, disease or condition, each short double-stranded cfDNA fragment having a nucleic acid sequence that is not present in a reference short double-stranded cfDNA fragment or a nucleic acid sequence that differs in abundance from the reference short double-stranded cfDNA fragment.

[0104] 12. The method of any one of embodiments 1 to 11, wherein the target physiological condition, disease or disorder is selected from the group consisting of:

[0105] i) cancer, including pancreatic ductal adenocarcinoma (PDAC), colorectal cancer (CRC), non-small cell lung cancer (NSCLC), bone cancer;

[0106] ii) fungal, parasitic, viral or bacterial infections, including sepsis, pancreatitis, Lyme disease, encephalitis;

[0107] iii) chronic inflammatory diseases, including chronic inflammatory bowel disease (IBD), such as ulcerative colitis or Crohn's disease;

[0108] iv) autoimmune diseases, including multiple sclerosis (MS), rheumatic diseases;

[0109] v) Metabolic disorders;

[0110] vi) irritable bowel syndrome;

[0111] vii) genetic diseases or developmental abnormalities, especially those affecting the fetus, including trisomy; and

[0112] viii) Gestation and developmental stages.

[0113] Preferably, the disease or disorder is cancer or an infectious disease, more preferably selected from pancreatic ductal adenocarcinoma (PDAC), colorectal cancer (CRC) and sepsis.

[0114] 13. The method of any one of embodiments 1 to 12, further comprising the step of mapping the determined nucleic acid sequence of the short double-stranded cfDNA fragments to a reference genome, preferably thereby obtaining a mapped sequence.

[0115] 14. The method of any one of embodiments 1 to 13, comprising identifying differentially enriched regions (DERs), typically by comparing the nucleic acid sequences of the short double-stranded cfDNA fragments of a sample with those of a reference.

[0116] 15. The method according to embodiment 13, wherein the reference genome is a vertebrate genome, preferably a mammalian genome, more preferably a human, goat, sheep, cow, pig, horse, dog, cat, hamster, rat, mouse, hamster or guinea pig genome, and even more preferably a human genome.

[0117] 16. The method of any one of embodiments 13 to 15, wherein the comparison in step d) comprises comparing the mapped sequences of the short double-stranded cfDNA fragments with the mapped sequences of a reference short double-stranded cfDNA fragments.

[0118] 17. The method of any one of embodiments 14 to 16, wherein at least one cfDNA-derived biomarker associated with a disease or condition is based on a combination of more than one DER.

[0119] 18. A method for assessing a target physiological condition, disease or condition in a sample from a subject suspected of being in the target physiological condition or suffering from a target disease or condition, comprising:

[0120] a) isolating cfDNA from a sample containing cfDNA from the subject;

[0121] b) selecting short double-stranded cfDNA fragments;

[0122] c) determining the nucleic acid sequence of the short double-stranded cfDNA fragments; and

[0123] d) determining the presence, absence or abundance of at least one cfDNA-derived biomarker associated with a target physiological condition, disease or disorder in the determined nucleic acid sequences of the short double-stranded cfDNA fragments.

[0124] 19. The method of embodiment 18, wherein at least one cfDNA-derived biomarker associated with the target physiological condition, disease, or disorder is identified or identifiable by the method of any one of embodiments 1 to 17.

[0125] 20. The method of any of embodiments 18 or 19, further comprising comparing the determined abundance of at least one cfDNA-derived biomarker associated with the target physiological condition, disease, or disorder to a reference.

[0126] 21. The method of any one of embodiments 18 to 20, further comprising the step of assessing the target condition, disease or disorder in the sample from the subject.

[0127] 22. A method for determining the efficacy of treating a target physiological condition, disease or condition in a sample from a subject known to be in the target physiological condition or to suffer from a target disease or condition, comprising:

[0128] a) isolating cfDNA from a sample containing cfDNA from the subject;

[0129] b) selecting short double-stranded cfDNA fragments;

[0130] c) determining the nucleic acid sequence of the short double-stranded cfDNA fragments;

[0131] d) determining the presence, absence or abundance of at least one cfDNA-derived biomarker associated with a target physiological condition, disease or disorder in the determined nucleic acid sequences of the short double-stranded cfDNA fragments;

[0132] e) repeating steps a) to d); and

[0133] f) determining the efficacy of the treatment based on the determination over time.

[0134] 23. The method of embodiment 22, wherein steps a) to d) are repeated at least once a year to several times a week.

[0135] 24. An apparatus for detecting at least one cfDNA-derived biomarker in a liquid biopsy sample, comprising at least:

[0136] 1) A sample collector and selector for isolating and selecting cfDNA;

[0137] 2) a sequencer, preferably for high-throughput sequencing; and

[0138] 3) An evaluation unit, which comprises at least a processor.

[0139] 25. The apparatus of embodiment 24, wherein the apparatus is suitable for performing the method of any one of embodiments 1 to 23.

[0140] 26. Use of cfDNA-derived biomarkers in a sample for assessing a disease or condition in a subject suspected of having a disease or condition, wherein the cfDNA-derived biomarkers are preferably obtained by the method of any one of embodiments 1 to 17.

[0141] 27. cfDNA-derived biomarkers, preferably cfDNA-derived biomarkers obtained by the method of any one of embodiments 1 to 17, which are used to diagnose a disease or condition in a subject suspected of having the disease or condition.

[0142] 28. A method for diagnosing a disease or condition in a subject suspected of having the disease or condition, the method comprising:

[0143] (I) determining the presence, absence or abundance of at least one cfDNA-derived biomarker associated with a target disease or condition in the determined nucleic acid sequences of the short double-stranded cfDNA fragments;

[0144] (II) diagnosing the disease or condition based on the determination in step (I).

[0145] All references cited in this specification are incorporated herein by reference, involving the entire disclosure content of the references and the disclosure content specifically mentioned in this specification. BRIEF DESCRIPTION OF THE DRAWINGS

[0146] Figure 1: Footprint DNA is enriched in gene regulatory regions and open chromatin. (a) Visual example of the comparison of footprint DNA (S03) and regular cfDNA (average of S05-S08). For footprint DNA (green), narrow and broad peaks are shown, while for regular cfDNA (purple), nucleosome-free regions (NFRs) are shown. RefSeq genes based on ChIP-Seq experiments, ENCODE promoter-like structures (PLS), and ENCODE transcription factor binding sites (TFBS) are included as references. (b) Pie charts show the proportion of annotated genomic features of footprint DNA (S03) broad peaks and regular cfDNA (S05-S08 combined) NFRs. Bar charts show the ratio between broad peaks and NFRs for each genomic feature. (c) Average coverage profile of footprint DNA (S03) and regular cfDNA (S06) along all annotated protein-coding genes. The gene bodies of all genes are scaled to 5 kilobases. Dashed vertical lines indicate the intervals shown in the subplots. Subplots show the average distribution of transcription start sites without rescaling the gene body. (d) Average coverage distribution plots of ATAC-seq data for footprint DNA (S03), regular cfDNA (S06), and publicly available DNase hypersensitive sites (DHS) derived from publicly available DNase-seq data. DNase-seq and ATAC-seq data were generated for the GM12878 cell line.

