Fragment omics in urine and plasma

Abstract

The fragment omics features provide various properties of a sample (e.g., urine or plasma) and / or a subject. The relative contribution or enrichment of clinically relevant DNA (e.g., the type of renal and non-renal urine cfDNA) is provided utilizing the fragment omics characteristics of urine free DNA. Such measurements may be used to reflect glomerular permeability and to monitor various diseases, such as kidney abnormalities. The fragmenomics characteristics may include corrected urine DNA concentration, length, terminal motif of the urine DNA molecule, and cfDNA molecules from open chromatin regions (OCRs) of one or more tissues. Furthermore, the nuclease activity or other fragmentation process of the cfDNA is determined based on the relative contributions associated with different cfDNA cleavage profiles, which are also used to determine the contribution, pathological grade and gestational age of the cfDNA from the tissue.

Description

Cross - reference to Related Applications

[0001] This application is a PCT application and claims the benefit of U.S. Provisional Patent Application No. 63 / 428,694, entitled "Fragmentomics in Urine and Plasma", filed on November 29, 2022, the entire content of which is incorporated herein by reference for all purposes. Background of the Invention

[0002] Urine - free DNA and plasma - free DNA molecules include DNA molecules released from different normal tissues / organs and malignant cells, but exhibit different fragmentation patterns. For example, compared with plasma - free DNA, urine - free DNA molecules generally have a shorter length distribution and are enriched in more distinct and sharper 10 - bp periodic peaks (Tsui et al., PLOS ONE 2012; 7: e48319). Using a mouse model with a deletion of the DNases gene, it was demonstrated that deoxyribonuclease 1 - like 3 (DNASE1L3) is the main contributor to the fragmentation of plasma - free DNA (Han et al. Am J Hum Genet. 2020; 106: 202–214). Compared with plasma samples, deoxyribonuclease 1 (DNASE1) is responsible for shaping the fragmentation characteristics of urine - free DNA (Chen et al. PLOS Genet. 2022; 18: e1010262).

[0003] Urine - free DNA can also include different types of DNA molecules with their respective characteristics. For example, there are "trans - renal" urine - free DNA molecules released from non - urinary systems (e.g., blood cells, liver, lung, colon, heart, brain, spleen, stomach, placental tissue) that reach the urinary system through the glomeruli of the kidneys. In addition to trans - renal free DNA, there are "non - trans - renal" urine - free DNA molecules that originate from the urinary system and are released directly from the urinary system, such as the renal tubules, bladder, urethra, etc. However, there is a lack of methods for identifying the characteristics or reflecting the extent of trans - renal and non - trans - renal free DNA from a given urine sample.

[0004] In addition, numerous studies have demonstrated that the use of plasma end motifs can inform the presence of various diseases ranging from autoimmune diseases to multiple cancer types (Chan et al. Am J Hum Genet. 2020; 107:882–894; Jiang et al. Cancer Discov. 2020; 10:664–73). Thus, comprehensively determining the usage levels of nucleases such as DNASE1L3, DNASE1, DNA fragmentation factor subunit beta (DFFB), etc. is clinically meaningful. We infer that using an end motif profile will allow inferring the extent of nucleases involved in the generation of cell-free DNA molecules (i.e., nuclease usage levels) and monitoring nuclease activities in different pathophysiological states. However, there is a lack of tools that allow for a comprehensive assessment of various DNA nucleases in a single assay. SUMMARY OF THE INVENTION

[0005] Methods, apparatuses, and systems are provided for determining the fragmentomics signature of a sample for various properties of a sample and / or a subject. The various methods can be used for urine samples and / or plasma samples.

[0006] As an example of a urine sample, the fragmentomics signature of urine cell-free DNA can be used to provide concentration fractions or enrichment of clinically relevant DNA (e.g., types of renal and non-renal urine cfDNA). Measurements of urine DNA fragmentomics can also be used to reflect glomerular permeability and monitor various diseases such as kidney abnormalities. The fragmentomics signature can include the corrected urine DNA concentration, length, end motifs of urine DNA molecules, and cfDNA molecules from open chromatin regions (OCRs) of one or more tissues.

[0007] In other embodiments (e.g., for urine, plasma, or other cell-free samples), nuclease activity or other fragmentation processes of cfDNA can be determined based on the relative contributions of different cfDNA cleavage profiles, which can also be used to determine concentration fractions, pathological grades, and gestational ages of cfDNA from tissues.

[0008] These and other embodiments of the present disclosure are described in detail below. For example, other embodiments relate to systems, devices, and computer-readable media associated with the methods described herein.

[0009] A better understanding of the nature and advantages of the embodiments of the present disclosure can be obtained with reference to the following detailed description and the accompanying drawings.

[0010] Referring to the remainder of the specification, including the drawings and claims, other features and advantages of the present disclosure will be recognized. Other features and advantages of the present disclosure, as well as the structure and operation of various embodiments of the present disclosure, will be described in detail below with reference to the drawings. In the drawings, the same reference numerals may represent the same or functionally similar components. Brief Description of the Drawings

[0011] Figure 1 Illustrates an exemplary overview of characterizing renal and non-renal DNA in urine samples according to some embodiments.

[0012] Figure 2 Shows a schematic diagram that illustrates the use of fragmentomics signatures to determine the renal urine cell-free DNA contribution.

[0013] Figure 3 Illustrates the process of obtaining sequencing data from urine samples according to some embodiments.

[0014] Figure 4 Shows a set of graphs that display the comparison of urine cell-free DNA concentrations before and after in vitro incubation in three urine collection groups.

[0015] Figure 5 Shows a set of graphs that display the comparison of urine cell-free DNA length distributions before and after in vitro incubation in the identification control, EDTA, and stabilizer groups.

[0016] Figure 6 Shows a graph that identifies the length difference between fetal DNA and maternal DNA in urine samples according to some embodiments.

[0017] Figure 7 Illustrates an exemplary schematic diagram that shows the biological process of converting plasma DNA into renal DNA according to some embodiments.

[0018] Figure 8 Shows a graph that identifies the relationship between glomerular basement membrane permeability and renal DNA length according to some embodiments.

[0019] Figure 9 Shows a graph that identifies different end motifs identified from fetal-specific cell-free DNA and common cell-free DNA in maternal urine.

[0020] Figure 10 Shows a set of graphs that identify the relationship between the concentration fraction of fetal DNA in urine cell-free DNA and the CC-terminal fragment and different lengths according to some embodiments.

[0021] Figure 11 Shows an exemplary graph that identifies various characteristics of DNA molecules from open chromatin regions according to some embodiments.

[0022] Figure 12 Shows an exemplary graph for determining the amount of urine-free DNA corresponding to open chromatin regions according to some embodiments.

[0023] Figure 13 Displays a set of box plots of the O / E ratios of fetal-specific free DNA molecules and common free DNA molecules in urine samples and plasma samples.

[0024] Figure 14 Shows a set of graphs that identify the enrichment of DNA molecules in open chromatin regions in fetal-specific DNA from urine samples according to some embodiments.

[0025] Figure 15 Shows a set of graphs that identify the non-enrichment of DNA molecules in open chromatin regions in fetal-specific DNA from plasma samples according to some embodiments.

[0026] Figure 16 Displays a graph showing the correlation between the fetal DNA fraction in maternal urine and the O / E ratio of all urine-free DNA fragments.

[0027] Figure 17 Displays a graph illustrating the correlation between the fetal DNA fraction in maternal urine and the O / E ratio of urine cfDNA fragments from placenta-specific DHS.

[0028] Figure 18 Displays a graph identifying the normalized end density of all urine-free DNA in the OCR of maternal urine and plasma samples.

[0029] Figure 19 Displays a set of graphs that identify the comparison between the end densities of urine-free DNA and plasma-free DNA for determining the concentration fraction of fetal DNA according to some embodiments.

[0030] Figure 20 Displays a set of graphs that identify the correlation between the fetal DNA fraction and the normalized end density of urine-free DNA of different lengths according to some embodiments.

[0031] Figure 21 Displays a flowchart for estimating the concentration fraction of clinically relevant DNA molecules in a subject's urine sample according to some embodiments.

[0032] Figure 22 Displays a set of graphs that identify the correlation between the fetal DNA fraction and the proportion of urine-free DNA fragments carrying the CC end.

[0033] Figure 23Describes techniques for enriching a collection of one or more terminal motifs using a probe.

[0034] Figure 24 Describes another technique for enriching a collection of one or more terminal motifs using a probe and beads.

[0035] Figure 25 Shows a flowchart for enriching clinically relevant DNA in a urine sample based on terminal motif signatures of urine-free DNA according to some embodiments.

[0036] Figure 26 Shows a set of graphs that identify enrichment of fetal DNA using urine-free DNA with various fragmentomic properties according to some embodiments.

[0037] Figure 27 Shows a bar graph that identifies enrichment of renal urine-free DNA using selective analysis of fragments with different fragmentomic characteristics.

[0038] Figure 28 Shows a flowchart for enriching clinically relevant DNA in a urine sample based on terminal motifs, open chromatin regions, and length of urine-free DNA according to some embodiments.

[0039] Figure 29 Shows a set of graphs that identify O / E ratio analysis in patients with RCC.

[0040] Figure 30 Shows fragmentomic analysis of renal DNA in proteinuric patients using blood-specific DHS.

[0041] Figure 31 Shows fragmentomic analysis of renal DNA in pregnant women with preeclampsia using DHS.

[0042] Figure 32 Shows a flowchart of a method for determining classification of kidney abnormalities based on urine-free DNA from open chromatin regions according to some embodiments.

[0043] Figure 33 Is fragmentomic analysis of renal DNA in patients with proteinuria and preeclampsia alone.

[0044] Figure 34 Shows a flowchart for determining classification of kidney abnormalities based on the length of urine-free DNA according to some embodiments.

[0045] Figure 35 Respectively shows analysis of urine cfDNA concentration of renal DNA in patients with proteinuria and preeclampsia.

[0046] Figure 36 is a flow chart for determining the classification of kidney abnormalities based on urine free DNA concentration according to some embodiments.

[0047] Figure 37 Shows ROC analysis using fragmentomics features of renal DNA to distinguish patients with proteinuria and preeclampsia from healthy controls.

[0048] Figure 38 Shows a figure ranking the frequencies of certain end motifs present in urine free DNA molecules according to some embodiments.

[0049] Figure 39 Shows a set of figures identifying the end motif profiles of murine plasma and urine free DNA molecules observed.

[0050] Figure 40 Shows a schematic workflow for exemplary nuclease usage level analysis of free DNA molecules.

[0051] Figure 41 Shows a figure identifying the proportional contribution of each F-profile (i.e., nuclease usage level) inferred from murine free DNA samples with different knockout genotypes using NMF analysis.

[0052] Figure 42 Shows a set of figures of six F-profiles (A-F) inferred from mouse plasma and urine free DNA using NMF analysis.

[0053] Figure 43 Shows a box plot identifying the proportional contribution of F-profile I on different types of samples according to some embodiments.

[0054] Figure 44 Shows the relative frequencies of free DNA molecules across 256 end motifs of F-profile I according to some embodiments.

[0055] Figure 45 Shows a box plot identifying the proportional contribution of F-profile II across different types of samples according to some embodiments.

[0056] Figure 46 Shows the relative frequencies of free DNA molecules across 256 end motifs of F-profile II according to some embodiments.

[0057] Figure 47 Shows a box plot identifying the proportional contribution of F-profile III across different types of samples according to some embodiments.

[0058] Figure 48Shows the relative frequencies of cell-free DNA molecules across 256 end motifs of the F-spectrum III, according to some embodiments.

[0059] Figure 49 Shows the relative frequencies of cell-free DNA molecules across 256 end motifs of the F-spectrum IV-VI, according to some embodiments.

[0060] Figure 50 Shows a schematic diagram comparing the end motif spectrum of a human subject with a reference F-spectrum determined based on murine samples, according to some embodiments.

[0061] Figure 51 Shows the proportional contributions across the F-spectrum of plasma and urine samples of a human subject, according to some embodiments.

[0062] Figure 52 Shows the proportional contributions across the F-spectrum of normal and DNASE1L3-deficient samples of a human subject, according to some embodiments.

[0063] Figure 53 Shows the proportional contributions across the F-spectrum of urine samples of pregnant human subjects, according to some embodiments.

[0064] Figure 54 Shows a flow chart for determining nuclease activity classification based on the F-spectrum of cell-free DNA molecules, according to some embodiments.

[0065] Figure 55 Shows a set of graphs for identifying the level of nuclease usage in cell-free DNA of pregnant women's urine.

[0066] Figure 56 Shows a flow chart for determining the concentration fraction of fetal DNA based on the F-spectrum of cell-free DNA molecules, according to some embodiments.

[0067] Figure 57 Shows a set of graphs for identifying the level of nuclease usage in cell-free DNA of pregnant women's plasma.

[0068] Figure 58 Shows a set of graphs for identifying the F-spectrum analysis and oxidative stress level of pregnant women.

[0069] Figure 59 Shows a flow chart for estimating gestational age based on the F-spectrum of cell-free DNA molecules, according to some embodiments.

[0070] Figure 60 Shows box plots of the F-spectrum I (DNASE1L3) levels of healthy subjects, patients with DNASE1L3 deficiency, and the parents of the patients.

[0071] Figure 61 Shows a set of graphs depicting nuclease usage levels in plasma free DNA of subjects with and without systemic lupus erythematosus (SLE).

[0072] Figure 62 Shows the proportional contributions of the F-profile in normal, HBV, and HCC plasma samples according to some embodiments.

[0073] Figure 63 Shows a set of graphs that depict nuclease usage levels in plasma free DNA of subjects with and without HCC.

[0074] Figure 64 Shows a bar graph that identifies the levels of oxidative stress in blood samples from controls and HCC patients.

[0075] Figure 65 Shows a set of graphs that provide F-profile analysis and levels of oxidative stress in CRC patients.

[0076] Figure 66 Shows a box plot that identifies the F-profile VI contributions in NPC patients before and during chemoradiation with cisplatin.

[0077] Figure 67 Shows a flowchart for the classification for determining the pathological grade based on the cell-free DNA-based F-profile according to some embodiments.

[0078] Figure 68 Illustrates a measurement system according to embodiments of the present disclosure.

[0079] Figure 69 Shows a block diagram of an exemplary computer system that can be used with the systems and methods according to embodiments of the present invention. The term

[0080] “Tissue” corresponds to a group of cells that aggregate together as a functional unit. More than one type of cell can be found in a single tissue. Different types of tissue can be composed of different types of cells (e.g., hepatocytes, alveolar cells, or blood cells), but can also correspond to tissue from different organisms (mother to fetus) or to healthy cells versus tumor cells. “Reference tissue” can correspond to the tissue used to assay the tissue-specific methylation level. Multiple samples of the same tissue type from different individuals can be used to determine the tissue-specific methylation level of that tissue type.

[0081] "Biological sample" refers to any sample taken from a subject (e.g., a human (or other animal), such as a pregnant woman, a person with cancer or other disease, or a person suspected of having cancer or other disease, an organ transplant recipient, or a subject suspected of having a disease process involving an organ (e.g., the heart in myocardial infarction, or the brain in stroke, or the hematopoietic system in anemia)) and containing one or more nucleic acid molecules of interest. A biological sample can be a body fluid, such as blood, plasma, serum, urine, vaginal fluid, fluid from the scrotum (e.g., ascites from the testis), vaginal lavage fluid, pleural fluid, ascites, cerebrospinal fluid, saliva, sweat, tears, sputum, bronchoalveolar lavage fluid, drainage fluid from the nipple, aspirated fluid from different parts of the body (e.g., thyroid, breast), intraocular fluid (e.g., aqueous humor), etc. Fecal samples can also be used. In various embodiments, most of the DNA in a biological sample that has been enriched for cell-free DNA (e.g., a plasma sample obtained by a centrifugation protocol) can be cell-free, e.g., greater than 50%, 60%, 70%, 80%, 90%, 95% or 99% of the DNA can be cell-free. The centrifugation protocol can include, for example, 3,000 g × 10 minutes to obtain the liquid fraction, and then centrifuging again at, e.g., 16,000 g for 10 minutes to remove residual cells. As part of the analysis of the biological sample, a statistically significant number of cell-free DNA molecules of the biological sample can be analyzed (to provide an accurate measurement). In some embodiments, at least 1,000 cell-free DNA molecules are analyzed. In other embodiments, at least 10,000 or 50,000 or 100,000 or 500,000 or 1,000,000 or 5,000,000 cell-free DNA molecules or more can be analyzed. At least the same number of sequence reads can be analyzed.

[0082] "Sequence read" refers to a string of nucleotides sequenced from any part or all of a nucleic acid molecule. For example, a sequence read can be a short string of nucleotides (e.g., 20 - 150 nucleotides) sequenced from a nucleic acid fragment, a short string of nucleotides at one or both ends of a nucleic acid fragment, or the sequencing of an entire nucleic acid fragment present in a biological sample. Sequence reads can be obtained in a variety of ways, such as using sequencing technologies or using probes, e.g., in hybridization arrays or capture probes that can be used for microarrays, or amplification technologies, such as polymerase chain reaction (PCR) or linear amplification using a single primer or isothermal amplification. As part of the analysis of a biological sample, at least 1,000 sequence reads can be analyzed. As other examples, at least 10,000 or 50,000 or 100,000 or 500,000 or 1,000,000 or 5,000,000 sequence reads or more can be analyzed. A certain amount of sequence reads can be used as a representative of the number of DNA fragments. To determine the number of DNA fragments from the amount of sequence reads, calculations can be performed to account for paired-end sequencing and / or biases in the sequencing technology.

[0083] Sequence reads can include "end sequences" associated with the ends of a fragment. The end sequences can correspond to the outermost N bases of the fragment, e.g., 2-30 bases at the end of the fragment. If the sequence read corresponds to an entire fragment, the sequence read can include two end sequences. When paired-end sequencing provides two sequence reads corresponding to the ends of a fragment, each sequence read can include one end sequence.

[0084] A "sequence motif" can refer to a short repeating pattern of bases in a DNA fragment (e.g., a cell-free DNA fragment). The sequence motif can occur at the end of the fragment and thus be part of or include the end sequence. An "end motif" can refer to a sequence motif of an end sequence that preferably occurs at the end of a DNA fragment that may be a particular type of tissue. The end motif can also occur before or after the end of the fragment and thus still correspond to the end sequence. A nuclease can have a particular cleavage preference for a particular end motif and a second most preferred cleavage preference for a second end motif.

[0085] The term "mapping" refers to the process of relating a sequence to a position or coordinates (e.g., genomic coordinates) in a reference (e.g., a reference genome) having a known reference sequence, where the sequence is similar to the known reference sequence at the position in the reference. The degree of similarity can be measured or reported according to "mapping quality". In one example of mapping quality used herein, a mapping quality X of a sequence relative to the position or coordinates reported in the reference indicates that the probability that the sequence maps to a different position is no greater than 10^(-X / 10). For example, a mapping quality of 30 indicates that the probability that the sequence maps to an alternative position is less than 0.1%.

[0086] A "reference genome" can be the entire genomic sequence of a reference organism, a portion of a reference genome, a consensus sequence of a number of reference organisms, a compiled sequence based on different components of different organisms, or any other suitable reference sequence. The reference can also include information about variations of the reference known to be found in a population of organisms.

[0087] The "ratio" at which DNA molecules terminate at a position is related to the frequency at which DNA molecules terminate at that position. Such a ratio can be referred to as "end density". The ratio can be based on the number of DNA molecules that terminate at that position normalized relative to the number of DNA molecules analyzed. The normalization can also be based on the average, median, or total number of ends in the surrounding region. The surrounding region used for normalization can include, but is not limited to, 500, 1000, 3000, 5000, etc. bp located upstream and / or downstream of that position.

[0088] "Relative frequency" (also referred to simply as "frequency") can refer to a proportion (e.g., percentage, fraction, or concentration). In particular, the relative frequency of a specific terminal motif (e.g., CCGA or just a single base) can provide the amount of free DNA fragments in a sample related to the terminal motif CCGA, e.g., by the proportion of terminal sequences having CCGA.

[0089] The terms "control", "control sample", "background sample", "reference", "reference sample", "normal", and "normal sample" may be used interchangeably to generally describe a sample that does not have a specific medical condition (renal abnormality) or is otherwise healthy. In one example, a reference sample is a sample taken from a subject without a medical condition. A reference sample can be obtained from a subject or from a database.

[0090] "Clinically relevant DNA" can refer to DNA of a specific tissue origin to be measured, e.g., to determine the concentration fraction of such DNA or to classify the phenotype of a sample (e.g., plasma). In the context herein, as opposed to non-renal DNA (e.g., from kidney or bladder), clinically relevant DNA can refer to renal DNA that was present prior to passing through the kidney. Examples of clinically relevant DNA are fetal DNA (e.g., from maternal plasma) or tumor DNA (e.g., from a patient's plasma). Another example includes measuring the amount of graft-related DNA in the urine of a transplant patient. Another example includes measuring the concentration fraction of liver DNA fragments (or other non-hematopoietic or hematopoietic tissues, e.g., blood cells) in a sample or the concentration fraction of brain DNA fragments in cerebrospinal fluid.

[0091] "Calibration sample" can correspond to a biological sample whose desired measurement value (e.g., nuclease activity, concentration fraction of clinically relevant nucleic acid, classification of a genetic disorder, or other desired property) is known or determined by a calibration method, such as ELISA for measuring the amount of nuclease or a determination of the rate of DNA digestion by a quantitative nuclease for measuring nuclease activity (exemplary methods can include fluorescence measurement or spectrophotometric measurement of the amount of DNA before and after adding the nuclease-containing sample, or in real time with the addition of the nuclease-containing sample; another example is using a radial enzyme diffusion method). The concentration fraction of clinically relevant DNA (e.g., tissue-specific DNA fraction) can be known, e.g., by a calibration method, such as using tissue-specific allele assays. For example, for tumors, fetuses, or transplants, alleles that are present in the tissue (e.g., the donor's genome) but not in the healthy / maternal / receptor's genome can be used as markers for the tissue corresponding to clinically relevant DNA. As another example, tissue-specific methylation patterns can be used. A calibration sample can have individual measurement values (e.g., the amount of fragments having a specific terminal motif or having a specific length), which can be determined to be related to the desired measurement value.

[0092] "Calibration data points" include "calibration values" (e.g., the quantity of fragments having a specific end motif or having a specific length) and measured or known values that are desired to be determined for other test samples. Calibration values can be determined from various types of data measured from the DNA molecules of a sample (e.g., the quantity of fragments having an end motif or having a specific length). The calibration values correspond to parameters related to the desired property, such as the classification of a genetic disorder, nuclease activity, or the efficacy of an anticoagulant dosage. For example, calibration values can be determined from measured values determined for a calibration sample for which the desired property is known. Calibration data points can be defined in a variety of ways, e.g., as discrete points or as a calibration function (also referred to as a calibration curve or calibration surface). The calibration function can be derived from additional mathematical transformations of the calibration data points.

[0093] The term "fetal DNA concentration fraction" can be used interchangeably with the terms "fetal DNA proportion" and "fetal DNA fraction", and refers to the proportion of fetal DNA molecules from a fetus present in a biological sample (e.g., a maternal plasma or serum sample) (Lo et al, Am J Hum Genet. 1998; 62:768-775; Lun et al, Clin Chem. 2008; 54:1664-1672). Similarly, a tumor fraction or tumor DNA fraction can refer to the concentration fraction of tumor DNA in a biological sample.

[0094] A "locus" (also referred to as a "genomic locus") corresponds to a single site, which can be a single base position or a set of related base positions, such as a CpG site, a TSS site, a DNase hypersensitivity site, or a larger set of related base positions. A "locus" can correspond to a region that includes multiple sites. A locus can include only one site, which would make the locus equivalent to the site in this context.

[0095] The term "open chromatin region (OCR)" refers to one or more sites corresponding to regions lacking nucleosomes (i.e., lacking histone-bound DNA). In some cases, OCRs include one or more DNase1 hypersensitive sites (DHSs) defined using DNase-seq (Meuleman et al. Nature. 2020;584:244-251). As an example, OCRs can be defined based on sites identified using DNase-seq, sites identified using assay for transposase-accessible chromatin using sequencing (ATAC-seq), transcription start sites (TSSs), CCCTC-binding factor (CTCF) sites, enhancer sites, regions marked by histone modifications (e.g., H3K27ac, H3K4me3, etc.), and other nuclease hypersensitive sites. In some cases, OCRs can be regions of relatively reduced nucleosome occupancy. In some cases, OCRs can be tissue-specific. In various embodiments, at least 100, 500, 1,000, 5,000, or 10,000 OCRs can be used in the embodiments described herein.

[0096] The term "kidney abnormality" refers to a condition that affects the kidney and possibly other organs. As an example, kidney abnormalities can include renal cell carcinoma (RCC), nephrotic syndrome, glomerulonephritis, Fabry disease, cystinosis, IgA nephropathy, IgM nephropathy, lupus nephritis, atypical hemolytic uremic syndrome (aHUS), polycystic kidney disease (PKD), Alport syndrome, interstitial nephritis, proteinuria, chronic kidney disease (CKD), acute kidney injury, preeclampsia, etc.

[0097] "End motif profile" can refer to the relationship of the end sequences (e.g., 1-30 bases) of free DNA fragments (also referred to as DNA fragments) in a sample. Various relationships can be provided, such as the amount of free DNA fragments having a specific end sequence (end motif), the relative frequency of free DNA fragments having a specific end sequence compared to one or more other end sequences. In some cases, other types of parameters (e.g., length) are used to determine the end motif profile. For example, the end motif profile can be provided in a variety of ways that illustrate the amount of free DNA fragments having one or more specific end sequences for a given length (single length or length range). A "reference end motif profile" or "F-profile" is an end motif profile that can be generated by applying a factorization algorithm (e.g., non-negative matrix factorization) to the relative frequencies of DNA molecules in a given biological sample spanning multiple end motifs (e.g., 256 end motifs).

[0098] The term "relative abundance" generally refers to the ratio of a first amount of nucleic acid fragments having a particular characteristic (e.g., a particular length, terminating at one or more particular coordinates / termination positions, or aligning (mapping) to a particular region of the genome) to a second amount of nucleic acid fragments having a particular characteristic (e.g., a particular length, terminating at one or more particular coordinates / termination positions, or aligning (mapping) to a particular region of the genome). In one example, relative abundance can refer to the ratio of the number of DNA fragments terminating at a first set of genomic positions (e.g., open chromatin regions) to the number (e.g., average or median) of DNA fragments terminating at a second set of genomic positions, which can be all genomic positions. Such relative abundance can be referred to as end density. In some aspects, "relative abundance" is a type of discrete value that relates the amount (one value) of cell-free DNA molecules terminating within one window of genomic positions to the amount (other values) of cell-free DNA molecules terminating within another window of genomic positions. The two windows can overlap, but can also have different lengths. In other embodiments, the two windows do not overlap. Additionally, the window can be one nucleotide in width and thus equivalent to one genomic position. End density is a type of relative abundance. In some cases, the observed-to-expected (O / E) ratio is another type of relative abundance.

[0099] As used herein, the term "classification" refers to any number or other character associated with a particular property of a sample. For example, the "+" symbol (or the word "positive") can indicate that the sample is classified as having a deletion or amplification. A classification can be binary (e.g., positive or negative) or have more levels of classification (e.g., on a scale from 1 to 10 or 0 to 1).

[0100] As used herein, the term "parameter" refers to a numerical value that characterizes a quantitative data set and / or a numerical relationship between quantitative data sets. For example, the ratio (or a function of the ratio) between a first amount of a first nucleic acid sequence and a second amount of a second nucleic acid sequence is a parameter. Parameters can be used to determine any of the classifications described herein, such as with respect to fetal, cancer, or transplantation analysis.

