Methods for longitudinally assessing donor-derived cell free DNA (DD-cfdna) in transplant recipients

Abstract

Disclosed herein includes a method for assessing an organ transplant by longitudinal monitoring of donor-derived cfDNA (dd-cfDNA) using a machine learning approach. Defining longitudinal dd-cfDNA trends is challenging due to differences in individual baselines and sampling frequency as well as disease-related fluctuations. Here we show that an unsupervised machine learning approach, based on dynamic time warping distances, enables differentiation of dd-cfDNA trajectories after rejection diagnosis into groups that correlate with outcomes, and provides an improved computerized systems for predicting an outcome of an organ transplant or treatment of organ transplant rejections. In addition, this disclosure provide approaches for combining dd-cfDNA measurements with clinical data to further assess organ transplant rejection and discriminate between rejection subtypes.

Description

N.069.W0.01METHODS FOR LONGITUDINALLY ASSESSING DONOR-DERIVED CELL FREE DNA (DD-CFDNA) IN TRANSPLANT RECIPIENTSCROSS-REFERENCE TO RELATED APPLICATIONS

[0001] The present application claims priority to U.S. Provisional Patent Application No.63 / 760,413, titled METHODS FOR LONGITUDINALLY ASSESSING DONOR-DERIVED CELL FREE DNA (DD-CFDNA) IN TRANSPLANT RECIPIENTS, filed February 19, 2025, which is hereby incorporated by reference herein in its entirety.BACKGROUND

[0002] The primary goal of organ transplantation is to improve the quality and life expectancy for patients with a dysfunctional organ. A major impediment to successful transplantation is persistent alloreactivity leading to graft dysfunction and graft loss. There is a lack of consensus on the optimal methods for monitoring of rejection. There is also substantial heterogeneity regarding the definition of rejection resolution and “successful” treatment. Evaluation of allograft function by serum creatinine levels can have a delayed response and be insensitive to the type of injury, while biopsy is not ideal for routine monitoring due to its invasive nature, risks of morbidity and considerable interobserver and sampling variability. Additionally, there is often discordance between functional and histological responses, which suggests that histopathology is not entirely reflective of allograft status.

[0003] Therefore, there remains a need for role of dd-cfDNA levels in monitoring organ transplants and responses to therapy of transplant rejection by identifying longitudinal dd-cfDNA trends.SUMMARY

[0004] In one aspect, the present disclosure provides a method for preparing a non-naturally occurring composition for assessing an organ transplant, comprising: longitudinally collecting two or more blood, plasma, serum or urine samples from an organ transplant recipient, extracting cell-free DNA (cfDNA) from the two or more blood, plasma, serum or urine samples or portions thereof from the organ transplant recipient, wherein the extracted cfDNA comprises aN.069.W0.01mixture of donor-derived cfDNA (dd-cfDNA) and recipient-derived cfDNA; preparing a sequencing library from the extracted cfDNA or its derivative and performing high-throughput sequencing on the sequencing library to obtain sequence reads; quantifying an amount of dd-cfDNA or a percentage of dd-cfDNA out of total cfDNA based on the sequence reads; and assessing the organ transplant by generating time series data of the amount of dd-cfDNA or the percentage of dd-cfDNA out of total cfDNA, wherein the time series data of the amount of dd-cfDNA or the percentage of dd-cfDNA out of total cfDNA are input for a machine learning (ML) model, wherein the ML model improves upon the use of a computerized system for determining likely outcomes by assigning the time series data for the organ transplant recipient to a cluster classification generated by the ML model from training data, and wherein the cluster classification provides a likelihood of a positive outcome of the organ transplant.

[0005] In some embodiments, a first sample of the two or more blood, plasma, serum or urine samples is collected prior to a scheduled biopsy or diagnostic test of the transplant recipient, and a second sample of the two or more blood, plasma, serum or urine samples is collected after the scheduled biopsy or diagnostic test.

[0006] In some embodiments, the organ transplant recipient suffers from a transplant rejection condition, wherein a first sample of the two or more blood, plasma, serum or urine samples is collected prior to starting a treatment for the transplant rejection condition, and a second sample of the two or more blood, plasma, serum or urine samples is collected after starting the treatment for the transplant rejection condition.

[0007] In some embodiments, the percentage of dd-cfDNA out of total cfDNA for each of the two or more longitudinally collected blood, plasma, serum or urine samples are categorized into two or more time periods.

[0008] In some embodiments, the two or more time periods comprise a first time period within 7 days prior to (i) a scheduled biopsy or diagnostic test for transplant rejection, or (ii) starting the treatment for the transplant rejection condition.

[0009] In some embodiments, a pairwise distance matrix of the percentage of dd-cfDNA out of total cfDNA values at each of the two or more time periods is generated.N.069.W0.01

[0010] In some embodiments, the pairwise distance matrix is the input for the ML model.

[0011] In some embodiments, the ML model is an unsupervised clustering algorithm.

[0012] In some embodiments, the pairwise distance matrix comprises a distance measure, and wherein the distance measure is dynamic time warping (DTW).

[0013] In some embodiments, the cluster classification was generated from time series data of a population of transplant recipients as training data.

[0014] In some embodiments, the cluster classification for the organ transplant recipient is performed by comparing the DTW of the time series data of the organ transplant recipient to the DTW of the time series data of the population of transplant recipients

[0015] In some embodiments, the cluster classification indicates a likelihood of a positive outcome or a negative outcome of the organ transplant or a treatment for transplant rejection.

[0016] In some embodiments, the unsupervised clustering algorithm is a k-Medoids clustering algorithm.

[0017] In some embodiments, a cluster number is based on a resultant silhouette score and a visual cluster separation.

[0018] In some embodiments, the cluster number is 2-10, preferably wherein the cluster number is 4.

[0019] In some embodiments, the cluster number is 4, and the clusters comprise i) low, ii) drop, iii) mid, and iv) high, wherein low and drop are indicative of a positive outcome, and wherein mid and high are indicative of a negative outcome.

[0020] In some embodiments, i) low indicates that the percentage of dd-cfDNA out of total cfDNA is below 1% in the first sample and remains below 1% thereafter, ii) drop indicates that the percentage of dd-cfDNA out of total cfDNA is above 1% in the first sample, drops below 1% within a time period, and remains below 1% thereafter, ii) mid indicates that the percentage of dd-cfDNA out of total cfDNA is above 1% in the first sample, drops below 1% within a timeN.069.W0.01period, and increases to above 1% after a second time period, and iv) high indicates that the percentage of dd-cfDNA out of total cfDNA is above 1% in the first sample and remains above 1% thereafter.

[0021] In some embodiments, the positive outcome is improved or restored organ function; and / or wherein the positive outcome comprises absence of death censored graft loss, subsequent rejection, and / or donor- specific antibodies (DSA).

[0022] In some embodiments, the negative outcome is an occurrence of death censored graft loss, a subsequent rejection, and / or a presence of donor- specific antibodies (DSA).

[0023] In some embodiments, the organ transplant is from a human.

[0024] In some embodiments, the organ transplant is a xenotransplant.

[0025] In some embodiments, the organ transplant is from a pig, a primate, a baboon, a cow, or a dog, preferably from a pig.

[0026] In some embodiments, preparing a sequencing library comprises performing universal amplification on the extracted cell-free DNA or its derivative.

[0027] In some embodiments, preparing a sequencing library comprises performing targeted enrichment on the extracted cell-free DNA or its derivative to enrich a plurality of polymorphic target loci, preferably a plurality of SNP loci.

[0028] In some embodiments, the organ transplant is one or more organs selected from the group consisting of heart, kidney, liver, lung, pancreas, and intestine.

[0029] In some embodiments, the organ transplant is kidney.

[0030] In some embodiments, the transplant recipient suffers from an antibody-mediated allograft rejection as classified by a histologic method or molecular methods. In some embodiments, the histologic method is the Banff 2019 Classification.

[0031] In some embodiments, the two or more blood, plasma, serum or urine samples are collected over a period of 8 weeks.N.069.W0.01

[0032] In some embodiments, the two or more time periods comprise a first time period comprising 7 days prior to a scheduled biopsy or diagnostic test of a transplant rejection, or starting the treatment for the transplant rejection condition, and the second time period is within 3 weeks after the scheduled biopsy or diagnostic test, or starting the treatment for the transplant rejection condition. In some embodiments, the method further comprises a third time period that is after the second time period. In some embodiments, longitudinally collecting two or more blood, plasma, serum or urine samples from an organ transplant recipient comprises collecting at a first time period comprising 7 days prior to a scheduled biopsy or diagnostic test of a transplant rejection, or starting the treatment for the transplant rejection condition, and a second time period within 3 weeks after the scheduled biopsy or diagnostic test, or starting the treatment for the transplant rejection condition.

[0033] In some embodiments, the quantifying step comprises: a) adding a tracer DNA to extracted DNA or its derivative to obtain a mixed composition; b) performing targeted multiplex amplification on the mixed composition comprising the extracted DNA or its derivative and the tracer DNA to amplify 100 to 20,000 different polymorphic target loci together in the same reaction volume using 100 to 20,000 different target- specific primers; c) sequencing the amplicons by high-throughput sequencing to generate sequence reads; and d) quantifying the amount of donor-derived cell-free DNA and the amount of total cell-free DNA from the sequence reads, wherein the amount of total cell-free DNA is quantified using sequence reads derived from the tracer DNA.

[0034] In some embodiments, the organ transplant recipient has a kidney transplant and has biopsy-proven acute rejection (BPAR). In some embodiments, the two or more samples are collected within one week prior to diagnosis of BPAR, and at weeks 2, 4, 6, and 8 after the BPAR diagnosis. In some embodiments, the organ transplant recipient received immunosuppressive treatment after the BPAR diagnosis. In some embodiments, the treatment for the transplant rejection condition comprises anti-CD38 monoclonal antibodies.

[0035] In some embodiments, the time series data input for the ML model is generated from the percentage of dd-cfDNA out of total cfDNA. In some embodiments, the time series data input for the ML model is generated from the amount of dd-cfDNA.N.069.W0.01

[0036] In some embodiments, the amount of dd-cfDNA or the percentage of dd-cfDNA out of total cfDNA is combined with clinical data to determine the likelihood of a positive outcome of the organ transplant. In some embodiments, the method further comprises determining T cell-mediated rejection (TCMR) rejection, antibody-mediated rejection (AB MR) rejection, and / or mixed rejection based on combining the amount of dd-cfDNA or the percentage of dd-cfDNA out of total cfDNA with clinical data.