[0147] Figure 2 : Liquid footprinting captures the binding of various transcription factors with a direct mechanism. (ae) Average coverage distribution plots of footprinted DNA (S03) and regular cfDNA (S06) for three transcription factors (NFE2, REST, and SPI1) from three different transcription factor superclasses are shown. The average distribution plots are based on the 1000 most robust binding sites annotated in the Gene Transcription Regulation Database (GTRD). In addition, the average curves of all ChIP-seq TFBS annotated in ENCODE3 show the comparison of footprinted DNA with regular cfDNA. Further average coverage distribution plots of footprinted DNase-seq and ATAC-seq of CTCF binding sites (CTCF BS) are shown. DNase-seq and ATAC-seq data were generated for the GM12878 cell line. (f) Speculated molecular model of the formation of footprinted DNA at a DNA binding protein (DBP) binding site surrounded by two nucleosomes on both sides. Arrows indicate the end points of the DNA fragments for the corresponding sequencing technology. The resulting theoretical coverage tracks for each sequencing technology are depicted.

[0148] Figure 3: Footprint DNA is enriched toward loci with markers of epigenetic activation and transcriptionally active regions. (a) ENCODE’s annotated promoter-like structures (PLS) are divided into two clusters based on publicly available cell-free H3K4me3 ChIP-Seq signal intensities, representing active (grey lines) or inactive promoters (black lines). The average coverage distribution of footprint DNA and normal cfDNA at active or inactive promoters shows the influence of promoter activation status. (b) Annotated CpG islands are divided into two clusters based on cell-free methyl CpG binding domain (cfMBD) sequencing signal intensities. The average coverage distribution of footprint DNA and normal cfDNA at methylated (grey lines) or unmethylated (black lines) CpG islands reveals the influence of methylation levels at CpG islands. (c) Histograms show the average expression levels of protein-coding genes in publicly available cell-free RNA sequencing data. For each category, 938 genes (5% of all analyzed genes) were selected: no expression (black), low expression (dark grey), medium expression (medium grey), and high expression (light grey). (d) Distribution of average coverage of transcription start sites, footprint DNA and regular cfDNA for the genomic categories defined in (c) reveals the impact of transcriptional activity. All data are from samples of healthy individuals. S03 was used as the footprint DNA dataset and S06 was used as the regular cfDNA dataset.

[0149] Figure 4: Comparison of liquid footprint data reveals condition-specific transcription factor binding sites and transcription factors. (a) Differential enrichment analysis of pancreatic ductal adenocarcinoma (PDAC) samples versus colorectal cancer (CRC) samples identified differentially occupied transcription factor binding sites (Adj.p value ≤ 0.1 and |FC| ≥ 2). (b) Differential enrichment analysis of sepsis samples versus postoperative (post-OP) samples identified differentially occupied transcription factor binding sites (Adj.p value ≤ 0.1 and |FC| ≥ 2). (c) Principal component analysis based on identified differential transcription factor binding sites separated all conditions. Four biological replicates were used for each condition. The centroid of each group is depicted as a larger data point in addition to the samples. Variance explained: PC1 = 29.6%, PC2 = 20.9%, PC3 = 8.6%. (d) Example of differentially enriched TFBS in CRC with novel TF binding in the LCP1 promoter. (e) Example of differentially enriched TFBS in sepsis with novel TF binding in the PLXNC1 promoter. (f)-(i) The top 10 differentially enriched TF motifs in the following: (f) CRC consensus peaks compared to PDAC consensus peaks. CTCF and CTCFL were also identified but not included in the figure. (g) PDAC consensus peaks compared to CRC consensus peaks. (h) Sepsis consensus peaks compared to post-OP consensus peaks. (i) Post-OP consensus peaks compared to sepsis consensus peaks. CTCF and CTCFL were also identified but not included in the figure (see Tables 4-7).

[0150] Figure 5 : Visual summary of footprint DNA extraction and sequencing in plasma. Cell-free DNA was extracted from patient plasma using an automated magnetic bead-based kit. Sequencing library preparation was performed by fragment end repair and adapter ligation. Due to PCR amplification, only intact double-stranded DNA fragments were enriched in the final sequencing library. Footprint DNA was enriched by size selection from the sequencing library. The footprint DNA sequencing library was sequenced in single-end mode on the Illumina platform. This figure was created with biorender.

[0151] Figure 6 : Description of the full footprint DNA enrichment process from cell-free DNA for next-generation sequencing. The steps shown correspond to: 1. Isolated cell-free DNA. 2. Double-stranded DNA sequencing library. 3. Size-selected sequencing library. 4. PCR amplified sequencing library. 5. Sequencing library after second size selection. Size selection is performed using an automated preparative gel electrophoresis instrument (selection window: 150bp to 200bp). The methods section describes the applied methods in detail. For each step, the fragment analyzer distribution map of S19 and the dilution factor are shown. Purple indicates DNA fragments that can be assigned to nucleosomes or conventional cfDNA, while green indicates DNA fragments that can be assigned to footprint DNA.

[0152] Figure 7 : Visual summary of the footprint DNA data processing pipeline. In short, the raw footprint DNA sequencing data is cleaned and filtered, mapped to the human reference genome, and filtered again before further downstream analysis. The individual steps are described in detail in Example 8.

[0153] Figure 8 : A histogram and fitted normal distribution of the lengths of the observed fully processed footprint DNA sequencing reads are depicted. One million random reads were selected from all 20 sequenced liquid footprint samples and the read lengths were plotted in the histogram (total n = 20 million). The observed distribution was fitted with a normal distribution represented by the red line (mean (μ) = 37.9, standard deviation (σ) = 6.6). The Kolmogorov-Smirnov test was used to assess the fit of the normal distribution to the given data, which produced a p-value of 0.056.

[0154] Fig. 9 : Average GC content of processed sequencing reads of conventional cfDNA and footprint DNA sequencing. Data of samples S01-S24 were used for this analysis (see Table 1). The bar length represents the mean, while the error bars represent the standard deviation.

[0155] Fig.10 : Comparison of liquid footprint data and single-stranded cfDNA sequencing data from Snyder et al. (2016). (a, c, e) Average coverage distribution of liquid footprint sequencing data (sequencing depth: short cfDNA (S03) = 8.29·10 6 , conventional cfDNA (S06) = 2.60·10 7 (b, d, f) Average coverage distribution of sequencing data generated by the single-stranded library preparation method of Snyder et al. (sequencing depth: short cfDNA = 1.33·10 8 , conventional cfDNA=3.84·10 8). These average coverage profiles are also plotted at a smaller scale to better visualize the dynamics of the data. (a and b) Average coverage profiles of annotated transcription start sites for short cfDNA and regular cfDNA. (c and d) Average coverage profiles of 1000 ChIP-Seq-validated TP53BP1 binding sites for short cfDNA and regular cfDNA. (ef) Average coverage profiles of 1000 ChIP-Seq-validated MYC binding sites for short cfDNA and regular cfDNA. Purple represents data from regular cfDNA, while green represents data from short cfDNA. The ultra-deep sequencing data of Snyder et al. were generated from plasma of healthy individuals using a single-stranded library preparation method. The raw sequencing data were retrieved from SRA (File ID = SRR2130051) and separated in silico into short cfDNA (35-80nt) and regular cfDNA (120-180nt).

[0156] Fig.11 :The relationship between cell-free MBD-Seq and cell-free H3K4me3 ChIP-Seq and gene expression levels. Based on the public cell-free RNA sequencing data, four groups of genes with different expression levels were defined (no expression (dark blue), low expression (blue), medium expression (red) and high expression (dark red). Figure 3 c). (a) Average coverage distribution of cell-free MBD-Seq reads at selected transcription start sites. (b) Average coverage distribution of H3K4me3 ChIP-Seq reads at selected transcription start sites.