[0101] The terms "threshold" and "threshold value" refer to a predetermined number used in an operation. For example, a cut-off length (or threshold length) can refer to the length above which fragments are excluded. A threshold value can be a value above or below which a specific classification is applied. Either of these terms can be used in any of these contexts. A threshold or threshold value can be a "reference value" or derived from a reference value that represents a specific classification or differentiates between two or more classifications. A threshold can be predetermined with or without reference to the characteristics of a sample or subject. For example, a threshold can be selected based on the age or gender of a test subject. A threshold can be selected after the output of test data and based on the output of the test data. For example, when the sequencing of a sample reaches a certain depth, a certain threshold can be used. As another example, reference subjects with known classifications and measured characteristic values (e.g., methylation levels, statistical length values, or counts) of one or more conditions can be used to determine a reference level to differentiate between different classifications of conditions and / or disorders (e.g., whether a subject has a condition). A reference value can be selected as a representative of a classification (e.g., an average) or a value between two clusters of a metric (e.g., selected to obtain a desired sensitivity and specificity). As another example, a reference value can be determined based on a statistical simulation of a sample. Any of these terms can be used in any of these contexts. Such reference values can be determined in various ways, as will be understood by those skilled in the art. For example, a metric can be determined for two different groups of objects with different known classifications, and a reference value can be selected as a representative of a classification (e.g., an average) or a value between two clusters of a metric (e.g., selected to obtain a desired sensitivity and specificity). As another example, a reference value can be determined based on a statistical simulation of a sample. Specific values of thresholds, threshold values, reference values, etc. can be determined based on a desired precision (e.g., sensitivity and specificity).

[0102] "Pathological grade" (or disorder grade) can refer to the amount, degree, or severity of pathology associated with an organism. One example is a cellular disorder expressing a nuclease. Another example of pathology is the rejection of a transplanted organ. Other exemplary pathologies can include autoimmune attacks (e.g., lupus nephritis or multiple sclerosis that damage the kidney), inflammatory diseases (e.g., hepatitis), fibrotic processes (e.g., cirrhosis), fatty infiltration (e.g., fatty liver disease), degenerative processes (e.g., Alzheimer's disease), and ischemic tissue damage (e.g., myocardial infarction or stroke). The healthy state of a subject can be considered a classification of no pathology. Pathology can be cancer.

[0103] The term "cancer grade" can refer to the presence or absence of cancer (i.e., present or absent), the stage of the cancer, the length of the tumor, the presence of metastases, the total tumor burden of the body, the response of the cancer to treatment, and / or other measures of the severity of the cancer (e.g., recurrence of the cancer). The cancer grade can be many or other markers, such as symbols, letters, and colors. The grade can be zero. The cancer grade can also include pre-deterioration or pre-cancerous conditions (states). The cancer grade can be used in various ways. For example, screening can check for the presence of cancer in people who were previously not known to have cancer. An assessment can survey people who have been diagnosed with cancer to monitor the progression of the cancer over time, study the effectiveness of treatment, or determine the prognosis. In one embodiment, the prognosis can be expressed as the chance that a patient will die of cancer, or the chance that the cancer will progress after a specific duration or time, or the chance or extent of cancer metastasis. Detection can refer to "screening" or can refer to checking whether someone with suggestive features of cancer (e.g., symptoms or other positive tests) has cancer.

[0104] The names of genes are usually written in italics. Human genes are usually also written in all capital letters. Mouse genes may not be capitalized after the first letter. Proteins are usually written in all capital letters and without italics. For example, a mouse can have the Dnase1l3 gene and the DNASE1L3 protein, while a human can have the DNASE1L3 gene and the DNASE1L3 protein.

[0105] A “machine learning model” (ML model) can refer to a software module configured to run on one or more processors to provide a classification or value of a property of one or more samples. Sample data (e.g., training data) can be used to generate an ML model to make predictions on test data. An example is an unsupervised learning model. Another example type of model is supervised learning that can be used with embodiments of the present disclosure. Exemplary supervised learning models can include different methods and algorithms, including analytical learning, statistical models, artificial neural networks, backpropagation, boosting (meta-algorithm), Bayesian statistics, case-based reasoning, decision tree learning, inductive logic programming, Gaussian process regression, genetic programming, group method of data handling, kernel estimators, learning automata, learning classifier systems, minimum message length (decision trees, decision graphs, etc.), multilinear subspace learning, naive Bayes classifier, maximum entropy classifier, conditional random field, nearest neighbor algorithm, probably approximately correct learning (PAC) learning, pulse rules, knowledge acquisition methods, symbolic machine learning algorithms, sub-symbolic machine learning algorithms, minimum complexity machines (MCM), random forests, ensembles of classifiers, sequential classification, data preprocessing, handling imbalanced datasets, statistical relational learning, or Proaftn (a multi-criterion classification algorithm). The model can include linear regression, logistic regression, deep recurrent neural networks (e.g., long short-term memory, LSTM), Bayesian classifiers, hidden Markov models (HMM), linear discriminant analysis (LDA), k-means clustering, density-based spatial clustering of applications with noise (DBSCAN), random forest algorithms, support vector machines (SVM), or any of the models described herein. Various cost / loss functions and various optimization techniques (e.g., using backpropagation, steepest descent, conjugate gradient, and Newton and quasi-Newton techniques) can be used to train the supervised learning model in various ways, the various cost / loss functions defining the error with known labels (e.g., the least squares and absolute differences with known classifications).

[0106] The terms "about" or "approximately" can refer to an acceptable error range of a particular value determined by a person skilled in the art, which will depend in part on how the value is measured or determined, i.e., the limitations of the measurement system. For example, "about" can refer to within 1 or more standard deviations according to the practice in the art. Optionally, "about" can mean a range of up to 20%, up to 10%, up to 5%, or up to 1% of a given value. Optionally, particularly for biological systems or processes, the terms "about" or "approximately" can refer to within one order of magnitude of a numerical value, within 5 times of a numerical value, and more preferably within 2 times of a numerical value. When a particular value is described in this application and the claims, unless otherwise specified, it should be assumed that the term "about" refers to an acceptable error range of that particular value. The term "about" can have the meaning commonly understood by a person skilled in the art. The term "about" can refer to ±10%. The term "about" can refer to ±5%. Detailed Description

[0107] Various fragmentomics features of a free sample (e.g., urine and plasma) are used to determine various properties of the sample and / or the subject.

[0108] As an example for urine samples, some embodiments can use the fragmentomics features of urine-free DNA to detect the contributions of renal and non-renal urine-free DNA. Such measurements can be used to reflect glomerular permeability and monitor various diseases, such as but not limited to kidney abnormalities, such as renal cancer and nephropathy, as well as proteinuria and preeclampsia, which can be classified as types of kidney abnormalities.

[0109] In addition, such measurements can also be used to determine the concentration fraction of clinically relevant free DNA molecules, as well as to enrich clinically relevant DNA in urine samples, including all types of renal DNA (e.g., from liver, lung, colon, heart, and blood sources, e.g., white blood cells), fetal DNA, tumor DNA, or DNA from a specific tissue different from the urethra, such as the kidney, ureter, and bladder. The relative contribution of renal-free DNA can be determined by measuring the relative abundance of free DNA molecules from open chromatin regions of one or more tissues (e.g., OCR (e.g., at least 100, 500, 1,000, 5,000, or 10,000 OCR) or a whole or representative sampling of one or more tissues contributing to any renal DNA). In other cases, the relative contribution or enrichment of renal-free DNA molecules can be determined based on free DNA molecules with specific lengths and / or end motifs from urine samples and the corrected urine concentration.

[0110] For example, for terminal motifs, some embodiments can utilize the presence of C at the ends of cfDNA molecules to enrich clinically relevant DNA of a sample. Thus, according to embodiments of the present disclosure, fragmentomics signatures (e.g., end tags and abundances of cfDNA from kidney-specific open chromatin regions in urine cell-free DNA) can be used to determine the contribution of kidney-derived cell-free DNA in urine.

[0111] In addition, terminal motifs can be used to simultaneously (i.e., together) analyze different types of cell-free DNA cleavage. The different types can be distinguished by representing different dimensions in a cleavage space that represents all nuclease activities that can occur in a subject. In the present disclosure, based on nuclease knockout mice and / or human subjects with various drug treatments, different types of cell-free DNA cleavage are associated with different fragmentation processes (including enzymatic and non-enzymatic breaks).

[0112] Compared to techniques that use one terminal motif or several of the top-ranked terminal motifs to focus on one specific nuclease activity at a time (Serpas et al. Proc. Natl. Acad. Sci. USA. 2019; 116:641-649; Chan et al. Am. J. Hum. Genet. 2020; 107:882-894; Chen et al. PLOS Genet. 2022; 18:e1010262), embodiments of the present disclosure can simultaneously evaluate multiple nuclease activities or other fragmentation processes that may be involved (e.g., induced by chemo- and radiotherapy) based on the putative relative contributions of different types of cell-free DNA cleavage. The contribution of each type of cell-free DNA cleavage can be determined by generating a set of F-spectra representing the relative frequencies of terminal motifs for a given biological sample. In some cases, the set of F-spectra can be generated by applying factorization (e.g., non-negative matrix factorization) to the relative frequencies. Analyzing the perturbed contributions can allow for the detection and monitoring of various diseases, but is not limited to cancer and immune diseases.

[0113] Thus, as described herein, terminal motifs (e.g., sample terminal motif spectra and reference spectra, referred to as reference F-spectra) can be used in various ways to determine the nature of a sample and / or the classification of a subject, such as determining the concentration fraction of clinically relevant DNA, the gestational age of a fetus, or the pathological grade of a subject. I. Review

[0114] The glomerular basement membrane (GBM) permits plasma-free DNA to pass through the kidney and become cell-free DNA in urine. Generally, smaller DNA molecules have greater GBM permeability than larger DNA molecules. For example, the permeability of the GBM decreases with increasing molecular length from plasma to urine (Lawrence et al. Proc Natl Acad Sci USA. 2017;114:2958-2963). Additionally, due to the association of histones with DNA, nucleosome-depleted DNA molecules (e.g., DNA molecules from nucleosome-depleted regions) can have a smaller molecular length than nucleosomal DNA with the same DNA length. In some embodiments, nucleosome-depleted DNA in enriched cell-free DNA in urine is used to determine the level of glomerular permeability.

[0115] Figure 1 An exemplary overview 100 of characteristics for identifying renal and non-renal DNA in a urine sample is shown according to some embodiments. As Figure 1 shown, urine-free DNA includes renal DNA 102 that comes from plasma and enters the urinary system through the kidney. For example, renal DNA can be derived from liver, blood cells, tumor, or fetal DNA. Urine-free DNA can also include non-renal DNA 104 from the urinary system. For example, non-renal free DNA can be derived from the urinary tract (e.g., the kidney and bladder). These two types of urine-free DNA may be associated with different characteristics. By identifying the differences between renal cell-free DNA (cfDNA) and non-renal free DNA, the contribution of renal free DNA or non-renal free DNA can be determined from a given urine sample. Renal cfDNA or certain types of renal DNA are generally considered clinically relevant DNA. Thus, from these differences, the contribution of total renal cfDNA or a specific type (e.g., fetal, tumor, or from a specific organ or blood cell type) of renal cfDNA can be determined.

[0116] Figure 2Figure 200 is shown, which illustrates the use of fragmentomics features to determine the contribution of renal urine-free DNA. In plasma 206, there are nucleosome-depleted free DNA 202 (i.e., without associated proteins such as histones) and free DNA molecules 204 of nucleosomes. When plasma DNA molecules 206 pass through the GBM in the kidney, the nucleosome-depleted free DNA molecules 202 have higher permeability compared to the nucleosome DNA molecules 204 with larger molecular lengths. At the same time, when entering the urine-free DNA pool, the renal urine-free DNA can still carry the end tags formed in plasma, such as the end tags mediated by DNASE1L3. Thus, according to embodiments of the present disclosure, the fragmentomics features (such as end tags) and relative abundances of cfDNA from renal-specific open chromatin regions of urine-free DNA can be used to determine the contribution of renal-free DNA in urine.

[0117] Figure 2 Two exemplary instances on the urine side are shown. The first urine sample 210 has a higher contribution of renal cfDNA, as shown by 3 out of 7 DNA fragments being renal urine cfDNA. The second urine sample 220 has a lower contribution of renal cfDNA, as shown by only 1 out of 5 DNA fragments being renal urine cfDNA. Various embodiments can distinguish these samples (and even distinguish different types of renal DNA), for example, as part of estimating the concentration fraction of clinically relevant DNA, determining the pathological classification (such as kidney abnormalities), or detecting preeclampsia or proteinuria (which can be classified as a type of kidney abnormality).

[0118] However, there are challenges. The fragmentomics of renal-free DNA is generally not well understood. In addition, the fragmentation processes of urine-free DNA and plasma-free DNA may involve different nucleases (Han et al. Am J Hum Genet. 2020; 106:202–214; Chen et al. PLOS Genet. 2022; 18:e1010262). For example, DNASE1L3 is the main nuclease that generates C-terminal fragments in plasma (Serpas et al. Proc Natl Acad Sci USA. 2019; 116:641-649), while DNASE1 is responsible for generating T-terminal fragments in urine (Chen et al. PLOS Genet. 2022; 18:e1010262).

[0119] Based on this difference, we hypothesize that the renal urine-free DNA molecules will carry the end-motif tags of those free DNAs present in plasma. In fact, the analysis of the end-motifs in urine-free DNA can be used to infer the contribution of renal-free DNA. For example, a higher amount of urine-free DNA carrying the C-terminus (i.e., a tag of plasma DNA) may imply a higher contribution of renal-free DNA.

[0120] In view of the above, it is necessary to understand the fragmentomics differences (e.g., length, end-motif) between renal and non-renal free DNAs. By identifying such differences, the contribution of renal-free DNA can be accurately estimated without any genetic or epigenetic information (e.g., SNPs from tumor tissue). The renal-free DNA contribution can then be applied to disease models. For example, a subject with renal function may have a higher or lower renal contribution than a normal subject. II. Exemplary Urine Sample Preparation

[0121] Urine samples for cfDNA analysis can be prepared using a variety of techniques. The techniques described below are merely examples that those skilled in the art will understand. A. Sample Collection

[0122] To determine the differences between renal and non-renal DNA, cell-free DNA molecules obtained from plasma and urine samples can be analyzed. For example, 192 human plasma and 18 urine cell-free DNA samples were sequenced using paired-end sequencing. Specifically, the plasma and urine cell-free DNA samples included: (i) urine cell-free DNA samples from pregnant women (n = 20), urine cell-free DNA samples from preeclampsia (n = 5), and plasma cell-free DNA samples from pregnant women (n = 11) (median paired-end reads: 129.5 million; range: 30.1 - 234.9 million); (ii) urine cell-free DNA samples from renal cell carcinoma (RCC) (n = 16), proteinuria (n = 24), and controls (n = 34) (median paired-end reads: 25.03 million; range: 13.34 - 75.02 million); (iii) plasma cell-free DNA samples from 8 healthy individuals, 10 patients with DNASE1L3 disease-related variants, and the parents of 3 patients with a mutant DNASE1L3 gene (median paired-end reads: 108 million; range: 40 - 162 million); (iv) plasma cell-free DNA samples from 24 SLE patients and 11 healthy individuals (median paired reads: 120 million; range: 18 - 208 million); (v) plasma cell-free DNA samples from 38 healthy individuals, 17 patients with chronic hepatitis B virus (HBV) but without hepatocellular carcinoma (HCC) (i.e., HBV carriers), and 34 patients with HCC (median paired-end reads: 38 million; range: 18 - 65 million); (vi) plasma cell-free DNA from 30 pregnant women in the first trimester (12 - 14 weeks; n = 10), second trimester (20 - 24 weeks; n = 10), and third trimester (38 - 40 weeks; n = 10) (median paired-end reads: 103 million; range: 52 - 186 million); (vii) plasma cell-free DNA samples from 15 healthy control subjects, 25 colorectal cancer (CRC) patients without liver metastasis, and 24 CRC patients with liver metastasis (median paired-end reads: 40 million; range: 16 - 89 million); and (viii) plasma cell-free DNA samples from nasopharyngeal carcinoma patients (n = 6) receiving cisplatin chemotherapy and paired pre-treatment patients (n = 6) (median paired-end reads: 5 million; range: 3 - 9 million). B. Using a stabilizer on urine samples

[0123] Figure 3 Process 300 according to some embodiments is shown, where sequencing data is obtained from a urine sample. At block 302, a urine sample from a pregnant woman (e.g., a sample from the third trimester) is used. Such a sample is used to determine whether the renal contribution can be correlated with the fetal DNA contribution.

[0124] Sequencing urine samples can be challenging because DNASE1 activity is very high in urine. If DNASE1 activity is not completely inhibited after urine collection, the in vitro continuous fragmentation caused by DNASE1 can confound the fragmentation pattern initially present in urine, which can reduce the fragmentomics signal of urine DNA fragments associated with specific diseases.

[0125] To address the above challenges, various collection and preservation methods can be used to better preserve the original characteristics of urine-free DNA. For example, as shown in box 304, a preservative can be added at box 306 to obtain a preserved sample. Different urine collection methods can be used, including adding ethylenediaminetetraacetic acid (EDTA) and adding stabilizers. EDTA can inhibit the cleavage activity of the DNASE1 family by chelating magnesium and calcium, which are essential ions required for DNASE1 digestion. Stabilizers can potentially stabilize urine DNA from degradation. Stabilizers can be, but are not limited to, preservatives provided by Collipee, diazolidinyl urea (DU), dimethyl urea, 2-bromo-2-nitropropane-1,3-diol, 5-hydroxymethoxymethyl-1-aza-3,7-dioxabicyclo(3.3.0)octane and 5-hydroxymethyl-1-aza-3,7-dioxabicyclo(3.3.0)octane and 5-hydroxypoly[methyleneoxy]methyl-1-aza-3,7-dioxabicyclo(3.3.0)octane, bicyclic oxazolidine (such as NuoSept95), DMDM hydantoin, imidazolidinyl urea (IDU), sodium hydroxymethylglycinate, hexamethylenetetramine chloroallyl chloride (quaternium-15), microbicides (such as Bioban, Preventol and Grotan), water-soluble zinc salts, EDTA, other metal ion chelators such as N,N'-bis-(dithiocarbamoyloxy)piperazine (BDP), diethyldithiocarbamate (salt) (DDTC), iminodisuccinic acid (IDS), polyaspartic acid, S,S-ethylenediamine-N,N'-disuccinic acid (EDDS), methylglycine diacetic acid (MGDA), etc.

[0126] Urine-free DNA can be better preserved in devices containing stabilizers compared to devices without stabilizers. As an illustrative example, urine-free DNA samples from 2 control subjects were collected by incubating in vitro at room temperature for different durations (e.g., 0 hours and 4 hours incubation) using different collection methods (no reagent added, EDTA and stabilizer groups). The free DNA concentration and length distribution were compared among the three urine collection groups.

[0127] Figure 4Shows a set of Figures 400, which show the comparison of urine-free DNA concentrations before and after in vitro incubation in three urine collection groups (a control group without adding any stabilizer, an EDTA and stabilizer group). The term "GE" refers to genomic equivalent. As Figure 4 shown, compared with the control group without adding any stabilizer (Sample 1: 6.04 GE / ml; Sample 2: 1.98 GE / ml) and the EDTA group (Sample 1: 2.37 GE / ml; Sample 2: 2.09 GE / ml), the fold change of urine-free DNA in the stabilizer group (Sample 1: 1.61 GE / ml; Sample 2: 1.15 GE / ml) after 4-hour incubation compared to 0-hour incubation is the smallest.

[0128] Before and after in vitro incubation, the autosomal DNA length distribution was further compared among the three collection groups.

[0129] Figure 5 Shows a set of Figures 500 that identify the comparison of the free DNA length distribution before and after in vitro incubation in the control, EDTA, and stabilizer groups. The first set of Figures 502, 504, and 506 correspond to the free DNA length distribution of the first sample, and the second set of Figures 508, 510, and 512 correspond to the free DNA length distribution of the second sample. In the control group (Figures 502, 508), compared with 0-hour incubation, the urine-free DNA length distribution changed greatly after 4-hour incubation. In the EDTA group (Figures 504, 510), compared with 0-hour incubation, the change in the urine-free DNA length distribution after 4-hour incubation was very small. In the stabilizer group (Figures 506, 512), compared with 0-hour incubation, the urine-free DNA length distribution showed the smallest change (i.e., almost unobservable) after 4-hour incubation. Figure 5 The results shown indicate that using a device containing a stabilizer can optimally maintain the fragmentation distribution of urine-free DNA molecules under room temperature conditions. C. Free DNA Extraction and Library Preparation

[0130] As Figure 3As shown, cell-free DNA can be extracted from urine (treated with stabilizer) using a Wizard Plus Minipreps DNA purification system (Promega) and guanidinium thiocyanate (Sigma-Aldrich). Cell-free DNA was extracted from plasma samples using a QIAamp circulating nucleic acid kit (Qiagen) according to the manufacturer's protocol. Indexed DNA libraries were constructed using a TruSeq DNA Nano Library Prep kit (Illumina) according to the manufacturer's instructions. The adapter-ligated DNA was enriched by PCR and then analyzed on an Agilent 4200 TapeStation (Agilent Technologies) for quality control and gel-based length determination. The libraries were quantified by Qubit dsDNA High Sensitivity Assay Kit (Thermo Fisher Scientific) prior to sequencing. D. DNA Sequencing and Alignment

[0131] As Figure 3 As further shown, multiplexed DNA libraries were sequenced on an Illumina platform for paired-end reads. Other sequencing technologies can be used, e.g., as described herein. For example, single reads of entire DNA fragments can be determined. Sequences were assigned to their corresponding samples based on their six-base index sequences. Paired-end reads from mouse plasma were aligned to the mouse reference genome (NCBI build 37 / UCSC MM9; non-repeat masked) or the human reference genome (NCBI build 37 / hg19) (Li et al. Bioinformatics 2009; 25:1966-1967) using the short oligonucleotide alignment program 2 (SOAP2). As will be appreciated by those skilled in the art, any other alignment tool can also be used. In some embodiments, up to two nucleotide mismatches are allowed. Only paired-end reads that aligned to the same chromosome in the correct orientation and spanned an insert fragment length of <600 bp were retained for downstream analysis. Paired-end reads with the same starting and ending genomic coordinates were considered PCR duplicates and discarded from downstream analysis.

[0132] For some uses, e.g., for plasma samples, the genotype of the buffy coat DNA from the mother can be paired with the corresponding placental sample. In fact, maternal and fetal genotypes can be determined. This genotype can be used to distinguish fetal and maternal DNA molecules, and thus a gold standard for the fetal DNA fraction in urine samples can be obtained. This actual fetal DNA fraction also allows us to establish a recalibration curve for estimating the degree of renal DNA or renal permeability, assuming that the higher the renal permeability, the more renal DNA. III. Length Characteristics of Urinary Cell-Free DNA

[0133] Analyze the length characteristics of urinary cfDNA to illustrate the effect of fragment length on the ability of renal cfDNA fragments to enter urine through the kidney. Smaller-sized molecules show an increased ability to pass from the blood through the kidney.

[0134] Figure 6 Figure 600 showing the identification of length differences between fetal DNA and maternal DNA in a urine sample according to some embodiments is presented. The length difference between fetal DNA and maternal DNA can be used as an example for determining renal DNA in a urine sample. For example, fetal DNA originates from the fetus and must pass through the kidney so that we can detect them in urine. In contrast, the detected maternal DNA can include non-renal DNA, which can be contributed by the kidney, bladder, etc. Based on this distinction, we can extend the characteristics of fetal and maternal DNA to determine the length difference between renal DNA and non-renal DNA.

[0135] As Figure 6 shown, most fetal-specific cell-free DNA molecules 620 (red) are less than 80 bp, which is significantly shorter than the common cell-free DNA molecules 610 (blue). The common cell-free DNA molecules 610 have the same alleles shared between the mother's haplotype and one of the fetal haplotypes. The fetal-specific cell-free DNA molecules 620 have fetal-specific alleles (inherited from the father) on one of the fetal haplotypes.

[0136] Based on this, it can be considered whether the length characteristics of renal DNA are related to the length characteristics of fetal DNA.

[0137] Figure 7 An exemplary schematic diagram 700 according to some embodiments is shown, which shows the biological process of converting plasma DNA into renal DNA. For example, to become renal DNA, plasma DNA from blood vessel 702 reaches the kidney 704 through various tissue membranes. For example, plasma DNA molecules pass through endothelial cells, the glomerular basement membrane (GBM), and podocytes. As Figure 7 shown, each of these biological structures has a different diameter. The renal structure can be related to the pore size of the kidney, through which plasma DNA molecules pass to become renal DNA. It can be considered that plasma DNA molecules small enough to pass through the pores can ultimately become renal DNA.

[0138] Figure 8 Figure 800 showing the relationship between glomerular basement membrane permeability and renal DNA length according to some embodiments is presented. The permeability of the glomerular basement membrane can be determined by molecular length. As Figure 8As shown, the x-axis is the radius of a given molecule, which corresponds to its length. And the y-axis identifies the renal permeability associated with the given molecule.

[0139] Specifically, the %GBM permeability represented on the y-axis determines the percentage of molecules of a specific length that pass through the GBM. For example, if the molecule is very small (e.g., 12 kDa), the permeability is estimated to be approximately 50%. In contrast, as the molecule gets larger (e.g., 150 kDa), the renal permeability will decrease significantly to approximately 10 - 15%. It is also known that nucleosomes typically have a size of 200 kDa / 5.5 nm (radius). Based on the size of nucleosomes, the putative DNA molecules wrapped in nucleosomes (and thus attached to proteins) will be associated with low GBM permeability compared to nucleosome-depleted DNA molecules. In fact, the length of renal DNA molecules passing through the GBM may be smaller compared to non-renal DNA molecules directly from the urinary system. IV. Terminal Motif Characteristics of Urinary Free DNA

[0140] In addition to length, the terminal sequences of urinary free DNA are analyzed to determine that the terminal motifs of renal urinary free DNA molecules are different from those of non-renal urinary free DNA molecules. In some embodiments, a 4-mer terminal motif is defined as the 4 nucleotides at the end of each 5’ fragment end of the free DNA molecule, for a total of 256 4-mer terminal motifs (i.e., 4 4 )). The median terminal motif frequency of the 256 terminal motifs is calculated and the fetal-specific fragments and common fragments in the maternal urine sample are ranked in descending order respectively. Other terminal motifs can be used, e.g., any K-mer terminal motif, where K is 1, 2, 3, 4, 5, 6, 7, 8, 9 or more. As described herein, the terminal motifs (e.g., the sample terminal motif spectrum and a reference spectrum, referred to as the reference F-spectrum) can be used in various ways to determine the properties of a urine sample and / or the classification of a subject, such as determining the concentration fraction of clinically relevant DNA, the gestational age of a fetus, or the pathological grade of a subject.

[0141] Figure 9 Figure 900 shows the identification of different terminal motifs identified from fetal-specific free DNA and common free DNA in maternal urine. As Figure 9 shown, the median terminal motif frequency of the 256 terminal motifs in fetal-specific common free DNA is calculated and ranked in descending order in the maternal urine sample. The x-axis identifies the ranking of the terminal motifs of the common fragments. The y-axis identifies the ranking of the motifs of the fetal-specific fragments. A higher ranking indicates a relatively higher frequency of the corresponding terminal motif (e.g., CCTG). The colored regions show the fetal-specific DNA fragments or common DNA fragments respectively, which have a preference for specific 4-mer terminal motifs.

[0142] Each of the top 10 motifs in fetal-specific cell-free DNA and common cell-free DNA is labeled with the corresponding terminal motif sequence. The top 10 motifs of fetal-specific urinary cell-free DNA and common urinary cell-free DNA are highlighted with red circles 902 and blue circles 904, respectively. The top 10 terminal motifs of fetal-specific cell-free DNA are dominated by C-terminal motifs (8 / 10), while the top 10 terminal motifs of common cell-free DNA are enriched in T-terminal motifs (4 / 10). DNASE1L3, which preferably cleaves C, has been previously identified as the dominant nuclease in plasma, while DNASE1, which preferably cleaves T, is the dominant nuclease in urine. Based on the above motif rankings, it can be determined that fetal DNA corresponds to renal DNA. These data suggest that motifs containing the C-terminus can be used to represent renal urinary cell-free DNA, which can then be used to distinguish fetal DNA from maternal urine samples. As described below, some embodiments can use the presence of C at the ends of cfDNA molecules to enrich the clinically relevant DNA of a sample, e.g., all renal DNA, fetal DNA, tumor DNA, or DNA from a specific tissue other than the kidney or bladder.

[0143] Figure 10 A set of graphs 1000 is shown that identify the relationship between the concentration fraction of fetal DNA in a urine sample and CC-terminal fragments according to some embodiments. As Figure 10 shown, the percentage of urinary cell-free DNA carrying the CC-terminus increases proportionally with the fetal DNA fraction in the urine sample. This linear relationship is further significant when the CC fragments are limited to fragments less than 80 base pairs in length. Thus, the terminal motifs of a urine sample can be used to determine the fraction of fetal DNA. V. Analysis of OCR Urinary Cell-Free DNA

[0144] Open chromatin regions can be used to determine the nature of a urine sample and / or the classification of a subject. In some cases, open chromatin regions can be associated with tissues that contribute to renal DNA or even specific cell type tissues (e.g., fetal, tumor, transplanted organ, or other tissues such as blood, liver, colon, etc., other than tissues from the urethra). For example, the abundance of cfDNA from such a set of regions can be used as part of estimating the concentration fraction of clinically relevant DNA, determining a pathological classification (e.g., kidney abnormalities), or detecting preeclampsia or proteinuria (which can be classified as a type of kidney abnormality).