[0037] In some embodiments, the method further comprises using a Two-Stage model comprising a first stage caller and a second stage caller, wherein the first stage caller provides a risk score of transplant rejection, and if the risk score from the first stage caller is above a threshold value, then the second stage caller provides a TCMR risk score for T cell-mediated rejection (TCMR) rejection, an AB MR risk score for antibody-mediated rejection (ABMR) rejection, and / or a Mixed risk score for mixed rejection. In some embodiments, the first stage caller comprises combining the amount of dd-cfDNA or the percentage of dd-cfDNA out of total cfDNA or longitudinal changes of the amount of dd-cfDNA or the percentage of dd-cfDNA out of total cfDNA with clinical data. In some embodiments, the second stage caller comprises combining longitudinal changes of the amount of dd-cfDNA or the percentage of dd-cfDNA out of total cfDNA with clinical data.

[0038] In some embodiments, the clinical data comprises prior transplant rejection history, demographic data, a plurality of immunological risk factors, vital signs, a plurality of organ function markers, imaging results, inflammatory markers, a gene expression profile indicative of transplant rejection, a plurality of serology test markers, and / or a plurality of biomarkers indicative of transplant rejection. In some embodiments, the demographic data comprises age or ethnicity. In some embodiments, the plurality of serology test markers comprises donor-specific antibodies (DSA). In some embodiments, the plurality of organ function markers comprises an amount of serum creatinine.

[0039] In some embodiments, a machine learning model is used to combine data obtained from the amount of dd-cfDNA or the percentage of dd-cfDNA out of total cfDNA with the clinical data. In some embodiments, the machine learning model provides (i) a risk score of transplant rejection, and / or (ii) a TCMR risk score for T cell-mediated rejection (TCMR) rejection, anN.069.W0.01ABMR risk score for antibody-mediated rejection (ABMR) rejection, and / or a Mixed risk score for mixed rejection. In some embodiments, the machine learning model comprises a temporal random forest model. In some embodiments, the first stage caller and the second stage caller are temporal random forest models.

[0040] In another aspect, the present disclosure relates to a method for preparing a non-naturally occurring composition for assessing an organ transplant, comprising: longitudinally collecting two or more blood, plasma, serum or urine samples from an organ transplant recipient, extracting cell-free DNA (cfDNA) from the two or more blood, plasma, serum or urine samples or portions thereof from the organ transplant recipient, wherein the extracted cfDNA comprises a mixture of donor-derived cfDNA (dd-cfDNA) and recipient-derived cfDNA; preparing a sequencing library from the extracted cfDNA or its derivative and performing high-throughput sequencing on the sequencing library to obtain sequence reads; quantifying an amount of dd-cfDNA or a percentage of dd-cfDNA out of total cfDNA based on the sequence reads, wherein the organ transplant is assessed by combining the amount of dd-cfDNA or the percentage of dd-cfDNA out of total cfDNA with clinical data from the organ transplant recipient.

[0041] In some embodiments, the method further comprises determining longitudinal dd-cfDNA changes based on the amount of dd-cfDNA or the percentage of dd-cfDNA out of total cfDNA. In some embodiments, the method further comprises determining a TCMR risk score for T cell-mediated rejection (TCMR) rejection, an ABMR risk score for antibody-mediated rejection (ABMR) rejection, and / or a Mixed risk score for mixed rejection based on combining (i) the amount of dd-cfDNA or the percentage of dd-cfDNA out of total cfDNA or longitudinal dd-cfDNA changes with (ii) clinical data from the organ transplant recipient.

[0042] In some embodiments, the second stage caller comprises combining longitudinal dd-cfDNA changes with clinical data. In some embodiments, the first stage caller comprises combining the amount of dd-cfDNA or the percentage of dd-cfDNA out of total cfDNA with prior transplant rejection history. In some embodiments, the second stage caller comprises combining longitudinal dd-cfDNA changes with demographic data and a plurality of organ function markers to determine a TCMR risk score for T cell-mediated rejection (TCMR) rejection. In some embodiments, the second stage caller comprises combining the amount of dd-N.069.W0.01cfDNA or the percentage of dd-cfDNA out of total cfDNA with prior ABMR history to determine an ABMR risk score. In some embodiments, combining the amount of dd-cfDNA or the percentage of dd-cfDNA out of total cfDNA with clinical data from the organ transplant recipient is performed by using a statistical model.

[0043] In some embodiments, determining TCMR rejection, ABMR rejection, and / or mixed rejection is performed simultaneously.BRIEF DESCRIPTION OF THE DRAWINGS

[0044] FIG. 1 shows a cluster classification of longitudinal trends in donor-derived cfDNA (dd-cfDNA) after treatment of transplant rejection. FIG. 1A shows a time series ribbon plot depicting percentage of donor-derived cfDNA (dd-cfDNA) out of total cfDNA from longitudinally collected samples pre and post biopsy of kidney transplant recipient. The time series were cluster classified by using a machine learning algorithm. FIG. IB is an abstraction of the trends shown in FIG. IB.

[0045] FIG.2 is a Sankey diagram showing associations of rejection types with dd-cfDNA trends, and dd-cfDNA trends and outcomes.

[0046] FIG. 3 is a graphic outline of the study timeline.

[0047] FIG.4 is a consort diagram of the study of kidney transplant recipients.

[0048] FIG. 5 is a box plot showing median dd-cfDNA (%) across biopsy findings at the time of index biopsy.

[0049] FIG.6 is a box plot showing dd-cfDNA levels at the time of index biopsy among allograft pathology classifications and sub-classifications.

[0050] FIG.7 is a box plot showing changes in dd-cfDNA (%) after rejection diagnosis and treatment.

[0051] FIG.8 shows longitudinal trends in dd-cfDNA (%) after rejection diagnosis. FIG. 8A) is a box plot showing longitudinal trends in dd-cfDNA (%) after rejection diagnosis in differentN.069.W0.01categories of rejection. FTG. 8B) is a time series plot showing differences in dd-cfDNA (%) over time after rejection diagnosis. Wilcoxon signed-rank test with FDR correction for multiple testing: ****P < le-04, ***P < 0.001, **P < 0.01, *P < 0.05, and ns: P > 0.05.

[0052] FIG. 9 is a bar graph showing proportion of samples of each rejection type assigned to each cluster.

[0053] FIG. 10 is a Sankey plot showing association of rejection types to clusters, and clusters to outcomes specifically for (A) AB MR, (B) TCMR and (C) Mixed Rejection cases.

[0054] FIG. 11 illustrates differences between Euclidean distance and dynamic time warping (DTW) distance as applied to time-series data (e.g., dd-cfDNA). Euclidean (linear) distance requires a one-to-one matching of samples, while DTW distance allows for comparison between sequences with differing lengths and sampling frequencies.

[0055] FIG. 12 presents a clustermap of calculated pairwise DTW distances between subjects’ dd-cfDNA trajectories. Warmer colors indicate greater similarity (shorter DTW distance) between the time series of two subjects, while cooler colors indicate lower similarity (larger DTW distance). Dendrograms indicate clusters of subjects with similar dd-cfDNA trajectories as identified by the k-Medoids clustering algorithm.

[0056] FIG. 13 shows a schematic illustration of the 2 stage prediction model. The left side of the vertical line shows the initial calling by dd-cfDNA measurements and using the two-threshold caller. The right side of the vertical line shows the 2 stage prediction model that combines dd-cfDNA measurements with additional data. *The additional data can include data from the 2TC model, patient age and ethnicity, rejection history, serum creatinine and donorspecific antibody (DSA) measurements.

[0057] FIG. 14 shows a schematic illustration of the Two-Stage Model Architecture. Stage 1 inputs and outputs are shown on the left side. Samples that are assigned high-risk by both the 2TC model and Stage 1 of the Two-Stage Model are further analyzed by Stage 2, which predicts TCMR and ABMR subtype rejections. The final output of the Two-Stage Model includes final labeling asN.069.W0.01No 2 Stage call, High-Risk (TCMR / ABMR / Type No-Call), and risk scores for 2 Stage Call, AB MR, TCMR.

[0058] FIG. 15 shows a schematic illustration of the information flow from the blood draws and 2TC calls at the left, inputting data from the 2TC assay into the Two-Stage Model, predicting rejection in Stage 1, and predicting subtype rejections in Stage 2 for samples deemed high risk by 2TC and Stagel.

[0059] FIG. 16 is depict ROC curves for the Two Stage Model applied to high risk kidney transplant recipients. The curves shows strong subtype discrimination.

[0060] FIG. 17 is a confusion matrix showing predicted classifications versus true classifications. The rows represent the true categories, and the columns represent the predicted categories. Values within the matrix indicate the number of instances for each true-predicted pair, with a color scale reflecting the magnitude of the counts. This confusion matrix illustrates that the integrated dd-cfDNA measurements and Two-Stage model can noninvasively and accurately differentiate between: no rejection, TCMR: T cell-mediated rejection, and ABMR: antibody-mediated rejection.DETAILED DESCRIPTION

[0061] The present disclosure is generally directed relates to a method for preparing a non-naturally occurring composition for assessing an organ transplant, comprising: longitudinally collecting two or more blood, plasma, serum or urine samples from an organ transplant recipient, extracting cell-free DNA (cfDNA) from the two or more blood, plasma, serum or urine samples or portions thereof from the organ transplant recipient, wherein the extracted cfDNA comprises a mixture of donor-derived cfDNA (dd-cfDNA) and recipient-derived cfDNA; preparing a sequencing library from the extracted cfDNA or its derivative and performing high-throughput sequencing on the sequencing library to obtain sequence reads; quantifying an amount of dd-cfDNA or a percentage of dd-cfDNA out of total cfDNA based on the sequence reads: and assessing the organ transplant by generating time series data of the amount of dd-cfDNA or the percentage of dd-cfDNA out of total cfDNA, wherein the time series data of the amount of dd-cfDNA or the percentage of dd-cfDNA out of total cfDNA are input for a machine learning (ML)N.069.W0.01model, wherein the ML model improves upon the use of a computerized system for determining likely outcomes by assigning the time series data for the organ transplant recipient to a cluster classification generated by the ML model from training data, and wherein the cluster classification provides a likelihood of a positive outcome of the organ transplant.