[0157] Fig.12 : The table shows the number of differentially enriched regions (DERs) from footprint DNA sequencing data for comparisons of pancreatic ductal adenocarcinoma (PDAC) vs colorectal cancer (CRC) and postoperative controls (post-OP) vs sepsis. The Venn diagram shows the overlap of the DERs for the two groups. Example

[0158] The following examples are intended to illustrate the present invention only and should not be construed as limiting the scope of the present invention in any way.

[0159] Example 1: Blood Sample

[0160] Blood samples from 9 healthy individuals, 4 patients with pancreatic ductal adenocarcinoma, 4 patients with colorectal cancer, 4 patients with sepsis, and 4 patients who had recently undergone surgery in the clinic were included in this study (Table 1). Blood from healthy individuals was commercially obtained from Biomex GmbH. Patients with sepsis participated in a previously published prospective observational clinical study that was conducted in the surgical intensive care unit of the University Hospital Heidelberg, Germany, from November 2013 to January 2015 (German Clinical Trial Registration Number: DRKS00005463). Patients who underwent surgery in the clinic and all carcinoma patients were part of the clinical study Immunophenotyping of Gastrointestinal Cancer Disease (ICD) at the University Clinic Erlangen, which was approved by the ethics committee. All experiments were performed in accordance with the study protocol approved by the ethics committee.

[0161] Table 1: Metadata for the generated datasets included in this study. The initial sequencing depth for all samples is given in the “Raw Reads” column, while the sequencing depth after the samples were fully processed is given in “Processed Reads”.

[0162]

[0163] Example 2: Cell-free DNA isolation

[0164] Plasma was prepared by centrifuging whole blood at 1,600 g and 4 °C for 10 min. Afterwards, plasma was centrifuged again at 16,000 g and 4 °C for 10 min. Then, 1.1 mL of the supernatant was transferred to a fresh 1.5 mL DNALoBind tube and stored at −80 °C. If necessary, the sample volume was filled with freshly prepared 1× phosphate-buffered saline solution. Cell-free DNA was isolated using the QIAsymphony SP DNA Preparation System and the QIAsymony DSP Circulating DNA Kit (Qiagen) according to the manufacturer's recommendations. The eluted cfDNA was quantified with the Qubit dsDNAHS Assay Kit (Life Technologies), and the cfDNA quality was assessed with the Fragment Analyzer High Sensitivity DNA Kit (Agilent).

[0165] Example 3: Conventional cell-free DNA sequencing

[0166] Sequencing libraries for conventional cell-free DNA were prepared using the NEBNext Ultra II DNA Library Preparation Kit (NEB) according to the manufacturer's protocol. 0.5 ng of isolated cfDNA was used as input. For all reactions, the NEBNext Adaptor was diluted 1:25. PCR was performed using 10 PCR cycles and 4 μL of primers. Sequencing was performed on the NextSeq 2000 (Illumina) using a 100 bp single-end kit.

[0167] Example 4: Conventional Cell-free DNA Sequencing Data Processing

[0168] After initial quality control of the raw sequencing reads using FastQC (v0.11.8), the following steps were performed38: (1) Sequencing adapters were removed, terminal polyG sequences (at least 10 consecutive Gs) were removed, reads shorter than 50 base pairs were removed, and quality trimming was performed (BBTools-bbduk.sh v38.67; Bushnell et al., 2017). (2) The processed reads were mapped to the human reference genome assembly GRCh37 (Sedlazeck et al., 2009, Li et al., 2013) using NextGenMap (v0.5.5) with default settings. (3) Mapped reads were deduplicated using samtools rmdup (v1.9), and reads in blacklisted regions were removed using bedtools intersect (v2.30.0; Amemiya et al., 2019; Li et al., 2009; Quinlan & Hall, 2010). Mapped reads were converted to BigWig format for visualization and other downstream analyses using deeptools (bin size = 10, normalization = counts per million (CPM), bamCoverage v3.5.1; Ramírez et al., 2016).

[0169] Example 5: Cell-free methyl CpG binding domain sequencing and cell-free methyl CpG binding domain sequencing data processing

[0170] Methylation enrichment was performed using the EpiMak Methylated DNA Enrichment Kit (NEB) according to the manufacturer's instructions. Two reactions were performed in parallel for each sample, with 5-15 ng of cfDNA input each time. The methylated cfDNA was eluted sequentially from both reactions using the same 50 μL of nuclease-free water to increase yield. Sequencing libraries were prepared using the NEBNext Ultra II DNA Library Preparation Kit (NEB) according to the manufacturer's protocol. 50 μL of methylated cfDNA from the EpiMark program was used as input. For all reactions, the NEBNext Adaptor was diluted 1:25. PCR was performed using 17 PCR cycles and 3 μL of primers. Sequencing was performed on the NextSeq2000 (Illumina) with a 100 bp single-end kit.

[0171] The cell-free methyl-CpG binding domain sequencing data were processed as conventional cell-free DNA sequencing data.

[0172] Example 6: Cell-free RNA sequencing data processing

[0173] Fastq files of five publicly available cell-free RNA sequencing datasets of different healthy individuals were obtained from the Sequencing Read Archive (SRA). The accession numbers and unique identifiers are listed in Table 2. All samples were sequenced in paired-end mode with a read length of 150 bp, averaging approximately 10 million reads per sample. After initial quality control of raw sequencing reads using FastQC (v0.11.8), the following steps were performed to obtain read counts for individual genes (Andrews, S. 2010): (1) Sequencing adapter removal, quality trimming, and removal of reads shorter than 100 base pairs (BBTools-bbduk.sh v38.67; Bushnell et al. (2017). (2) Processed reads were mapped to the human reference genome assembly GRCh37 using NextGenMap (v0.5.5) with default settings, retaining only reads with MAPQ values ​​greater than 2 (Sedlazeck et al., 2013). (3) Gene quantification was performed in the raw read counts using featureCounts (v2.0.1) with UCSC gene annotations (https: / / hgdownload.soe.ucsc.edu / goldenPath / hg19 / bigZips / genes / hg19.known Gene.gtf.gz) (Liao et al., 2014). Only protein-coding genes on autosomes were included in the final read count matrix to reduce confounding by sample donor gender. (4) If necessary, raw read counts were converted to transcripts per million (TPM) using R.

[0174] Table 2 - Metadata of public datasets used in this study. Cell-free H3K4me3 ChIP-Seq data were obtained from Zenodo: https: / / zenodo.org / record / 4277001 / files / Analysis.tgz?download=1. ATAC-seq data from ENCODE were downloaded together with the GRCh38 assembly and converted locally to the GRCh37 assembly with liftOver.