[0145] The permeability of the renal membrane (e.g., GBM) favors shorter DNA fragments. As a result, renal-free DNA passing through the renal membrane is shorter than non-renal DNA fragments directly derived from the urinary system. In addition, free DNA molecules bound to nucleosomes may have difficulty passing through the GBM because the permeability of nucleosomes is estimated to be approximately 10 - 15%. In contrast, nucleosome-depleted free DNA molecules derived from open chromatin regions are not bound by any nucleosomes and can pass through the GBM with a higher permeability. Based on the above characteristics, the identification of free DNA molecules derived from open chromatin regions can be used to detect renal DNA in urine samples. In addition, the contribution of free DNA molecules from open chromatin regions can also be used to predict the classification of certain diseases and determine the concentration fraction of fetal DNA. A. Correlation of Renal DNA with DNA from OCRs

[0146] According to some embodiments, renal DNA can be correlated with DNA from open chromatin regions. Free DNA molecules depleted of nucleosomes in plasma have a smaller molecular length, which allows them to pass through the GBM and be converted into renal DNA. Based on this characteristic, it can be determined whether renal DNA is enriched in open chromatin regions, which correspond to nucleosome-depleted regions. Such enrichment is described in later sections, for example, for all renal DNA or for certain tissue types.

[0147] Figure 11 Exemplary FIG. 1100 shows various characteristics for identifying DNA molecules from open chromatin regions according to some embodiments. Open chromatin regions can be identified in various ways, for example, based on the position of DNase1 hypersensitive sites (DHSs) because DNase1 has a cleavage preference within genomic regions relatively lacking bound histones. For example, a given open chromatin region can be identified as a genomic region where the ends of fragments digested by DNase1 are aggregated. There are approximately 1 million DHS sites, and the median length of the regions is approximately 200 base pairs. Such open chromatin regions contribute 9.46% of the genome. Such DHSs (and corresponding OCRs) can be specific to a particular tissue, such as placenta-specific DHSs.

[0148] As another example of identifying OCRs, to obtain DNA molecules from open chromatin regions, DNase-Seq can be used. Specifically, DNA molecules from urine samples can be digested with DNase1, which preferentially cuts DNA molecules with hypersensitive sites and sequences them. Then, the sequence reads can be considered as DNA molecules from open chromatin regions. As a further example, OCRs can be identified from sites identified using, but not limited to, DNase-seq, sites identified using Assay for Transposase-Accessible Chromatin using sequencing (ATAC-seq), and transcription start sites (TSSs). B. Identifying tissue-specific OCRs

[0149] OCRs can be determined and generally used for all tissues, for tissues specific to the kidney, or for specific tissues. Different tissues typically have different regions of open chromatin. Thus, a specific set of OCRs can be identified, for example, depending on the clinically relevant DNA desired. As an example, whether a class of cfDNA is enriched or depleted in one or more OCRs can be determined, for example, to identify OCRs associated with one or more tissues.

[0150] Figure 12 Exemplary graph 1200 for determining the amount of urine-free DNA corresponding to open chromatin regions according to some embodiments is shown. To analyze the fragmentomics characteristics of urine-free DNA based on open chromatin regions, 14 urine samples from pregnant women were collected and the properties of free DNA in pregnant women's urine were studied. The characteristics of fetal-specific free DNA molecules and common free DNA molecules can represent the characteristics of renal-free DNA and non-renal-free DNA, respectively. This representation can be made because renal-free DNA is shorter than non-renal-free DNA.

[0151] In Figure 12In the example shown, cfDNA having any fetal-specific allele (e.g., single nucleotide polymorphism, SNP) can be identified. Then, the amount of such cfDNA in a window spanning the genome (e.g., 10, 20, 30, 40, 50, or 60 bp) can be determined by aligning sequence reads to a reference genome. The expected value can be determined as the number of fetal-specific SNPs in a region (where a region can include one or more windows) divided by the average or median of the fetal-specific SNPs in all windows / regions. The observed value can be determined as the number of fetal-specific reads in a region divided by the average or median of all windows / regions. If the observed ratio is greater than the expected ratio, the region can be identified as an OCR because smaller fetal DNA fragments are more prevalent. Such OCRs determined using fetal DNA are specific to fetal tissues. However, DHS sites for various tissues can be used, or more generally, OCRs for various tissues can be used. Thus, DHSs for a specific tissue or tissue collection can be used.

[0152] In one exemplary embodiment, to evaluate cell-free DNA in maternal urine in pregnant women, we obtained a median fetal fraction of 0.31% (range: 0.20% - 9.00%) in maternal urine samples. The fetal DNA fraction in maternal urine cell-free DNA was estimated by a single nucleotide polymorphism (SNP)-based method (Yu et al. Clin Chem. 2013;59:1228 - 1237). The contribution of nucleosome-depleted cell-free DNA can be indicated by the amount of DNA molecules from open chromatin regions (OCRs), where OCRs correspond to regions lacking nucleosomes (i.e., DNA lacking histone binding). For illustrative purposes, we used DNase1 hypersensitive sites (DHSs) defined by DNase-seq to represent OCRs (Meuleman et al. Nature. 2020;584:244 - 251).

[0153] From the OCRs, the amount of nucleosome-depleted DNA molecules is determined by the number of sequenced cell-free DNA molecules aligned to the OCRs. The amount of nucleosome-depleted DNA molecules can be a normalized value. For example, the amount of nucleosome-depleted DNA molecules can be converted to a percentage by dividing by the total number of sequenced molecules. Additionally or alternatively, the amount of nucleosome-depleted DNA molecules can be calculated by dividing the number of sequenced molecules (O) within the observed OCR (also referred to as the observed OCR-related DNA contribution) by the expected OCR-related value (E). This measurement is defined herein as the "O / E ratio".

[0154] The desired OCR-related DNA contribution can correspond to the theoretical percentage of OCRs in the reference genome. For example, the observed value (O) can include the percentage of fragments that align to OCRs among all fragments, and the expected value (E) is the theoretical percentage of OCRs in the reference genome (e.g., the human reference genome). In some cases, the desired OCR-related DNA contribution of fetal-specific DNA can be calculated by normalizing the number of fetal-specific single nucleotide polymorphisms (SNPs) in OCRs by the number of fetal-specific SNPs in all genomic regions. Such relative frequencies (e.g., percentages) can provide the desired percentage, which can be compared to the percentage of cell-free DNA molecules observed to align to OCRs. SNPs can be obtained by genotyping assays. In some cases, the desired OCR-related DNA contribution corresponds to the percentage of DNA molecules randomly sampled that fall within OCRs.

[0155] The OCR-related DNA contribution observed in fetal DNA can be calculated by normalizing the number of fetal-specific molecules that align to OCRs by the number of fetal-specific molecules that align to all genomic regions. For O / E ratio analysis in non-renal urine cell-free DNA, molecules carrying shared alleles between the fetal and maternal genomes are analyzed according to embodiments of the present disclosure. To use the O / E ratio to determine OCR enrichment, if the O / E ratio is close to 1, no OCR enrichment is found. If the O / E ratio is greater than 1, the contribution of OCR-related DNA increases. A higher O / E ratio may imply a higher contribution of nucleosome-depleted DNA, potentially indicating higher glomerular permeability. Other tissue-specific regions can be identified in a similar manner. C. Quantifying Renal cfDNA from OCRs in Urine and Plasma

[0156] The amount of cfDNA in OCR regions can be used to quantify renal DNA in urine samples. Fetal cfDNA is used as an example of renal DNA, but other examples can also be applied to tumors and other tissues that produce renal DNA.

[0157] Figure 13 A set of box plots 1300 showing the O / E ratios of fetal-specific cell-free DNA molecules and shared cell-free DNA molecules in open chromatin regions in urine samples (box plot 1302) and plasma samples (box plot 1304) are shown. These OCRs are not tissue-specific but are a general sampling of OCRs across different tissues. Specifically, all known DHS sites are used.

[0158] As shown in box plot 1302, in urine samples with a fetal fraction higher than 0.44%, the median O / E ratio of fetal-specific free DNA (which is renal urine free DNA) is 1.84 (range: 1.68 - 2.13), which is 1.67 times higher than the median O / E ratio of the common free DNA mainly of non-renal origin (median: 1.10; range: 1.08 - 1.19). In contrast, in plasma samples (box plot 1304), no significant enrichment of the O / E ratio was observed in fetal-specific free DNA (median of O / E ratio: 1.048; range: 1.023 - 1.126) and common free DNA (median: 1.058; range: 1.033 - 1.124).

[0159] In summary, compared to non-renal DNA molecules, the contribution of OCR-related DNA in fetal DNA (an example of renal DNA molecules) is increased. These data suggest that the amount of OCR-related DNA can be used to estimate the concentration fraction of renal free DNA in urine. The higher the O / E ratio, the higher the concentration fraction of clinically relevant DNA (e.g., one or more tissues that use its OCR). When using the OCRs of different renal tissues, the concentration fraction will correspond to the average concentration of those tissues (e.g., weighted by the amount and length of the corresponding OCRs). The concentration fraction can approach the renal DNA concentration, where more OCRs of different tissues can provide higher precision for approaching the renal DNA concentration.

[0160] To estimate the concentration fraction, calibration (training) samples with known concentration fractions of clinically relevant DNA can be used. The calibration value can correspond to the relative abundance of the calibration sample, where the calibration value and the known concentration fraction include calibration data points. If the new sample has a higher relative abundance, the new sample has a higher concentration fraction than the calibration sample. If the new sample has a lower relative abundance, the new sample has a lower concentration fraction than the calibration sample. By using multiple calibration samples, the range of the concentration fraction can be determined. In other embodiments, the calibration function (also called the calibration curve) can be determined via function fitting (e.g., linear or non-linear regression) of the calibration data points.

[0161] Figure 14 A set of graphs 1400 is shown, which identifies the enrichment of DNA molecules in open chromatin regions in fetal-specific DNA from urine samples according to some embodiments. Graph 1402 identifies the expected and observed percentages of fetal-specific urine DNA molecules from OCR regions. Graph 1404 identifies the expected and observed percentages of common urine DNA molecules from OCR regions. The expected percentage can be determined based on the length of the OCR region as a proportion of the genome.

[0162] As shown in FIG. 1402, fetal-specific urine DNA is enriched in open chromatin regions because the observed values are substantially greater than the expected values. In addition, FIG. 1404 shows that the expected and observed values of total urine DNA molecules show a smaller decrease. Based on FIGS. 1402 and 1404, it is shown that urine DNA is enriched in open chromatin regions. These results also suggest that the renal filtration mechanism contributes to the enrichment of kidney-filtered DNA in open chromatin regions. Accordingly, embodiments can enrich clinically relevant DNA in urine samples, for example, by selecting cfDNA from open chromatin regions specific to one or more kidney-filtered tissues (e.g., typically fetal, tumor, transplanted, or kidney-filtered tissues).

[0163] Figure 15 A set of FIGS. 1500 is shown that identifies the non-enrichment of DNA molecules in open chromatin regions in fetal-specific DNA from plasma samples according to some embodiments. FIG. 1502 identifies the expected and observed percentages of fetal-specific plasma DNA molecules from OCR regions. FIG. 1504 identifies the expected and observed percentages of total plasma DNA molecules from OCR regions. As shown in FIGS. 1502 and 1504, neither fetal-specific plasma DNA nor total plasma DNA is enriched in open chromatin regions. In addition, the O / E ratios shown in FIG. 1506 also indicate no enrichment of plasma DNA from open chromatin regions. Accordingly, the enrichment of cell-free urine DNA in open chromatin regions that is observed can be used to determine the fraction of fetal-specific DNA, and such determination is not feasible for cell-free plasma DNA. IV. Identification of Clinically Relevant DNA in Urine

[0164] Since different fragmentation patterns can be identified between kidney-filtered and non-kidney-filtered cell-free urine DNA, we hypothesized that the contribution of kidney-filtered cell-free DNA could be enriched by selectively analyzing the fragmentomic characteristics of kidney-filtered cell-free urine DNA. Fragmentomic characteristics can include, but are not limited to, end motifs (e.g., CC ends), genomic regions (e.g., OCRs), and lengths (e.g., <= 80 bp). In addition, the accuracy of determining the contribution of kidney-filtered cell-free DNA can be further improved by identifying cell-free urine DNA molecules from open chromatin regions. A. Estimating the Amount of Clinically Relevant DNA Using the Abundance of OCRs

[0165] As previously shown in Figure 13 it was shown that greater enrichment of OCR-related DNA was observed in fetal-specific cell-free urine DNA compared to total cell-free DNA in urine samples. This is different from the non-enrichment of open chromatin regions in plasma DNA molecules.

[0166] Thus, the fetal DNA fraction (or fraction of other clinically relevant DNA) can be determined in a urine sample using the O / E ratio of all urine-free DNA fragments. To calculate the O / E ratio of all urine-free DNA fragments, the observed OCR-related DNA contribution can be determined as the percentage of fragments aligned with the OCR among all fragments. The expected OCR-related DNA contribution can be defined as the theoretical percentage of OCRs in a reference genome (e.g., the human reference genome). 1. O / E ratio

[0167] Figure 16 FIG. 1600 showing the correlation between the fetal DNA fraction in maternal urine and the O / E ratio of all urine-free DNA fragments from OCR regions. All OCRs corresponding to DHS sites were used. Here, we define the observed value (O) as the percentage of fragments aligned with the OCR among all fragments, and the expected value (E) as the theoretical percentage of OCRs in the reference genome. The higher the O / E ratio, the more enriched the fragments from the OCR region. As Figure 16 shown, the concentration fraction of fetal DNA in maternal urine increases proportionally with the O / E ratio of all urine-free DNA fragments (Pearson's R = 0.866; p-value < 0.001). Thus, the concentration fraction of fetal DNA in a urine sample can be estimated by determining the O / E ratio of DNA molecules from OCRs.

[0168] Figure 17 FIG. 1700 showing the correlation between the fetal DNA fraction in maternal urine and the O / E ratio of urine cfDNA fragments from placenta-specific DHSs. When all DHS sites were used, the higher the O / E ratio, the more enriched the fragments from the OCR region. As Figure 17 shown, the concentration fraction of fetal DNA in maternal urine increases proportionally with the O / E ratio of all urine-free DNA fragments in placenta-specific DHSs (Pearson's R = 0.820; p-value < 0.001). Thus, the concentration fraction of fetal DNA in a urine sample can be estimated by determining the O / E ratio of DNA molecules from tissue-specific OCRs (e.g., placenta-specific OCRs). 2. Use of normalized end density and length

[0169] As another example of using relative abundance to determine the concentration fraction of clinically relevant DNA, the end density of all cell-free DNA in the OCR is used to determine the fetal fraction in a urine sample. End density can identify the ratio of DNA molecules that terminate at a specific location (e.g., a DNase1 hypersensitive site). For example, for each DNase1 hypersensitive site, the normalized end density can be calculated at a distance of 0 bp from the central genomic location. A higher normalized end density at the OCR (at a distance of 0 bp from the central genomic location of the OCR) can be associated with a higher fraction of renal cell-free DNA (e.g., fetal DNA) in the urine.

[0170] To determine the end density of the OCR region, we analyzed 14 maternal urine samples and 11 maternal plasma samples. The 5' and 3' ends of DNA fragments within 1 kb upstream and 1 kb downstream of the central genomic location of the OCR were analyzed. The normalized end density is defined as the count of fragment ends located within a window around the OCR (e.g., 1 kb upstream and 1 kb downstream) divided by the median or mean count of one or more loci / regions adjacent (e.g., flanking) to the OCR used. Other windows upstream or downstream, e.g., at least 100, 200, 300, 400, 500, 600, 700, 800, 900, or greater than 1,000 bp, can also be used. As an example, the adjacent loci can be outside the window used to define the OCR and can have various lengths, e.g., as described above.

[0171] Figure 18 FIG. 1800 shows the normalized end density of all cell-free DNA in the OCRs of maternal urine and plasma samples. All OCRs corresponding to DHS sites were used. Maternal urine and maternal plasma samples are represented by red and yellow lines, respectively. As shown in FIG. 1800, the normalized end density of the OCR is significantly more enriched in maternal urine samples than in maternal plasma samples.

[0172] Figure 19 FIG. 1900 shows a set of plots identifying a comparison between the end density of cell-free DNA in urine and cell-free DNA in plasma for determining the concentration fraction of fetal DNA according to some embodiments. For urine samples, the end density of DNA molecules from the OCR can be used to determine the concentration fraction of fetal DNA. This determination cannot be made for plasma samples. FIG. 1902 shows that the fetal DNA fraction increases proportionally with the normalized end density of cell-free DNA in urine, while FIG. 1904 does not show this increase for cell-free DNA in plasma. Thus, the relative abundance at the OCR in urine samples can be used to estimate the concentration fraction of clinically relevant DNA in urine samples.

[0173] Figure 20Shows a set of graphs 2000 that identify the correlation between fetal DNA fraction and the normalized end density of urine-free DNA of different lengths according to some embodiments. In Figure 20 Figure 2002 shows the normalized end density of all free DNA fragments in a maternal urine sample, while Figure 2004 shows the normalized end density of fragments in a maternal urine sample that are equal to or less than 80 bp in length. As described herein, other length thresholds (length thresholds) may be used in addition to 80 bp.

[0174] In Figure 2002, the fetal DNA fraction in maternal urine was significantly correlated with the normalized end density of OCR (Pearson's R = 0.926; p-value < 0.001). The correlation between the fetal DNA fraction and the normalized end density in maternal urine could be further enhanced by selecting fragments <= 80 bp (Pearson's R = 0.960; P-value < 0.001) (Figure 2004). The results suggest that the use of molecules from OCR can inform the extent of renal urine-free DNA. 3. Methods

[0175] Figure 21 Shows a flowchart of method 2100 for estimating the concentration fraction of clinically relevant DNA molecules in a subject's urine sample according to some embodiments. The urine sample may include free clinically relevant DNA and other DNA. In other instances, the biological sample may not include clinically relevant DNA, and the estimated concentration fraction may indicate zero or a low percentage of clinically relevant DNA.

[0176] The urine sample may include a mixture of free DNA molecules from one or more tissue types, such as the heart, lungs, and liver. For example, the urine sample may be obtained from a pregnant woman containing maternal free DNA molecules and fetal free DNA molecules. The urine sample may contain tumor-specific free DNA molecules as well as other tissue-specific free DNA molecules. Clinically relevant DNA molecules may include fetal DNA. In some embodiments, clinically relevant DNA includes tumor DNA. Aspects of method 2100 and any other methods described herein may be performed by a computer system.

[0177] In some cases, urine samples are treated with a DNA stabilizer prior to obtaining cell-free DNA molecules. Different DNA stabilizers can be used, such as EDTA and Collipee stabilizer. EDTA can inhibit the cleavage activity of the DNASE1 family by chelating magnesium and calcium, which are essential ions required for DNASE1 digestion. The stabilizer can potentially stabilize urine DNA against degradation. The stabilizer can be, but is not limited to, preservatives provided by Collipee, diazolidinyl urea (DU), dimethyl urea, 2-bromo-2-nitropropane-1,3-diol, 5-hydroxymethoxymethyl-1-aza-3,7-dioxabicyclo(3.3.0)octane and 5-hydroxymethyl-1-aza-3,7-dioxabicyclo(3.3.0)octane and 5-hydroxypoly[methyleneoxy]methyl-1-aza-3,7-dioxabicyclo(3.3.0)octane, bicyclic oxazolidine (such as NuoSept95), DMDM hydantoin, imidazolidinyl urea (IDU), sodium hydroxymethylglycinate, hexamethylenetetramine chloroallyl chloride (quaternium-15), microbicides (such as Bioban, Preventol and Grotan), water-soluble zinc salts, EDTA, other metal ion chelators such as N,N'-bis-(dithiocarbamoyloxy)piperazine (BDP), diethyldithiocarbamate (salt) (DDTC), iminodisuccinic acid (IDS), polyaspartic acid, S,S-ethylenediamine-N,N'-disuccinic acid (EDDS), methylglycine diacetic acid (MGDA), etc.

[0178] In block 2102, multiple cell-free DNA molecules from a urine sample are analyzed. In some cases, multiple cell-free DNA molecules in a urine sample are analyzed to obtain sequence reads. As an example, sequence reads can be obtained using sequencing or probe-based techniques, either of which can include enrichment, such as by amplification or capture probes.

[0179] Sequencing can be performed in a variety of ways, for example, using massively parallel sequencing or next-generation sequencing, using single molecule sequencing, and / or using double-stranded or single-stranded DNA sequencing library preparation protocols. Those skilled in the art will understand the various sequencing techniques that can be used. As part of the sequencing, some sequence reads may correspond to cellular nucleic acids.

[0180] In some cases, analyzing the plurality of cell-free DNA molecules includes: (i) determining the locations of the plurality of cell-free DNA molecules; and (ii) based on the locations, identifying a set of cell-free DNA molecules that are from open chromatin regions of one or more tissues associated with the clinically relevant DNA molecules. All OCRs or only a subset of the OCRs can be used. For example, OCRs specific to the tissues that generate (e.g., contribute to) renal DNA can be used. Any one or more renal-specific OCRs can be used in the embodiments of the present disclosure. Such regions can be referred to as renal open chromatin regions.

[0181] The location of the cfDNA molecules can be determined by aligning (mapping) one or more corresponding sequence reads to a reference genome. As another example, the location can be defined based on the probes used, e.g., by emitting a signal, such as the color of a fluorescent dye, for identification. In this way, it can be determined whether the cfDNA molecules are within the renal OCR.

[0182] As will be understood by those skilled in the art in light of the present disclosure, OCRs can be identified in various ways. Open chromatin regions can include one or more DNase1 hypersensitive sites (DHSs) defined using DNase-seq (Meuleman et al. Nature. 2020; 584:244-251). Open chromatin regions can include sites identified using DNase-seq, sites identified using assay for transposase-accessible chromatin using sequencing (ATAC-seq), transcription start sites (TSSs), CCCTC-binding factor (CTCF) sites, enhancer sites, and other nuclease hypersensitive sites.

[0183] In some embodiments, analyzing the plurality of cell-free DNA molecules further includes identifying a set of cell-free DNA molecules that are from open chromatin regions of one or more tissues and have a length less than a specified length threshold. For example, as Figure 20As shown in Figure 2004, the relative abundance of clinically relevant DNA molecules can be calculated based on shorter DNA fragments from open chromatin regions (e.g., fragments less than 80 base pairs in length). The length threshold can filter shorter DNA fragments when determining the relative abundance of clinically relevant DNA molecules, as the characteristics of renal DNA (e.g., DNA molecules through the GBM of the kidney) can lie in its shorter length. As described herein, a certain amount of renal DNA can be used to identify clinically relevant DNA molecules in urine samples. As an example, the length threshold can be 40 base pairs, 50 base pairs, 60 base pairs, 70 base pairs, 80 base pairs, 90 base pairs, 100 base pairs, 110 base pairs, 120 base pairs, 130 base pairs, 140 base pairs, 150 base pairs, or 160 base pairs, which can be used in any embodiment using the length threshold as described herein.

[0184] A statistically significant number of free DNA molecules can be analyzed to provide an accurate determination of the concentration fraction. In some embodiments, at least 1,000 free DNA molecules are analyzed. In other embodiments, at least 5000 or 10,000 or 50,000 or 100,000 or 500,000 or 1,000,000 or 5,000,000 or more free DNA molecules can be analyzed. In some cases, the collection of free DNA molecules includes at least 1000, 2000, 3000, 4000, 5000, 10,000, 50,000, or 100,000 free DNA molecules.

[0185] To identify a collection of free DNA molecules from open chromatin regions, DNA fragments from OCRs in urine samples can be enriched (e.g., targeted sequencing), resulting in an enriched sample. For example, a biological sample can be enriched for DNA fragments from open chromatin regions of one or more tissues, such as CTCF sites, TSS sites, DNase1 hypersensitive sites, or PolII regions. Enrichment can include using capture probes that bind to a portion or the entire genome as defined by, for example, a reference genome. As another example, enrichment can use primers to amplify certain regions of the genome (e.g., by PCR, rolling circle amplification, or multiple displacement amplification (MDA)). In some cases, enrichment includes using a collection of free DNA molecules from open chromatin regions of one or more tissues and having a length less than a specified length threshold. In some embodiments, free DNA molecules in urine samples with multiple fragmentomics features are enriched, including free DNA molecules that: (i) are from open chromatin regions of one or more tissues; (ii) have a length less than a specified length threshold (e.g., 80 base pairs); and / or (iii) have one or more end sequences corresponding to sequence end tags (e.g., CC ends).

[0186] In block 2104, a collection of cell-free DNA molecules is used to determine the relative abundances of multiple cell-free DNA molecules from open chromatin regions of one or more tissues. In some cases, the relative abundances can include normalized end densities. For example, the normalized end density can be calculated by dividing the count of fragment ends of a collection of DNA molecules within windows of different lengths centered around an OCR (e.g., 1 kb upstream and 1 kb downstream or other as described herein) by the median or average count across all loci flanking the OCRs. OCRs can be defined in various ways, e.g., by CTCF sites, TSS sites, DNase1 hypersensitive sites, Pol II regions.

[0187] Thus, the end density can include a first quantity of a collection of cell-free DNA molecules from open chromatin regions of one or more tissues divided by a second quantity of multiple cell-free DNA molecules from one or more other regions, such as regions adjacent to one or more OCRs, potentially all OCRs used. The second quantity can be the quantity of all multiple cell-free DNA molecules, and thus the first quantity can be a subset of the second quantity.

[0188] In some embodiments, as previously shown in Figure 16 - 17 a urine sample from a healthy subject can be expected to show a specific quantity of DNA molecules from open chromatin regions, or a specific ratio (e.g., O / E ratio) between a first relative frequency (e.g., percentage) (observed value) of urine DNA molecules from OCRs and a second relative frequency (expected value) of a reference sequence of a reference genome of open chromatin regions of one or more tissues. Thus, the expected OCR-related DNA contribution can correspond to the theoretical percentage of OCRs in the reference genome (e.g., the human reference genome). For example, the observed value (O) can include the percentage of fragments that align with OCRs among all fragments and the expected value (E) as the theoretical percentage of OCRs in the reference genome. In some cases, the expected value is determined based on the relative frequencies of single nucleotide variants of a reference genome of open chromatin regions of one or more tissues. In various instances, the relative abundance can be a ratio between the first relative frequency and the second relative frequency or a ratio of one of the frequencies to the sum of the two values.

[0189] In block 2106, the concentration fraction of clinically relevant DNA molecules in a biological sample is estimated by comparing the relative abundances to one or more calibration values determined from one or more calibration samples, the concentration fractions of clinically relevant DNA molecules of which are known. As Figure 19 and 20As shown, fetal and maternal DNA have different relative abundances. A sample having a mixture of both will have a relative abundance that depends on the ratio of fetal / maternal DNA in the sample. The concentration fraction of the calibration sample can be determined in other ways, e.g., using a locus on the Y chromosome of a male fetus or a fetal-specific marker (e.g., an allele inherited from the father or a fetal-specific epigenetic marker).

[0190] The calibration data points can include the relative abundance of clinically relevant DNA and the measured / known fraction. This comparison can include comparing the relative abundance to a calibration curve (composed of the calibration data points), so this comparison can identify the point on the curve having the measured relative abundance of the test sample. For example, the relative abundance can be compared to the calibration curve by inputting the relative abundance into a calibration function representing the calibration curve. Then the concentration fraction corresponding to the identified point can be used to estimate the concentration fraction. For example, the relative abundance can be provided as an input to a calibration function (e.g., a linear or non-linear fit) to obtain an output of the concentration fraction.

[0191] Thus, comparing the relative abundance to one or more calibration values can include comparing the relative abundance to a calibration curve including one or more calibration values. And to obtain calibration data points, for each calibration sample of one or more calibration samples, some embodiments can measure the concentration fraction of clinically relevant DNA molecules in the calibration sample and measure the relative abundance of free DNA molecules from open chromatin regions of one or more tissues from the calibration sample. As described above, measuring the concentration fraction of clinically relevant DNA molecules can use tissue-specific alleles or tissue-specific methylation patterns.