[0062] Longitudinal collection of samples

[0063] The present methods include longitudinally collecting two or more blood, plasma, serum or urine samples from an organ transplant recipient. In some embodiments, a first sample of the two or more blood, plasma, serum or urine samples is collected prior to a scheduled biopsy or diagnostic test of the transplant recipient, and a second sample of the two or more blood, plasma, serum or urine samples is collected after the scheduled biopsy or diagnostic test. In some embodiments, a sample is collected every week for 8 weeks starting one week prior to the scheduled biopsy or diagnostic test for transplant rejection. In some embodiments, a sample is collected every two weeks for 8 weeks starting one week prior to the scheduled biopsy or diagnostic test for transplant rejection. In some embodiments, a sample is collected every three weeks for 8 weeks starting one week prior to the scheduled biopsy or diagnostic test for transplant rejection. In some embodiments, a sample is collected every week for up to 12 weeks starting one week prior to the scheduled biopsy or diagnostic test for transplant rejection. In some embodiments, a sample is collected every two weeks for up to 12 weeks starting one week prior to the scheduled biopsy or diagnostic test for transplant rejection. In some embodiments, a sample is collected every three weeks for up to 12 weeks starting one week prior to the scheduled biopsy or diagnostic test for transplant rejection.

[0064] In some embodiments, the organ transplant recipient suffers from a transplant rejection condition, wherein a first sample of the two or more blood, plasma, serum or urine samples is collected prior to starting a treatment for the transplant rejection condition, and a second sample of the two or more blood, plasma, serum or urine samples is collected after starting the treatment for the transplant rejection condition. In some embodiments, the organ transplant recipient suffers from a transplant rejection condition, wherein a first sample of the two or more blood, plasma, serum or urine samples is collected within one-two weeks or seven to 14 days prior to starting a treatment for the transplant rejection condition, and a second sample of the two or more blood,N.069.W0.01plasma, serum or urine samples is collected within one week, two weeks, or three weeks after starting the treatment for the transplant rejection condition. In some embodiments, the organ transplant recipient suffers from a transplant rejection condition, wherein a first sample of the two or more blood, plasma, serum or urine samples is collected within seven days prior to starting a treatment for the transplant rejection condition, and a second sample of the two or more blood, plasma, serum or urine samples is collected within three weeks after starting the treatment for the transplant rejection condition.

[0065] In some embodiments, the two or more time periods comprise a first time period comprising 7 days prior to a scheduled biopsy or diagnostic test of a transplant rejection, or starting the treatment for the transplant rejection condition, and the second time period is within 3 weeks after the scheduled biopsy or diagnostic test, or starting the treatment for the transplant rejection condition. In some embodiments, the method further comprises a third time period that is after the second time period.

[0066] In some embodiments, the two or more blood, plasma, serum or urine samples are collected over a period of 4 weeks, 6, weeks, 8 weeks, 10 weeks, 12 weeks, 6 months, or 12 months. In some embodiments, the two or more blood, plasma, serum or urine samples are collected over a period of 8 weeks.

[0067] Transplant Recipient

[0068] In some embodiments, the organ transplant is from a human. In some embodiments, the organ transplant is a xenotransplant. In some embodiments, the organ transplant is from a pig, a primate, a baboon, a cow, or a dog, preferably from a pig. In some embodiments, preparing a sequencing library comprises performing universal amplification on the extracted cell-free DNA or its derivative. In some embodiments, preparing a sequencing library comprises performing targeted enrichment on the extracted cell-free DNA or its derivative to enrich a plurality of polymorphic target loci, preferably a plurality of SNP loci. In some embodiments, the organ transplant is one or more organs selected from the group consisting of heart, kidney, liver, lung, pancreas, and intestine. In some embodiments, the organ transplant is kidney.N.069.W0.01

[0069] In some embodiments, the transplant recipient suffers from an antibody-mediated allograft rejection as classified by a histologic method or molecular methods. In some embodiments, the histologic method is the Banff 2019 Classification. In some embodiments, the organ transplant recipient has a kidney transplant and has biopsy-proven acute rejection (BPAR). In some embodiments, the two or more samples are collected within one week prior to diagnosis of BPAR, and at weeks 2, 4, 6. and 8 after the BPAR diagnosis. In some embodiments, the organ transplant recipient received immunosuppressive treatment after the BPAR diagnosis. In some embodiments, the treatment for the transplant rejection condition comprises anti-CD38 monoclonal antibodies. In some embodiments, organ transplant has been treated or reconditioned prior to the transplantation.

[0070] Transplant rejection occurs when a recipient's immune system attacks a transplanted organ or tissue, causing it to malfunction due to the body recognizing the new tissue as foreign; symptoms can include pain or tenderness around the transplant site, fever, flu-like symptoms, decreased organ function, swelling, and fatigue, depending on the transplanted organ, and often require prompt medical attention to manage with increased immunosuppressant medication: diagnosis can involve a biopsy of the transplanted tissue to confirm rejection.

[0071] Hyperacute rejection (HR) can occur rapidly within minutes to hours due to pre-existing antibodies against the donor tissue. Acute rejection (AR) can happen within days to weeks after transplant, often presenting with noticeable symptoms like fever, pain, and decreased organ function. Chronic rejection (CR) develops slowly over time, leading to gradual decline in organ function with less obvious symptoms.

[0072] Kidney transplant rejection can cause decreased urine output, swelling in legs, pain in the kidney area. Liver transplant rejection can be associated with Jaundice, abdominal pain, and nausea. Heart transplant can be associated with shortness of breath, chest pain, fatigue. Lung transplant can be associated with cough, difficulty breathing, chest tightness

[0073] Transplant rejection conditions can include ABMR (Antibody-Mediated Rejection) or acute T cell-mediated rejection (TCMR). In TCMR, the immune system's T cells attack the transplanted organ, causing damage and potential loss of function. ABMR is primarily driven byN.069.W0.01antibodies produced by the immune system. Both conditions can be diagnosed through kidney biopsy and require specific immunosuppressive therapies to manage.

[0074] Analytical methods

[0075] The longitudinally collected samples provide data for generating time series. The data can be “bucketed” or categorized into time periods. For example, any sample collected within 7 days prior to a scheduled biopsy or diagnostic test of the transplant recipient can be categorized as an index sample or a first sample. In some embodiments, the percentage of dd-cfDNA out of total cfDNA for each of the two or more longitudinally collected blood, plasma, serum or urine samples are categorized into two or more time periods. In some embodiments, the two or more time periods comprise a first time period within 7 days prior to (i) a scheduled biopsy or diagnostic test for transplant rejection, or (ii) starting the treatment for the transplant rejection condition.

[0076] Based on the time periods, a pairwise distance matrix of the percentage of dd-cfDNA out of total cfDNA values at each of the two or more time periods can be generated.

[0077] In some embodiments, the pairwise distance matrix is the input for the ML model. In some embodiments, the ML model is an unsupervised clustering algorithm. In some embodiments, the pairwise distance matrix comprises a distance measure, and wherein the distance measure is dynamic time warping (DTW). In some embodiments, the cluster classification was generated from time series data of a population of transplant recipients as training data. In some embodiments, the cluster classification for the organ transplant recipient is performed by comparing the DTW of the time series data of the organ transplant recipient to the DTW of the time series data of the population of transplant recipients.

[0078] In some embodiments, the cluster classification indicates a likelihood of a positive outcome or a negative outcome of the organ transplant or a treatment for transplant rejection. In some embodiments, the unsupervised clustering algorithm is a k-Medoids clustering algorithm.

[0079] In some embodiments, a cluster number is based on a resultant silhouette score and a visual cluster separation.N.069.W0.01

[0080] In some embodiments, the cluster number is 2-10, preferably wherein the cluster number is 4.

[0081] In some embodiments, the cluster number is 4, and the clusters comprise i) low, ii) drop, iii) mid, and iv) high, wherein low and drop are indicative of a positive outcome, and wherein mid and high are indicative of a negative outcome. In some embodiments, i) low indicates that the percentage of dd-cfDNA out of total cfDNA is below 1% in the first sample and remains below 1% thereafter, ii) drop indicates that the percentage of dd-cfDNA out of total cfDNA is above 1% in the first sample, drops below 1% within a time period, and remains below 1% thereafter, ii) mid indicates that the percentage of dd-cfDNA out of total cfDNA is above 1% in the first sample, drops below 1% within a time period, and increases to above 1% after a second time period, and iv) high indicates that the percentage of dd-cfDNA out of total cfDNA is above 1% in the first sample and remains above 1% thereafter.

[0082] Definition of outcomes: A composite endpoint, termed “Negative Outcome”, was defined as one or more of the following 4 components: 1) the presence of death-censored graft loss, 2) the development of a subsequent BPAR episode, 3) the presence of persistently elevated donor specific antibodies (DSA) levels at the time of biopsy, or detection of de novo DSA during the follow-up period, and 4) lack of resolution of renal dysfunction, defined as either a lack of improvement of in eGFR (at least 5% from the baseline measurement at the time of BPAR), or a proteinuria / creatinine ratio >0.5 mg / mg. A “Positive Outcome” was defined by the absence of all 4 events in the composite endpoint. eGFR was calculated using the 2021 CKD-EPI formula without race. Inker et al., Am J Kidney Dis.; 78(5):736-749 (2021). The outcome can be determined at 1 year after the biopsy or diagnostic test of transplant rejection, or 1 year after starting treatment of transplant rejection condition. The outcome can be determined at 6 months after the biopsy or diagnostic test of transplant rejection, or 6 months after starting treatment of transplant rejection condition. The outcome can be determined at 18 months after the biopsy or diagnostic test of transplant rejection, or 18 months after starting treatment of transplant rejection condition. The outcome can be determined at 6-24 months after the biopsy or diagnostic test of transplant rejection, or 6-24 months after starting treatment of transplant rejection condition. The outcome can be determined at 12-24 months after the biopsy or diagnostic test of transplant rejection, or 12-24 months after starting treatment of transplant rejection condition. The outcomeN.069.W0.01can be determined at 18-24 months after the biopsy or diagnostic test of transplant rejection, or 18-24 months after starting treatment of transplant rejection condition. The outcome can be determined at 6-36 months after the biopsy or diagnostic test of transplant rejection, or 6-36 months after starting treatment of transplant rejection condition. In some embodiments, “outcomes” or “likely outcomes” refers to a current or ongoing state of the organ. As used herein, “outcomes” are not limited to an event after the test is run.