[0175]

[0176]

[0177] Example 7: Footprint DNA Sequencing

[0178] Sequencing libraries were prepared from 3 ng to 15 ng of cfDNA using the NEXTFLEX Cell-Free DNA Sequencing Kit (V2) (NOVA-5150-02) according to the manufacturer's recommendations, depending on the clinical condition, with one exception: the final library after PCR amplification was eluted in 20 μL of nuclease-free water. Library generation was performed using a Biomek FXP workstation (Beckman Coulter). Library quality was assessed by a fragment analyzer high sensitivity DNA kit (Agilent), and concentration was measured by a Qubit dsDNA HS assay kit (Life Technologies). Size selection of cfDNA libraries was performed using a BluePippin instrument (SageScience). To select footprint DNA portions in the range of 150-200 bp (corresponding to the range of short cfDNA fragments attached to sequencing primers), samples were applied to a 3% agarose BluePippin box according to the manufacturer's protocol. In brief, the sample was hydrated to 30 μL and mixed with 10 μL of the internal markers provided (100 bp to 250 bp). As described in the step C PCR amplification of the NEXTFLEX cell-free DNA sequencing kit (V2) (NOVA-5150-02), 23 μL of the eluted selected size sample was re-amplified with 25 μL NEXTFLEX PCRMaster Mix 2.0 and 2 μL of 1:10 diluted NEXTFLEX Primer Mix 2.0. Then, the sample was purified with 1.2 volumes of AMPureXP beads (Beckman Coulter) according to the manufacturer's recommendations. The size selection performance was evaluated by the fragment analyzer high sensitivity DNA kit (Agilent). If the sample still contained fragments outside the 150 bp to 200 bp target range, the size selection was repeated as described above, except that the input sample was adjusted to 30 μL with water and the re-amplification step was performed with 5 to 8 cycles. After size selection, sequencing was performed on the NextSeq 2000 (Illumina) using a 100 bp single-end kit.

[0179] Example 8: Footprint DNA Sequencing Data Processing

[0180] After initial quality control of raw sequencing reads using FastQC (v0.11.8), the following steps were performed in sequence to remove sequencing artifacts: (1) Removal of sequencing adapters, terminal polyG sequences (at least 10 consecutive Gs), and quality fine-tuning (BBTools-bbduk.sh v38.67). (2) Removal of terminal single nucleotide A added during library preparation (BBTools-bbduk.sh v38.67). (3) Size selection of sequencing fragments allowed read lengths greater than 20 bp and less than 60 bp (BBTools-bbduk.sh v38.67). (4) Removal of sequencing reads with low complexity, i.e., dust scores less than 7 (prinseq-litev0.20.4) (Andrews, S. 2010; Bushnell et al., 2017; Schmieder & Edwards, 2011). (5) The processed reads were mapped to the human reference genome assembly GRCh37 using NextGenMap (v0.5.5) with default settings, Sedlazeck et al. (2013). (6) The mapped reads were deduplicated using samtools rmdup (v1.9), reads in blacklisted regions were removed using bedtools intersect (v2.30.0), and reads with MAPQ values ​​below 1 were removed using samtools view (v1.9) (Amemiya et al., 2019; Li et al., 2009; Quinlan & Hall, 2010). The reads were converted to BigWig format for visualization and other downstream analyses using deep tools (bin size = 10, normalization = counts per million (CPM), bamCoverage v3.5.1; Ramírez et al., 2016). Figure 7 The workflow is summarized graphically.

[0181] Example 9: Processing of Cell-Free DNA Sequencing Data by Snyder et al.

[0182] After preliminary quality control of raw sequencing reads using FastQC (v0.11.8), the following steps were performed: (1) Removal of sequencing adapters and quality fine-tuning (cutadapt v4.0; Martin, M. 2011). (2) Fine-tuned reads were mapped to the human reference genome assembly GRCh37 using NextGenMap (v0.5.5) with a minimum map quality of 10; Sedlazeck et al. (2013). (3) Size selection of mapped reads in silico: 35-80 nt for footprint DNA and 120-180 nt for regular cfDNA. (4) Removal of low-complexity reads and deduplication (prinseq-lite v0.20.4; Schmieder and Edwards, 2011). (5) Reads in blacklisted regions were removed using bedtools intersect (v2.30.0; Quinlan & Hall, 2010). (6) For downstream analysis and visualization, BigWig files were generated using deep tools (bin size = 10, normalization = counts per million (CPM), bamCoverage v3.5.1; Ramírez et al., 2016).

[0183] Example 10: Peak calling and nucleosome-free region calling

[0184] For footprint DNA sequencing data, narrow and broad peaks were called using MACS2 call peaks (narrow: --nomodel-extsize32-callpeaks-minlength30-q 0.05, broad: --nomodel--extsize 32-wide-maxgap100-minlength500-widecutoff0.05)48. Consensus peaks were identified using R and defined as genomic sites where narrow peaks were identified in at least three of the four samples. In addition, consensus peaks that were less than 31 nucleotides apart were merged. For regular cfDNA, nucleosome-free regions were called on the merged dataset of four biological replicates to increase genome-wide coverage and thus reliability. Nucleosome-free regions were identified using the R package NucDyn with nucleR under default settings; Buitrago et al. (2019). Peaks and nucleosome-free regions were annotated as genomic features in R using the ChIPseeker package with default settings; Yu et al. (2015).

[0185] Example 11: Average coverage curve and heat map

[0186] Average coverage distribution plots and heatmaps were created from BigWig files using deep tools (computeMatrix, plotHeatmap, and plotProfile; v3.5.1; Ramírez et al., 2016). Different genomic reference positions were used for the average coverage distribution plots. Figure 2 and Figure 6 , validated transcription factor binding sites were retrieved from the Gene Transcription Regulation Database (GTRD; Kolmykov et al., 2021). In this paper, transcription factor binding sites were ranked according to their robustness across different cell lines and tissues, i.e., the number of tissues and cell lines in which the corresponding binding site was found. Only the top 1000 binding sites were used for the average coverage distribution map. The promoter-like structures of ENCODE were converted from GRCh38 to GRCh37 using the liftOver tool from UCSC and used as Figure 3 The reference region in a (available at https: / / www.encodeproject.org / files / ENCFF379UDA / ). Figure 3 b was used as the reference region (available on the UCSC TableBrowser, track name = cpgIslandExt). Figure 3 The heatmaps in a and b are clustered into two groups using k-means clustering based on line-wise averages. Missing data in the heatmaps are plotted in black.

[0187] Example 12: Transcription factor motif enrichment analysis

[0188] Enriched transcription factor motifs were identified using the AME tool from the MEME suite (average score—Fisher method; v5.4.1; McLeay & Bailey, 2010). Input DNA sequences were retrieved from the consensus peak sets and analyzed for enrichment of motifs listed in the HOCOMOCOv11_core_HUMAN_mono database; Kulakovskiy et al. (2018). DNA sequences of consensus peaks in footprint DNA sequencing data from healthy individuals were compared to control sequences generated by rotating the letters in the input (--shuffle) while preserving the frequency of k-mers. The proportions of identified transcription factor motif classes and their respective superclasses were summarized in a dendrogram using R and the dendrogram package (v.2.4.3). Footprint DNA sequencing data from states other than healthy were compared to each other. DNA sequences under each consensus peak set were used as control sequences, e.g., consensus peak DNA sequences from CRC as input compared to consensus peak DNA sequences from PDAC as control.

[0189] Example 13: Differential enrichment analysis

[0190] Differentially enriched regions (DERs) were identified using raw read counts in the combined common peaks of the two compared conditions with TMM normalization, glmLRT and dispersion = 0 using the edgeR method implemented in the R package DEBrowser (v1.2.0) (Kucukural et al., 2019; Robinson et al., 2010). Genes were considered differentially expressed between the two conditions (four biological replicates per condition) with an adjusted p-value less than 0.1 and a fold change ≤-2 or ≥2. Volcano plots for differential enrichment analysis were generated using the R package EnhancedVolcano (v1.8.0). DERs identified in both comparisons were used for principal component analysis (PCA) and the first three principal components were visualized in R using pca3d (v0.10.2). Samples for each condition were connected to the centroid of the corresponding condition with a line.