[0192] The concentration fraction is a quantitative value and can be a range of values. For example, the concentration fraction can determine that the quantitative value is greater than or less than a specified value. In other embodiments, the concentration fraction can have an upper and lower limit, which can correspond to the resolution at which the concentration fraction can be determined. B. Enriching Urine Samples Using Terminal Motif Features

[0193] In some embodiments, the fraction of renal urine free DNA in a urine sample can be determined using certain terminal motifs, e.g., as described in Part IV and elsewhere herein. The terminal motifs can include, but are not limited to, terminal sequences having certain lengths (e.g., 1-mer, 2-mer, 3-mer, 4-mer, 5-mer, 6-mer, 7-mer, 8-mer, 9-mer). Further data are provided below. 1. Relationship between C-Terminal Terminal Motifs and Fetal Fraction

[0194] Figure 22A set of graphs 2200 showing the correlation between the fetal DNA fraction and the proportion of urine-free DNA fragments carrying the CC terminus is presented. In Figure 22 Figure 2202 shows the correlation between the fetal DNA fraction and the proportion of all urine-free DNA fragments carrying the CC terminus in a urine sample. Curve 2202 shows the correlation between the fetal DNA fraction and the proportion of free DNA fragments carrying the CC terminus, where the CC terminus is equal to or less than 80 bp.

[0195] In Figure 2202, the fetal DNA fraction in maternal urine was significantly correlated with the contribution of the proportion of urine-free DNA fragments carrying the "CC-terminus" in all fragments (Pearson's R = 0.637; P-value = 0.006). In Figure 2204, after selecting fragments <= 80 bp, a further increase in the correlation between the fetal DNA fraction in maternal urine and the proportion of urine-free DNA fragments carrying the "CC-terminus" was observed (Pearson's R = 0.807; P-value < 0.001). 2. Exemplary enrichment protocol

[0196] Figure 23 A technique 2300 for enriching a set of one or more terminal motifs using probes is illustrated. Technique 2300 can be used to enrich the C-terminal motif to enrich clinically relevant DNA in urine samples, such as Figure 9 、 10 as shown in 22.

[0197] As shown, the free DNA molecules 2302 have different terminal motifs, for example, 1-mer terminal motifs in this example.

[0198] In step 2304, cfDNA fragments with different terminal motifs are ligated to a common sequence 2305 (e.g., an artificial sequence). More than one sequence can be used, and using one annotation sequence can be more effective. The length of the artificial sequence should be ≥ a specified length, such as 16 bp (or 17, 18, 19, 20, 21, 22, 23, 24, 25, or 26 bp) to ensure the specificity of probe binding (4 16 > 3×10 9 [human genome length]). The artificial sequence at the end of the DNA fragment can facilitate the probe recognition of specific DNA terminal motifs.

[0199] In step 2306, the DNA molecules with the common sequence 2305 are denatured to separate the two strands, resulting in single-stranded cfDNA 2308 with different terminal motifs and the common sequence 2305. Various denaturation protocols can be used, such as using temperature, as will be understood by those skilled in the art.

[0200] In step 2310, a surface 2312 (e.g., a chip surface) is immobilized with a plurality of probe sequences 2316. The probe sequences 2316 have two components, which include a sequence complementary to a common sequence 2305 and a motif sequence 2318 complementary to a targeted end motif sequence (e.g., "G" complementary to a target C-terminal motif). Only fragments having the targeted end motif (e.g., having an end-C motif) can bind to the probe, leaving fragments ending with other motifs unbound (i.e., unbound fragments 2320). The complementary motif sequences 2318 can be a set of different end motifs, e.g., if 2-mers or higher oligomers are used. For example, for 2-mers, four different probes can be used for the four different 2-mers ending with C.

[0201] In step 2314, the unbound fragments 2320 are washed away. The remaining bound fragments 2322 can be detected or further analyzed in various ways. For example, only probes having a complementary motif sequence 2318 that binds to the fragment can be extended (e.g., by ligating a nucleotide conjugated to a fluorescent dye by a DNA polymerase). In this way, a fluorescent signal can be detected when a fragment carrying the targeted motif is present. As other examples of detection, the reaction can extend the bound cfDNA fragment by a biotin-labeled nucleotide. The biotin can be detected by streptavidin conjugated to a fluorophore. As another alternative, the reaction can extend a dinitrophenyl-labeled nucleotide. The dinitrophenyl can be detected by an anti-DNP antibody labeled with a fluorophore. In other embodiments, the bound fragments can be sequenced in a separate process.

[0202] Figure 24 Another technique 2400 for enriching a set of one or more end motifs using probes and beads is illustrated. Similar to Figure 23 , cfDNA fragments having different end motifs can be ligated to an artificial sequence. The artificial sequence at the end of the DNA fragment can facilitate probe recognition of a specific DNA end motif. The double-stranded DNA fragment can be denatured to become single-stranded DNA.

[0203] In the example shown, a probe targeting a DNA fragment having a specific end motif has three components: a biotin that can bind to a streptavidin bead, a complementary sequence of the common sequence, and a complementary motif sequence of the specific end motif sequence (e.g., "G" targeting the C-terminal motif). The probe hybridizes to the DNA fragment. Only fragments having the specific end motif can bind to the probe, leaving fragments having other end motifs unbound.

[0204] Streptavidin beads can capture the probes because of the high affinity between biotin and streptavidin. Only fragments with a specific terminal motif can be captured by streptavidin beads. Unbound fragments are washed away. As a result, cfDNA fragments with a specific terminal motif can be captured by this design.

[0205] Fragments that bind to the complementary motif sequence using Technique 2400 can be detected or further analyzed in the same manner as Technique 2300.

[0206] Instead of washing away unbound fragments to enrich for the target terminal motif, the target terminal motif can be amplified. For example, primers containing a common sequence and the target terminal motif can be added to the reaction along with nucleotides, and an amplification process (such as PCR or rolling circle) can be carried out. 3. Method

[0207] Figure 25 A flowchart of Method 2500 for enriching clinically relevant DNA in a urine sample based on the terminal motif signature of urine-free DNA is shown according to some embodiments. Aspects of Method 2500 and other methods described herein can be carried out in a manner similar to Method 2100, e.g., sample preparation and analysis of DNA molecules. The urine sample can include free clinically relevant DNA and other DNA. In other instances, the biological sample may not include clinically relevant DNA, and the estimated concentration fraction can indicate zero or a low percentage of clinically relevant DNA. Aspects of Method 2500 and any other method described herein can be performed by a computer system.

[0208] In block 2502, a plurality of free DNA molecules from a urine sample are analyzed. Aspects of block 2502 can be carried out in a manner similar to block 2102 of Method 2100. For example, a plurality of free DNA molecules in the urine sample can be analyzed to obtain sequence reads. The sequence reads can include terminal sequences corresponding to the ends of a plurality of free DNA fragments. As an example, the sequence reads can be obtained using sequencing or probe-based techniques, either of which can include enrichment, e.g., by amplification or capture probes.

[0209] Sequencing can be carried out in various ways, e.g., using massively parallel sequencing or next-generation sequencing, using single-molecule sequencing, and / or using double-stranded or single-stranded DNA sequencing library preparation protocols. Those skilled in the art will understand the various sequencing techniques that can be used. As part of the sequencing, some sequence reads may correspond to cellular nucleic acids.

[0210] In some embodiments, analyzing multiple cell-free DNA molecules further includes identifying a set of cell-free DNA molecules that are located in a set of one or more sequence motifs that include C-terminal nucleotides. The sequence end tags can be part of a K-mer end motif, such as a 2-mer, 3-mer, 4-mer, etc. For example, the set of cell-free DNA molecules is further identified based on having a CC end. Additionally, the end sequences can be required to be at both ends of the DNA fragment, or a specific pair of different end motifs can be used to select a specific set of DNA fragments.

[0211] When sequencing, identifying a set of multiple cell-free DNA molecules can include identifying sequence reads that have end sequences in a set of one or more sequence motifs. Thus, the enriched sample can correspond to sequence reads that have end sequences in a set of one or more sequence motif groups. As an alternative to sequencing, to identify DNA molecules that have one or more end sequences, one or more probe molecules can be attached to a surface and the sequence motifs in the end sequences can be detected by hybridization.

[0212] In some embodiments, the set of cell-free DNA molecules is further identified based on their respective lengths (e.g., fragments less than 80 base pairs in length). As Figure 22 shown, a length threshold can filter shorter DNA fragments because renal DNA (e.g., DNA molecules passing through the GBM of the kidney) can be characterized by its shorter length. As described herein, a certain amount of renal DNA can be used to identify clinically relevant DNA molecules in urine samples. The specified length threshold can include 40 base pairs, 50 base pairs, 60 base pairs, 70 base pairs, 80 base pairs, 90 base pairs, 100 base pairs, 110 base pairs, 120 base pairs, 130 base pairs, 140 base pairs, 150 base pairs, or 160 base pairs.

[0213] As described herein, a statistically significant number of cell-free DNA molecules can be analyzed.

[0214] In block 2504, an enriched sample can be generated by using a set of cell-free DNA molecules in a set of one or more sequence motifs. Thus, the enriched sample includes a higher concentration of clinically relevant DNA compared to the urine sample. The enriched sample can be a computer-simulated sample because only measurements of certain cfDNA molecules are used. In other instances, the enriched sample can be a physical sample.

[0215] Enrichment can include using capture probes that bind to a collection of one or more sequence motifs. For example, identifying a collection of free DNA molecules or generating an enriched sample can include subjecting a plurality of free DNA molecules to one or more probe molecules that detect a collection of one or more sequence motifs in the end sequences of the plurality of free DNA molecules. A collection of free DNA molecules can be obtained using such probe molecules. As Figure 23 and 24 described, some embodiments can ligate a common sequence to a plurality of free DNA molecules. Then, one or more probe molecules can include a complementary sequence of the common sequence.

[0216] In some cases, generating an enriched sample includes using probe molecules to capture a collection of free DNA molecules and discarding other free DNA molecules among the plurality of free DNA molecules. In other cases, generating an enriched sample can include amplifying a collection of free DNA molecules using one or more probe molecules.

[0217] Capture probes can also bind (target) a portion or the entire genome as defined, for example, by a reference genome. As another example, enrichment can use primer amplification (e.g., by PCR, rolling circle amplification or multiple displacement amplification (MDA)) of certain regions of the genome.

[0218] In some embodiments, enriching free DNA molecules in a urine sample that have a plurality of fragmentomics features includes the following free DNA molecules: (i) open chromatin regions from one or more tissues; (ii) having a length less than a specified length threshold (e.g., 80 base pairs); and / or (iii) having one or more end sequences corresponding to sequence end tags (e.g., the C-terminus or CC-terminus of a K-mer end motif).

[0219] In block 2506, properties associated with clinically relevant DNA in the enriched urine sample are determined. As an example, properties of clinically relevant DNA in the urine sample can be (1) the concentration fraction of clinically relevant DNA or (2) the pathologic grade of the subject from whom the biological sample was obtained, e.g., where the pathologic grade is associated with the clinically relevant DNA. Those skilled in the art will appreciate the various properties that can be determined using a collection of cell-free DNA molecules having terminal sequences, e.g., fetal inheritance of haplotypes, detection of mutations, copy number aberrations (e.g., aneuploidy), methylation properties, various base modifications, genomic interactions, protein binding states, fragmentomic signatures, etc., the terminal sequences being in a collection of one or more sequence motifs including the C-terminal nucleotide, as variously described in U.S. Publication Nos. 2009 / 0087847, 2009 / 0029377, 2011 / 0276277, 2011 / 0105353, 2013 / 0040824, 2014 / 0100121, 2014 / 0080715, and 2020 / 0199656. C. Enriching the urine sample using features such as OCR

[0220] The above fragmentomic signatures (e.g., terminal motifs, lengths, enrichment of open chromatin regions) can be combined to estimate the renal DNA contribution. For example, fetal DNA molecules can be enriched with CC termini. Based on this correlation, the renal DNA contribution can be estimated based on the proportion of urine cell-free DNA having CC termini in the urine sample. If the proportion of urine cell-free DNA having CC termini and a length (e.g., fragments shorter than 80 bp) is used together, the estimation of the renal DNA contribution in the urine sample can become more precise. Indeed, the accuracy of estimating the fetal DNA fraction can also be improved.

[0221] Figure 26 A set of graphs is shown that depict the enrichment of fetal DNA using urine cell-free DNA having various fragmentomic properties according to some embodiments. As Figure 26 shown, DNA molecules filtered for CC termini (FIG. 2610), OCR (FIG. 2620), and lengths less than 80 base pairs (FIG. 2630) can result in a significant increase in the enrichment of fetal DNA in the urine sample. The use of the combination (FIG. 2640) results in a further significant enrichment compared to using the above fragmentomic signatures alone.

[0222] Figure 27Displays a bar chart 2700 showing the enrichment of cell-free DNA in renal urine using selective analysis of fragments with different fragmentomics characteristics. In the absence of selection 2702, in CC-end-based selective analysis 2704, in OCR-based selective analysis 2706, in length-based selection (<=80bp) 2708, and in the case of a combination of features corresponding to the end motif, OCR, and length 2710, the percentage of fetal DNA fraction in cell-free urine DNA increases. To show the increase in the fetal DNA fraction, we calculated the average increase in the fetal DNA concentration fraction after different criteria selections.

[0223] As Figure 27 shown, if fragments are filtered for having a CC end, within the OCR region, or a length equal to or less than 80bp, the fetal DNA fraction in a given urine sample increases by 78.6%, 60.1%, and 223.8% respectively. In other words, filtering DNA molecules using the CC end and OCR region criteria results in an approximately 1-fold increase in the fetal DNA concentration fraction. If the length (fragments equal to or less than 80bp) criterion is used as a filter for DNA molecules, the fetal DNA concentration fraction increases by approximately 2-fold. If these three fragmentomics characteristics are combined, the fetal DNA fraction in urine can be further increased by 836.8% (more than 8-fold). Thus, Figure 27 the data in

[0224] show that target renal cell-free urine DNA can be enriched based on the selective analysis of cell-free DNA molecules according to different combinations of fragmentomics characteristics.

[0225] Figure 28Disclosed is a flowchart of method 2800 for enriching clinically relevant DNA in a urine sample based on terminal motifs and open chromatin regions according to some embodiments. Aspects of method 2800 and other methods described herein may be performed in a manner similar to method 2100 and / or method 2500, e.g., sample preparation and analysis of DNA molecules. The urine sample may include free clinically relevant DNA and other DNA. In other instances, the biological sample may not include clinically relevant DNA, and the estimated concentration fraction may indicate zero or a low percentage of clinically relevant DNA. Aspects of method 2800 and any other method described herein may be performed by a computer system.

[0226] At block 2802, a plurality of free DNA molecules from a urine sample are analyzed. Aspects of block 2802 may be performed in a manner similar to block 2102 or block 2502. For example, a plurality of free DNA molecules in the urine sample are analyzed to obtain sequence reads. The sequence reads may include terminal sequences corresponding to the ends of a plurality of free DNA fragments. As an example, the sequence reads may be obtained using sequencing or probe-based techniques, either of which may include, e.g., enrichment via amplification or capture probes, as described in method 2500.

[0227] In some embodiments, analyzing the plurality of free DNA molecules further includes identifying a set of free DNA molecules that: (i) are from open chromatin regions of one or more tissues; (ii) have a length less than a specified length threshold (e.g., 80 base pairs); and / or (iii) have one or more terminal sequences corresponding to sequence end tags (e.g., the C-terminus or CC-terminus of a K-mer terminal motif). The open chromatin regions may be identified in a manner similar to that described herein.

[0228] As Figure 26 and 27 shown, a set of free DNA molecules may be identified based on their respective lengths (e.g., fragments less than 80 base pairs). The length threshold may filter shorter DNA fragments since transrenal DNA (e.g., DNA molecules passing through the GBM of the kidney) may be characterized by its shorter length. As described herein, a certain amount of transrenal DNA may be used to identify clinically relevant DNA molecules in a urine sample. The specified length threshold may include 40 base pairs, 50 base pairs, 60 base pairs, 70 base pairs, 80 base pairs, 90 base pairs, 100 base pairs, 110 base pairs, 120 base pairs, 130 base pairs, 140 base pairs, 150 base pairs, or 160 base pairs.

[0229] In some embodiments, various methods (e.g., gel electrophoresis) can be used to determine the lengths of multiple free DNA molecules. For example, gel electrophoresis, filtration, length-selective precipitation, or hybridization can be used to measure the lengths of multiple free DNA molecules. Additionally or alternatively, sequence reads can be used to measure the lengths of multiple free DNA fragments. For example, sequence reads can be obtained from sequencing of multiple free DNA molecules from a biological sample (e.g., massively parallel sequencing, single molecule real-time sequencing, nanopore sequencing). To measure the length of a free DNA molecule, the number of nucleotides can be counted for each sequence read. Those skilled in the art will understand the various sequencing techniques that can be used. As part of the sequencing, some sequence reads may correspond to cellular nucleic acids. In some cases, DNA fragments in a urine sample that are less than a predetermined length threshold (e.g., 80 bp) can be enriched.

[0230] A collection of free DNA molecules can be further identified based on having one or more end sequences corresponding to sequence end tags. The sequence end tags can be part of an end motif, such as a 2-mer, 3-mer, etc. For example, a collection of free DNA molecules is further identified based on having a CC end. Additionally, it can be required that the end sequences be at both ends of the DNA fragment, or a specific pair of different end motifs can be used to select a specific collection of DNA fragments. In addition to sequencing, to identify DNA molecules having one or more end sequences, one or more probe molecules can be attached to a surface or bead and the sequence motifs in the end sequences can be detected by hybridization.

[0231] As described herein, a statistically significant number of free DNA molecules can be analyzed.

[0232] At block 2804, an enriched sample can be generated by using a collection of free DNA molecules that: (i) are from open chromatin regions of one or more tissues; (ii) have a length less than a specified length threshold (e.g., 80 base pairs); and / or (iii) have one or more end sequences corresponding to sequence end tags (e.g., CC end). Aspects of block 2804 can be performed in a manner similar to block 2504. Thus, the enriched sample includes a higher concentration of clinically relevant DNA compared to a urine sample. For example, a biological sample can be enriched for DNA fragments from open chromatin regions of one or more tissues, such as DNA fragments at CTCF sites, TSS sites, DNase1 hypersensitive sites, or Pol II regions. Enrichment can include using capture probes that bind to a portion or the entire genome as defined by, for example, a reference genome. As another example, enrichment can use primers to amplify certain regions of the genome (e.g., by PCR, rolling circle amplification, or multiple displacement amplification (MDA)).

[0233] In some cases, the enriched sample includes capturing a collection of free DNA molecules using probe molecules and discarding other free DNA molecules among the plurality of free DNA molecules. The enriched sample can be a computer-simulated sample.

[0234] At block 2806, a property associated with clinically relevant DNA in the enriched urine sample is determined. Aspects of block 2806 can be performed in a manner similar to block 2506. As an example, the property of the clinically relevant DNA in the urine sample can be (1) the concentration fraction of the clinically relevant DNA or (2) the pathological grade of the subject from whom the biological sample is obtained, e.g., where the pathological grade is associated with the clinically relevant DNA. Those skilled in the art will recognize the various properties that can be determined, e.g., as described above for method 2500. VII. Classification of Urinary cfDNA for Abnormalities

[0235] In some embodiments, cancer can be detected and monitored by using renal urine free DNA molecules. For example, renal cell carcinoma (RCC) is a disease in which malignant cells are found in the lining of the renal tubules. If renal function is affected, the concentration fraction of renal free DNA can be altered. Indeed, patients with renal cancer will exhibit an abnormality in the concentration fraction of renal urine free DNA compared to subjects without renal cancer. Other renal abnormalities (other than RCC) can also affect renal urine free DNA in the urine sample, e.g., based on length or region, such as OCR. Other examples include proteinuria and preeclampsia. A. Classification Based on OCR

[0236] It is expected that urine samples from healthy subjects will exhibit a specific amount of DNA molecules from open chromatin regions of one or more tissues, or a specific ratio between the observed frequency of urine DNA molecules from open chromatin regions and the expected frequency of the reference sequence of the reference genome of open chromatin regions of one or more tissues. However, if the permeability of the GBM is disrupted in some subjects (e.g., subjects with nephrotic syndrome, glomerulonephritis), the above ratio may increase or decrease. If such a change relative to the normal amount exceeds a predetermined threshold, it can be determined that the subject has a disease or other abnormal condition affecting renal permeability.

[0237] The amount of DNA molecules in open chromatin regions of one or more tissues of a control subject can be measured. Then, a significant deviation from the measured amount of DNA molecules can be used to determine whether a given subject has a kidney abnormality. For example, a blood sample can include cell-free DNA molecules derived from different organs (e.g., heart, lung, liver). The amount of cell-free DNA molecules corresponding to open chromatin regions of the liver (e.g.) can be determined for a urine sample (e.g., using targeted sequencing of OCR). If there is a statistically significant difference between the determined amount of cell-free DNA molecules and the calibrated amount of cell-free DNA molecules corresponding to open chromatin regions of a healthy subject, the classification of the kidney abnormality can be determined. More than one tissue-specific region can be used. Overall, the measurement can be for all OCRs or for OCRs specific to one or more that contribute renal DNA or for OCRs specific to one or more cell types, as described in the previous section. 1. Renal cell carcinoma

[0238] For illustrative purposes, we analyzed the O / E ratio of urine cell-free DNA from 15 control subjects and 16 renal cell carcinoma (RCC) patients. To calculate the O / E ratio for all urine cell-free DNA fragments, the observed OCR-related DNA contribution can be determined as the percentage of fragments aligned with OCRs among all fragments. The expected OCR-related DNA contribution can be defined as the theoretical percentage of OCRs in the human genome.

[0239] Figure 29 A set of graphs 2900 identifying the O / E ratio analysis in patients with RCC is shown. Box plot 2902 shows the box plot of the O / E ratio between control subjects and patients with RCC. As shown in box plot 2902, the O / E ratio of RCC patients (median: 1.378; range: 1.174 - 1.863) was significantly higher than that of the control group (median: 1.288; range: 1.190 - 1.571) (Mann-Whitney, P-value < 0.001). In addition, we also performed receiver operating characteristic (ROC) analysis on these samples. ROC 2904 shows the ROC for the performance level used to distinguish patients with RCC from control subjects. The area under the curve (AUC) of ROC 2904 was 0.964 in the discrimination between RCC patients and control subjects. 2. Proteinuria

[0240] Proteinuria, also known as albuminuria, is an elevation of proteins in the urine and can be considered a kidney abnormality. Since kidney function is not working well, it will allow more proteins to enter the urine, so it is a type of kidney abnormality.

[0241] Since the urine of patients with proteinuria contains excessive protein, we hypothesized that we could identify proteinuria patients from healthy controls using the fragmentomics features of urine cfDNA. We used the abundance of OCRs. Any kidney-specific OCR can be used. As with other embodiments of the present disclosure, for any given use case (e.g., classification of kidney abnormalities, estimation of concentration fractions, or enrichment), only OCRs are used, although two separate determinations can be made and then combined.

[0242] Since there is no placental DNA in the urine of healthy controls and subjects with proteinuria, we used blood-related regions to represent kidney-related genomic locations / regions.

[0243] Figure 30 A fragmentomics analysis of kidney DNA in proteinuria patients is shown. Box plot 3002 shows the O / E ratio of urine cfDNA fragments in OCRs (blood-specific DHSs) in healthy controls and proteinuria patients.

[0244] In the O / E ratio analysis, proteinuria patients had significantly lower O / E ratios for fragments in OCRs (blood-specific DHSs) (Mann-Whitney U test, P value = 0.0052). These results indicate a reduced proportion of fragments from OCRs in proteinuria patients. ROC analysis is provided later. 3. Preeclampsia

[0245] We hypothesized that we could use the fragmentomics features of urine cfDNA to identify pregnant women with preeclampsia. Pregnant women with preeclampsia are typically diagnosed with elevated protein levels in the urine, indicating impaired GBM function in the kidneys. We speculated that if large-length plasma molecules such as proteins can pass through the GBM and enter the urine, then large-length DNA molecules from plasma (e.g., long DNA molecules bound to histones or DNA molecules) can also enter the urine.

[0246] We used DHSs to represent OCRs. Other ways to identify OCRs are described elsewhere in the present disclosure.

[0247] Figure 31 A fragmentomics analysis of kidney DNA in pregnant women with preeclampsia is shown. Box plot 3110 shows the O / E ratio of urine cfDNA fragments in OCRs (all DHSs) in healthy pregnant women and pregnant women with preeclampsia. Box plot 3120 shows the O / E ratio of urine cfDNA fragments in OCRs (placenta-specific DHSs) in healthy pregnant women and pregnant women with preeclampsia.

[0248] The use of the O / E ratio of fragments to distinguish between healthy pregnant women and pregnant women with preeclampsia showed better performance in placenta-specific DHS (box plot 3120) than in all DHS (box plot 3110) (Mann-Whitney U test, P-values: 0.0011 vs 0.0118). Therefore, we used tissue-specific regions in the urine of subjects with preeclampsia and proteinuria for O / E ratio analysis. Pregnant women with preeclampsia had significantly lower O / E ratios in fragments of OCR (placenta-specific DHS) compared to healthy pregnant women (Mann-Whitney U test, P-value = 0.0011). These data indicate a reduced proportion of fragments from OCR in patients with preeclampsia.

[0249] When the kidney abnormality is preeclampsia, additional factors can be used. For example, the determination of the presence of hypertension can also be used. For example, blood pressure can be compared to a threshold to determine whether a subject has hypertension. Another factor can be the presence of protein in the urine, such as proteinuria. 4. Methods

[0250] Figure 32 A flowchart of a method 3200 for determining the classification of kidney abnormalities based on urine-free DNA from open chromatin regions according to some embodiments is shown. Aspects of method 3200 and other methods described herein can be performed in a manner similar to the above-described methods, e.g., sample preparation and analysis of DNA molecules. The urine sample can include a mixture of free DNA molecules from one or more tissue types, such as the heart, lung, and liver. The urine sample can contain tumor-specific free DNA molecules as well as other tissue-specific free DNA molecules. For example, the urine sample can include free DNA molecules specific for renal cell carcinoma (RCC). Aspects of method 3200 and any other method described herein can be performed by a computer system.

[0251] In block 3202, a plurality of free DNA molecules from the urine sample are analyzed. Aspects of block 3202 can be performed in a manner similar to the similar blocks of other methods (e.g., block 2102 of method 2100), as can be done for other methods herein. For example, a plurality of free DNA molecules in the urine sample are analyzed to obtain sequence reads. As an example, the sequence reads can be obtained using sequencing or probe-based techniques, either of which can include enrichment, e.g., by amplification or capture probes.

[0252] In some cases, analyzing the plurality of cell-free DNA molecules includes: (i) determining the locations of the plurality of cell-free DNA molecules; and (ii) based on the locations, identifying a set of cell-free DNA molecules that are from open chromatin regions of one or more tissues associated with the clinically relevant DNA molecules. The one or more tissues may include at least one of the heart, lung, or liver. Open chromatin regions may include one or more DNase1 hypersensitive sites (DHSs), which are defined using DNase-seq (Meuleman et al. Nature. 2020;584:244-251). Open chromatin regions may be identified as described herein.

[0253] As described herein, a statistically significant number of cell-free DNA molecules can be analyzed.

[0254] To identify a set of cell-free DNA molecules from open chromatin regions, DNA fragments from open chromatin regions of a urine sample can be enriched (e.g., targeted sequencing), resulting in an enriched sample. For example, a biological sample can be enriched for DNA fragments from open chromatin regions of one or more tissues, such as CTCF sites, TSS sites, DNase1 hypersensitive sites, or Pol II regions. Enrichment can include using capture probes that bind to a portion or the entire genome as defined by, for example, a reference genome. As another example, enrichment can use primers to amplify certain regions of the genome (e.g., by PCR, rolling circle amplification, or multiple displacement amplification (MDA)). In some cases, enrichment includes using a set of cell-free DNA molecules from open chromatin regions of one or more tissues and having a length less than a specified length threshold.