[0083] In some embodiments, the positive outcome is improved or restored organ function; and / or wherein the positive outcome comprises absence of death censored graft loss, subsequent rejection, and / or donor- specific antibodies (DSA).

[0084] In some embodiments, the negative outcome is an occurrence of death censored graft loss, a subsequent rejection, and / or a presence of donor- specific antibodies (DSA).

[0085] The above methods can be implemented on a computer or a computerized system. Some embodiments include electronic components of a computer or computer system, such as microprocessors, storage and memory that store computer program instructions in a computer-readable storage medium. Many of the features described in this specification can be implemented as processes that are specified as a set of program instructions encoded on a computer-readable storage medium. Computer systems used in connection with embodiments of the present disclosure can have other capabilities not specifically described here. When these program instructions are executed by one or more processing units, they cause the processing unit(s) to perform various operation indicated in the program instructions. Examples of program instructions or computer code include machine code, such as is produced by a compiler, and files including higher-level code that are executed by a computer, an electronic component, or a microprocessor using an interpreter.

[0086] Extraction of cfDNA

[0087] In some embodiments, the methods described herein comprise extracting cell-free DNA from a liquid sample or portion thereof of an organ transplant recipient, wherein the extracted cell-free DNA comprises a mixture of donor-derived cell-free DNA and recipient-derived cell-free DNA. Methods that are particularly useful in exemplary embodiments include methods forN.069.W0.01isolating circulating free DNA (cfDNA) from a liquid sample, and in illustrative embodiments from a blood, serum, plasma, or urine.

[0088] In certain illustrative embodiments, isolation of cfDNA from a liquid (e.g., blood or blood derivative sample such as a plasma sample) can involve binding DNA molecules from a sample to a matrix and isolating the DNA molecules in the presence of a solvent. In some embodiments, the method further comprises incubating the biological sample comprising DNA molecules with a protease, prior to contacting the DNA molecules to the matrix. In some embodiments, the method can further include the steps of washing the matrix with a wash buffer to remove impurities, and optionally, drying the matrix. Enriched nucleic acid samples can be eluted from the matrix with an elution buffer.

[0089] Other methods for nucleic acid isolation, for example cfDNA isolation, and optional enrichment of certain cfDNA can include ion exchange columns, or microfluidic devices, such as solid phase isolation, based on DNA capture by immobilized beads or functionalized surface. Additional methods include liquid phase isolation, utilizing an electric field, or chemical reagents, instead of a functionalized surface. In illustrative embodiments herein, isolation of cfDNA from a patient sample is performed using a DNA isolation kit (e.g., QIAamp Circulating Nucleic Acid kit (Qiagen)).

[0090] In some embodiments, cfDNA or their derivatives of certain sizes can be enriched before or after subjecting the cfDNA to methods herein. In some embodiments, size selection can be performed before the sequencing library preparation. In some embodiments, size selection can be performed after the sequencing library preparation and before sequencing. In some embodiments, size selection is performed on a sequencing-ready pool. Enriched cfDNA molecules can be, for example 50 to 1200 base pairs in length, 70 to 500 base pairs in length, 100 to 200 base pairs in length, or 130 to 170 base pairs in length. In some embodiments, the enriched cfDNA molecules are from 50 to 200 bp in length. In some embodiments, the enriched cfDNA molecules are between 60 and 200 bp in length, between 60 and 150 bp in length, or between 80 and 140 bp in length, in length before the enriched cfDNA molecules, or derivatives thereof, are ligated to adapters in methods herein. In some embodiments, the enriched cfDNA molecules are about or less than 200, 175, 166, 150, 100 bp in length before they are ligated toN.069.W0.01adapters. Such enrichment methods can be performed for example using the methods described in WO2018 / 156418 and WO2019161244, each of which is incorporated herein by reference in its entirety.

[0091] Preparation of Sequencing Library

[0092] In some embodiments, the method described herein comprises preparing a sequencing library from the extracted cell-free DNA or its derivative. The preparation of the sequencing library can comprise, for example, appending an adapter to the extracted cell-free DNA or its derivative. The preparation of the sequencing library can comprise, for example, performing a universal amplification on the extracted cell-free DNA or its derivative. The preparation of the sequencing library can comprise, for example, performing targeted enrichment on the extracted cell-free DNA or its derivative to enrich a plurality of polymorphic target loci, preferably a plurality of SNP loci.

[0093] Library Preparation

[0094] In some embodiments, the method further comprises appending an adapter to the extracted cfDNA or derivative thereof and generating adapted DNA before performing targeted enrichment (e.g., with a panel of oligonucleotide probes). In some embodiments, the adapter is a Y-adapter. In some embodiments, the adapter comprises a universal priming site.

[0095] In some embodiments, the adapter comprises a molecular barcode or index sequence, wherein sequence reads generated from the high-throughput sequencing can be grouped together using the molecular barcode or index sequence. In some embodiments, the adapters do not comprise a molecular barcode or index sequence, wherein sequence reads generated from the high-throughput sequencing can be grouped together using the fragment-end sequence of the extracted cell-free DNA or DNA derived therefrom. In some embodiments, the sequence reads that are grouped together using the molecular barcode or index sequence and / or the fragment-end sequence can be subject to error correction to correct sequencing errors and generate a consensus sequence.N.069.W0.01

[0096] In some embodiments, the method further comprises amplifying the adapted DNA using a primer that binds to the universal primer binding site and generating adapted- amplified DNA before performing targeted enrichment (e.g., with a panel of oligonucleotide probes). In some embodiments, the adapted- amplified DNA further comprises a sequencing adapter sequence or sequencing primer binding site for high-throughput sequencing. In some embodiments, the adapted- amplified DNA further comprises a sample barcode or index sequence, which allows multiplexed sequencing of pooled sequencing libraries (e.g., multiplexed sequencing of sequencing libraries generated from multiple samples).

[0097] Typically, methods herein include a step of appending nucleic acid adapters to extracted cfDNA molecules or to nucleic acid derivatives generated therefrom. For example, adapters may be appended on to the DNA molecules by ligation. The extracted cfDNA molecules in illustrative embodiments are extracted from a sample of a subject. In some embodiments, appending nucleic acid adapters is performed after the extracted cfDNA molecules are fragmented to form fragmented DNA molecules. Typically, methods include exposing the extracted cfDNA molecules to one or more polymerases or kinases, such as Klenow Large Fragment Polymerase and T4 polynucleotide kinase (PNK), as well as a ligase, such as T4 ligase. In some embodiments, the extracted cfDNA molecules or the fragmented DNA molecules are exposed to one or more polymerases and / or kinases to generate the nucleic acid derivatives. In some embodiments, the method further comprises appending adapters to the nucleic acid derivatives generated therefrom. In some embodiments, the extracted cfDNA molecules are not fragmented prior to appending nucleic acid adapters thereto.

[0098] In some embodiments, adapters are ligated to the extracted cfDNA molecules. In some embodiments, before such ligation, the extracted cfDNA molecules can be modified to form sample nucleic acid derivatives, for example to make them more amenable to adapter ligation. For example, extracted cfDNA molecules can be blunt ended, nucleotides can be added to the extracted cfDNA molecules or blunted-ended derivative therefrom, and / or phosphate moieties can be added or removed from the ends of DNA molecules or derivatives thereof. In some embodiments, prior to ligation, the extracted cfDNA molecules may be blunt ended, and then a single adenosine base can be added to the 3’ end. In some embodiments, prior to ligation the DNA may be cleaved using a restriction enzyme or some other cleavage method. In someN.069.W0.01embodiments, during ligation the 3’ adenosine of the DNA fragments and the complementary 3’ thymidine overhang of an adapter can enhance ligation efficiency. In some embodiments, adapter ligation is performed using a T4 ligase.

[0099] In some embodiments, adapters containing one or more universal priming sequences are utilized in methods herein. In some embodiments, the adapters are Y adapters, for example in methods involving NGS sequencing. In some embodiments, the adapters each comprises a universal priming site. In some embodiments, the adapter may comprise a barcode or index sequence. Thus, multiple samples can be analyzed in the same sequencing reaction. The sample barcode or index sequence can be used to process data according to the sample from which the data was generated.

[0100] In some embodiments, the adapter may comprise a molecular barcode or index sequence. In some embodiments, the number of adapters having different molecular barcode or index sequences is between 10 to 1,000, and wherein the ratio of the total number of template nucleic acid or cfDNA molecules to the number of different molecular barcode or index sequences in the ligation reaction is at least 1,000:1. The number of different molecular barcode or index sequences in the ligation reaction, in certain embodiments, ranges fromlO to 50, 10 to 100, 50 to 200, 100 to 300, 200 to 500, 300 to 600, 500 to 700, 600 to 800 or 700 to 1,000. In some embodiments, there are at least 1, 10, 20, 30, 40, 50, or at least 100; 200, 300, 400, 500, 600, 700, 800, 900, or 1000 different molecular barcode or index sequences in the ligation reaction. In some embodiments, the ratio of the total number of template nucleic acid or cfDNA molecules to the number of different molecular barcode or index sequences in the ligation reaction is at least 10,000:1. In some embodiments, the ratio of the total number of template nucleic acid or cfDNA molecules to the number of different molecular barcode or index sequences in the ligation reaction ranges from 50,000: 1 to 50:1, from 25,000: 1 to 100:1, from 10,000:1 to 100:1, from 10:000:1 to 8,000:1 to 500:1, from 5,000:1 to 200:1, from 10,000:1 to 50:1. In some embodiments, the methods disclosed herein result in at least 100; 200; 500; 750; 1,000; 2,000; 5,000; 7,500; 10,000; 20,000; 25,000; 30,000; 40,000; 50,000 different molecular barcode or index sequences to each one template nucleic acid or cfDNA molecules.N.069.W0.01

[0101] Exemplary library preparation protocols are provided in Abbosh et al., Nature 545:446-451 (2017); and Sigdel et al., J. Clin. Med. 8(1): 19 (2019), each of which is incorporated herein by reference in its entirety.