[0191] Example 14: Results

[0192] Enrichment of cell-free DNA footprint fragments at regulatory genomic regions

[0193] Gel electrophoresis-based size selection was used to enrich for short, double-stranded cfDNA with high information content. Our protocol for enriching short, double-stranded cfDNA involves extracting total cell-free DNA from plasma using a magnetic bead-based kit, followed by double-stranded DNA-specific library preparation ( Figure 5 To select double-stranded DNA fragments up to 60 bp in length, two subsequent size selection steps were performed using a preparative gel electrophoresis apparatus ( Figure 6 After high-throughput sequencing of the size-selected libraries, the reads were quality checked to ensure a normal size distribution between 20 bp and 60 bp ( Figure 7 The scheme showed that the normal distribution of sequencing reads had an average read length of 37.9 bp (SD = 6.6 bp, n = 2x10 7 Readings were uniformly sampled from 20 individuals: 4 healthy individuals, 4 patients with pancreatic ductal adenocarcinoma, 4 patients with colorectal cancer, 4 patients with sepsis, and 4 patients after surgery; Figure 8 The average GC content of the size-selected short double-stranded cfDNA reads (from now on referred to as "footprint DNA") was increased to 57.8% (SD = 1.9%, Fig. 9), while the average GC content of human genomic DNA is 40.9%; Piovesan et al. (2019). To determine the genomic location and origin of footprint DNA, we mapped footprint DNA to human chromosomes and compared it with sequencing data of conventional cfDNA from four other healthy individuals (Table 1). The distribution patterns of footprint DNA and conventional cfDNA are very different, because footprint DNA tends to accumulate in the form of a single narrow peak or clusters of narrow peaks (i.e., broad peaks) compared with conventional cfDNA, and these narrow peak clusters are usually located at the transcription start site (TSS) of the gene or the ChIP-Seq-validated transcription factor binding site (Table 3). Table 3 lists the transcription factors whose DNA motifs of the four healthy individuals are significantly enriched in the footprint DNA consensus peaks. More than 200 transcription factor motifs of all nine transcription factor superclasses in the database were identified as enriched in the footprint DNA peaks.

[0194] Table 3 Transcription factors whose DNA motifs are significantly enriched in the consensus peaks of footprint DNA of four healthy individuals. The enriched transcription factor motifs were identified using the AME tool in the MEME suite with the HOCOMOCOv11_core_HUMAN_mono database. The enrichment of consensus peak DNA sequences was analyzed compared to random k-mers of the consensus peaks.

[0195]

[0196]

[0197] *DBD – DNA binding domain

[0198] To compare broad peaks of footprint DNA with those of regular cfDNA, nucleosome-free regions (NFRs) should be analyzed, as the absence of regular cfDNA is an indicator of the presence of other DNA-binding proteins. Assignment of broad peaks of footprint DNA and NFRs of regular cfDNA to annotated functional elements of the genome showed that the proportion of broad peaks was approximately fourfold and sixfold higher in promoters (>1Kb upstream of TSS) and 5'UTRs of genes, respectively. In addition, the proportion of broad peaks assigned to exons was approximately fourfold higher than that of NFRs ( Figure 1 b). A more detailed examination of the chromosomal coverage of protein-coding genes revealed an inverse occupancy profile between footprint DNA and regular cfDNA. While footprint DNA was strongly enriched at TSSs, regular cfDNA was largely depleted at these sites. Furthermore, footprint DNA had an opposite pattern compared to regular cfDNA, with footprint DNA exhibiting an oscillatory pattern that reversed 1 Kb downstream of the TSS in regular cfDNA ( Figure 1c). In addition to enrichment at TSS, footprint DNA also showed significant signal at reference annotated DNase hypersensitive sites (DHS) in a B cell line (GM12878), whereas regular cfDNA signal oscillated at adjacent genomic locations and was depleted at DHS ( Figure 1 d) Overall, footprint DNA accumulates at open chromatin or TSS of genes, whereas regular cfDNA reads, i.e. nucleosomes, are clearly depleted in these regions.

[0199] To compare the information of footprint DNA with the broad peak of regular cfDNA, the nucleosome-free region (NFR) should be analyzed, as the absence of regular cfDNA is indicative of the presence of other DNA-binding proteins.

[0200] Liquid DNA footprinting to detect binding of various transcription factors

[0201] Since footprint DNA can be protected from nuclease digestion by binding to transcription factors, the peak regions were examined for their potential to contain transcription factor binding motifs. Therefore, a consensus peak set was defined from the narrow peaks in the liquid footprint sequencing data of four healthy individuals. Transcription factor motif enrichment analysis at the genomic locations of these consensus peaks revealed significant enrichment of 203 of the 401 transcription factor binding motifs in the reference motif database (HOCOMOCO). The 203 transcription factor motifs identified belonged to 46 different transcription factor families, including nine transcription factor superclasses. In addition, ChIP-Seq-validated transcription factor binding sites from ENCODE3 revealed clear enrichment signals for nuclear factor erythroid 2 (NFE2), RE1 silencing transcription factor (REST), and Spi-1 proto-oncogene (SPI1) in footprint DNA, while regular cfDNA was depleted in these binding sites. Overall, the average distribution map of all ChIP-seq-based transcription factor binding sites (TFBS) in ENCODE3 showed a clear enrichment in footprint DNA compared to regular cfDNA (Table 3). The binding sites of CCCTC binding factor (CTCF) verified by ChIP-Seq showed more obvious footprint DNA signals ( Figure 2 e). In contrast, near these binding sites, regular cfDNA exhibited regular, high-frequency oscillatory signals. Consistently, DNase-seq data showed a combination of footprint peak signals with adjacent oscillatory patterns, whereas ATAC-seq detected broad peaks representing open chromatin at CTCF binding sites ( Figure 2 e). Our data show that footprinted DNA most specifically reveals DNA binding at regulatory sites with high resolution, exemplified by NFE2, REST or SPI1 ( Figure 2ac, f). Previous studies have shown that a significant portion of small cfDNA exists as short single-stranded DNA. To find out the relationship between single-stranded and double-stranded short cfDNA, we compared our footprint DNA signals at transcription start sites and transcription factor binding sites with the short cfDNA sequencing method of Snyder et al. (single-stranded library preparation). We clearly found a superior signal-to-noise ratio, allowing the detection of binding signals that were undetectable in the short single-stranded DNA of Snyder et al., 2016, such as TP53BP1 ( Fig.10 ). Furthermore, our data suggest that footprint DNA is not a degradation product of conventional cfDNA but rather represents its own biological entity, as these fragments predominantly occur in locations where conventional cfDNA is absent.