[0255] In block 3204, the relative abundances of the plurality of cell-free DNA molecules from open chromatin regions of one or more tissues are determined. Aspects of block 3204 can be performed in a manner similar to similar blocks of other methods (e.g., block 2104 of method 2100), as can be done for other methods herein. In some cases, the relative abundances can include a normalized end density. For example, the normalized end density can be calculated by dividing the count of fragment ends of a set of DNA molecules located within 1 kb upstream and 1 kb downstream of an OCR (e.g., a CTCF site, TSS site, DNase1 hypersensitive site, Pol II region) by the median count across all loci flanking the OCR.

[0256] In some embodiments, as previously in Figure 29 - 31As shown, urine samples from healthy subjects can be expected to show a specific amount of DNA molecules from open chromatin regions, or a specific ratio (e.g., O / E ratio) between the first relative frequency (e.g., percentage) (observed value) of urine DNA molecules from OCRs and the second relative frequency (expected value) of the reference sequence of the reference genome of open chromatin regions from one or more tissues. Thus, the desired OCR-related DNA contribution can correspond to the theoretical percentage of OCRs in the reference genome (e.g., the human reference genome). For example, the observed value (O) can include the percentage of fragments aligned with OCRs among all fragments and the expected value (E) which is the theoretical percentage of OCRs in the reference genome. In some cases, the expected value is determined based on the relative frequency of single nucleotide variants of the reference genome of open chromatin regions from one or more tissues. However, if the permeability of GBM is disturbed in some subjects, the above ratio may increase or decrease. If such a change relative to the normal amount exceeds a predetermined threshold, it can be determined that the subject has a disease or other abnormal condition (such as nephrotic syndrome, glomerulonephritis) affecting renal permeability.

[0257] At block 3206, the relative abundance value is compared with a reference value. The reference value can correspond to another relative abundance determined based on cell-free DNA molecules from open chromatin regions of one or more reference samples, where the one or more reference samples are associated with known classifications of kidney abnormalities. For example, the reference value can correspond to the relative abundance determined from healthy subjects. In some cases, the reference value is a calibration value or is determined by the calibration value of a calibration sample. As with other reference values, the specific value chosen can depend on the trade-off between specificity and sensitivity. In some embodiments, a machine learning model can be used to perform the comparison.

[0258] At block 3208, a classification of the subject with a kidney abnormality is determined based on the comparison. In some embodiments, comparing the relative abundance with the reference value includes: (1) determining whether the relative abundance differs from the reference value by at least a threshold amount or the difference is less than the threshold amount; (2) determining whether the relative abundance is at least a threshold amount less than the reference value; or (3) determining whether the relative abundance is greater than the reference value by at least a threshold amount. As an example, kidney abnormalities can include renal cell carcinoma (RCC), nephrotic syndrome, glomerulonephritis, Fabry disease, cystinosis, IgA nephropathy, IgM nephropathy, lupus nephritis, atypical hemolytic uremic syndrome (aHUS), polycystic kidney disease (PKD), Alport syndrome, interstitial nephritis, proteinuria, chronic kidney disease, acute kidney injury, proteinuria, preeclampsia, etc. In some cases, the classification of a subject with a kidney abnormality includes an increased permeability level related to the glomerular basement membrane of the kidney.

[0259] The classification of kidney abnormalities can be determined using machine learning trained with a training dataset. The training dataset can include training samples. The training samples can be associated with known classifications of kidney abnormalities. In another example, a machine learning model can be used to make a comparison with a reference value. The machine learning model can be applied to relative abundances to produce a classification of kidney abnormalities. Machine learning models can include, but are not limited to, convolutional neural networks (CNNs), linear regression, logistic regression, deep recurrent neural networks (e.g., fully connected recurrent neural networks (RNNs), gated recurrent units (GRUs), long short-term memory (LSTMs)), transformer-based methods (e.g., XLNet, BERT, XLM, RoBERTa), Bayesian classifiers, hidden Markov models (HMMs), linear discriminant analysis (LDA), k-means clustering, density-based spatial clustering of applications with noise (DBSCAN), random forest algorithms, adaptive boosting (AdaBoost), extreme gradient boosting (XGBoost), support vector machines (SVMs), or composite models that include one or more of the models presented above. B. Length-based classification

[0260] Additionally, or from classification using OCR, the length of urine cfDNA can be used for classification. The classification of kidney abnormalities can be performed in a similar manner, but instead using the statistics of the length distribution of cfDNA in urine samples. 1. Proteinuria and preeclampsia

[0261] Figure 33 Is the fragmentomics analysis of renal DNA in patients with proteinuria and preeclampsia alone 3300.

[0262] Box plot 3310 shows the proportion of urine cfDNA >80bp in healthy controls and patients with proteinuria. We observed a higher proportion of long urine cfDNA fragments (i.e., >80bp) in patients with proteinuria than in healthy controls (Mann-Whitney U test, P value = 0.0256).

[0263] Box plot 3320 shows the proportion of urine cfDNA >80bp in healthy pregnant women and pregnant women with preeclampsia. We observed a higher proportion of long urine cfDNA fragments (i.e., >80bp) (Mann-Whitney U test, P value = 0.0021) in pregnant women with preeclampsia than in healthy pregnant women. 2. Methods

[0264] Figure 34Shows a flowchart of method 3400 for determining a classification of kidney abnormalities based on the length of urine-free DNA according to some embodiments. Aspects of method 3400 and other methods described herein can be performed in a manner similar to the methods described above, e.g., sample preparation and analysis of DNA molecules. A urine sample can include a mixture of free DNA molecules from one or more tissue types, such as the heart, lung, and liver. A urine sample can contain tumor-specific free DNA molecules as well as other tissue-specific free DNA molecules. For example, a urine sample can include free DNA molecules specific to renal cell carcinoma (RCC). Aspects of method 3400 and any other method described herein can be performed by a computer system.

[0265] At block 3402, a plurality of free DNA molecules from a urine sample are analyzed. Aspects of block 3402 can be performed in a manner similar to similar blocks of other methods (e.g., block 2102 of method 2100), as can be done for other methods herein. For example, a plurality of free DNA molecules in a urine sample are analyzed to obtain sequence reads. As an example, sequence reads can be obtained using sequencing or probe-based techniques, either of which can include enrichment, e.g., by amplification or capture probes.

[0266] In some embodiments, analyzing a plurality of free DNA molecules includes determining the lengths of the plurality of free DNA molecules. Various methods (e.g., gel electrophoresis) can be used to determine the lengths of the plurality of free DNA molecules. For example, gel electrophoresis, filtration, size-selective precipitation, or hybridization can be used to measure the lengths of the plurality of free DNA molecules. Additionally or alternatively, the lengths of the plurality of free DNA fragments can be measured using sequence reads. For example, sequence reads can be obtained from sequencing of a plurality of free DNA molecules from a biological sample (e.g., massively parallel sequencing, single molecule real-time sequencing, nanopore sequencing). Then, to measure the length of a free DNA molecule, the number of nucleotides can be counted for each sequence read. Those skilled in the art will understand the various sequencing techniques that can be used. As part of the sequencing, some sequence reads can correspond to cellular nucleic acids. In some cases, DNA fragments in a urine sample having a length less than a predetermined length threshold (e.g., 80 bps) can be enriched.

[0267] As described herein, a statistically significant number of free DNA molecules can be analyzed.

[0268] At block 3404, statistical values for a collection of cell-free DNA molecules are determined. The statistical values can be determined based on the lengths of multiple cell-free DNA molecules. The lengths can form a length distribution. Various statistical values can be used, for example, the mean, average, median, or mode of the length distribution can be used. As another example, the ratio of cfDNA in a first length range to a second length range can be used, where the length ranges are different but can overlap. The second length range can be all lengths, i.e., all cfDNA molecules.

[0269] In one example, the relative amount of renal DNA in a urine sample (an example of a statistical value) can be characterized by DNA fragments having a length less than 80 base pairs. If the permeability of the GBM is disrupted in some subjects, the relative amount of shorter DNA fragments in the urine sample can increase or decrease. If such a change relative to the normal amount exceeds a threshold (reference value), it can be determined that the subject has a disease or other abnormal condition (such as nephrotic syndrome, glomerulonephritis) affecting renal permeability.

[0270] For example, the statistical value can be the length ratio of a first amount of cell-free DNA molecules having a length less than a length threshold (e.g., 80 bp) to a second amount corresponding to multiple cell-free DNA molecules. As an example, the length threshold can be 40 base pairs, 50 base pairs, 60 base pairs, 70 base pairs, 80 base pairs, 90 base pairs, 100 base pairs, 110 base pairs, 120 base pairs, 130 base pairs, 140 base pairs, 150 base pairs, or 160 base pairs, which can be used in any implementation using a length threshold as described herein.

[0271] In some cases, determining the statistical value includes the ratio of a collection of cell-free DNA molecules having lengths within a length range to multiple cell-free DNA molecules in a urine sample. The length range can have a lower limit and an upper limit, for example, selected from lower limits of 0, 5, 10, 15, 20, 30, 35, 40, 45, 50, 55, or 60 bases and any one of 40, 50, 60, 70, 80, 90, 100, 110, 120, 130, 140, 150, or 160 bases.

[0272] At block 3406, the statistical value is compared to a reference value. The reference value can correspond to another statistical value determined based on the measured lengths of cell-free DNA molecules from one or more reference samples, where the one or more reference samples are associated with known classifications of renal abnormalities. For example, the reference value can be determined based on the lengths of cell-free DNA molecules in healthy urine samples. In some cases, the reference value is a calibration value or is determined from the calibration value of a calibration (training) sample.

[0273] At block 3408, a classification of a subject with a kidney abnormality is determined based on a comparison. In some embodiments, comparing a statistical value to a reference value includes: (1) determining whether the statistical value differs from the reference value by at least a threshold amount or whether the difference is less than the threshold amount; (2) determining whether the statistical value is less than the reference value by at least the threshold amount; or (3) determining whether the statistical value is greater than the reference value by at least the threshold amount. Kidney abnormalities can include renal cell carcinoma (RCC), nephrotic syndrome, glomerulonephritis, Fabry disease, cystinosis, IgA nephropathy, IgM nephropathy, lupus nephritis, atypical hemolytic uremic syndrome (aHUS), polycystic kidney disease (PKD), Alport syndrome, interstitial nephritis, proteinuria, chronic kidney disease, acute kidney injury, and the like. In some cases, the classification of a subject with a kidney abnormality includes an increased permeability level associated with the glomerular basement membrane of the kidney.

[0274] The classification of kidney abnormalities can be determined using machine learning trained using a training dataset. The training dataset can include training samples. The training samples can be associated with known classifications of kidney abnormalities. In another example, a machine learning model can be used to perform the comparison to the reference value. The machine learning model can be applied to the statistical value to generate a classification of the kidney abnormality. Machine learning models can include, but are not limited to, convolutional neural networks (CNNs), linear regression, logistic regression, deep recurrent neural networks (e.g., fully connected recurrent neural networks (RNNs), gated recurrent units (GRUs), long short-term memory (LSTMs)), transformer-based methods (e.g., XLNet, BERT, XLM, RoBERTa), Bayesian classifiers, hidden Markov models (HMMs), linear discriminant analysis (LDA), k-means clustering, density-based spatial clustering of applications with noise (DBSCAN), random forest algorithms, adaptive boosting (AdaBoost), extreme gradient boosting (XGBoost), support vector machines (SVMs), or composite models including one or more of the models presented above. C. Classification Based on Urine cfDNA Concentration

[0275] We also evaluated the differences in urine cfDNA concentration between healthy pregnant women and preeclamptic pregnant women. Since the cfDNA concentration in urine samples depends on the hydration status of the subject, the urine cfDNA concentration was corrected (normalized).

[0276] In some embodiments, the correction of urine concentration can be performed using creatinine. For example, the amount of urine DNA (e.g., measured by mass per volume, such as ng / mL) can be corrected by the amount of creatinine (e.g., mmol). In one embodiment, the correction value is calculated by dividing the urine cfDNA concentration per milliliter of urine sample (e.g., determined by Qubit assay) by the creatinine concentration, expressed as nanograms per milliliter of cfDNA per millimole of creatinine (ng / ml / mmol Cr). Myocytes produce creatinine at a constant rate, and all creatinine filtered by the glomerulus is excreted in the urine. Thus, the representation of urine cfDNA concentration based on millimoles of creatinine will minimize the variation in urine cfDNA concentration caused by differences in the hydration status of the subject.

[0277] Figure 35 Analysis of urine cfDNA concentration of renal DNA in patients with proteinuria and preeclampsia is shown, respectively.

[0278] Box plot 5310 shows the urine cfDNA concentration of healthy controls and patients with proteinuria. We observed a higher urine cfDNA concentration in patients with proteinuria compared to healthy controls (Mann-Whitney U test, P value = 0.0015).

[0279] Box plot 3520 shows the urine cfDNA concentration of healthy pregnant women and pregnant women with preeclampsia. We observed a higher urine cfDNA concentration in pregnant women with preeclampsia compared to healthy pregnant women (Mann-Whitney U test, P value = 0.0190).

[0280] Figure 36 is a flowchart of a method 3600 for determining a classification of renal abnormalities based on urine cell-free DNA concentration according to some embodiments. Method 3600 can detect renal abnormalities using a urine sample of a subject, wherein the urine sample includes cell-free DNA molecules.

[0281] In block 3602, a first amount of cell-free DNA molecules in the urine sample is determined. As an example, a fluorometer, spectrophotometer, PCR, or sequencing measurement can be used to determine the first amount. The first amount can be filtered to obtain cfDNA molecules that meet one or more criteria. For example, the cfDNA can have a specified length, such as greater than a length threshold, which can be 40 - 200 bp. Examples of length thresholds are provided herein and include 40, 50, 60, 70, 80, 90, 100, 110, 120, 130, 140, 150, 160, 170, 180, 190, or 200 bp. The specified length can be a length range with an upper and lower limit, only a lower limit, or only an upper limit.

[0282] In block 3604, an initial concentration is determined using a first amount and volume of a urine sample. When using length as a criterion, the initial concentration can be the proportion of free DNA molecules in the urine sample within a specified range. The specified range can be greater than a size threshold, as described above, for example.

[0283] In block 3606, a corrected concentration is determined using a second amount of a specific compound in the urine sample. The compound can be a digestive waste and thus a compound that occurs naturally in the subject. As an example, the specific compound can be creatinine. Creatinine is a waste product from the digestion of proteins in food and the normal breakdown of muscle tissue (such as creatine).

[0284] In block 3608, the corrected concentration is compared with a reference value. The reference value can be determined from one or more reference subjects of known classification, such as the presence or absence of a kidney abnormality or the specific severity of a kidney abnormality.

[0285] In block 3610, a classification of a subject with a kidney abnormality is determined based on the comparison. Examples of kidney abnormalities are provided herein, including preeclampsia and proteinuria.

[0286] Additional details of an example manner for determining the first amount are provided. The NanoDrop spectrophotometer is based on the principle that nucleic acids (i.e., DNA and RNA) absorb ultraviolet light and have a peak at a wavelength of 260 nanometers (nm). A light detector measures the light passing through the sample. The more light absorbed by the nucleic acids, the less light will strike the photodetector, resulting in a higher optical density (OD) and a higher nucleic acid concentration in the sample.

[0287] The Qubit fluorometer quantifies DNA concentration by detecting a fluorescent dye in the sample. The fluorescent dye specific to the DNA substrate exhibits extremely low fluorescence before binding to the DNA target. When bound to DNA, the dye molecule increases fluorescence by several orders of magnitude through intercalation between DNA bases.

[0288] The quantitative polymerase chain reaction (qPCR) assay quantifies DNA concentration by detecting the fluorescent signal of DNA products during real-time PCR. qPCR monitors the amplification of target DNA molecules during PCR using a fluorescent dye or a DNA probe labeled with a fluorescent reporter. As a result, the amount of amplified product is correlated with the fluorescence intensity.

[0289] Digital polymerase chain reaction (dPCR) assays involve partitioning a PCR solution into tens of thousands of nanoliter-sized droplets, in which individual PCR reactions of single DNA molecules occur. DNA probes with fluorescent reporters will facilitate the detection of target DNA in the droplets, and the fraction of droplets containing the target DNA can be converted into a DNA quantity. Further details can be found in the following three publications, all of which used NanoDrop, Qubit, and qPCR for DNA concentration determination: Simbolo et al., PLOS ONE 2013;8:e62692 Heydt et al., PLOS ONE 2014;9:e104566 Ponti et al., Clinica Chimica Acta 2018;479:14-19. Further details of dPCR for DNA concentration determination can be found in Gai et al., Clin.Chem.2018;64:1239-1249. D. Technical Comparison

[0290] Figure 37 ROC analysis showing the separation of patients with proteinuria and preeclampsia from healthy controls using the fragmentomics signature of renal DNA. Figure 37 ROC analysis showing the use of the above techniques.

[0291] ROC 3710 showed AUCs of 0.76, 0.68, 0.73, and 0.75 in differentiating patients with proteinuria from healthy controls using the O / E of urine cfDNA concentration, length (i.e., >80 bp), and OCR (blood-specific DHS). By combining these fragmentomics signatures with the support vector machine (SVM) method, the AUC could be further increased to 0.85 in differentiating proteinuria.

[0292] We further performed ROC analysis on these samples, in which the AUCs were 0.84, 0.84, and 0.85 in differentiating pregnant women with preeclampsia from healthy pregnant women using the O / E of urine DNA concentration, cfDNA length (i.e., >80 bp), and OCR (placenta-specific DHS), respectively. When combining these three fragmentomics signatures using SVM, an improved performance (AUC: 0.93) in the differentiation between preeclampsia and healthy subjects could be observed.

[0293] SVM provides separation of samples in a higher dimension. The number of features input into the SVM will provide the number of dimensions in the SVM. In the example above, we used three features and thus three dimensions. Additional features can be used, which will result in more than three dimensions. VIII. F-profile of cfDNA and Nuclease Activity

[0294] Different types of cell-free DNA cleavage are associated with different fragmentation processes, including enzymatic and non-enzymatic breaks. When using one end motif or several of the top-ranked end motifs each time, there are techniques that focus on one specific nuclease activity. This approach may be effective, but may not provide a comprehensive view of the nuclease activities occurring in a given sample (e.g., plasma sample, urine sample).

[0295] To address the above deficiencies, the putative relative contributions associated with different types of cell-free DNA cleavage can be used to simultaneously evaluate multiple nuclease activities or other fragmentation processes. For example, the relative frequencies of DNA molecules corresponding to 256 end motifs can be determined for subjects with a known disease diagnosis (e.g., HCC). The relative frequencies of the DNA molecules can be decomposed into a set of "F-profiles" that identify the relationships of the end sequences (e.g., 1 - 30 bases) of the cell-free DNA fragments (also known as DNA fragments) in the sample. Then the set of F-profiles can be used for the deconvolution of the relative frequencies of DNA molecules obtained from another subject to predict the fraction of clinically relevant DNA molecules, the classification of diseases, etc.

[0296] Figure 38 Figure 3800 is shown, which identifies the ranking of the frequencies of certain end motifs present in urine cell-free DNA molecules. Figure 38 Corresponding to Figure 9 . As Figure 38 shown, fetal-specific transrenal DNA mainly consists of fragments with a C terminus, which is usually associated with the DNASE1L3 cleavage preference. The total non-transrenal DNA mainly consists of fragments with a T terminus, which is usually associated with the DNASE1 cleavage preference.

[0297] Although focusing on certain end motifs can be beneficial for determining fetal DNA (for example), Figure 3800 shows additional end motif information that can provide further insights: the relative frequencies of DNA molecules spanning most of the 256 end motifs are different between fetal-specific DNA and total DNA. Therefore, it is advantageous to incorporate the relative frequencies of DNA molecules spanning all 256 end motifs to determine the fetal DNA fraction or to determine the disease classification of a subject. A. End motif spectrum characteristics across different murine samples

[0298] Figure 39 A set of Figures 3900 is shown that identify the end motif spectra of the observed murine plasma and urine cell-free DNA molecules. As Figure 39As shown, the frequencies of 256 4-mer end motifs in plasma and urine cell-free DNA of mice with different nuclease knockout genotypes are organized alphabetically to form an end motif spectrum. Motifs starting with adenine (A), cytosine (C), guanine (G), and thymine (T) are highlighted in blue, red, green, and yellow, respectively.

[0299] We observed certain distinct patterns in the end motif spectra of different mice. Figures 3902, 3904, 3906, and 3908 show the observed end motif frequencies in plasma cell-free DNA from WT mice, Dnase1l3 - / - mice, Dnase1 - / - mice, and Dffb - / - mice, respectively. The observed end motif frequencies in urine cell-free DNA from WT mice, Dnase1l3 - / - mice, and Dnase1 - / - mice are shown in Charts 3910, 3912, and 3914, respectively.

[0300] Compared to WT mice, plasma cell-free DNA from Dnase1l3 - / - mice shows periodic spikes in the end motif spectrum, typically among those end motifs with A, C, and G ends. For urine cell-free DNA from WT mice, the abundance of motifs with T ends is significantly elevated compared to Dnase1 - / - mice (P < 0.0001, Mann-Whitney U test). Although it is visually difficult to discern differences when comparing plasma cell-free DNA from WT mice to Dnase1 - / - mice, or urine cell-free DNA from WT mice to Dnase1l3 - / - mice, we hypothesize that subtle differences in the 256-dimensional end motif spectrum can be delineated when using a reference spectrum, such as by factorization into reference spectra. In some embodiments, non-negative matrix factorization (NMF) is used to consider the 256 motifs as a whole, rather than focusing on one or a few specific motif types.

[0301] The end motif spectrum can be a K-mer, where K can have various values, such as 1, 2, 3, 4, 5, 6, or more. As Figure 39 shown, K is 4. B. NMF for determining the F-spectrum of urine cfDNA

[0302] Figure 40Shows a schematic workflow for the exemplary nuclease usage level analysis of cell-free DNA molecules. Ninety-three murine cell-free DNA samples were sequenced, including 60 plasma cell-free DNA samples and 33 urine cell-free DNA samples. From 27 wild-type (WT) mice, 10 mice with a deletion of the Dnase1 gene (Dnase1 - / - ), 18 mice with a deletion of the Dnase1l3 gene (Dnase1l3 - / - ), and 5 mice with a deletion of the Dffb gene (Dffb - / - ), murine plasma cell-free DNA samples were collected, with a median of 50 million paired-end reads (range: 16 - 243 million). Additionally, whole-genome sequencing data of murine urine cell-free DNA samples were obtained from 14 WT mice, 10 Dnase1 - / - mice, and 9 Dnase1l3 - / - mice (median of paired-end reads: 43 million; range: 2 - 134 million).

[0303] In block 4002, 93 murine cell-free DNA samples, including WT mice and nuclease-deficient mice (e.g., Dnase1l3 - / - , Dnase1 - / - , Dffb - / - ), were assayed for the terminal 4 nucleotides at each end of the 5’ fragment (i.e., 4-mer end motif; n = 256).

[0304] For each murine sample, 256 4-mer end motifs of cell-free DNA molecules were then used to infer their respective nuclease usage levels.

[0305] In block 4004, 6 classes of reference end motif profiles, called F profiles, were determined from the cell-free DNA molecules of each murine sample. In some embodiments, non-negative matrix factorization (NMF) analysis of the relative frequencies of DNA molecules terminated with 4-mer end motifs was performed to determine potentially distinct types of cell-free DNA cleavage.

[0306] We applied non-negative matrix factorization (NMF) (Daniel et al. Nature 1999; 401:788-791; Stein-O’Brien et al. Trends Genet. 2018; 34:790–805) analysis to decompose the relative frequencies of cell-free DNA molecules into multiple F-profiles. A total of 93 murine cell-free DNA samples with DNA nucleases knocked out with different genotypes were used for this NMF analysis, including 60 plasma cell-free DNA samples and 33 urine cell-free DNA samples. After obtaining the terminal motif frequencies, a data matrix (M) was constructed in such a way that each row represents a cell-free DNA sample (a total of 93 murine cell-free DNA samples), and each column represents the type of terminal motif (a total of 256 terminal motifs), thus having a dimension of 93×256. NMF analysis was performed on the data matrix to obtain two matrices W and F. The mathematical relationship among M, W, and F is shown as follows: M = WF.

[0307] M is the result of the product of W and F, where W is the relative weight of each F-profile in a 93×n matrix, where n corresponds to the number of F-profiles. F represents the F-profiles in an n×256 matrix. W and F were determined by minimizing the following objective function: ‖M - WF‖, subject to W≥0 F≥0.

[0308] Singular value decomposition (SVD) was used to initialize the process of NMF. This factorization analysis was implemented in the Python language by using sklearn.decomposition.NMF (v1.1.1) (Pedregosa et al. J. Mach. Learn. Res. 2011; 12:2825–2830).

[0309] To estimate the optimal number of F-profiles, a 5-fold cross-validation pre-analysis was performed. This factorization analysis can generate many different types of cell-free DNA cleavage. In this example, six F-profiles (i.e., F-profile I, II, III, IV, V, and VI) were determined by considering the trade-off between the reproducibility of the factorization components and the value of the objective function (i.e., the terminal motif spectrum reconstruction error). The number of different types of cell-free DNA cleavage can be but is not limited to 2, 3, 4, 5, 6, 7, 8, 9, 10, 15, 20, 30, 40, 50, 100, etc., with corresponding numbers of reference terminal motif spectra. In Figure 40 , F-profile I, II, and III can be related to the cleavage preferences of DNASE1, DNASE1L3, and DFFB, respectively.

[0310] In block 4006, the nuclease usage analysis framework learned from murine cfDNA can be extrapolated to human cfDNA analysis to inform the proportional contributions of different nuclease activities in murine cfDNA samples and human cfDNA samples. The observed end motif spectra can be reconstructed by iteratively adjusting the proportional contributions of each F-spectrum. In other words, using the F-spectra generated from murine cfDNA, the proportional contributions of the F-spectra of any cfDNA sample are inferred. In some embodiments, this inferred proportional contribution of the F-spectra is used to reflect the nuclease activity or nuclease usage level in any cfDNA sample.

[0311] Additionally or alternatively, this inferred proportional contribution of the F-spectra can be used to reflect other types of fragmentation that a patient may be involved in, such as but not limited to oxidative stress-induced DNA damage, drug treatment-induced DNA damage, radiation-induced DNA damage, etc. The presence, absence, and alteration of F-spectrum contributions can imply that a patient has a disease or is at risk of developing a disease. In some embodiments, another mathematical algorithm is used for factorization, such as but not limited to principal component analysis (PCA), t-distributed stochastic neighbor embedding, uniform manifold approximation and projection, etc. C. F-Spectra in Plasma and Urine Samples

[0312] Figure 41 Figure 4100 is shown, which identifies the proportional contributions of each F-spectrum (i.e., nuclease usage level) inferred from murine cfDNA samples with different knockout genotypes using NMF analysis. Each F-spectrum can indicate a pattern of the relative frequencies of cfDNA molecules spanning 256 end motifs. Six F-spectra can be used as labels for the end motif frequencies of cfDNA molecules in the corresponding samples. The proportional contribution of each F-spectrum in each cfDNA sample can be determined when the minimum error is reached between the observed end motif spectrum and the sum of the F-spectra weighted by their proportional contributions.

[0313] As Figure 41 shown, the proportional contributions of F-spectra in murine samples tend to be similar based on their respective nuclease activity levels. For example, the WT sample shows a significant contribution of F-spectrum I, while the Dnase1l3 - / - sample shows little contribution of F-spectrum I. In another example, the DnaseI - / - sample shows a significantly less contribution of F-spectrum II, while the Dffb - / - sample shows little contribution of F-spectra III, IV, and V. Using the similar pattern of F-spectrum contributions, the nuclease activity level of another sample from a test subject can be evaluated. D. Fragmentomics Characteristics of F-Spectra

[0314] As described above, these six F-spectra are associated with possible DNA nuclease activities. To demonstrate this feasibility, we investigated the typical terminal motifs in the F-spectra and measured the changes in their proportional contributions when specific nuclease activities were depleted or enhanced.

[0315] Figure 42 A set of plots showing six F-spectra (A-F) inferred from mouse plasma and urine cell-free DNA using NMF analysis is presented. Each of the F-spectra 4202-4212 contains a 4-mer terminal motif spectrum, 1-mer terminal motif frequencies, and sequence preferences at each position of the 4-mer terminal motifs. F-spectrum I 4202 shows a predominance of the C-terminal motif (55%), characterized by a "CC"-starting motif, which is consistent with the DNASE1L3 cleavage properties confirmed in our previous studies (Serpas et al. Proc Natl Acad Sci USA. 2019;116:641-649; Jiang et al. Cancer Discov. 2020;10:664–73). We observed that the contribution of F-spectrum I in plasma cell-free DNA of Dnase1l3 - / - mice was significantly lower than that in WT mice (median: 2.7% vs 35.4%; range: 0.0 - 4.6% vs 19.5 - 47.9% (P < 0.0001, Mann-Whitney U test)). Therefore, F-spectrum I is considered a DNASE1L3-related F-spectrum, which can be used to reflect the nuclease usage level of DNASE1L3.