[0102] Targeted Enrichment

[0103] In some embodiments, the method described herein comprises performing targeted enrichment on the extracted cell-free DNA or DNA derived therefrom to enrich a plurality of polymorphic loci (e.g., SNP loci), wherein at least some of said polymorphic loci can distinguish donor-derived cfDNA (dd-cfDNA) from recipient-derived cfDNA (rd-cfDNA). In some embodiments, the percentage of dd-cfDNA out of total cfDNA in the plasma fraction of the blood sample of the transplant recipient can be estimated using sequence reads at the plurality of enriched polymorphic or SNP loci.

[0104] The plurality of polymorphic or SNP loci being enriched can comprise, for example, at least 100 polymorphic or SNP loci, at least 200 polymorphic or SNP loci, at least 500 polymorphic or SNP loci, at least 1,000 polymorphic or SNP loci, at least 2,000 polymorphic or SNP loci, between 100 and 20,000 polymorphic or SNP loci, between 100 and 500 polymorphic or SNP loci, between 500 and 2,000 polymorphic or SNP loci, or between 2,000 and 20,000 polymorphic or SNP loci. Exemplary lists of polymorphic or SNP loci include those targeted by an exemplary 1,200-plex primer library, an exemplary 2,686-plex primer library, and an exemplary 10,984-plex primer library disclosed in U.S. Pat. No. 9,677,118 including sequence listing thereof, which is incorporated herein by reference in its entirety.

[0105] In some embodiments, the targeted enrichment enriches a plurality of polymorphic or SNP loci each having a minor allele frequency (MAF) of at least 1%, at least 2%, at least 3%, at least 4%, at least 5%, at least 6%, at least 7%, at least 8%, at least 9%, or at least 10%. At an average MAF of 5%, 1,000 polymorphic or SNP loci may translate to 50 informative SNPs.

[0106] In some embodiments, the targeted enrichment comprises preforming targeted multiplex amplification to enrich the polymorphic loci. In some embodiments, the targeted enrichment comprises preforming targeted probe capture to enrich the polymorphic loci. In some embodiments, the targeted enrichment comprises preforming linked target capture to enrich theN.069.W0.01polymorphic loci. Exemplary target enrichment protocols based on multiplex PCR amplification are provided in Zimmermann et al., Prenat. Diagn. 32:1233-1241 (2012); and Sigdel et al., J. Clin. Med. 8(1): 19 (2019), each of which is incorporated herein by reference in its entirety.

[0107] Targeted amplification

[0108] In some embodiments, the targeted enrichment technique can involve targeted multiplex amplification (e.g., PCR or isothermal amplification). Methods in some aspects herein include performing one or, in some embodiments, two or more amplifications. Such amplifications in certain illustrative embodiments include at least one targeted amplification wherein at least one primer and in certain embodiments both primers of a primer pair, one or more primer pairs, or a set of primer pairs used for the amplification are each designed to bind to a specific nucleic acid sequence at or near, typically within, a genomic region of interest comprising a polymorphic target loci (i.e. are target- specific primers) to generate target amplicons. In some embodiments, methods herein include one or more universal amplifications.

[0109] A number of amplification technologies can be used with methods herein. For example, such amplification can be an isothermal amplification (e.g., recombinase polymerase amplification (RPA) (Kersting et al. 2014 Microchim Acta 181 (13-14), 1715-1723, (incorporated by reference in its entirety)), a ligase-based amplification, PCR, or a combination thereof (e.g., ligation-mediated PCR). In some illustrative embodiments, the targeted amplification is a targeted PCR(s) that is performed using a PCR reaction mixture that includes one primer pair, or in illustrative embodiments a set of primer pairs, and at least a portion of the library of DNA molecules comprising the extracted cfDNA or DNA derived therefrom (e.g., adapted DNA, adapted-amplified DNA).

[0110] Typically, at least one primer of a primer pair used for targeted amplification herein, is a target- specific primer designed to bind to a specific nucleic acid sequence in or near, typically within, a genomic region of interest comprising the polymorphic target loci. A target- specific primer can be designed to bind to any sequence within or near a polymorphic target loci for amplification of the polymorphic target loci. One of the advantages of the methods described herein is increased flexibility in primer / probe design for targeted amplification or enrichment. In some embodiments, one primer of the one or more primer pairs or the set of primer pairs in theN.069.W0.01reaction mixture used for a targeted amplification is a universal primer and binds to a primer binding site on an adapter. Thus, for example, in such embodiments a universal primer that binds an adapter primer binding site can be used for an amplification reaction along with a targetspecific primer that binds a primer binding site on a sample DNA region.

[0111] Target-specific primers typically define the ends of target amplicons, which typically encompass at least a portion of the polymorphic target loci. In some embodiments, a PCR can be performed using two target-specific primers. The target amplicon in such embodiment would extend from the sample DNA region bound by target- specific primer on a 5’ end to the sample DNA region bound by primer on the 3’ end. In some embodiments, a PCR can be performed using a universal primer and a target-specific primer. The target amplicon in such embodiment would extend from the sample DNA region bound by target-specific primer on a 3’ end of one strand to the end of the sample DNA fragment on the 5’ end of that strand.

[0112] In some methods herein, a universal amplification of the library of DNA molecules comprising the extracted cfDNA or DNA derived therefrom can be performed before the targeted amplification. Such universal amplification can be performed for example using a universal primer pair that binds primer binding sites in the adapter. Thus, in some embodiments, the methods herein include performing a universal PCR using the adapted DNA molecules, and a universal PCR primer pair comprising primers designed to bind universal primer binding sequences on the adapters, before performing one or more targeted PCRs.

[0113] The one or more primer pairs in illustrative embodiments is a set of primer pairs. In some embodiments, the set of primer pairs is a set of between 20 and 100,000, between 50 and 50,000, between 100 and 20,000, between 100 and 500. between 500 and 2,000, or between 2,000 and 20,000 primer pairs.

[0114] In some embodiments, at least one of the primer pairs comprises a universal primer and a target- specific primer. In some embodiments, at least one of the primer pairs comprises two target- specific primers. In some embodiments, at least one of the primers comprises a sequencing tag. In some embodiments, at least one of the primers comprises a sample index. In some embodiments, at least one of the primers comprises biotin modification. In some embodiments, performing a PCR further comprises using primers comprising a sequencing tag. In someN.069.W0.01embodiments, performing a PCR further comprises using primers comprising a sample index. Tn some embodiments, the primers of the primer pairs are probe-dependent primers and the amplification is a target capture polymerase chain reaction.

[0115] In some embodiments, one or both of the primer binding sites of a primer pair can include at least a portion of one of the adapter sequences. In some embodiments, one of the primer binding sites of a primer pair can include at least a portion of the adapter sequences. In some embodiments, both of the primer binding sites of a primer pair can include at least a portion of the adapter sequences. In some embodiments, neither of the primer binding sites of a primer pair include any of the adapter sequences.

[0116] Methods as described herein, in some embodiments, can include multiple amplification cycles (e.g., multiple PCR temperature cycles), and in some embodiments can include several sequential PCR reactions performed during the same set of temperature cycles. In some embodiments, amplification cycles can include at least 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, or 20 cycles. In some embodiments, amplification cycles can include at least 7, 8, 9, or 10 cycles. In illustrative embodiments, amplification cycles can include at least 11, 12, 13, 14, 15, 16, or 17 cycles.

[0117] Typically, in embodiments described herein, PCR amplification is performed by adding a PCR reaction mixture to the DNA template followed by addition of a polymerase enzyme, and then amplified through multiple amplification cycles. In some embodiments, the PCR reaction mixture contains one or more primer pairs, deoxynucleotides (dNTPs), PCR reaction buffer, and deionized water. In some embodiments, the dNTPs can comprise a mixture of dATP, dCTP, dGTP and dTTP. In some embodiments, the final concentration of each dNTP in the reaction mixture can range from 0.05 mM to 0.5 mM dNTPs, for example, from 0.05 mM to 0.5 mM, 0.05 mM to 0.1 mM, 0.05 mM to 0.15 mM, 0.05 to 0.2 mM, 0.05 mM to 0.25 mM, 0.05 mM to 0.3 mM, 0.05 mM to 0.35 mM, 0.05 to 0.4 mM 0.05 mM to 0.45 mM, from .1 mM to 0.5 mM, 0.15 mM to 0.5 mM, or 0.2 mM to 0.5 mM, 0.25 to 0.5 mM, 0.3 mM to 0.5 mM, 0.35 mM to 0.5 mM, 0.4 mM 0.5 mM, or from 0.45 mM to 0.5 mM. In illustrative embodiments, the final concentration of each dNTP in the reaction mixture is between 0.15 mM and 0.25 mM. In some embodiments, the final concentration of each dNTP in the reaction mixture is 0.2 mM.N.069.W0.01

[0118] PCR buffer solution creates a suitable environment for the polymerase chain reaction and can contain many different components, including magnesium chloride (MgC12), potassium chloride (KC1), dimethyl sulfoxide (DMSO), and glycerin or bovine serum albumin (BSA). In some embodiments, the concentration of KC1 can be between 25 and 50 mM, between 25 and 75 mM, between 25 and 100 mM, between 30 and 100 mM, between 50 and 100 mM, or between 70 and 100 mM. In some embodiments, the MgC12 concentration can be in the range of 0.5 mM to 5 mM, between 0.5 mM to 4.5 mM, 0.5 to 4.0 mM, 0.5 to 3.5 mM, 0.5 to 3.0 mM, 0.5 to 2.5 mM, 0.5 to 2.0 mM, 0.5 to 1.5 mM, 1.0 to 5 mM, 1.5 to 5mM, 2.0 to 5 mM, 2.5 to 5 mM, 3.0 to 5 mM, 3.5 to 5 mM, 4.0 to 5 mM, or 4.5 to 5 mM. In some embodiments, the concentration of MgC12 is 2.0 mM.