[0202] Relationships between footprint DNA signals, markers of epigenetic activation, and transcriptional activity

[0203] Given that footprinted DNA fragments most likely originate from active binding of regulatory proteins and are overrepresented in open chromatin regions, it is critical to better understand the occurrence of liquid DNA footprinting in the context of epigenetic and transcriptional regulation. To determine the relationship between footprinted DNA signal intensity and promoter activation status, annotated promoter-like structures were assigned to active and inactive promoters based on cell-free histone 3 lysine 4 trimethylation (H3K4me3) signals in publicly available ChIP-Seq data from healthy individuals11. Potentially active promoters with strong H3K4me3 signals showed significantly higher footprinted DNA occupancy than promoters with weak or no H3K4me3 signals, while regular cfDNA showed the exact opposite behavior ( Figure 3 a). Furthermore, the strongest signals for H3K4me3, representing nucleosome positions, occur at local minima of footprinted DNA and vice versa. It is well known that DNA methylation is an important regulator of gene activity, associated with transcription factor binding and therefore potentially related to DNA footprint signals. Therefore, annotated CpG islands were clustered into methylated and unmethylated CpG islands based on signal intensity from cell-free methyl CpG binding domain sequencing (MBD-Seq) data of healthy individuals. Strongly methylated CpG islands showed weaker accumulation of footprinted DNA compared to unmethylated CpG islands, while regular cfDNA again exhibited the opposite behavior ( Figure 3b), clearly showing a strong relationship between DNA methylation and footprint DNA. To further analyze the connection between local footprint DNA signals and downstream gene transcription, we utilized publicly available cell-free RNA-seq data from five healthy individuals, Zhu, Y. et al. (2021). Four gene subgroups were defined based on the average transcript abundance levels of protein-coding genes: “no expression”, “low expression”, “medium expression”, and “high expression” genes ( Figure 3 c). Consistent with the definition of these gene subgroups, as the expression level at the corresponding TSS of the gene increases, the H3K4me3 histone modification increases and the DNA methylation level decreases ( Fig.11 ). Genes with moderate or high expression showed considerable enrichment of footprint DNA reads at their respective TSSs, whereas genes with low or no expression showed no substantial enrichment. Conventional cfDNA again exhibited the opposite behavior to footprint DNA and showed less pronounced differences between active expression (high and moderate) and low or no expression ( Figure 3 d). In summary, the signal intensity of footprint DNA is higher in active promoters than in inactive promoters, higher in unmethylated CpG islands than in methylated CpG islands, and higher in the TSS of actively transcribed genes than in the TSS of untranscribed genes. Therefore, footprint DNA is enriched at loci with markers of epigenetic activation and transcriptionally active genomic regions.

[0204] Liquid footprinting technology enables detection of disease-specific transcription factor binding sites in liquid biopsies

[0205] To evaluate the potential of liquid footprinting to identify disease-specific loci, footprint DNA data were generated for four biological replicates of four different clinical indications: two types of gastrointestinal cancer (pancreatic ductal adenocarcinoma (PDAC) and colorectal cancer (CRC)), as well as sepsis and postoperative controls (post-OP, Table 1, Tables 4-7). Postoperative clinical patients were selected as controls because they showed cfDNA concentration levels comparable to sepsis patients but without septic infection. Comparison of the consensus peak sets between PDAC and CRC revealed 731 different loci with at least a two-fold change with an adjusted p-value ≤ 0.1 ( Figure 4 a and Fig.12 For 1107 loci, significant differential enrichment was detected when comparing sepsis to post-OP ( Figure 4 b and Fig.12 ). Principal component analysis (PCA) based on all differentially enriched regions (DERs) showed that the first three principal components clearly separated all four clinical indications ( Figure 4 c). In the context of a larger locus with transcription factor binding sites, two exemplary DERs show distinct differential DNA footprints near the TSS ( Figure 4 d and e). Differentially occupied transcription factor binding sites were detected in the promoter of the lymphocyte cytoprotein 1 (LCP1) gene in CRC patients compared with PDAC patients ( Figure 4 d), while another binding transcription factor binding site was detected near the promoter of Plexin C1 gene (PLXNC1) in sepsis patients compared with post-OP patients ( Figure 4 e). In addition to differential analysis of signal intensities at defined consensus peaks, differential enrichment of transcription factor binding motifs in all consensus peaks was analyzed. To this end, the relative abundance of transcription factor motifs in consensus peaks of one condition was compared to consensus peaks of the corresponding condition to be compared. Enrichment of 14 different transcription factor motifs was detected in PDAC compared with CRC, 19 in CRC compared with PDAC, 14 in sepsis compared with post-OP, and 126 in post-OP compared with sepsis (E value ≤ 10, Figure 4 fi). For example, the most strongly enriched binding motif in PDAC patients compared with CRC patients belonged to the recombination signal binding protein of the immunoglobulin kappa J region (RBPJ; Figure 4 f). One of the physiological roles of the transcription factor RBPJ is the regulation of early pancreatic cell development. In general, liquid footprinting technology enables the detection of condition-specific occupancy of loci or transcription factor binding sites in liquid biopsies, which can be used to distinguish different diseases for diagnostic purposes. Footprint DNA sequencing may also help identify transcription factors that may have physiological relevance to the condition.

[0206] Table 4: Genomic locations of the highest DERs for PDAC identified when comparing PDAC to CRC on the human genome. The top 20 DERs are listed with reference to the human genome GenBank accession number = GCA_000001405.1

[0207]

[0208]

[0209] Table 5: Genomic locations of the highest DERs in CRC identified when comparing PDAC to CRC on the human genome. The top 20 DERs are listed with reference to the human genome GenBank accession number = GCA_000001405.1

[0210]

[0211] Table 6: Genomic locations of the highest DERs for sepsis identified when comparing sepsis to POPs on the human genome. The top 20 DERs are listed based on the human genome GenBank accession number = GCA_000001405.1

[0212]

[0213] Table 7: Genomic positions of the highest DERs for POPs identified in comparison of sepsis and POPs on the human genome. The top 20 DERs are listed based on human genome GenBank accession number = GCA_000001405.1

[0214]

[0215]

[0216] In this study, we present a novel approach for comprehensive DNA footprinting in liquid biopsies by analyzing short double-stranded plasma cell-free DNA. Our liquid footprinting assay protocol consists of preparative gel electrophoresis to specifically enrich for cfDNA fragments with an average length of approximately 40 bp, combined with high-throughput next-generation sequencing ( Figure 5 , 6 , 8). We observed a strong enrichment of short cfDNA at gene TSSs, while nucleosomal depletion of regular cfDNA was observed ( Figure 1 ac). A closer examination of the average signal at the TSS also showed that at the 5'UTR, the oscillation pattern was opposite to that of regular cfDNA. These findings suggest that nucleosomes regularly shift from the promoter of a gene to the 5'UTR and reveal the binding of regulatory DBPs between displaced nucleosomes through footprint DNA occupancy. In line with the enrichment of footprint DNA at the TSS, we also found that footprint DNA fragments showed a higher average GC content than the human genome (footprint DNA = 57.8%, human genome = 40.9%). Human gene regulatory elements have increased GC content, Benjamini and Speed ​​(2012); Zhang et al. (2004). Therefore, footprint DNA fragments and transcription factor binding sites enriched in these regions of the genome show elevated GC content. In addition to enrichment at the TSS, footprint DNA also accumulates at DNAse hypersensitive sites. At these open chromatin locations, regular cfDNA exhibits overlapping signals from high-frequency and low-frequency nucleosome positioning ( Figure 1 d), which was previously described by Ulz et al., 2019. Consistent with this, we observed two different types of footprint DNA signals generated by DBP with narrow binding signals, such as CTCF ( Figure 2 ae), and multiple adjacent protein binding events leading to a broader signal ( Fig.10 ). The narrow signals observed for regulatory proteins such as CTCF, SPI1 or REST can be explained by their ability to initiate chromatin remodeling or nucleosome replacement in closed chromatin, which is why they are called pioneer factors; Heinz et al. (2010); Barozzi et al. (2014); Fu et al. (2008); Vanzan et al. (2021). Pioneer factors can bind directly to closed chromatin without the need or presence of auxiliary proteins. In contrast, binding sites of regulatory proteins such as MYC or TP53BP1 show broader signals. These regulatory proteins appear to bind to loci with open chromatin in the presence of other regulatory proteins ( Fig.10 ). Notably, we also found a significant correlation between DNA footprinting signals and markers representing transcriptional regulation. In this context, we determined that promoters with the H3K4me3 histone modification, a signature of active gene transcription, showed stronger footprint DNA signals than promoters without the H3K4me3 histone modification ( Figure 3 a). In addition, we observed a correlation between the methylation level of CpG and the corresponding occupancy of footprint DNA ( Figure 3 b). Since unmethylated CpG islands are considered markers of epigenetic activation, depletion of conventional cfDNA and presence of DNA footprinting signals indicate binding of regulatory activators at these loci. Regarding gene expression, footprint DNA signals at TSS are increased at actively transcribed genes compared to lowly expressed genes. However, footprint DNA does not strictly capture the stepwise dynamics of gene expression levels, but rather exhibits a binary switch between high and low expression ( Figure 3 cd). Overall, footprint DNA was significantly associated with DNA methylation signatures, H3K4me3 histone modification, and transcriptional activity of downstream loci. Notably, we found that footprint DNA was consistently enriched in locations where conventional cfDNA was depleted. We propose that the molecular origin of footprint DNA is most likely cfDNA, which is protected from enzymatic digestion by binding to regulatory proteins including transcription factors ( Figure 2 e and f). Consistent with this model, we found that footprint DNA consensus peaks contained a large number of transcription factor motifs, and known transcription factor binding sites showed significant enrichment of footprint DNA (Table 3). Taken together, our data provide strong evidence that footprint DNA represents a characteristic subset of cfDNA that is distinct from nucleosomal cfDNA and does not result from the degradation of conventional cfDNA, which may help infer regulatory DNA binding events at the genome-wide level.