[0316] F-spectrum II 4204 exhibits a major preference for the T-terminal motif (51%), with a preference for "TG"-starting motifs. In WT mice, the contribution of F-spectrum II in urine cell-free DNA was significantly higher than that in plasma cell-free DNA (median: 43.4% vs 11.6%; range: 31.8 - 50.1% vs 0.0 - 22.1% (P < 0.0001, Mann-Whitney U test)). Notably, for WT mice, the DNASE1 activity in urine is much higher than that in plasma (Chen et al. PLOS Genet. 2022;18:e1010262)). In addition, compared with their WT counterparts, the median contribution of F-spectrum II in plasma and urine cell-free DNA of Dnase1 - / - mice was reduced by approximately 8-fold. Therefore, it is inferred that F-spectrum II is related to DNASE1 activity.

[0317] F-spectrum III 4206 includes a substantial proportion of A-terminal motifs (40%), which are characterized by a preference for C and T nucleotides at the third and fourth positions of the 4-mer motif, respectively, in the 5’ to 3’ direction. Compared to its counterpart in WT mice (median: 10.1%; range: 0.0 - 26.9%) (P = 0.0004, Mann-Whitney U test), the contribution of F-spectrum III to plasma-free DNA in Dffb - / - mice is significantly reduced (median: 0.0%; range: 0.0 - 0.5%). Thus, F-spectrum III is thought to be associated with DFFB activity.

[0318] Although F-spectrum IV 4208 exhibits a high C-terminal preference (50%) (which is to some extent a hint of F-spectrum I), it has several distinct features, e.g., the absence of a CC-terminal preference. F-spectrum IV also shows a "G" base preference at the second, third, and fourth positions in the 4-mer motif. F-spectrum V 4210 exhibits a strong G-terminal preference (50%). These results suggest that F-spectrum IV and V are not directly attributed to the known nucleases involved in free DNA fragmentation, meaning that some other enzymatic and / or non-enzymatic processes may play a role in the free DNA fragmentation process. In addition, F-spectrum VI 4212 shows a relatively uniform distribution over 256 motifs with no obvious sequence preference, suggesting a possibility that other DNA nucleases or other factors may also cause non-specific cleavage.

[0319] Figure 43 Box plot 4300 is shown, which identifies the proportional contribution of F-spectrum I on different types of samples according to some embodiments. The x-axis shows free DNA molecules obtained from different types of samples, including wild-type (WT) murine samples (n = 27), Dnase1 - / - and Dnase1l3 - / - murine samples (n = 2), Dnase1l3 - / - murine samples (n = 18), and Dnase1l3 - / - and Cd401g - / - murine samples (n = 5). In addition, the x-axis further shows various types of pregnant murine samples, including WT murine samples (n = 2), Dnase1l3 - / - maternal samples in which fetal DNA shows Dnase1l3 + / - (n = 4)) and Dnase1l3 - / - maternal samples in which fetal DNA shows Dnase1l3 - / - (n = 3)). The proportional contribution of F-spectrum I is determined relative to the contributions from other F-spectra II - VI.

[0320] As Figure 43 shown, for murine samples with DNASE1L3 knocked out, the proportional contribution of F-spectrum I was significantly reduced. Additionally, for pregnant samples, a significant reduction in the contribution of F-spectrum I also seemed to be similar. Interestingly, the F-spectrum I contribution of fetal Dnase1l3 + / - samples was slightly increased compared to that of fetal Dnase1l3 - / - samples. Based on the proportional contribution of F-spectrum I, it can be determined that F-spectrum I corresponds to the tag related to DNASE1L3.

[0321] Figure 44 shows the relative frequency 4400 of cell-free DNA molecules spanning 256 end motifs of F-spectrum I according to some embodiments. Figure 4410 shows the end motif spectrum using 4-mer as the end motif. Figure 4420 shows the end motif spectrum using 1-mer (single nucleotide) as the end motif. As Figure 44 shown in the figure, F-spectrum I has a cleavage preference for the C-terminus, followed by the T-terminus, and then the A-terminus and G-terminus. This cleavage preference is basically similar to that of DNASE1L3, which supports the Figure 43 findings shown.

[0322] Figure 45 shows box plot 4500, which identifies the proportional contribution of F-spectrum II across different types of samples according to some embodiments. The x-axis shows cell-free DNA molecules obtained from different types of murine samples, including wild-type (WT) plasma samples (n = 27), WT urine samples (n = 14), Dnase1 - / - plasma samples (n = 10), Dnase1 / Dnase1l3 double knockout plasma samples (n = 2), and Dnase1 - / - urine samples (n = 10). The proportional contribution of F-spectrum II is determined relative to the contributions from other F-spectra I and III-VI.

[0323] As Figure 45 shown, the proportional contribution of F-spectrum II in urine samples is significantly higher compared to plasma samples. Additionally, the proportional contribution of F-spectrum II does not show a significant change among different types of plasma samples. However, the proportional contribution of F-spectrum II in Dnase1 - / - urine samples shows a significant reduction compared to WT urine samples. Based on the proportional contribution of F-spectrum II, it can be determined that F-spectrum II corresponds to the tag related to DNASE1. In fact, DNase1 nuclease activity seems to be more active in urine samples, which is related to the high F-spectrum II contribution of WT urine samples.

[0324] Figure 46Shows the relative frequencies 4600 of cell-free DNA molecules across 256 end motifs of the F-spectrum II, according to some embodiments. As Figure 46 shown, the F-spectrum II has a cleavage preference for the T end, followed by the C end, and the A and G ends. This cleavage preference is substantially similar to that of DNASE1, which supports the Figure 45 findings shown.

[0325] Figure 47 Shows a box plot 4700 that identifies the proportional contributions of the F-spectrum III across different types of samples, according to some embodiments. The x-axis shows cell-free DNA molecules obtained from samples including wild-type (WT) samples (n = 27) and Dffb - / - samples (n = 5). The proportional contribution of the F-spectrum III is determined relative to the contributions from the other F-spectra I, II, and IV-VI. As Figure 47 shown, the proportional contribution of the F-spectrum III in Dffb - / - samples is significantly reduced relative to the WT samples. Based on the proportional contribution of the F-spectrum III, it can be determined that the F-spectrum III corresponds to a tag associated with DFFB.

[0326] Figure 48 Shows the relative frequencies 4800 of cell-free DNA molecules across 256 end motifs of the F-spectrum III, according to some embodiments. As Figure 48 shown, the F-spectrum III has a cleavage preference for the A end, followed by the G end, and the C and T ends. This cleavage preference is substantially similar to that of DFFB, which supports the Figure 47 findings shown.

[0327] In addition to F-spectra I-III, we also deconvolve F-spectra IV-VI.

[0328] Figure 49 Shows the relative frequencies 4900 of cell-free DNA molecules across 256 end motifs of the F-spectrum VI-VI, according to some embodiments. Each of the F-spectra IV-VI shows its own cleavage preference. For example, F-spectrum IV 4902 shows a cleavage preference for the C end, F-spectrum V 4904 shows a cleavage preference for the G end, and F-spectrum VI 4906 does not show a specific cleavage preference. Based on the above, it can be seen that F-spectrum VI may be related to a cleavage pattern that is not caused by the specific nucleases we studied, but is related to other types of fragmentation factors. E. Analysis of F-spectra of different samples from human subjects

[0329] We then explored whether the murine F-profile of DNASE-mediated cell-free DNA cleavage could be applied to human subjects.

[0330] Figure 50 Figure 5000 shows a schematic of comparing the end motif profiles of human subjects to a reference F-profile determined based on murine samples, according to some embodiments. To make the motif patterns directly comparable between humans and mice, the frequencies of 4-mer end motifs associated with human and murine cell-free DNA can be normalized by the genomic background of the human and murine genomes, respectively. For example, the expected 4-mer end motif frequencies can be used in the normalization step, where the expected end motif frequencies are determined by simulating 4-mer end motifs from the reference genome using a 4-bp sliding window spanning each chromosome. The normalized end motif frequency is calculated as the ratio of the observed frequency to the expected frequency, and then divided by the sum of all 256 normalized motif frequencies. The total normalized end motif frequency can equal 100%. The end motif frequencies referred to in this NMF-based nuclease usage analysis are called normalized end motif frequencies.

[0331] Once normalization is complete, the proportional contributions of the F-profile can be determined for the normalized end frequencies of the human sample. The proportional contributions can be determined by applying deconvolution to the normalized end frequencies. For example, a data matrix M of dimension W by F can be used, where: (i) M can represent the normalized end frequencies on 256 end motifs of a given biological sample; (ii) F can represent the end frequencies of the reference F-profile obtained from murine samples; and (iii) W can represent the relative weights corresponding to the proportional contributions of each F-profile.

[0332] The F end frequencies can be determined based on the proportion of cell-free DNA molecules in a set of reference F-profiles. The proportional contributions can be determined by solving for the W relative weights using non-negative least squares (NNLS) on the values from the data matrix M and the reference F-profile. The proportional contributions determined using deconvolution can be used to identify the extent of nuclease activity levels in certain human biological samples (e.g., a relatively reduced contribution of F-profile I).

[0333] Figure 51 Figure 5100 shows the proportional contributions of the F-profile across plasma and urine samples of human subjects, according to some embodiments. First, the Figure 51 deconvolution process in Figure 51 is applied to plasma and urine samples from human subjects. As

[0334] Figure 51 The data shown is consistent with the experiments, namely that DNASE1L3 makes a major contribution to the fragmentation pattern of free DNA molecules in plasma samples, while DNASE1 makes a major contribution to the fragmentation pattern of free DNA molecules in urine samples. Thus, it has been shown that the deconvolution process using the reference F-spectrum from murine samples can be effectively used to identify the fragmentation cleavage patterns of free DNA molecules in human samples.

[0335] Figure 52 Shows the proportional contributions 5200 of the F-spectrum across normal and DNASE1L3-deficient samples of a human subject, according to some embodiments. Figure 52 The deconvolution process in Figure 52 is applied to normal and DNASE1L3-deficient samples from a human subject. As Figure 52 shown, plasma samples from control subjects include a relatively high contribution (greater than about 40%) of F-spectrum I. In contrast, DNASE1L3-deficient samples include a significantly lower contribution (less than about 15%) of F-spectrum I.

[0336] Figure 53 Shows the proportional contributions 5300 of the F-spectrum across urine samples of pregnant human subjects, according to some embodiments. Figure 42 The deconvolution process in Figure 53 is applied to samples from pregnant human subjects. As Figure 53 shown, urine samples include a relatively high contribution of F-spectrum II that has previously been associated with DNASE1 activity. Figure 38 The data shown is consistent with the experiments, namely that DNASE1 makes a major contribution to the fragmentation pattern of free DNA molecules in urine samples. Additionally, a higher proportional contribution of F-spectrum IV may indicate a relatively high number of free DNA fragments with C termini. This observation is consistent with the graph data in

[0337] Figure 54 Shows a flow chart of a method 5400 for determining a nuclease activity classification based on the F-spectrum of free DNA molecules, according to some embodiments. Exemplary biological samples can be free samples comprising free DNA, such as blood, plasma, serum, urine, and saliva.

[0338] In block 5402, a set of reference F-spectra is stored. For each nucleotide in the set of nucleotides, each reference F-spectrum of the set identifies the proportion of free DNA molecules that terminate at that nucleotide. In addition, each reference F-spectrum is associated with a type of fragmentation factor. The type of fragmentation factor identifies a specific enzyme (e.g., DNASE1L3, DNASE1), protein (e.g., DFFB), or other biological component or process that causes fragmentation of free DNA molecules. In some cases, the set of reference F-spectra of the set includes 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 15, 20, 25, 30, 40, 45, 50, or more than 50 F-spectra. For example, the set of reference F-spectra of the set can include six F-spectra I-VI.

[0339] Each reference F-spectrum and the sample end motif spectrum can have a separate proportion for each K-mer end motif in the set of K-mer end motifs. For example, Figure 39 Figure 4420 shows the spectrum where K = 4, resulting in 256 different proportion values. Figure 44 Figure 4420 shows the spectrum where K = 1, resulting in 4 different proportion values. Thus, each reference F-spectrum in the set of reference F-spectra can specify the proportion of free DNA molecules that terminate at each K-mer end motif in the set of K-mer end motifs, where K is 1 or 2 or greater.

[0340] In some cases, one or more reference samples are used to determine the set of reference F-spectra. The reference samples can be obtained from non-human subjects (e.g., murine samples), and the genetic disorder classification of the non-human subjects is known (e.g., WT, DNASE1L3 - / - , DNASE1 - / - ). To determine the reference F-spectra of the set of reference spectra, a factorization algorithm (e.g., NMF, PCA) is used to factorize the relative frequencies of free DNA molecules in the reference samples into several F-spectra. For example, reference free DNA samples with DNA nuclease knockouts of different genotypes are selected. After obtaining the end motif frequencies of the reference samples, a data matrix (M) is constructed in such a way that each row indicates a free DNA sample (e.g., a total of 93 murine free DNA samples) and each column represents the type of end motif (e.g., a total of 256 4-mer end motifs), resulting in a dimension of 93×256. Then NMF analysis can be performed on the data matrix to obtain two matrices W and F. M = WF.

[0341] M is the result of the product of W and F, where W is the relative weight of each F-spectrum in a 93×n matrix, and n corresponds to the number of F-spectra. F represents the F-spectra in an n×256 matrix. W and F are determined by minimizing the following objective function: ‖M-WF‖, subject to W≥0 F≥0.

[0342] In block 5404, multiple free DNA molecules from a biological sample are analyzed to obtain sequence reads. The sequence reads include end sequences corresponding to the ends of the multiple free DNA molecules. The sequence reads can include end sequences corresponding to the ends of multiple free DNA fragments. As an example, the sequence reads can be obtained using sequencing or probe-based techniques, either of which can include enrichment, such as by amplification or capture probes.

[0343] Sequencing can be performed in a variety of ways, for example, using massively parallel sequencing or next-generation sequencing, using single-molecule sequencing, and / or using double-stranded or single-stranded DNA sequencing library preparation protocols. Those skilled in the art will understand the various sequencing techniques that can be used. As part of the sequencing, some sequence reads can correspond to cellular nucleic acids.

[0344] Sequencing can be targeted sequencing as described herein. For example, DNA fragments from a specific region of a biological sample can be enriched. The enrichment can include using capture probes that bind to a portion or the entire genome as defined, for example, by a reference genome.

[0345] A statistically significant number of free DNA molecules can be analyzed to provide an accurate determination of the concentration fraction. In some embodiments, at least 1,000 free DNA molecules are analyzed. In other embodiments, at least 10,000 or 50,000 or 100,000 or 500,000 or 1,000,000 or 5,000,000 free DNA molecules or more can be analyzed.

[0346] In block 5406, a sample end motif profile of a subject is determined by determining the proportion of multiple free DNA molecules that terminate with each nucleotide in a set of nucleotides based on the end sequences. The sample end motif map identifies the relative frequencies of multiple end motifs corresponding to the end sequences of the multiple free DNA molecules. Multiple end motifs can correspond to all possible combinations of N base positions. For example, if multiple end motifs correspond to 4-mers, the multiple end motifs of the sample end motif profile can include 256 combinations of 4-mers (e.g., CCCA, TTCC).

[0347] To determine the sample end motif spectrum, for each of a plurality of cell-free DNA molecules, the end motif of each of one or more end sequences of the cell-free DNA molecule is determined. The end motif can include N base positions (e.g., 1, 2, 3, 4, 5, 6, etc.). As an example, the end motif can be determined by analyzing sequence reads at the ends corresponding to the ends of the DNA molecule, correlating signals to specific motifs (e.g., when using probes), and / or aligning the sequence reads to a reference genome.

[0348] For example, after sequencing by a sequencing device, the sequence reads can be received by a computer system, which can be communicatively coupled to the sequencing device performing the sequencing, e.g., via wired or wireless communication or via a removable memory device. In some embodiments, the receiving includes one or more sequence reads from both ends of a nucleic acid fragment. The location of the DNA molecule can be determined by mapping (aligning) one or more sequence reads of the DNA molecule to the corresponding part of the human genome, e.g., a specific region. Additionally or alternatively, a specific probe (e.g., after PCR or other amplification) can indicate the location or a specific end motif, e.g., via a specific fluorescence color. The identification can be that the cell-free DNA molecule corresponds to one of a plurality of end motifs.

[0349] Then, the relative frequencies of a plurality of end motifs corresponding to the end sequences of the plurality of cell-free DNA molecules are determined to determine the sample end motif spectrum of the subject. The relative frequency of a sequence motif can provide the proportion of the plurality of cell-free DNA molecules having an end sequence corresponding to the sequence motif.

[0350] At block 5408, the proportional contributions of a set of reference F spectra are determined, the proportional aggregation of which provides the sample end motif spectrum. The sum of the proportional contributions of the set of reference F spectra is one. The proportional contributions can be determined by applying deconvolution to the sample end motif spectrum of the subject. For example, a data matrix M of dimension W by F can be used, where: (i) M can represent the normalized end frequencies of 256 end motifs spanning the sample end motif spectrum; (ii) F can represent the end frequencies of a reference F-spectrum obtained from a murine sample; and (iii) W can represent the relative weights corresponding to the proportional contributions of each F spectrum. The proportional contributions can be determined by solving for W using non-negative least squares (NNLS) based on the values from the data matrix M and the reference F spectra F. The proportional contributions determined using deconvolution can be used to identify the level of fragmentation factor activity in the subject (e.g., a relative decrease in the contribution of F spectrum I).

[0351] In some cases, the frequencies of 4-mer end motifs of the end motif spectra of samples from subjects (e.g., human subjects) and those of the end motifs of a reference sample (e.g., a murine sample) are normalized by the genomic background of their respective genomes. For example, expected 4-mer end motif frequencies can be used in the normalization step, where the expected end motif frequencies are determined by simulating 4-mer end motifs from the reference genome using a 4-bp sliding window spanning each chromosome. The normalized end motif frequency is calculated as the ratio of the observed frequency to the expected frequency, and then divided by the sum of all 256 normalized motif frequencies. The total normalized end motif frequency can equal 100%.

[0352] At block 5410, a classification of the nuclease activity of a particular type of nuclease is determined based on a proportional contribution associated with a particular type of fragmentation factor. For example, the classification of the nuclease activity of a particular type of nuclease can include a classification of reduced nuclease activity associated with a particular nuclease. The classification of nuclease activity can be used to determine whether a subject has a nuclease activity defect or a classification of a genetic disorder of a nuclease-related gene. The genetic disorder can be a disorder of the DNASE1L3 gene. The genetic disorder can include disorders of one or more of the following genes: DNASE1, DFFB, TREX1 (trimer repair exonuclease 1), AEN (apoptosis-enhancing nuclease), EXO1 (exonuclease 1), DNASE2 (deoxyribonuclease 2), ENDOG (endonuclease G), APEX1 (apurinic / apyrimidinic endodeoxyribonuclease 1), FEN1 (flap structure-specific endonuclease 1), DNASE1L1 (deoxyribonuclease 1-like 1), DNASE1L2 (deoxyribonuclease 1-like 2), and EXOG (exonuclease / endonuclease G).

[0353] In some cases, a reduction in the level of nuclease activity associated with a particular type of nuclease is determined based on a proportional contribution of a set of reference F-spectra. For example, the proportional contribution associated with one of the set of reference F-spectra can be compared to a threshold. Based on the comparison (e.g., if the proportional contribution exceeds the threshold), a reduction in the level of nuclease activity can be determined. In some cases, the threshold is determined using one or more reference samples with a known classification of nuclease activity. IX. Concentration fractions of F-spectra of cfDNA

[0354] Since cell-free DNA molecules in urine still retain the DNASE1L3 cleavage tags of plasma cell-free DNA, we infer that the level of nuclease usage in urine may represent the amount of cell-free DNA in urine. We hypothesized that NMF-based analysis of nuclease usage levels could be used to determine the fractional contribution of cell-free DNA in urine samples. To this end, we applied the nuclease usage level analysis to 14 maternal urine samples.

[0355] Figure 55 shows a set of graphs 5500 for identifying nuclease usage levels in free DNA of pregnant women's urine. Graph 5502 shows the correlation between fetal DNA fraction and the level of F-profile I (DNASE1L3). Graph 5504 shows the correlation between fetal DNA fraction and the level of F-profile IV. As Figure 55 shown, we found that the proportional contributions of F-profile I (DNASE1L3) and F-profile IV were significantly correlated with the fetal DNA fraction estimated by SNP-based methods (Pearson's r = 0.60, P = 0.025) in maternal urine free DNA. Thus, the nuclease usage level analysis presented in this disclosure can be used to monitor the proportion of free DNA passing through the kidney in urine samples.

[0356] Figure 56 shows a flowchart of a method 5600 for determining the concentration fraction of fetal DNA based on the F-profile of free DNA molecules according to some embodiments. Exemplary biological samples can be free samples including free DNA, such as blood, plasma, serum, urine, and saliva. The biological sample can include free clinically relevant DNA and other DNA. In other instances, the biological sample may not include clinically relevant DNA, and the estimated concentration fraction can indicate zero or a low percentage of clinically relevant DNA.

[0357] The biological sample can include a mixture of free DNA molecules from one or more tissue types, such as the heart, lungs, and liver. For example, the biological sample can be obtained from a pregnant woman containing maternal free DNA molecules and fetal free DNA molecules. The biological sample can contain tumor-specific free DNA molecules as well as other tissue-specific free DNA molecules. Clinically relevant DNA molecules can be any DNA of any tissue type described herein, such as fetal DNA, tumor DNA, or graft DNA. Aspects of method 5600 and any other method described herein can be performed by a computer system. Aspects of method 5600 can be performed in a manner similar to Figure 54 method 5400.

[0358] At block 5602, a set of reference F-spectra is stored. For each nucleotide in the nucleotide set, each reference F-spectrum of the set can identify the proportion of free DNA molecules that terminate at that nucleotide. Additionally, each reference F-spectrum can be associated with a type of fragmentation factor. The type of fragmentation factor can identify a specific enzyme (e.g., DNASE1L3, DNASE1), protein (e.g., DFFB), or other biological component or process that causes fragmentation of free DNA molecules. In some cases, the set of reference F-spectra includes 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 15, 20, 25, 30, 40, 45, 50, or more than 50 F-spectra. For example, the set of reference F-spectra can include six F-spectra I-VI. Block 5602 can be performed in a manner similar to Figure 54 block 5402.

[0359] At block 5604, multiple free DNA molecules from a biological sample are analyzed to obtain sequence reads. The sequence reads include end sequences corresponding to the ends of the multiple free DNA molecules. The sequence reads can include end sequences corresponding to the ends of multiple free DNA fragments. Block 5604 can be performed in a manner similar to Figure 54 block 5404.

[0360] At block 5606, a sample end motif spectrum of a subject is determined by determining the proportion of multiple free DNA molecules that terminate at each nucleotide of the nucleotide set based on the end sequences. The sample end motif spectrum identifies the relative frequencies of multiple end motifs corresponding to the end sequences of the multiple free DNA molecules. Multiple end motifs can correspond to all possible combinations of N base positions. For example, if multiple end motifs correspond to 4-mers, the multiple end motifs of the sample end motif spectrum can include 256 combinations of 4-mers (e.g., CCCA, TTCC). Block 5606 can be performed in a manner similar to Figure 54 block 5406.

[0361] At block 5608, a proportional contribution of the set of reference F-spectra is determined, the proportional aggregation of which provides the sample end motif spectrum. The sum of the proportional contributions of the set of reference F-spectra pairs is one. Block 5608 can be performed in a manner similar to Figure 54 block 5408.

[0362] The set of reference F-spectra includes a first reference F-spectrum related to the concentration fraction of clinically relevant DNA molecules, e.g., as determined using a calibration sample with a known concentration fraction. Figure 55 An example is provided where the first reference F-spectrum is F-spectrum I (corresponding to DNASE1L3) or F-spectrum IV. The first proportional contribution to the set of reference F-spectra can correspond to the first reference F-spectrum.

[0363] At block 5610, the concentration fraction of clinically relevant DNA molecules in a biological sample is estimated by comparing a first proportional contribution corresponding to a first reference F-spectrum with one or more calibration values determined from one or more calibration samples, where the concentration fraction of clinically relevant DNA molecules in the one or more calibration samples is known. Aspects of block 5610 can be performed in a manner similar to block 2106 of method 2100. The first reference F-spectrum can correspond to a particular type of nuclease, such as DNASE1L3.

[0364] As Figure 55 shown, the fetal DNA fraction increases as the proportional contribution of one or more reference F-spectra (e.g., F-spectrum I, F-spectrum IV) increases. Any one of the one or more reference F-spectra can be the first reference F-spectrum, so long as it is related to the concentration fraction of clinically relevant DNA. Thus, the known proportional contributions of one or more reference F-spectra in the calibration samples can be used as calibration data points for determining the concentration fraction of clinically relevant DNA molecules in a biological sample.

[0365] For each of the one or more calibration samples, some embodiments can measure the concentration fraction of clinically relevant DNA molecules in the calibration sample and measure the proportional contribution of the first reference end-motif spectrum of the calibration sample, thereby determining one or more calibration data points. The proportional contribution of the first reference end-motif spectrum can be used as a calibration value. The total proportional contributions of a set of reference F-spectra can be determined, thereby determining multiple calibration values, e.g., when multiple reference F-spectra are used to estimate the concentration fraction. Those skilled in the art will understand that the concentration fraction can be measured in various ways, some of which are described herein, such as using tissue-specific alleles or tissue-specific methylation patterns.

[0366] In some cases, the proportional contributions determined for one or more reference F-spectra (including the first proportional contribution of the first reference F-spectrum) are compared with a calibration curve (composed of calibration data points), and thus the comparison can identify the point on the curve having the known proportional contribution of the biological sample. The concentration fraction corresponding to the identified point can then be used to estimate the concentration fraction. For example, the determined proportional contributions can be provided as inputs (e.g., linear or non-linear fits) to a calibration function to obtain an output of the concentration fraction.

[0367] In some embodiments, as described above, multiple proportional contributions can be used. In such cases, the calibration curve can be a two-dimensional or more-dimensional calibration surface. Thus, estimating the concentration fraction of clinically relevant DNA molecules in a biological sample can include comparing one or more additional proportional contributions with one or more additional calibration values determined from one or more calibration samples, where the concentration fraction of clinically relevant DNA molecules in the calibration samples is known. X. Estimating Gestational Age Based on the F-Spectrum in cfDNA

[0368] The nuclease usage levels identified from the factorization analysis of cell-free DNA molecules can be used to estimate the gestational age of the fetus in a sample obtained from a pregnant woman. For example, the proportional contributions of the F-spectrum obtained from samples of known gestational age can be determined. The determined proportional contributions can then be used as calibration data points to estimate the gestational age of another pregnancy sample.

[0369] As further described below, there is a correlation between the proportional contribution of F-spectrum I and the gestational age of the fetus. This correlation can also indicate that, based on the activity level of DNASE1L3, the gestational age can be affected because F-spectrum I represents the cleavage preference of DNASE1L3. A. Estimating Gestational Age

[0370] Figure 57 A set of plots 5700 identifying the nuclease usage levels in cell-free DNA from pregnant women's plasma is shown. Box plots 5702 and 5704 show the expression levels of DNASE1L3 in the placentas of pregnant women in different trimesters. Box plot 5706 shows the contribution of F-spectrum I (DNASE1L3) in the cell-free DNA of the pregnant women's maternal plasma in the first, second, and third trimesters. Figure 5708 shows the correlation between the fetal DNA fraction and the level of F-spectrum I (DNASE1L3). As shown in box plots 5702 and 5704, in placental tissue, an upregulation of the DNASE1L3 gene expression level with gestational age was observed based on transcriptome data (Mikheev et al. Reprod. Sci. 2008; 15:866 - 877; Sitras et al. PLOS One 2012; 7:e33294).