[0119] In some embodiments, the buffer solution is a Q5® Reaction Buffer (B9027S, New England Biolabs, Inc.). In some embodiments, the reaction buffer is Taq Reaction Buffer (B9014S. New England Biolabs, Inc). In some embodiments, the reaction buffer is a Taq (Mg-free) Reaction Buffer (B9015S, New England Biolabs, Inc.).

[0120] In some embodiments, a DNA polymerase is used to produce DNA amplicons using DNA as a template. In some embodiments, the polymerase is a Q5® DNA Polymerase, such as Q5® High-Fidelity DNA Polymerase (M0491S, New England BioLabs, Inc.) or Q5® Hot Start High-Fidelity DNA Polymerase (M0493S, New England BioLabs, Inc.). Q5® High-Fidelity DNA polymerase is a high-fidelity, thermostable, DNA polymerase with 3' > 5' exonuclease activity, fused to a processivity-enhancing Sso7d domain. Q5® High-Fidelity DNA polymerase lacks 5'— ► 3 'exonuclease activity and strand displacement activity.

[0121] In some embodiments, the polymerase is a T4 DNA polymerase (M0203S. New England BioLabs, Inc.). T4 DNA Polymerase catalyzes the synthesis of DNA in the 5'— ► 3' direction and requires the presence of template and primer. This enzyme has a 3'^ 5' exonuclease activity which is much more active than that found in DNA Polymerase I. T4 DNA polymerase lacks 5'^- 3' exonuclease activity and strand displacement activity.

[0122] In some embodiments of any of the aspects herein, the length of the primers can be between 10 to 100 nucleotides, such as between 10 to 75 nucleotides, 10 to 40 nucleotides, 10 to 35 nucleotides. 10 to 30 nucleotides, 10 to 20 nucleotides, 15 to 100 nucleotides. 20 to 100N.069.W0.01nucleotides, from 25 to 100 nucleotides, from 30 to 100 nucleotides from 35 to 100 nucleotides, from 40 to 100 nucleotides, from 45 to 100 nucleotides, from 50 to 100 nucleotides, from 55 to 100 nucleotides, from 60 to 100 nucleotides, from 65 to 100 nucleotides, from 70 to 100 nucleotides, or from 75 to 100 nucleotides. In some embodiments, the range of the length of the primers is between 5 to 50 nucleotides, such as 5 to 40 nucleotides, 5 to 20 nucleotides, or 5 to 10 nucleotides, hr some embodiments, the primers are between 5 and 50 bp in length, between 10 and 40 bp in length, between 15 and 30 bp in length, between 15 and 25 bp in length, between 20 and 40 bp in length, between 25 and 50 bp in length, or between 30 and 50 bp in length. In some embodiments, the primers are between 25 and 100 bp in length, between 35 and 100 bp in length, between 45 and 100 bp in length, between 55 and 100 bp in length, between 65 and 100 bp in length, or between 75 and 100 bp in length.

[0123] In some embodiments of any of the aspects or embodiments herein, the number of primer pairs can range from 20 to 100.000 primer pairs that each bind to one or more primer binding sequences. In some embodiments, the primer pairs are a part of a set of primer pairs. In some embodiments, the set of primers range from 50 to 50,000, from 100 to 20,000, from 100 to 500, from 500 to 2,000, or from 2,000 to 20,000 primer pairs. In some embodiments, the number of primer pairs can range from 10 to 10,000, 10 to 1,000, 10 to 100, 10 to 50, 10 to 40, 10 to 30, 15 to 30, or 15 to 25 primer pairs.

[0124] In some embodiments, PCR is used to generate short amplicons. The fragment sizes of dd-cfDNA may be distributed in approximately a Gaussian fashion with a mean of 160 bp, a standard deviation of 15 bp, a minimum size of about 100 bp, and a maximum size of about 220 bp. Because cfDNA fragments are short, the likelihood of both primer sites being present the likelihood of a fragment of length L comprising both the forward and reverse primers sites is the ratio of the length of the amplicon to the length of the fragment. Under ideal conditions, assays in which the amplicon is 45, 50, 55, 60, 65, or 70 bp will successfully amplify from 72%, 69%, 66%, 63%, 59%, or 56%, respectively, of available template fragment molecules. Thus, in some embodiments target amplicons generated in method herein are between 40 and 100, 40 and 75, or 45 and 70 bp in length. The amplicon length is the distance between the 5 -prime ends of the forward and reverse priming sites. In an embodiment, a substantial fraction of the amplicons areN.069.W0.01between 25 on the low end of the range, and 100 bp, 90 bp, 80 bp, 70 bp, 65 bp, 60 bp, 55 bp, 50 bp, or 45 bp on the high end of the range.

[0125] In some embodiments, targeted amplification is performed using direct multiplexed PCR, sequential PCR, nested PCR, doubly nested PCR, one-and-a-half sided nested PCR, fully nested PCR, one sided fully nested PCR, one-sided nested PCR, hemi-nested PCR, hemi-nested PCR, triply hemi-nested PCR, semi-nested PCR, one sided semi-nested PCR, reverse semi-nested PCR method, or one-sided PCR, which are described in U.S. Publication No 2012 / 0270212, U.S. Publication No. 2013 / 0123120, and U.S. Publication No. 2015 / 0322507, each of which is incorporated herein by reference in their entirety.

[0126] In some embodiments, the method of amplifying target loci in a nucleic acid sample involves (i) contacting the nucleic acid sample with a library of primers that simultaneously hybridize to at least 100; 200; 500; 1,000; 2.000; 5.000; 10,000; 20,000; 50.000; or 100,000 different target loci to produce a single reaction mixture; and (ii) subjecting the reaction mixture to primer extension reaction conditions (such as PCR conditions) to produce amplified products that include target amplicons. In some embodiments, at least 50, 60, 70, 80, 90, 95, 96, 97, 98, 99, or 99.5% of the targeted loci are amplified. In various embodiments, less than 60, 50, 40, 30, 20, 10, 5, 4, 3, 2, 1, 0.5, 0.25, 0.1, or 0.05% of the amplified products are primer dimers. In some embodiments, the primers are in solution (such as being dissolved in the liquid phase rather than in a solid phase). In some embodiments, the primers are in solution and are not immobilized on a solid support. In some embodiments, the primers are not part of a microarray.

[0127] In certain embodiments, the multiplex amplification reaction is performed under limiting primer conditions for at least 1 / 2 of the reactions. In some embodiments, limiting primer concentrations are used in 1 / 10, 1 / 5, 1 / 4, 1 / 3, 1 / 2, or all of the reactions of the multiplex reaction. Provided herein are factors to consider in achieving limiting primer conditions in an amplification reaction such as PCR.

[0128] In certain embodiments, the multiplex amplification reaction can include, for example, between 50 and 50,000 multiplex reactions. In certain embodiments, the following ranges of multiplex reactions are performed: between 100, 200, 250, 500, 1000, 2500, 5000, 10,000, 20,000,N.069.W0.0125000, 50000 on the low end of the range and between 200, 250, 500, 1000, 2500, 5000, 10,000, 20,000, 25000, 50000, and 100,000 on the high end of the range.

[0129] In an embodiment, a multiplex PCR assay is designed to amplify potentially heterozygous SNP or other polymorphic or non-polymorphic loci on one or more chromosomes and these assays are used in a single reaction to amplify DNA. The number of PCR assays may be between 50 and 200 PCR assays, between 200 and 1,000 PCR assays, between 1,000 and 5,000 PCR assays, or between 5,000 and 20,000 PCR assays (50 to 200-plex, 200 to 1,000-plex, 1,000 to 5.000-plex, 5,000 to 20,000-plex, more than 20,000-plex respectively). In an embodiment, a multiplex pool of at least 10,000 PCR assays (10,000-plex) are designed to amplify potentially heterozygous SNP loci a single reaction to amplify cfDNA obtained from a blood, plasma, serum, or bronchoalveolar lavage sample. The SNP frequencies of each locus may be determined by clonal or some other method of sequencing of the amplicons. In another embodiment the original cfDNA samples is split into two samples and parallel 5,000-plex assays are performed. In another embodiment the original cfDNA samples is split into n samples and parallel (~10,000 / n)-plex assays are performed where n is between 2 and 12, or between 12 and 24, or between 24 and 48, or between 48 and 96.

[0130] In an embodiment, a method disclosed herein uses highly efficient highly multiplexed targeted PCR to amplify DNA followed by high throughput sequencing to determine the allele frequencies at each target locus. One technique that allows highly multiplexed targeted PCR to perform in a highly efficient manner involves designing primers that are unlikely to hybridize with one another. The PCR probes, typically referred to as primers, are selected by creating a thermodynamic model of potentially adverse interactions between at least 100, at least 200, at least 500, at least 1,000, at least 2,000. at least 5,000, at least 10,000, at least 20,000, or at least 50,000 potential primer pairs, or unintended interactions between primers and sample DNA, and then using the model to eliminate designs that are incompatible with other the designs in the pool. Another technique that allows highly multiplexed targeted PCR to perform in a highly efficient manner is using a partial or full nesting approach to the targeted PCR. Using one or a combination of these approaches allows multiplexing of at least 100, at least 200, at least 500, at least 1,000, at least 2,000, at least 5.000, at least 10,000, at least 20,000, or at least 50,000 primers in a single pool with the resulting amplified DNA comprising a majority of DNA molecules that, when sequenced, will map to targeted loci. Using one or a combination of these approaches allowsN.069.W0.01multiplexing of a large number of primers in a single pool with the resulting amplified DNA comprising greater than 50%, greater than 80%, greater than 90%, greater than 95%, greater than 98%, or greater than 99% DNA molecules that map to targeted loci.

[0131] Bioinformatics methods for analyzing the sequence data obtained from multiplex PCR are described in U.S. Publication No 2012 / 0270212, U.S. Publication No. 2013 / 0123120, and U.S. Publication No. 2015 / 0322507, each of which is incorporated herein by reference in its entirety.