[0217] In contrast to footprint DNA, which is characterized by short double-stranded DNA (short dsDNA), publications by Snyder et al. (2016), Hudecova et al. (2022), and Hisano et al. (2021) reported the presence of short single-stranded DNA (short ssDNA). In contrast to short dsDNA, short ssDNA fragments may account for a significant portion of total cfDNA, comprising up to 20% of the sequencing library. However, short ssDNA showed very low relative signal intensities at transcription start sites or TFBS of proteins such as MYC compared to short dsDNA ( Fig.10 ). Therefore, we hypothesize that the cfDNA initially bound by regulatory DNA binding proteins is short dsDNA, while the short ssDNA may be of other biological origin or may represent further degradation products with a higher signal-to-noise ratio. Hudecova et al. and Hisano et al. proposed that a large fraction of short ssDNA may originate from non-canonical DNA structures such as G4 quadruplexes, which may largely superimpose the signal of regulatory protein-DNA binding events.

[0218] Transcription factor motifs of footprinted DNA revealed differentially enriched sets of transcription factors between conditions, including samples from patients with sepsis, colorectal cancer (CRC), and pancreatic adenocarcinoma (PDAC). Figure 4fi). Sepsis is characterized by a complex interplay of inflammatory and anti-inflammatory processes orchestrated by regulators of the immune system. Seven of the top 10 differentially enriched transcription factors identified in sepsis samples are associated with regulatory pathways of the immune system. For example, transcription factor 4 (TCF4, also known as E2-2 or ITF-2) regulates genes that regulate the differentiation of dendritic cells into interferon-producing plasmacytoid dendritic cells; Reizis (2010). TCF4 also regulates immune responses as a downstream target of TLR2 signaling to induce the expression of immune regulatory genes such as interleukin 10 (IL-10; Manoharan et al., 2014). The identified transcription factor ZFX is known to be involved in the maintenance of peripheral T cells and the expansion and maintenance of B cells during development and peripheral homeostasis, SmithRaska et al. (2018). These findings may reveal a possible link between transcription factors specific to immune system regulation and sepsis. For CRC patients, specificity protein 1 (SP1) was identified as a top hit, and several Fos / Jun family transcription factors were identified as differentially enriched. SP1 is a ubiquitous transcription factor that is a mediator of key physiological pathways including cell cycle, proliferation, and metastasis. SP1 plays a key role in regulating genes involved in colorectal cancer growth and metastasis26. Two members of the Fos / Jun family, FOS-like 1 (FOSL1) and FOS-like 2 (FOSL2), are known to promote tumorigenesis and metastasis in colon cancer; Liu et al. (2021). For PDAC patients, immunoglobulin kappa J region recombination signal binding protein (RBPJ) was identified as a corresponding top hit. RBPJ is known to form a heterocomplex with pancreas-related transcription factor 1a (PTF1A), and their interaction is required in the early stages of pancreatic growth, morphogenesis, and lineage fate determination (Masui et al., 2007). In addition, peroxisome proliferator-activated receptor gamma (PPARγ) and retinoid X receptor alpha (RXRα) have been identified and are known to form heterocomplexes; Lehrke & Lazar (2005). PPARγ is a key regulator of adipocyte differentiation, regulating the production and secretion of insulin and adipokines, and has been implicated in PDAC. In addition, RXRα promotes pancreatic cancer cell proliferation and inhibits apoptosis; Chen et al. (2019). In addition, we identified additional disease-specific differentially occupied loci that may be associated with underlying physiological changes in the disease and clearly separated patients using PCA ( Figure 4ae). For example, in patients with colorectal cancer, we found differential binding of DBP near the TSS of the lymphocytic cytoplasmic protein 1 (LCP1 also known as plastin-2) gene compared to patients with PDAC. LCP1 is associated with colorectal cancer progression, prognosis, and metastasis, and patients with advanced colorectal cancer have higher plastin-2 expression levels; Ning et al. (2014). In patients with sepsis, we found strong footprint DNA signals near the TSS of the Plexin C1 gene (PLXNC1) compared to postoperative controls. PLXNC1 enhances neutrophil adhesion, migration, and activation during acute lung injury and is induced during acute inflammatory responses (Granja et al., 2014).

[0219] In summary, analysis of short double-stranded fragments may provide the most accurate picture of genome-wide transcription factor footprints in liquid biopsies. With the ability to identify disease-specific transcription factor binding site occupancy for patient stratification, we see great potential for footprint DNA sequencing in liquid biopsy applications.

[0220] The results reflected in this study are based on relatively small sample numbers. Future validation studies with larger cohort sizes will aid clinical validation. In addition to clinical validity and practicality, cost-effectiveness is also an important factor in clinical diagnostic approaches. Although our workflow already reflects a resource-saving alternative to ultra-deep high-throughput sequencing of total cfDNA, a targeted and even more economical approach may be more conducive to implementation. By using new liquid footprinting methods that directly map regulatory protein binding to the genome, we hope to add a new tool to liquid biopsy diagnostics that can improve cancer detection or enable clinically relevant differential diagnosis of cancer types.

[0221] References:

[0222] Amemiya, et al. (2019) Sci Rep 9,9354; 10.1038 / s41598-019-45839-z.

[0223] Andrews, S. (2010) FASTQC. Babraham Bioinformatics, Babraham Institute, Cambridge, United Kingdom, 2010).

[0224] Aucamp et al. (2018) Biol. Rev.93,1649-1683.

[0225] Barozzi et al.(2014)Molecular cell 54,844-857.

[0226] Benjamini&Speed(2012)Nucleic Acids Research 40,e72.

[0227] Buitrago et al.(2019)Nucleic Acids Research 47,9511-9523.

[0228] Bushnell et al.(2017)PloS one 12,e0185056.

[0229] Burnham,et al.(2016)Scientific Reports 6,27859;10.1038 / srep27859.

[0230] Chen et al.(2019)European review for medical and pharmacologicalsciences 23.

[0231] Fu et al(2008)PLoS genetics 4,e1000138.