[0371] The NMF-based nuclease usage level analysis can be used to estimate gestational age based on certain F-spectra of cell-free DNA. We analyzed the nuclease usage levels based on the maternal plasma end motifs using a previously published cohort of 30 pregnant women (10 in each trimester) (Jiang et al. Clin. Chem. 2017; 63:606–608). As shown in box plot 5706, we observed that the level of F-spectrum I (DNASE1L3) in the cell-free DNA of the maternal plasma gradually increased with gestational age in the first trimester (median: 40.2%; range: 38.5 - 42.7%), the second trimester (median: 41.3%; range: 36.2 - 42.8%), and the third trimester (median: 43.1%; range: 34.5 - 44.0%).

[0372] The nuclease usage level analysis disclosed herein can also be used to determine the contribution fraction of fetal DNA in plasma samples. As shown in FIG. 5708, the level of F-profile I (DNASE1L3) in maternal plasma cell-free DNA is significantly correlated with the fetal DNA fraction estimated by SNP-based methods (Pearson's r = 0.40, P = 0.027). Thus, nuclease usage level analysis can be used to monitor physiological states such as pregnancy. B. Relationship between Oxidative Stress and Gestational Age before

[0373] In addition to cancer patients, we also studied plasma from pregnant women in the first trimester (n = 10), second trimester (n = 10), and third trimester (n = 10). Previous studies have elucidated that oxidative stress in the placenta has been reported to decrease with increasing gestational age (Basu et al. Obstet Gynecol Int 2015; 2015:276095).

[0374] Figure 58 A set of FIGS. 5800 identifying F-profile analysis and oxidative stress levels in pregnant women is shown. FIG. 5802 shows the oxidative stress levels in placental tissues from pregnant women in different trimesters. Box plot 5804 shows the F-profile VI contributions of fetal-specific DNA in the plasma of pregnant women during different trimesters. Box plot 5806 shows the F-profile VI contributions of maternal-specific DNA in the plasma of pregnant women during different trimesters. As Figure 58 shown in the figures, as the pregnancy progresses, the median level of F-profile VI in fetal-specific DNA decreases significantly (first: 26.7%; second: 23.7%; third: 22.0%) (P = 0.014, Kruskal-Wallis test), while the level of F-profile VI in maternal-specific DNA does not change significantly. These data suggest that the F-profile VI level can indicate the contribution of cell-free DNA derived from oxidative stress-induced fragmentation. C. Method for Estimating Gestational Age Based on F-Profile of Cell-Free DNA

[0375] Figure 59 A flowchart of a method 6800 for estimating gestational age based on the F-profile of cell-free DNA molecules according to some embodiments is shown. A biological sample is obtained from a female subject carrying a fetus. The biological sample can be a sample having cell-free DNA molecules from the female subject and the fetus, such as a plasma, serum, urine, saliva, cerebrospinal fluid, pleural fluid, amniotic fluid, peritoneal fluid, or ascites sample. Aspects of method 5900 and any other method described herein can be performed by a computer system. Aspects of method 5900 can be performed in a manner similar to Figure 54 method 5400 as

[0376] At block 5902, a set of reference F-spectra is stored. For each nucleotide in the nucleotide set, each reference F-spectrum in the set identifies the proportion of free DNA molecules that terminate at that nucleotide. Additionally, each reference F-spectrum is associated with a type of fragmentation factor. The type of fragmentation factor identifies a specific enzyme (e.g., DNASE1L3, DNASE1), protein (e.g., DFFB), or other biological component or process that causes fragmentation of free DNA molecules. In some cases, the set of reference F-spectra includes 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 15, 20, 25, 30, 40, 45, 50, or more than 50 F-spectra. For example, the set of reference F-spectra can include six F-spectra I-VI. Block 5902 can be performed in a manner similar to Figure 54 block 5402.

[0377] At block 5904, a plurality of free DNA molecules from a biological sample are analyzed to obtain sequence reads. The sequence reads include end sequences corresponding to the ends of the plurality of free DNA molecules. The sequence reads can include end sequences corresponding to the ends of a plurality of free DNA fragments. Block 5904 can be performed in a manner similar to Figure 54 block 5404.

[0378] At block 5906, a sample end motif spectrum of a subject is determined by determining the proportion of a plurality of free DNA molecules that terminate at each nucleotide in the nucleotide set based on the end sequences. The sample end motif spectrum identifies the relative frequencies of a plurality of end motifs corresponding to the end sequences of the plurality of free DNA molecules. The plurality of end motifs can correspond to all possible combinations of N base positions. For example, if the plurality of end motifs correspond to 4-mers, the plurality of end motifs of the sample end motif spectrum can include 256 combinations of 4-mers (e.g., CCCA, TTCC). Block 5906 can be performed in a manner similar to Figure 54 block 5406.

[0379] At block 5908, a proportional contribution of the set of reference F-spectra is determined, the proportional aggregation of which provides the sample end motif spectrum. The sum of the proportional contributions of the set of reference F-spectra is one. Block 5908 can be performed in a manner similar to Figure 54 block 5408.

[0380] The set of reference F-spectra includes a first reference F-spectrum related to gestational age, e.g., as determined using a calibration sample of known gestational age. Figure 57 and 58 Examples are provided where the first reference F-spectrum is F-spectrum I (corresponding to DNASE1L3) or F-spectrum IV. The first proportional contribution to the set of reference F-spectra can correspond to the first reference F-spectrum.

[0381] At block 5910, the gestational age of the fetus is estimated by comparing the first proportional contribution corresponding to the first reference F-spectrum with one or more calibration values determined from one or more calibration samples having known gestational ages. Aspects of block 5910 can be performed in a manner similar to block 2106 and block 5610. As Figure 57 shown, the first reference F-spectrum can correspond to a particular type of nuclease, such as DNASE1L3. As Figure 58 shown, a reference F-spectrum that does not correspond to a nuclease (e.g., F-spectrum VI) can also be used.

[0382] As an example, the proportional contributions of the reference F-spectrum of the calibration data points (e.g., F-spectrum I representing DNASE1L3) can be plotted on a graph and form clusters for different gestational ages, and the determined proportional contributions of the biological sample can also be plotted on the graph to determine the cluster into which the biological sample falls. Any one of the one or more reference F-spectra can be the first reference F-spectrum as long as it is related to the gestational age. Thus, the known proportional contributions of one or more reference F-spectra in the calibration sample can be used as calibration data points for determining the gestational age.

[0383] Thus, for each calibration sample of the one or more calibration samples, some embodiments can measure the gestational age in the calibration sample and measure the proportional contribution of the first reference end-motif spectrum determined for the calibration sample. For example, menstrual history and ultrasound examination are two methods for measuring gestational age. For example, the gestational age can be estimated based on the date of the last menstrual period. It can be assumed that conception occurs on day 14 of the cycle, which is affected by ovulation variations between menstrual cycles and between individuals. Ultrasonic measurement of the embryo or fetus in the first trimester can be the most accurate method for establishing gestational age. Using various parameters, such as mean sac diameter (MSD), crown-rump length (CRL), biparietal diameter (BPD), and head circumference (HC), the gestational age can be estimated from the ultrasound.

[0384] The proportional contribution of the first reference end-motif spectrum can be used as a calibration value. The total proportional contributions of the set of reference F-spectra can be determined, thereby determining multiple calibration values, e.g., when multiple reference F-spectra are used to estimate the gestational age. Those skilled in the art will understand that the gestational age can be measured in various ways.

[0385] In some cases, the proportional contributions determined for one or more reference F-spectra (including the first proportional contribution of the first reference F-spectrum) can be compared with a calibration curve (composed of calibration data points), and thus, this comparison can identify the point on the curve having the known proportional contribution of the biological sample. Then, the gestational age corresponding to the identified point can be used to estimate the gestational age. For example, the determined proportional contribution can be provided as an input to a calibration function (e.g., linear or non-linear fit) to obtain an output of the gestational age.

[0386] In some embodiments, as described above, multiple proportional contributions may be used. In such cases, the calibration curve may be a two-dimensional or more-dimensional calibration surface. Thus, estimating gestational age may include comparing one or more additional proportional contributions to one or more additional calibration values determined from one or more calibration samples of known gestational age. XI. Classification of Pathology Based on the F-Spectrum in cfDNA

[0387] The F-spectrum can also be used to classify the pathology grade in a subject. Examples of pathologies are autoimmune diseases (such as SLE) and cancers. A. Systemic Lupus Erythematosus (SLE)

[0388] Analysis of nuclease usage levels can be used to distinguish human subjects with or without DNase1L3 deficiency based on certain F-spectra of cell-free DNA. Human subjects with DNase1L3 deficiency will develop systemic lupus erythematosus (SLE)-like symptoms with childhood onset, which is also known as familial SLE (Chan et al. Am. J. Hum. Genet. 2020; 107:882-894). We studied nuclease usage levels by analyzing plasma cell-free DNA from patients with two copies of the DNase1L3 gene carrying a genetic mutation (i.e., DNase1L3-deficient) (n = 10), the parents of these patients carrying one copy of the mutant DNase1L3 gene (i.e., the other copy being functional) (n = 3), and healthy control subjects (n = 8) (Chan et al. Am. J. Hum. Genet. 2020; 107:882-894).

[0389] Figure 60 Box plot 6000 of the F-spectrum I (DNase1L3) levels of healthy subjects, patients with DNase1L3 deficiency, and the parents of the patients is shown. As Figure 60 shown, the F-spectrum I (DNase1L3) in the plasma cell-free DNA of patients with DNase1L3 deficiency appears to be significantly decreased compared to their parents (median: 51.4%; range: 47.4-51.9%) and healthy subjects (median: 52.9%; range: 47.3-58.2%) (median: 7.3%; range: 3.8-20.5%) (P < 0.0001, Kruskal-Wallis test).

[0390] Analysis of nuclease usage levels can distinguish SLE patients from non-SLE patients.

[0391] Figure 61Shows a set of graphs 6100 that identify the levels of nuclease usage in plasma-free DNA of subjects with and without systemic lupus erythematosus (SLE). Box plot 6102 shows the levels of F-profile I (DNASE1L3) in plasma-free DNA of healthy control subjects, patients with inactive SLE, and patients with active SLE. ROC curve 6104 shows the discrimination between SLE patients and non-SLE patients using the F-profile I (DNASE1L3) assessment.

[0392] Figure 6106 shows the correlation between the Systemic Lupus Erythematosus Disease Activity Index (SLEDAI) and the F-profile I level (DNASE1L3) in SLE patients. In a cohort consisting of 10 healthy controls, 13 and 11 patients with active and inactive sporadic SLE, respectively (Chan et al. Proc Natl Acad Sci USA. 2014;111:E5302–E5311), box plot 6102 shows that the levels of DNASE1L3 usage gradually decreased in healthy subjects (median: 39.8%; range: 38.0 - 42.3), patients with inactive SLE (median: 33.3%; range: 31.4 - 41.0%), and patients with active SLE (median: 29.7%; range: 14.9 - 34.2%) (P<0.0001, Kruskal-Wallis test). As shown by ROC curve 6104, the measure of DNASE1L3 usage level (F-profile I) enables discrimination between human individuals with and without SLE, with an AUC of 0.97.

[0393] Furthermore, as shown in Figure 6106, the DNASE1L3 usage level was negatively correlated with the Systemic Lupus Erythematosus Disease Activity Index (SLEDAI) (Pearson's r: -0.43; P = 0.036). Thus, the measure of DNASE1L3 usage level (F-profile I) will inform the presence of an autoimmune disease and facilitate the monitoring of disease progression. B. Cancer

[0394] In addition to SLE, nuclease usage level analysis is publicly available for discriminating between human subjects with and without hepatocellular carcinoma (HCC). It has been reported that HCC patients are affected by DNASE1L3 activity (Jiang et al. Cancer Discov. 2020;10:664–73). Combining the relationship between DNASE1L3 and HCC, nuclease usage level analysis was applied to a cohort consisting of 38 healthy controls, 17 HBV carriers without HCC, and 34 previously studied HCC patients (Jiang et al. Cancer Discov. 2020;10:664–73).

[0395] Figure 62 shows the proportional contributions of the F-spectra in normal, HBV, and HCC plasma samples according to some embodiments. The Figure 50 deconvolution process in Figure 62 is applied to normal, HBV, and HCC plasma samples from human subjects. As Figure 62 shown, samples from control subjects include a relatively high contribution of F-spectrum I. Similarly, samples obtained from HBV patients also include a relatively high contribution of F-spectrum I. In contrast, HCC samples include a relatively low contribution from F-spectrum I.

[0396] Figure 63 shows a set of graphs that show the nuclease usage levels in plasma-free DNA of subjects with and without HCC. Box plots of F-spectrum I 6302 and VI 6304 show the nuclease levels in plasma-free DNA of patients with and without HCC. The ROC curve 6306 shows the evaluation of the discrimination between the non-HCC and HCC groups using different metrics, including the motif diversity score and six F-spectra. Compared to healthy controls, box plot 6302 shows that a median decrease of 6.9% in the usage level of F-spectrum I (DNASE1L3) was indeed found in HCC patients, while no significant change was observed in HBV carriers.

[0397] As shown in box plot 6304, we also found a gradual increase in the usage level of F-spectrum VI in HBV carriers and HCC patients. In addition, ROC curve 6306 shows that among the six F-spectra, the most discriminative in detecting HCC patients is F-spectrum VI (AUC: 0.97), which appears to be randomly distributed across 256 terminal motifs (i.e., no obvious preference among the terminal motifs). This performance is superior to the previously reported motif diversity score (AUC: 0.86) (P = 0.019, DeLong test), which is used to quantify the uniformity of the total terminal motif frequency (Jiang et al. Cancer Discov. 2020;10:664–73). These data suggest that the analysis of nuclease usage levels by simultaneously considering the involvement of multiple nucleases may improve the signal-to-noise ratio for disease detection. C. Relationship between disease and oxidative stress

[0398] Since the F-spectrum VI showed promising discriminatory ability between patients with and without HCC, any biological implications related to the F-spectrum VI were considered. Due to the nature of the F-spectrum showing no obvious preference in the frequencies of 256 4-mer motifs, a possible speculation is that the fragmentation of cell-free DNA occurring in cancer patients may preferentially involve DNA breaks different from those induced by the classical apoptotic pathway.

[0399] Figure 64 Bar graph 6040 showing the levels of oxidative stress in blood samples from controls and HCC patients is presented. As Figure 64 shown, it has been reported that the level of oxidative stress in blood samples with HCC is higher than that in normal controls (Arsian et al. J. Cancer Ther. 2014; 5: 192 - 197). Based on the above, it can be considered that the F-spectrum VI is related to the degree of oxidative stress, such that the contribution of the F-spectrum VI in HCC patients is significantly increased (see Figure 63 box plot 6304).

[0400] To verify the above hypothesis, we utilized those clinical models in which certain tissues were reported to have higher / lower levels of oxidative stress. We first analyzed the cell-free DNA in plasma from 15 controls, 25 colorectal cancer (CRC) patients without liver metastasis, and 24 CRC patients with liver metastasis.

[0401] Figure 65 A set of graphs 6500 providing F-spectrum analysis and levels of oxidative stress in CRC patients is presented. Box plot 6502 shows the contribution of the F-spectrum VI in controls, CRC patients with and without liver metastasis. Bar graph 6504 shows the levels of oxidative stress in colon tissues from controls and CRC patients at different stages (II - IV). As shown in box plot 6502, we did find a significant increasing trend of the F-spectrum VI levels from controls (median: 24.3%; range 16.8 - 33.3%), CRC patients without liver metastasis (median: 30.5%; range 23.4 - 34.5%), to CRC patients with liver metastasis (median: 34.5%; range 18.1 - 43.5%) (Kruskal - Wallis test, P < 0.0001). This finding is consistent with the reported increase in oxidative stress in colorectal cancer patients and its further enhancement in advanced patients, as shown in bar graph 6504 (Skrzydlewska et al. World J Gastroenterol. 2005; 11: 403 - 406). D. After treatment of the disease

[0402] In addition to CRC patients, we also analyzed the F-profiles of plasma DNA in 6 patients with nasopharyngeal carcinoma (NPC) before and after chemoradiotherapy with cisplatin.

[0403] Figure 66 Box plot 6600 showing the contribution of F-profile VI in identifying NPC patients before and during chemoradiotherapy with cisplatin is presented. An increase in oxidative stress during chemoradiotherapy has been reported to be further enhanced (Conklin et al. Integr. Cancer Ther. 2004; 3:294–300). As Figure 66 shown, the level of F-profile VI in NPC patients increased during chemoradiotherapy with cisplatin (median: 22.1%; range: 18.5–22.8%) compared to paired patients before treatment (median: 23.6%; range: 21.8–25.5%) (P = 0.04, Kruskal-Wallis test). E. Classifying pathologic grade using F-profile cfDNA

[0404] Figure 67 A flowchart of method 6700 for determining pathologic grade based on the F-profile of cell-free DNA is shown according to some embodiments. Exemplary biological samples can be cell-free samples including cell-free DNA, such as blood, plasma, serum, urine, and saliva. Pathologies can include cancers (e.g., hepatocellular carcinoma, lung cancer, breast cancer, gastric cancer, glioblastoma multiforme, pancreatic cancer, colorectal cancer, nasopharyngeal carcinoma, head and neck squamous cell carcinoma, etc.) and autoimmune disorders (e.g., systemic lupus erythematosus). Aspects of method 6700 can be performed in a manner similar to Figure 54 method 5400.

[0405] In block 6702, a set of reference F-profiles is stored. For each nucleotide in the nucleotide set, each reference F-profile of the set identifies the proportion of cell-free DNA molecules that terminate at that nucleotide. In addition, each reference F-profile is associated with a type of fragmentation factor. The type of fragmentation factor identifies a specific enzyme (e.g., DNASE1L3, DNASE1), protein (e.g., DFFB), or other biological component or process that causes fragmentation of cell-free DNA molecules. In some cases, the set of reference F-profiles includes 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 15, 20, 25, 30, 40, 45, 50, or more than 50 F-profiles. For example, the set of reference F-profiles can include six F-profiles I-VI. Block 6702 can be performed in a manner similar to Figure 54 block 5402.

[0406] At block 6704, multiple free DNA molecules from a biological sample are analyzed to obtain sequence reads. The sequence reads include end sequences corresponding to the ends of the multiple free DNA molecules. The sequence reads may include end sequences corresponding to the ends of multiple free DNA fragments. Block 6704 can be performed in a manner similar to Figure 54 block 5404.

[0407] At block 6706, a sample end motif profile of a subject is determined by determining the proportion of multiple free DNA molecules that terminate with each nucleotide of a nucleotide set based on the end sequences. The sample end motif profile identifies the relative frequencies of multiple end motifs corresponding to the end sequences of the multiple free DNA molecules. Multiple end motifs may correspond to all possible combinations of N base positions. For example, if multiple end motifs correspond to 4-mers, the multiple end motifs of the sample end motif profile may include 256 combinations of 4-mers (e.g., CCCA, TTCC). Block 6706 can be performed in a manner similar to Figure 54 block 5406.

[0408] At block 6708, the proportional contributions of a set of reference F profiles are determined, and the proportional aggregation of the reference F profiles provides the sample end motif profile. The sum of the proportional contributions of the set of reference F profiles is one. Block 6708 can be performed in a manner similar to Figure 54 block 5408.

[0409] At block 6710, a classification of the pathological grade of a subject can be determined based on a determination that at least one of the determined proportional contributions exceeds a predetermined threshold. The predetermined threshold may correspond to the proportional contribution of a specific reference F profile (e.g., F profile I, F profile IV). For example, a classification of the pathological grade of a subject can be determined based on a determination that one of the determined proportional contributions is less than the predetermined threshold, as shown in Figure 63 box plot 6302. In another example, a classification of the pathological grade of a subject can be determined based on a determination that one of the determined proportional contributions is greater than the predetermined threshold, as shown in Figure 63 box plot 6304.

[0410] The pathological grade may include no cancer, early stage, middle stage or late stage. Then, the classification can select one of the grades. Thus, the classification can be determined from multiple cancer grades including multiple cancer stages. As an example, the cancer can be hepatocellular carcinoma, lung cancer, breast cancer, gastric cancer, glioblastoma multiforme, pancreatic cancer, colorectal cancer, nasopharyngeal cancer, and head and neck squamous cell carcinoma. For example, the autoimmune disease can be systemic lupus erythematosus.

[0411] In other instances, the pathology grade corresponds to a concentration fraction of clinically relevant DNA associated with the pathology. For example, the pathology grade can be cancer, and the clinically relevant DNA can be tumor DNA. The reference value can be a calibration value determined from a calibration sample. XII. Treatment A. Treatment Options

[0412] Embodiments of the present disclosure can accurately predict disease recurrence, thereby facilitating early intervention and selection of appropriate treatment to improve disease outcomes and the overall survival rate of the subject. For example, intensified chemotherapy can be selected for a subject if their corresponding sample predicts disease recurrence. In another instance, a biological sample of a subject who has completed initial treatment can be sequenced to identify viral DNA that predicts disease recurrence. In such an instance, an alternative treatment regimen (e.g., higher dose) and / or a different treatment can be selected for the subject because the subject's cancer may have become tolerant to the initial treatment.

[0413] Embodiments can also include treating a subject in response to determining a classification of pathologic recurrence. For example, surgery can be selected as a possible treatment if the prediction corresponds to a local regional failure. In another instance, chemotherapy can be additionally selected as a possible treatment if the prediction corresponds to distant metastasis. In some embodiments, the treatment includes surgery, radiotherapy, chemotherapy, immunotherapy, targeted therapy, hormone therapy, stem cell transplantation, or precision medicine. Based on the determined recurrence classification, a treatment plan can be developed to reduce the risk of harm to the subject and increase the overall survival rate. Embodiments can further include treating the subject according to the treatment plan. B. Types of Treatment

[0414] Embodiments can further include treating the pathology of a patient after determining the classification of the subject. Treatment can be provided based on the determined pathology grade, concentration fraction of clinically relevant DNA, or the tissue of origin. For example, identified mutations can be targeted with specific drugs or chemotherapy. The tissue of origin can be used to guide surgery or any other form of treatment. And the pathology grade can be used to determine how invasive any type of treatment will be, which can also be determined based on the pathology grade. The pathology (e.g., cancer) can be treated by chemotherapy, drugs, diet, therapy, and / or surgery. In some embodiments, the more the value of a parameter (e.g., amount or length) exceeds the reference value, the more aggressive the treatment may be.

[0415] Treatment can include resection. For bladder cancer, treatment can include transurethral resection of bladder tumor (TURBT). This method is used for diagnosis, staging, and treatment. During TURBT, the surgeon inserts a cystoscope into the bladder through the urethra. Then a tool with a small wire loop, laser, or high-energy electricity is used to remove the tumor. For patients with non-muscle invasive bladder cancer (NMIBC), TURBT can be used to treat or eliminate the cancer. Another treatment can include radical cystectomy and lymph node dissection. Radical cystectomy is the removal of the entire bladder and possibly surrounding tissues and organs. Treatment can also include urinary diversion. Urinary diversion is creating a new path for urine to flow out of the body when the bladder is removed as part of the treatment.

[0416] Treatment can include chemotherapy, which is the use of drugs to destroy cancer cells, usually by keeping the cancer cells from growing and dividing. The drugs can include, for example, but are not limited to, mitomycin-C (which is available as a common drug), gemcitabine (Gemzar), and thiotepa (Tepadina) for intravesical chemotherapy. Systemic chemotherapy can include, for example, but is not limited to, cisplatin and gemcitabine, methotrexate (Rheumatrex, Trexall), vinblastine (Velban), doxorubicin, and cisplatin.

[0417] In some embodiments, treatment can include immunotherapy. Immunotherapy can include immune checkpoint inhibitors that block a protein called PD-1. The inhibitors can include, but are not limited to, atezolizumab (TECENTRIQ), nivolumab (Opdivo), avelumab (Bavencio), durvalumab (Imfinzi), and pembrolizumab (Keytruda).

[0418] Treatment embodiments can also include targeted therapy. Targeted therapy is a treatment that targets specific genes and / or proteins that contribute to the growth and survival of cancer. For example, erdafitinib is an orally administered drug that is approved to treat people with locally advanced or metastatic urothelial carcinoma having FGFR3 or FGFR2 genetic mutations, where the locally advanced or metastatic urothelial carcinoma has cancer cells that continue to grow or spread.

[0419] Some treatments can include radiation therapy. Radiation therapy is the use of high-energy X-rays or other particles to destroy cancer cells. In addition to each individual treatment, combinations of these treatments described herein can also be used. In some embodiments, when the value of a parameter exceeds a threshold (which itself exceeds a reference value), a combination of treatments can be used. Information about treatments in the references is incorporated herein by reference. XIII. Exemplary System

[0420] Figure 68 Illustrated is a measurement system 6800 in accordance with an embodiment of the present disclosure. The system shown includes a sample 6805, such as free DNA molecules within an assay device 6810, where an assay 6808 can be performed on the sample 6805. For example, the sample 6805 can be contacted with reagents of the assay 6808 to provide a signal of a physical characteristic 6815. An example of an assay device can be a flow cell or a tube through which droplets (the droplets including the assay) containing assay probes and / or primers move. The physical characteristic 6815 (e.g., fluorescence intensity, voltage, or current) from the sample is detected by a detector 6820. The detector 6820 can make measurements at intervals (e.g., periodic intervals) to obtain data points that constitute a data signal. In one embodiment, an analog-to-digital converter converts the analog signal from the detector into digital form at multiple times. The assay device 6810 and the detector 6820 can form an assay system, e.g., a sequencing system that performs sequencing according to the embodiments described herein. The data signal 6825 is sent from the detector 6820 to a logic system 6830. As an example, the data signal 6825 can be used to determine a sequence and / or location in a reference genome of a DNA molecule. The data signal 6825 can include multiple measurements taken simultaneously, such as different electrical signals of different samples 6805 of fluorescent dyes or different molecules, so the data signal 6825 can correspond to multiple signals. The data signal 6825 can be stored in a local memory 6835, an external memory 6840, or a storage device 6845.

[0421] The logic system 6830 can be or can include a computer system, an ASIC, a microprocessor, a graphics processing unit (GPU), etc. It can also include a display (e.g., a monitor, an LED display, etc.) and a user input device (e.g., a mouse, a keyboard, a button, etc.) or be coupled to a display (e.g., a monitor, an LED display, etc.) and a user input device (e.g., a mouse, a keyboard, a button, etc.). The logic system 6830 and other components can be part of a stand-alone or network-connected computer system, or they can be directly connected to or incorporated within a device (e.g., a sequencing device) that includes the detector 6820 and / or the assay device 6810. The logic system 6830 can also include software executed in a processor 6850. The logic system 6830 can include a computer-readable medium storing instructions for controlling the measurement system 6800 to perform any of the methods described herein. For example, the logic system 6830 can provide commands to a system including the assay device 6810 to perform sequencing or other physical operations. Such physical operations can be performed in a specific order, e.g., adding and removing reagents in a specific order. Such physical operations can be performed by a robotic system, e.g., including a robotic arm, which can be used to obtain samples and perform assays.

[0422] The measurement system 6800 may also include a treatment device 6860, which may provide treatment to an object. The treatment device 6860 may determine the treatment and / or be used to perform the treatment. Examples of such treatments may include surgery, radiotherapy, chemotherapy, immunotherapy, targeted therapy, hormone therapy, and stem cell transplantation. The logic system 6830 may be connected to the treatment device 6860, for example, to provide the results of the methods described herein. The treatment device may receive inputs from other devices, such as imaging devices and user inputs (e.g., to control the treatment, such as control on a robotic system).

[0423] Any computer system mentioned herein may utilize any suitable number of subsystems. Examples of such subsystems are shown in Figure 69 computer system 10. In some embodiments, the computer system includes a single computer device, where the subsystems may be components of the computer device. In other embodiments, the computer system may include multiple computer devices, each computer device being a subsystem with internal components. The computer system may include desktop and laptop computers, tablet computers, mobile phones, and other mobile devices.

[0424] Figure 69 The subsystems shown are interconnected by a system bus 75. Additional subsystems, such as a printer 74, a keyboard 78, a storage device 79, a monitor 76 (e.g., a display screen, such as an LED), coupled to a display adapter 82, and other subsystems are shown. Peripheral devices and input / output (I / O) devices coupled to the I / O controller 71 may be connected to the computer system by any number of means known in the art, such as input / output (I / O) ports 77 (e.g., USB, ). For example, the I / O port 77 or an external interface 81 (e.g., Ethernet, Wi-Fi, etc.) may be used to connect the computer system 10 to a wide area network, such as the Internet, a mouse input device, or a scanner. The interconnection via the system bus 75 allows the central processor 73 to communicate with each subsystem and control the execution of multiple instructions from the system memory 72 or the storage device 79 (e.g., a fixed disk such as a hard disk drive or an optical disk), as well as the exchange of information between the subsystems. The system memory 72 and / or the storage device 79 may be embodied as a computer-readable medium. Another subsystem is a data acquisition device 85, such as a camera, a microphone, an accelerometer, etc. Any data mentioned herein may be output from one component to another and may be output to the user.