[0132] Hybrid Capture

[0133] In some embodiments, the targeted enrichment technique can involve fragment capture by hybridization (i.e., hybrid capture). Although any hybrid capture method can be used to perform methods herein that include a targeted enrichment step, in some embodiments, a method of the present disclosure may involve using any of the hybrid capture methods disclosed herein to enrich cfDNA having one or more polymorphic target loci. In some embodiments described herein, the targeted enrichment steps can be performed after cfDNA molecules are extracted. In some embodiments described herein, the targeted enrichment steps can be performed after appending adapters to the extracted cfDNA molecules. In some embodiments described herein, the targeted enrichment steps can be performed after universal PCR amplification of the adapted DNA.

[0134] In capture by hybridization, hybrid capture oligonucleotide probes complementary to one or both strands of specific target DNA sequences, or DNA derived therefrom in a sample, are utilized, i.e., the probes may be strand specific. The specific target DNA sequence in illustrative embodiments overlaps with or is found within a target region of a sample DNA molecule such as a cfDNA. Thus, hybrid capture probes when used in methods herein can be designed to bind to a DNA molecule that contains at least one target region or a portion thereof. In some embodiments, the hybrid capture probes can be designed to bind to a target DNA sequence within or overlapping a target region. In other examples, the hybrid capture probes can be designed to bind to a common region that is flanking but not overlapping the target region and that can be a common region that was added to some, most, almost all or all of the DNA in a sample, or added to all amplicons using a common sequence on at least one primer of a primerN.069.W0.01pair. Tn illustrative embodiments, a hybrid capture probe or set thereof, are designed to bind to a target DNA sequence within target region, or set of target regions, respectively.

[0135] Hybrid capture probes may be added to a prepared sample and hybridized through a denature-reannealing process to form duplexes of exogenous-endogenous fragments (e.g., hybrid capture probes bound to sample DNA molecules, or DNA derived therefrom). These duplexes may then be physically separated from the sample by various means. In some embodiments, once the hybrid capture probes are removed, the sample DNA molecules, or DNA derived therefrom can be amplified. Some ways to physically remove the hybrid capture probes are by covalently bonding the hybrid capture probes to a solid support, for example a magnetic bead, or a chip. Another way to physically remove the hybrid capture probes is by covalently bonding them to a molecular moiety with a strong affinity for another molecular moiety. An example of such a molecular pair is biotin and streptavidin, such as is used in SURE SELECT (Agilent). Thus, hybrid capture probes, for example that bind to a target DNA sequence within or overlapping a target region of a DNA molecule obtained or derived from a sample, can be covalently attached to a biotin molecule, and after hybridization with sample DNA or DNA derived therefrom, a solid support with streptavidin affixed can be used to pull down the biotinylated hybrid capture probes, which are hybridized to DNA molecules obtained or derived from a sample that include a target region that includes the target DNA sequence recognized by the hybrid capture probes. Thus, in some embodiments, the hybrid capture probes are immobilized, directly or indirectly to a solid support. In some embodiments, the hybrid capture probes include a binding partner, for example biotin.

[0136] In some embodiments of any of the aspects herein, the hybrid capture probes can be a part of a set of at least two hybrid capture probes. In some embodiments, the set includes at least one hybrid capture probe for each polymorphic target loci. In some embodiments, the set includes two or more hybrid capture probes for each polymorphic target loci. In some embodiments, the set includes three or more hybrid capture probes for each polymorphic target loci. In some embodiments, the set includes four or more hybrid capture probes for each polymorphic target loci.N.069.W0.01

[0137] In some embodiments of any of the aspects herein, the hybrid capture probes can have a length in the range of 30 bases to 170 bases, 30 bases to 160 bases, 30 bases to 150 bases, 30 bases to 140 bases, 30 bases to 130 bases, 30 bases to 120 bases, 30 bases to 110 bases, 30 bases to 100 bases, 30 bases to 90 bases, 30 bases to 80 bases, 30 bases to 70 bases, 30 bases to 60 bases, 30 bases to 50 bases, 40 bases to 160 bases, 40 bases to 150 bases, 40 bases to 140 bases, 40 bases to 130 bases. 40 bases to 120 bases, 40 bases to 110 bases. 40 bases to 100 bases, 40 bases to 90 bases, 40 bases to 80 bases, 40 bases to 70 bases, 40 bases to 60 bases, 50 bases to 150 bases, 50 bases to 140 bases, 50 bases to 130 bases, 50 bases to 120 bases, 50 bases to 110 bases, 50 bases to 100 bases, 50 bases to 90 bases, 50 bases to 80 bases, 50 bases to 70 bases, 60 bases to 140 bases, 60 bases to 130 bases, 60 bases to 120 bases, 60 bases to 110 bases, 60 bases to 100 bases, 60 bases to 90 bases, 60 bases to 80 bases, 70 bases to 130 bases. 70 bases to 120 bases, 70 bases to 110 bases, 70 bases to 100 bases, 70 bases to 90 bases, 80 bases to 120 bases, 80 bases to 110 bases, 80 bases to 100 bases, 90 bases to 120 bases, 90 bases to 110 bases, 100 bases to 165 bases, 100 bases to 150 bases, 100 bases to 140 bases, 100 bases to 130 bases, 100 bases to 120 bases, 110 bases to 150 bases, 110 bases to 140 bases, 110 bases to 130 bases, 120 bases to 150 bases, or 130 bases to 160 bases.

[0138] In some embodiments, the targeted probe capture is performed using tiling probes providing at least 2X, at least 3X, or at least 4X tiled coverage of each polymorphic target loci. In some embodiments, the hybrid capture panel was designed as a 4X tiling probe set (e.g., 4 baits per base with ~90 bp overlap for ~120bp probes). In some embodiments, the hybrid capture panel was designed as a 2X tiling probe set (2 baits per base with ~60 bp overlap for ~120bp probes).

[0139] Linked target capture using Probe-dependent primers

[0140] In some embodiments, the targeted enrichment technique can involve probe-dependent primers. Probe-dependent primers (PDPs) have been disclosed (Pel, et al. “Rapid and highly-specific generation of targeted DNA sequencing libraries enabled by linking capture probes with universal primers” PLoS ONE 13(12):e0208283 (2018); WO 2017168332A1 “Linked duplex target capture”, which are hereby incorporated by reference in their entirety). Such embodiments can be considered linked target capture (LTC) methods. Briefly, in an LTC method, a target-N.069.W0.01specific probe is linked to a universal primer. The target specific probe is designed to hybridize to a target of interest such as one of the polymorphic target loci. The universal primer linked to the probe is designed to hybridize to a universal priming site in the adapters that have been ligated to the cell-free DNA. The binding of the probe to the target brings the linked universal primer into proximity with the universal priming site and in fact, the ability of the universal primer to bind to the universal priming site and be extended depends on the probe binding to its target. Results to-date have shown that the universal primers do not hybridize to adapters attached to fragments that do not include the target. The linked target capture is highly target specific and primer extension depends on successful probe binding. For that reason, the universal primers linked to the probes are "probe-dependent primers". LTC may be performed using only one PDP, e.g., for a linear amplification, but preferably uses paired forward and reverse PDPs (as shown in Fig. 1(b) of Pel 2018 as target-capture PCR1) to amplify the fragment exponentially. The bound probe does not interfere with primer extension to copy the entire fragment (including the distal adapter) when a strand-displacing polymerase is used. It is noted that the cell-free DNA fragment is copied by primers extended from within the ligated adapters, so the entirety of the fragment is copied into amplicons. Due to the probe, LTC gives the target specificity of conventional PCR with gene-specific primers while due to the priming sites in the adapters, the entire fragment is amplified. The resulting amplification products will include a copy of the entire target-containing fragment with adapters at both ends. Preferably, PDPs are designed to incorporate non-extendable capture probes linked 5’ to 5’ with a primer. Multiple linker types are possible as discussed below. Typically, probes of PDPs can be between 30 to 70 nucleotides in length, and include or comprise a 3’ inverted dT base or 3’ C3 spacer to inhibit polymerase extension. In some embodiments, probes are designed to cover the desired region with overlap between forward and reverse probes. In some embodiments, the probes are between 20 and 100 nucleotides in length. In some embodiments, the size of the probe can be between 20 and 40 nucleotides, between 30 and 50 nucleotides, between 40 and 60, between 50 and 70 between 60 and 80, between 70 and 90, 80 and 100, 90 and 110, 100 and 120 nucleotides in length. In some embodiments, at least one of the probes of a PDP pair comprises a sample index.

[0141] In PDPs, forward and reverse probes can be designed to bind to nucleic acid sequences within or near a polymorphic target loci on a sample DNA molecule to enrich nucleic acidN.069.W0.01molecules comprising the polymorphic target loci of interest or copies thereof. Typically, the primer portion of a PDP is a universal primer designed to bind to a universal primer site on the appended adapter. In some embodiments, the PDP is designed with a sequencer binding sequence, such as an Illumina flow cell binding sequence, incorporated therein. In some embodiments, the sequencer flow cell binding sequence is between the probe and universal primer, and adjacent to the primer. Linked primers of the invention may also include sequencing tags to ensure that all cluster reads originate from the same linked template molecule. The lengths of the primers can be extended or shortened at the 5' end or the 3' end to produce primers with desired melting temperatures. Also, the annealing position of each primer pair can be designed such that the sequence and length of the primer pairs yield the desired melting temperature. In illustrative embodiments, the primer is a low melting temperature universal primer complementary to a portion of the ligated adapter.

[0142] The primer can be tailed or untailed depending on the specific requirements. In some embodiments, the universal primer comprises an A tail. The length of the primers of the PDP can range from 5 to 40 nucleotides in length. In certain embodiments, the PDP primers are between 10 and 25 nucleotides long. In embodiments, the primers of the PDP can range from 5 to 15 nucleotides, from 10 to 25 nucleotides, from 15 to 35 nucleotides, or from 25 to 40 nucleotides in length.

[0143] Typically, probe dependent primers comprise a linker between the probe and the primer. Probe and primer portions of the PDP are typically linked by a polyethylene glycol derivative, an oligosaccharide, a lipid, a hydrocarbon, a polymer, or a protein. In some embodiments, the linker is a PEG molecule, or derivative thereof. In some embodiments, the linker is an oligosaccharide. In some embodiments, the linker is a lipid. In some embodiments, the linker is a hydrocarbon. In some embodiments, the linker is a polymer. In some embodiments, the linker is a protein, or portion thereof. Linkers based on click chemistry is described in WO2017 / 168332A1, which is incorporated herein by reference in its entirety.