[0232] Granja et al.(2014)Mucosal immunology 7,879-891.

[0233] Heinz et al(2010)Molecular cell 38,576-589.

[0234] Heitzer et al.(2020)Trends in molecular medicine 26,519-528.

[0235] Hisano et al.(2021)BMC biology 19,225.

[0236] Hudecova et al.(2022)Genome research 32,215-227.

[0237] Kolmykov et al.(2021)Nucleic Acids Research 49,D104-D111.

[0238] Kucukural et al.(2019)BMC genomics 20,6.

[0239] Kulakovskiy et al.(2018)Nucleic Acids Research 46,D252-D259.

[0240] Lehrke & Lazar (2005) Cell 123,993-999.

[0241] Li et al.(2009)Bioinformatics(Oxford,England)25,2078-2079.

[0242] Liu et al.(2021)Pharmacological research 165,105405.

[0243] Liao et al.(2014)Bioinformatics(Oxford,England)30,923–930.

[0244] Lo et al.(2021)Science 372;10.1126 / science.aaw3616.

[0245] Manoharan et al.(2014)Journal of immunology 193,4203-4213.

[0246] Martin,M.(2011)EMBnet j.17,10.

[0247] Masui et al.(2007)Genes&development 21,2629-2643.

[0248] McLeay&Bailey(2010)BMC bioinformatics 11,165.

[0249] Ning.et al.(2014)Molecular cancer therapeutics 13,528-539.

[0250] Piovesan et al.(2019)BMC research notes 12,106.

[0251] Ramírez et al.(2016)Nucleic Acids Research 44,W160-5.

[0252] Reizis(2010)Current opinion in immunology 22,206-211.

[0253] Robinson et al.(2010)Bioinformatics(Oxford,England)26,139-140.

[0254] Schmieder&Edwards(2011)Bioinformatics(Oxford,England)27,863–864.

[0255] Sedlazeck,et al.(2013)Bioinformatics(Oxford,England)29,2790-2791.

[0256] Smith-Raska et al.(2018)Frontiers in immunology 9,1482.

[0257] Snyder et al.(2016)Cell 164,57-68;10.

[0258] Sun,et al.(2015)PNAS 112,E5503-12.

[0259] Quinlan&Hall(2010)Bioinformatics(Oxford,England)26,841-842.

[0260] Ulz et al.(2019)Nat Commun 10,4666;10.1038 / s41467-019-12714-4.

[0261] Vanzan et al.(2021)Nat Commun 12,3337.

[0262] Yu et al.(2015)Bioinformatics(Oxford,England)31,2382-2383.Zhang etal.(2004)PNAS 101,16855–16860.

[0263] Zhu,Y.et al.(2021)Theranostics 11,181-193。

Claims

1. A method for identifying a cfDNA-derived biomarker for a target physiological condition, disease, or disorder, comprising: a) isolating cfDNA from a sample containing cfDNA from a subject known to be in a target physiological condition or to have a target disease or condition; b) selecting short double-stranded cfDNA fragments; c) determining the nucleic acid sequence of the short double-stranded cfDNA fragments; d) comparing the nucleic acid sequence of the short double-stranded cfDNA fragments to a reference; as well as e) identifying, based on said comparing in step d), at least one cfDNA-derived biomarker associated with a target physiological condition, disease, or disorder.

2. The method of claim 1, wherein the short double-stranded cfDNA fragments are double-stranded cfDNA fragments with a fragment size of less than 100 bp.

3. The method of claim 1 or 2, wherein the sample is a body fluid sample.

4. The method of claim 1 or 2, wherein the selection of short double-stranded cfDNA fragments comprises gel electrophoresis size separation of double-stranded cfDNA, gradient gel electrophoresis, PCR-based methods, hybridization and capture, and / or selection using binding protein cross-linking and capture.

5. The method of claim 4, wherein the selecting short double-stranded cfDNA fragments comprises gel electrophoresis size separation of a double-stranded cfDNA sequencing library.

6. The method of claim 1 or 2, wherein the determined nucleic acid sequence comprises a protein binding footprint DNA sequence, and the protein binding footprint DNA sequence includes a protein-DNA interaction site, such as a transcription factor (TF) binding site, a structural protein binding site, a regulatory protein binding site, a transcription start site or a gene promoter region, a CpG island.

7. The method of claim 1 or 2, wherein the reference comprises one or more short double-stranded cfDNA fragments, wherein the one or more short double-stranded cfDNA fragments (i) are derived from a body fluid sample of a healthy subject or from a body fluid sample of a subject who is not in a target physiological condition or does not suffer from a target disease or condition, or (ii) are known short double-stranded cfDNA fragments.

8. The method of claim 1 or 2, wherein the target physiological condition, disease or disorder is selected from the group consisting of: i) cancer, including pancreatic ductal adenocarcinoma (PDAC), colorectal cancer (CRC), non-small cell lung cancer (NSCLC), bone cancer; ii) fungal, parasitic, viral or bacterial infections, including sepsis, pancreatitis, Lyme disease, encephalitis; iii) chronic inflammatory diseases, including chronic inflammatory bowel disease (IBD), such as ulcerative colitis or Crohn's disease; iv) autoimmune diseases, including multiple sclerosis (MS), rheumatic diseases; v) Metabolic disorders; vi) irritable bowel syndrome; vii) genetic diseases or developmental abnormalities, especially those affecting the fetus, including trisomy; and viii) Pregnancy and development stage.

9. The method of claim 1 or 2, further comprising a step of mapping the determined nucleic acid sequences of the short double-stranded cfDNA fragments to a reference genome.

10. The method of claim 1 or 2, comprising identifying differentially enriched regions (DERs), typically by comparing the nucleic acid sequences of the short double-stranded cfDNA fragments of a sample to a reference.

11. The method of claim 9, wherein at least one cfDNA-derived biomarker associated with a disease or condition is based on a combination of more than one DER.

12. A method for assessing a target physiological condition, disease or condition in a sample from a subject suspected of being in the target physiological condition or suffering from a target disease or condition, comprising: a) isolating cfDNA from a sample containing cfDNA from the subject; b) selecting short double-stranded cfDNA fragments; c) determining the nucleic acid sequence of the short double-stranded cfDNA fragments; and d) determining the presence, absence or abundance of at least one cfDNA-derived biomarker associated with a target physiological condition, disease or disorder in the determined nucleic acid sequences of the short double-stranded cfDNA fragments.

13. A method for determining the efficacy of treating a target physiological condition, disease or condition in a sample from a subject known to be in the target physiological condition or to suffer from a target disease or condition, comprising: a) isolating cfDNA from a sample containing cfDNA from the subject; b) selecting short double-stranded cfDNA fragments; c) determining the nucleic acid sequence of the short double-stranded cfDNA fragments; d) determining the presence, absence or abundance of at least one cfDNA-derived biomarker associated with a target physiological condition, disease or disorder in the determined nucleic acid sequences of the short double-stranded cfDNA fragments; e) repeating steps a) to d); and f) determining the efficacy of the treatment based on the determination over time.

14. An apparatus for detecting at least one cfDNA-derived biomarker in a liquid biopsy sample, comprising at least: 1) A sample collector and selector for isolating and selecting cfDNA; 2) Sequencer; and 3) An evaluation unit, which comprises at least a processor.

15. Use of cfDNA-derived biomarkers in a sample for assessing a disease or condition in a subject suspected of having the disease or condition.

16. cfDNA-derived biomarkers for use in diagnosing a disease or condition in a subject suspected of having the disease or condition.

Images