[0425] A computer system may include multiple identical components or subsystems, e.g., connected together via an external interface 81, via an internal interface, or via a removable storage device that can be connected from one component to another and removed from another component. In some embodiments, a computer system, subsystem, or device may communicate via a network. In such a case, one computer may be considered a client and another computer may be considered a server, where each computer may be part of the same computer system. The client and server may each include multiple systems, subsystems, or components.

[0426] Aspects of the embodiments may be implemented using hardware circuitry (e.g., an application specific integrated circuit or a field programmable gate array) and / or using computer software stored in memory in a modular or integrated manner and a general purpose programmable processor in the form of control logic, and thus the processor may include a memory storing software instructions configuring the hardware circuitry, and an FPGA having configuration instructions or an ASIC. As used herein, a processor may include a single core processor, a multi-core processor on the same integrated chip, or multiple processing units on a single circuit board or networked processors, as well as dedicated hardware. Based on the present disclosure and the teachings provided herein, those skilled in the art will know and understand other ways and / or methods of implementing the embodiments of the present disclosure using hardware as well as combinations of hardware and software.

[0427] Any of the software components or functions described in this application may be implemented as software code to be executed by a processor using any suitable computer language (e.g., Java, C, C++, C#, Objective-C, Swift, or a scripting language such as Perl or Python) using, for example, conventional or object-oriented techniques. The software code may be stored as a series of instructions or commands for storage and / or transmission on a computer-readable medium. Suitable non-transitory computer-readable media may include random access memory (RAM), read-only memory (ROM), magnetic media such as a hard disk drive or a floppy disk, or optical media such as a compact disc (CD), a digital versatile disc (DVD), or a Blu-ray disc, flash memory, etc. The computer-readable media may be any combination of such devices. Additionally, the order of operations may be rearranged. When the operations of a process are complete, the process may be terminated, but may also have additional steps not included in the figure. A process may correspond to a method, function, program, subroutine, subprogram, etc. When a process corresponds to a function, its termination may correspond to the return of the function to the calling function or the main function.

[0428] Such programs can also be encoded and transmitted using carrier signals adapted to be transmitted via wired, optical, and / or wireless networks (including the Internet) that comply with various protocols. Accordingly, computer-readable media can be created using data signals encoded with such programs. A computer-readable medium encoded with program code can be packaged with a compatible device or provided separately from other devices (e.g., downloaded via the Internet). Any such computer-readable medium can reside on or within a single computer product (e.g., a hard disk drive, a CD, or an entire computer system), and can exist on or within different computer products in a system or network. A computer system can include a monitor, a printer, or other suitable display for providing any of the results mentioned herein to a user.

[0429] Any of the methods described herein can be performed, in whole or in part, by a computer system that includes one or more processors, which can be configured to perform the steps. Any operations performed with a processor (e.g., aligning, determining, comparing, computing, calculating) can be performed in real time. The term "real time" can refer to a computing operation or process that is completed within a certain time constraint. The time constraint can be 1 minute, 1 hour, 1 day, or 7 days. Accordingly, embodiments can be directed to a computer system configured to perform the steps of any of the methods described herein, which computer system may have different components for performing the corresponding steps or groups of corresponding steps. Although presented as numbered steps, the method steps herein can be performed simultaneously or at different times or in different orders. Additionally, portions of these steps can be used with portions of other steps from other methods. Moreover, all or part of the steps can be optional. Additionally, any of the steps in any of the methods herein can be performed with a module, unit, circuit, or other means of a system for performing these steps.

[0430] Without departing from the spirit and scope of the embodiments of the present disclosure, the specific details of particular embodiments can be combined in any suitable manner. However, other embodiments of the present disclosure can be directed to particular embodiments that relate to each individual aspect, or particular combinations of these individual aspects.

[0431] The foregoing description of the exemplary embodiments of the present disclosure has been presented for purposes of illustration and description. The foregoing description is not intended to be exhaustive or to limit the present disclosure to the precise forms described, and many modifications and variations are possible in light of the above teachings.

[0432] Unless the contrary is specifically stated, the recitation of "a" or "an" or "the" is intended to mean "one or more." Unless the contrary is specifically stated, the use of "or" is intended to mean "inclusive or" and not "exclusive or." The reference to a "first" component does not necessarily require the provision of a second component. Further, unless expressly stated, the reference to a "first" or "second" component does not limit the referenced component to a particular location. The term "based on" is intended to mean "at least partially based on."

[0433] Claims may be drafted to exclude any elements that may be optional. Thus, this statement is intended to serve as a basis for the use of exclusive terms such as "solely," "only," etc. in connection with the recitation of claim elements or the use of "negative" limitations.

[0434] It should be understood that the invention is not limited to the specific embodiments described, as these may vary. It should also be understood that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting, since the scope of the invention will be defined only by the appended claims. Efforts have been made to ensure the accuracy of the numbers used (e.g., amounts, temperatures, etc.), but some experimental error and deviation should be accounted for. Unless otherwise indicated, parts are parts by weight, molecular weights are weight average molecular weights, temperatures are in degrees Celsius, and pressures are at or near atmospheric pressure.

[0435] All patents, patent applications, publications, and descriptions mentioned herein are hereby incorporated by reference in their entirety for all purposes. None are admitted to be prior art. In the event of a conflict between this application and the references provided herein, this application will govern.

Claims

1. A method for estimating the concentration fraction of clinically relevant DNA molecules in a urine sample of a subject, the urine sample comprising free clinically relevant DNA molecules and other DNA molecules, the method comprising: Analyze multiple free DNA molecules from a urine sample, wherein analyzing the multiple free DNA molecules comprises: Determine the locations of the multiple free DNA molecules; and Based on the locations of the multiple free DNA molecules, identify a set of free DNA molecules from open chromatin regions of one or more tissues associated with the clinically relevant DNA molecules; Use the set of free DNA molecules to determine the relative abundances of the set of free DNA molecules, which are from open chromatin regions of the one or more tissues; and Estimate the concentration fraction of the clinically relevant DNA molecules in the urine sample by comparing the relative abundances with one or more calibration values determined from one or more calibration samples, wherein the concentration fractions of the clinically relevant DNA molecules in the calibration samples are known.

2. The method according to claim 1, wherein comparing the relative abundance with the one or more calibration values comprises comparing the relative abundance with a calibration curve comprising the one or more calibration values.

3. The method according to any one of claims 1 to 2, further comprising: For each calibration sample of the one or more calibration samples: Measure the concentration fraction of the clinically relevant DNA molecules in the calibration sample; and Measure the relative abundances of the free DNA molecules from open chromatin regions of the one or more tissues from the calibration sample.

4. The method according to claim 3, wherein measuring the concentration fraction of the clinically relevant DNA molecules uses tissue-specific alleles or tissue-specific methylation patterns.

5. A method for enriching clinically relevant DNA molecules in a urine sample, the urine sample comprising free clinically relevant DNA molecules and other DNA molecules, the method comprising: Analyze multiple free DNA molecules from the urine sample, wherein analyzing the multiple free DNA molecules comprises: Identify a set of free DNA molecules from the multiple free DNA molecules, which are from open chromatin regions of one or more tissues associated with the clinically relevant DNA molecules; and Use the set of free DNA molecules from open chromatin regions of the one or more tissues to generate an enriched sample, wherein the enriched sample has a higher concentration of clinically relevant DNA compared to the urine sample.

6. The method according to claim 5, further comprising determining a property associated with the clinically relevant DNA molecules in the enriched sample, wherein the property associated with the clinically relevant DNA molecules in the urine sample is (1) the concentration fraction of the clinically relevant DNA molecules or (2) the pathological grade of the subject from whom the urine sample is obtained, the pathological grade being associated with the clinically relevant DNA molecules.

7. The method according to any one of claims 5 to 6, wherein generating the enriched sample further comprises using a collection of free DNA molecules from open chromatin regions of the one or more tissues and having a length less than a specified length threshold.

8. The method according to claim 7, wherein the specified length threshold is 40 base pairs, 50 base pairs, 60 base pairs, 70 base pairs, 80 base pairs, 90 base pairs, 100 base pairs, 110 base pairs, 120 base pairs, 130 base pairs, 140 base pairs, 150 base pairs or 160 base pairs.

9. The method according to any one of claims 5 to 6, wherein generating the enriched sample further comprises using a collection of free DNA molecules from open chromatin regions of the one or more tissues and having one or more end sequences corresponding to end sequence features.

10. The method according to any one of claims 5 to 6, wherein identifying the collection of free DNA molecules or generating the enriched sample comprises: Subject the multiple free DNA molecules to probe molecules having sequences from the open chromatin regions, thereby obtaining the set of free DNA molecules.

11. The method according to claim 10, wherein generating the enriched sample comprises: Amplify the set of free DNA molecules using the one or more probe molecules.

12. The method according to claim 10, wherein generating the collection of free DNA molecules comprises: Capture the set of free DNA molecules using the one or more probe molecules; and Discard other free DNA molecules among the multiple free DNA molecules.

13. The method according to claim 10, wherein one or more probe molecules are attached to a surface and a collection of one or more sequence motifs in the end sequences is detected by hybridization.

14. A method for enriching clinically relevant DNA in a urine sample, the urine sample comprising the free clinically relevant DNA and other DNA, the method comprising: Analyze multiple free DNA molecules from the urine sample, wherein analyzing the multiple free DNA molecules comprises: Identify a set of free DNA molecules having terminal sequences from the multiple free DNA molecules, wherein the terminal sequences are in a set of one or more sequence motifs including a C-terminal nucleotide; and Use the set of free DNA molecules having terminal sequences in the set of one or more sequence motifs to generate an enriched sample, wherein the enriched sample has a higher concentration of clinically relevant DNA compared to the urine sample.

15. The method according to claim 14, wherein generating the enriched sample further comprises using a collection of free DNA molecules that are within a collection of one or more sequence motifs including C-terminal nucleotides and have a length less than a specified length threshold.

16. The method according to claim 15, wherein the specified length threshold is 40 base pairs, 50 base pairs, 60 base pairs, 70 base pairs, 80 base pairs, 90 base pairs, 100 base pairs, 110 base pairs, 120 base pairs, 130 base pairs, 140 base pairs, 150 base pairs or 160 base pairs.

17. The method according to any one of claims 14 to 16, wherein identifying the set of free DNA molecules or generating the enriched sample comprises: Subject the multiple free DNA molecules to one or more probe molecules that detect a set of one or more sequence motifs in the terminal sequences of the multiple free DNA molecules, thereby obtaining the set of free DNA molecules.

18. The method according to claim 17, further comprising: Attach a common sequence to the multiple free DNA molecules, wherein the one or more probe molecules include sequences complementary to the common sequence.

19. The method according to claim 17, wherein generating the enriched sample comprises: Amplify the collection of free DNA molecules using the one or more probe molecules.

20. The method according to claim 17, wherein generating the set of free DNA molecules comprises: Capture the collection of free DNA molecules using the one or more probe molecules; and Discard other free DNA molecules among the plurality of free DNA molecules.

21. The method according to claim 17, wherein one or more probe molecules are attached to a surface and a set of one or more sequence motifs in the terminal sequences is detected by hybridization.

22. The method according to any one of claims 14 to 21, wherein analyzing the plurality of free DNA molecules comprises receiving sequence reads obtained from sequencing of the plurality of free DNA molecules, and wherein identifying the set of the plurality of free DNA molecules comprises identifying sequence reads having terminal sequences in the set of one or more sequence motifs.

23. The method according to claim 22, wherein the enriched sample corresponds to sequence reads having terminal sequences in the set of one or more sequence motifs.

24. The method according to any one of claims 1, 5 or 17, wherein the clinically relevant DNA molecules are renal DNA molecules.

25. The method according to any one of claims 1, 5 or 17, wherein the clinically relevant DNA molecules comprise fetal DNA or tumor DNA.

26. A method for detecting kidney abnormalities using a urine sample of a subject, the urine sample comprising free DNA molecules, the method comprising: Analyze a plurality of free DNA molecules from the urine sample, wherein analyzing the plurality of free DNA molecules includes: Determine the positions of the plurality of free DNA molecules; and Based on the positions of the plurality of free DNA molecules, identify a collection of free DNA molecules from open chromatin regions of one or more tissues; Determine the relative abundance of the collection of free DNA molecules from the open chromatin regions of the one or more tissues compared to other free DNA molecules among the plurality of free DNA molecules; Compare the relative abundance with a reference value; and Based on the comparison, determine the classification of a subject with a kidney abnormality.

27. The method according to claim 26, wherein the reference value corresponds to another relative abundance determined based on free DNA molecules in open chromatin regions from one or more reference samples, wherein the one or more reference samples are associated with known classifications of kidney abnormalities.

28. The method according to claim 26, wherein determining the classification further comprises applying a machine learning model to the relative abundance to generate an output indicative of the classification of the subject having the kidney abnormality, wherein the machine learning model is trained using a training data set determined from one or more training samples having known classifications of the kidney abnormality.

29. The method according to claim 26, wherein the set of free DNA molecules terminates at one or more positions in a window around open chromatin regions of the one or more tissues.

30. The method according to claim 1 or claim 26, wherein determining the relative abundance comprises: Determine a first relative frequency of the collection of free DNA molecules from the open chromatin regions of the one or more tissues; Determine a second relative frequency of a reference sequence of a reference genome from the open chromatin regions of the one or more tissues; and Based on the first relative frequency and the second relative frequency, determine the relative abundance.

31. The method according to claim 30, wherein determining the second relative frequency comprises identifying single nucleotide variants of a reference genome in open chromatin regions from the one or more tissues.

32. The method according to claim 30, wherein the relative abundance is a ratio between the first relative frequency and the second relative frequency.

33. The method according to claim 1 or claim 26, wherein the relative abundance is the end density of the plurality of free DNA molecules terminating at open chromatin regions of the one or more tissues.

34. The method according to claim 33, wherein the end density comprises a first quantity of the set of free DNA molecules from open chromatin regions of the one or more tissues divided by a second quantity of the plurality of free DNA molecules from one or more other regions.

35. The method according to claim 1 or claim 26, wherein determining the positions of the plurality of free DNA molecules comprises aligning sequence reads of the plurality of free DNA molecules with a reference genome.

36. The method according to any one of claims 1, 5, and 26, wherein the one or more tissues comprise at least one of heart, lung, colon, liver, or white blood cells.

37. The method according to any one of claims 1, 5, and 26, wherein the set of free DNA molecules terminates at one or more positions in a window around open chromatin regions of the one or more tissues.

38. The method according to any one of claims 1, 5, and 26, wherein the open chromatin region comprises Dnase1 hypersensitive sites.

39. The method according to any one of claims 1 to 38, wherein the set of free DNA molecules comprises at least 5,000 free DNA molecules.

40. A method for detecting kidney abnormalities using a urine sample from a subject, the urine sample comprising free DNA molecules, the method comprising: Analyze a plurality of free DNA molecules from the urine sample, wherein analyzing the plurality of free DNA molecules includes determining the lengths of the plurality of free DNA molecules; Use the lengths of the plurality of free DNA molecules to determine a statistical value; Compare the statistical value with a reference value; and Based on the comparison, determine the classification of a subject with a kidney abnormality.

41. The method according to claim 40, wherein determining the statistical value comprises: Identify a collection of free DNA molecules having lengths within a length range from the plurality of free DNA molecules; and Further determine the statistical value based on the amount of the collection of free DNA molecules.

42. The method according to claim 41, wherein the statistical value is the proportion of the set of free DNA molecules relative to the plurality of free DNA molecules from the urine sample.

43. The method according to claim 41 or claim 42, wherein the length range has an upper limit selected from at least 80 bases, at least 90 bases, at least 100 bases, at least 110 bases, or at least 120 bases.

44. The method according to any one of claims 40 to 43, wherein one or more reference samples with known classifications of kidney abnormalities are used to determine the reference value.

45. The method according to any one of claims 40 to 44, wherein determining the length of the plurality of free DNA molecules comprises using gel electrophoresis, filtration, size - selection precipitation, or hybridization.

46. The method according to any one of claims 40 to 45, wherein determining the length of the plurality of free DNA molecules comprises: Receive sequence reads obtained from sequencing the plurality of free DNA molecules from the urine sample; and For each sequence read, count the number of nucleotides of the sequence read.

47. The method according to claim 46, wherein sequencing the plurality of free DNA molecules comprises performing massively parallel sequencing, single - molecule real - time sequencing, or nanopore sequencing on the plurality of free DNA molecules.

48. The method according to any one of claims 26 - 47, wherein the kidney abnormality is preeclampsia or proteinuria.

49. The method according to any one of claims 26 - 47, wherein the classification of a subject having a kidney abnormality comprises an increased level of permeability associated with the glomerular basement membrane of the kidney.

50. The method according to any one of claims 1 to 49, wherein the urine sample is treated with a DNA stabilizer before obtaining the plurality of free DNA molecules from the urine sample.

51. The method according to any one of claims 1 to 50, wherein analyzing the plurality of free DNA molecules comprises receiving sequence reads obtained from sequencing the plurality of free DNA molecules.

52. A method for detecting a kidney abnormality using a urine sample of a subject, the urine sample comprising free DNA molecules, the method comprising: Determine a first amount of free DNA molecules in the urine sample; Use the first amount and volume of the urine sample to determine an initial concentration; Use a second amount of a specific compound in the urine sample to determine a corrected concentration; Compare the corrected concentration with a reference value; and Based on the comparison, determine the classification of a subject with a kidney abnormality.

53. The method according to claim 52, wherein determining the first amount comprises measuring using a fluorometer, a spectrophotometer, or PCR.

54. The method according to claim 52, wherein the initial concentration is the proportion of free DNA molecules in the urine sample that are greater than a length threshold.

55. The method according to claim 54, wherein the length threshold is 40 - 200 bp.

56. The method according to claim 52, wherein the specific compound is creatinine.

57. The method according to claim 52, wherein the kidney abnormality is preeclampsia or proteinuria.

58. A method for determining the level of nuclease activity in a biological sample of a subject, the method comprising: Store a collection of reference F spectra, wherein each reference F spectrum in the collection: For each nucleotide in a set of nucleotides, identify the proportion of free DNA molecules terminated by the nucleotide; and Is associated with a type of fragmentation factor; Analyze a plurality of free DNA molecules in the biological sample to obtain sequence reads, the sequence reads including terminal sequences corresponding to ends of the plurality of free DNA molecules; Determine a sample end motif spectrum by determining, based on the terminal sequences, the proportion of the plurality of free DNA molecules terminated by each nucleotide in the set of nucleotides; Determine the proportional contributions of the collection of reference F spectra whose proportions aggregate to provide the sample end motif spectrum, wherein the proportional contributions sum to one; and Classifying nuclease activity of a specific type of nuclease based on proportional contributions associated with a reference F-spectrum of a set of reference F-spectra.

59. The method according to claim 58, wherein determining the classification of nuclease activity comprises: Determining a reduced nuclease activity level associated with a specific type of nuclease.

60. The method according to claim 59, wherein determining the reduced nuclease activity level comprises: Comparing a proportional contribution associated with one of the set of reference F-spectra to a threshold; and Determining the reduced nuclease activity level based on the comparison.

61. The method according to claim 60, wherein the threshold is determined using one or more reference samples of known classification having nuclease activity.

62. The method according to any one of claims 58 to 61, further comprising determining the classification of the genetic disorder of the subject based on the classification of the nuclease activity of the specific type of nuclease.

63. The method according to any one of claims 58 - 62, wherein the specific type of nuclease comprises one of DNASE1, DNASE1L3, DFFB, TREX1 (trinucleotide repair exonuclease 1), AEN (apoptosis-enhancing nuclease), EXO1 (exonuclease 1), DNASE2 (deoxyribonuclease 2), ENDOG (endonuclease G), APEX1 (apurinic / apyrimidinic endodeoxyribonuclease 1), FEN1 (flap structure-specific endonuclease 1), DNASE1L1 (deoxyribonuclease 1-like 1), DNASE1L2 (deoxyribonuclease 1-like 2), or EXOG (exonuclease / endonuclease G).

64. A method for estimating the concentration fraction of clinically relevant DNA molecules in a biological sample of a subject, the biological sample comprising the free clinically relevant DNA molecules and other DNA molecules, the method comprising: Storing a set of reference F-spectra, wherein each reference F-spectrum of the set: For each nucleotide in a set of nucleotides, identifying the proportion of free DNA molecules that terminate with the nucleotide; and Is associated with a type of fragmentation factor; Analyzing a plurality of free DNA molecules in the biological sample to obtain sequence reads, the sequence reads including end sequences corresponding to ends of the plurality of free DNA molecules; Determining a sample end motif spectrum by determining the proportion of the plurality of free DNA molecules that terminate with each nucleotide of the set of nucleotides based on the end sequences; Determining a proportional contribution of the set of reference F-spectra whose proportions aggregate to provide the sample end motif spectrum, wherein the set of reference F-spectra includes a first reference F-spectrum having a first proportional contribution, and wherein the proportional contributions sum to one; and Estimating a concentration fraction of the clinically relevant DNA molecules in the biological sample by comparing the first proportional contribution to one or more calibration values determined from one or more calibration samples, the concentration fractions of the clinically relevant DNA molecules of the one or more calibration samples being known.

65. The method according to claim 64, wherein the clinically relevant DNA molecules comprise fetal DNA or tumor DNA.

66. The method according to claim 64 or claim 70, wherein comparing the first proportional contribution with the one or more calibration values comprises comparing the first proportional contribution with a calibration curve comprising the one or more calibration values.

67. The method according to any one of claims 64 to 66, further comprising: For each calibration sample of the one or more calibration samples: Measuring the concentration fraction of the clinically relevant DNA molecules in the calibration sample; and Measuring the proportional contribution of a first reference end motif spectrum determined for the calibration sample, thereby determining one or more calibration values.

7. The method according to claim 6, wherein measuring the concentration fraction of the clinically relevant DNA molecule uses tissue-specific alleles or tissue-specific methylation patterns.

69. The method according to any one of claims 64 to 77, wherein estimating the concentration fraction of the clinically relevant DNA molecule in the biological sample comprises comparing one or more additional proportional contributions with one or more additional calibration values determined from the one or more calibration samples, the concentration fractions of the clinically relevant DNA molecules in the one or more calibration samples being known.

70. A method of analyzing a biological sample from a female subject carrying a fetus, the biological sample comprising cell-free DNA molecules from the female subject and the fetus, the method comprising: Storing a set of reference F-spectra, wherein each reference F-spectrum of the set: For each nucleotide in a set of nucleotides, identifying the proportion of free DNA molecules that terminate with the nucleotide; and Is associated with a type of fragmentation factor; Analyzing a plurality of free DNA molecules in the biological sample to obtain sequence reads, wherein the sequence reads include end sequences corresponding to ends of the plurality of free DNA molecules; Determining a sample end motif spectrum by determining the proportion of the plurality of free DNA molecules that terminate with each nucleotide of the set of nucleotides based on the end sequences; Determining a proportional contribution of the set of reference F-spectra whose proportions aggregate to provide the sample end motif spectrum, wherein the set of reference F-spectra includes a first reference F-spectrum having a first proportional contribution, and wherein the proportional contributions sum to one; and Estimating the gestational age of a fetus by comparing the first proportional contribution to one or more calibration values determined from one or more calibration samples having known gestational ages.

71. The method according to claim 70, wherein comparing the first proportional contribution with the one or more calibration values comprises comparing the first proportional contribution with a calibration curve comprising the one or more calibration values.

72. The method according to claim 70 or claim 71, further comprising: For each calibration sample of the one or more calibration samples: Estimating the gestational age in the calibration sample; and Measuring the proportional contribution of a calibration end motif spectrum determined for the calibration sample.

73. The method according to any one of claims 70 to 72, further comprising: Identifying that the plurality of free DNA molecules are from a fetus.

74. The method according to claim 73, wherein identifying the plurality of cell-free DNA molecules as being from the fetus uses a fetus-specific allele or a fetus-specific epigenetic marker.

75. The method according to any one of claims 70 to 74, wherein estimating the gestational age comprises comparing one or more additional proportional contributions with one or more additional calibration values determined from the one or more calibration samples, the gestational age of the one or more calibration samples being known.

76. The method according to any one of claims 64 and 70, wherein the first reference F-spectrum corresponds to a specific type of nuclease.

77. The method according to claim 76, wherein the specific type of nuclease is DNASE1L3.

78. A method for classifying a pathological grade in a biological sample of a subject, the biological sample comprising cell-free DNA, the method comprising: Storing a set of reference F-spectra, wherein each reference F-spectrum of the set: For each nucleotide in a set of nucleotides, identify the proportion of free DNA molecules that terminate with said nucleotide; and is associated with the type of fragmentation factor; Analyze a plurality of free DNA molecules in the biological sample to obtain sequence reads, wherein the sequence reads include end sequences corresponding to the ends of the plurality of free DNA molecules; Determine a sample end motif profile by determining, based on the end sequences, the proportion of the plurality of free DNA molecules that terminate with each nucleotide in the set of nucleotides; Determine the proportional contributions of a set of reference F-profiles whose proportional aggregations provide the sample end motif profile, wherein the sum of the proportional contributions is one; and Determine a classification of the pathological grade of the subject based on a determination that at least one of the proportional contributions exceeds a predetermined threshold.

79. The method according to claim 78, wherein the predetermined threshold is determined using one or more reference samples having a known classification of the pathological grade.

80. The method according to claim 78 or claim 79, wherein the pathology is cancer or an immune disorder.

81. The method according to claim 80, wherein the cancer is hepatocellular carcinoma, lung cancer, breast cancer, gastric cancer, glioblastoma multiforme, pancreatic cancer, colorectal cancer, nasopharyngeal cancer, and head and neck squamous cell carcinoma.

82. The method according to claim 80, wherein the classification is determined from a plurality of cancer grades including a plurality of cancer stages.

83. The method according to any one of claims 58 to 82, wherein the biological sample is a plasma or urine sample.

84. The method according to any one of claims 58 to 82, wherein each reference F-spectrum of the set of reference F-spectra specifies the proportion of cell-free DNA molecules terminated by each K-mer end motif of a set of K-mer end motifs, where K is two or more.

85. The method according to any one of claims 58 to 83, wherein the set of reference F-spectra is determined using cell-free DNA molecules obtained from one or more reference samples.

86. The method according to any one of claims 58 to 85, wherein analyzing the plurality of cell-free DNA molecules from the biological sample comprises: Receive sequence reads obtained from sequencing a plurality of free DNA molecules from the biological sample.

87. The method according to claim 86, wherein sequencing the plurality of cell-free DNA molecules comprises performing massively parallel sequencing, single molecule real-time sequencing, or nanopore sequencing on the plurality of cell-free DNA molecules.

88. The method according to any one of claims 58 to 87, wherein determining the proportional contribution of the set of reference F spectra comprises: Apply deconvolution to the sample end motif profile of the subject to determine the proportional contributions of a set of reference F-profiles.

89. The method according to claim 88, wherein, Applying the deconvolution includes: For a data matrix M of size W×F, where: (i) M represents the end frequencies of end motifs spanning the sample end motif profile, wherein the end frequencies of the data matrix M are determined based on the proportions of a plurality of free DNA molecules of the sample end motif profile; (ii) F represents the end frequencies of a set of reference F-profiles obtained from murine samples, wherein the F end frequencies are determined based on the proportions of free DNA molecules of the set of reference F-profiles; and (iii) W represents the relative weights corresponding to the proportional contributions of each of the set of reference F-profiles: Determine the proportional contributions of the set of reference F-profiles by determining W relative weights from the data matrix M and the set of reference F-profiles.

90. The method according to any one of claims 58 to 89, wherein analyzing the plurality of cell-free DNA molecules comprises receiving sequence reads obtained from sequencing the plurality of cell-free DNA molecules.

91. The method according to any one of claims 1 to 90, wherein the plurality of cell-free DNA molecules comprises at least 5,000 cell-free DNA molecules.

92. A computer product comprising a non-transitory computer-readable medium storing a plurality of instructions which, when executed, cause a computer system to perform the method according to any one of the preceding claims.

93. A system comprising: The computer product according to claim 92; and One or more processors for executing instructions stored on a computer-readable medium.

94. A system comprising means for performing any one of the above methods.

95. A system comprising one or more processors configured to perform any one of the above methods.

Images