[0144] Sequencing to Generate Sequence Reads

[0145] In some embodiments, the method described herein comprises performing sequencing on the sequencing library to generate sequence reads, and using the sequence reads to quantify anN.069.W0.01amount of donor-derived cell-free DNA and / or a percentage of donor-derived cell-free DNA out of total cell-free DNA.

[0146] In some embodiments, the sequencing is next-generation sequencing or high-throughput sequencing. In some embodiments, the sequencing is untargeted shotgun sequencing (e.g., whole-genome sequencing). In some embodiments, the sequencing is targeted sequencing performed on enriched DNA (e.g., enriched at a plurality of polymorphic loci that can distinguish between the lung transplant and the transplant recipient) or DNA derived thereof.

[0147] DNA sequencing techniques, particularly high throughput next- generation sequencing techniques (often referred to as massively parallel sequencing techniques) such as those employed NOVASEQ (ILLUMINA), MISEQ (ILLUMINA), HISEQ (ILLUMINA), ION TORRENT (LIFE TECHNOLOGIES), GENOME ANALYZER ILX (ILLUMINA), GS FLEX+(ROCHE 454) etc., can be used for determining the sequences of the enriched DNA or DNA derived thereof to elucidate the sequence of the original cfDNA. High throughput genetic sequencers are amenable to the use of barcoding (i.e., sample tagging with distinctive nucleic acid sequences) so as to identify specific samples from individuals thereby permitting the simultaneous analysis of multiple samples in a single run of the DNA sequencer. Methods as described herein that utilize NGS detection, in some embodiments can have an average or median depth of read of at least 10, 20, 50, 100, 200, 500, 1000, 2000, 5000, 10,000, 20,000, 50,000, 100,000, 150,000, or 200,000.

[0148] Methods herein can include analyzing data obtained from next- generation sequencing techniques. In some embodiments of methods herein, the enriched DNA or DNA derived thereof can be subjected to sequencing using next- generation sequencing techniques. For a skilled artisan, algorithm design tools are available that can be used and / or adapted to analyze the sequencing data. In addition, those skilled in the art can determine appropriate parameters for measuring alignment to a consensus sequence and / or to a known target region sequence, including any algorithms needed to achieve maximal alignment over the length of the sequences being compared.

[0149] Sequence reads can be demultiplexed using an in-house tool and mapped using the Burrows-Wheeler alignment software, Bwa mem function (BWA, Burrows-Wheeler AlignmentN.069.W0.01Software (see Li H. and Durbin R. (2010) Fast and accurate long-read alignment with Burrows-Wheeler Transform. Bioinformatics.) in single end or paired end mode to a version of reference genome. The reference genome used can be hgl9 or hg38. Amplification statistics QC can be performed by analyzing one or more of, but not limiting to, total reads, number of mapped reads, number of mapped reads on target, and number of reads counted.

[0150] Methods herein can include a background error model that can be constructed using normal, healthy, or non-diseased liquid samples, in illustrative embodiments, normal, healthy, or non-diseased plasma samples, which are sequenced on the same sequencing ran to account for run-specific artifacts. In some embodiments, 5, 10, 15, 20, 25, 30, 40, 50, 100, 150, 200, 250, or more than 250 normal, healthy, or non-diseased liquid samples, in illustrative embodiments, plasma samples can be analyzed on the same sequencing run. The number of samples that can be sequenced on the same sequencing ran can be in the range of 5 to 500, 5 to 400, 5 to 300, 5 to 250, 20 to 250. 30 to 250. 50 to 250, 75 to 250, 100 to 250, 50 to 500. or 100 to 500. Sample barcodes are used in illustrative embodiments. In some illustrative embodiments, 20, 25, 40, or 50 normal samples (e.g., plasma samples) can be analyzed on the same sequencing run. Outlier samples can be iteratively removed from the model to account for noise and contamination. In some embodiments, samples with a Z score of greater than 5, 6, 7, 8, 9, or 10 are removed from the data analysis. For each base substitution of every genomic loci, the DOR weighted mean and standard deviation of the error can be calculated.

[0151] Methods herein can include calculating percent identity that can be calculated by determining the number of matched positions in aligned DNA sequences, dividing the number of matched positions by the total number of aligned DNA sequences, and multiplying by 100. A matched position refers to a position in which identical nucleotides occur at the same position in aligned DNA sequences. The percent identity over a particular length can be determined by counting the number of matched positions over that length and dividing that number by the length followed by multiplying the resulting value by 100. A non-limiting example for calculating the percent identity, can be, if (i) a 500-nucleotide DNA target sequence is compared to a subject DNA sequence, (ii) an alignment program presents 200 nucleotides from the target DNA sequence aligned with a region of the subject DNA sequence where the first and last nucleotides of that 200-nucleotide region are matches, and (iii) the number of matches over thoseN.069.W0.01200 aligned nucleotides is 180, then the 500-nucleotide nucleic acid target sequence contains a length of 200 and a sequence identity over that length of 90 percent (i.e., 180, 200x100=90).

[0152] In some embodiments, the uniformity in DOR can be measured using standard methods such as, but not limiting to, DOR slope, normalized median depth of read (nmDOR), or breadth of read (BOR). DOR slope represents the slope of the line in the linear portion of a list of loci sorted in descending DOR order. Closer to zero is better, as it represents a flat line. In some embodiments, the uniformity in DOR can be measured using the percent of reads in the 90th-95th percentile. For this measurement, the loci are sorted in descending DOR order. In illustrative embodiments, a DOR distribution using the 90th-95th percentile contains 5 percent of reads. The reads of all loci between the 90th percentile and 95th percentile can be counted and divided by the total reads for all loci.

[0153] In some embodiments, the magnitude of the DOR slope can be less than 0.005, 0.001. 0.0005, 0.0001, 0.00005, 0.00001, 0.000005, or 0.000001. The magnitude of the DOR slope can be between 0 and 0.005, such as 0.000001 to 0.005. such as between 0.000005 to 0.00001, 0.00001 to 0.00005, 0.00005 to 0.0001, 0.0001 to 0.0005, 0.0005 to 0.001, or 0.001 to 0.005. The percent of reads in the 90th-95th percentile can be between 0.2 and 9 percent, such as between 0.2 to 8 percent, 0.2 to 7 percent, 0.2 to 6 percent, 0.4 to 9 percent, 0.4 to 8 percent, 0.4 to 7 percent, 0.4 to 6 percent, 1 to 9 percent, 1 to 8 percent, 1 to 7 percent, 1 to 6 percent, 2 to 9 percent, 2 to 8 percent, 2 to 7 percent, 2 to 6 percent, 3 to 9 percent, 3 to 8 percent, 3 to 7 percent, 3 to 6 percent, 0.2 to 1.0 percent, 1 to 2 percent, 2 to 3 percent, 2 to 4 percent, 3 to 4 percent, 4 to 5 percent, 5 to 6 percent, or 6 to 8 percent, or 7 to 9 percent. In some embodiments of methods herein, the method can produce a composition comprising at least 100 different amplicons (e.g„ at least 300, 500, 750. 1,000, 2.000, 5.000, 7,500. 10.000, 15,000. 20.000, 25,000, 30,000, 40,000, 50,000, 75,000, or 100,000 non-identical amplicons) with the magnitude of the DOR slope in any of the ranges herein, or with a percent of reads in the 90th-95th percentile in any of the ranges herein. In some embodiments, different amplicons can range in between 100 to 500,000, 100 to 400,000, 100 to 300,000, 100 to 200,000, 100 to 100,000, 100 to 75,000, 100 to 50,000, 100 to 40.000, 100 to 30,000, 100 to 25,000, 100 to 20,000, or 100 to 15,000 non-identical amplicons.N.069.W0.01

[0154] In some embodiments, the high-throughput sequencing is performed with a median depth of read of at least 100, at least 200, at least 500, at least 1,000, at least 2,000, at least 5,000, at least 10,000, at least 20.000, at least 50,000. at least 100.000, about 100-10,000, about 200-10,000, about 500-10,000, or about 1,000-100,000 per polymorphic locus.

[0155] The use of sequence reads generated by sequencing to quantify an amount of donor-derived cell-free DNA and / or a percentage of donor-derived cell-free DNA out of total cell-free DNA are described in Sigdel et al., J. Clin. Med. 8(1): 19 (2019); W02020 / 010255;WO2021 / 243045; and WO2022 / 182878, each of which is incorporated herein by reference in its entirety. In some embodiments, the time series data input for the ML model is generated from the percentage of dd-cfDNA out of total cfDNA. In some embodiments, the time series data input for the ML model is generated from the amount of dd-cfDNA. In some embodiments, the time series data input for the ML model is generated from an absolute amount of dd-cfDNA. In some embodiments, the time series data input for the ML model is generated from an absolute amount of dd-cfDNA normalized to a reference such as tracer DNA further described below. For example, the absolute amount of dd-cfDNA can be calculated by multiplying the percentage of donor- derived cell-free DNA with the number of reads of total cell-free DNA divided by the number of reads of tracer DNA per plasma volume. A reference used for normalization can refer to a value obtained from a reference molecule such as tracer DNA used to compare data, a standard used to calibrate data, or a population used to normalize data.

[0156] Tracer DNA

[0157] Use of tracer DNA to improve quantification of the amount of total cell-free DNA is described in WO2021 / 243045 and WO2022 / 182878. each of which is incorporated herein by reference in its entirety. Tracer DNA (or internal calibration DNA) refers to a composition of DNA for which one or more of the following features are known - length, sequence, nucleotide composition, quantity, or biological origin. The tracer DNA can be added to a biological sample derived from a human subject to help estimate the amount of total cfDNA in said sample.

[0158] In some embodiments, the method further comprises adding a tracer DNA to extracted DNA or its derivative to obtain a mixed composition. In some embodiments, the tracer DNAN.069.W0.01comprises synthetic double-stranded DNA molecules. Tn some embodiments, the Tracer DNA comprises DNA molecules of non-human origin.

[0159] In some embodiments, the tracer DNA comprises DNA molecules having a length of about 50-500 bp, or about 75-300 bp, or about 100-250 bp, or about 125-200 bp, or about 125 bp, or about 160 bp, or about 200 bp, or about 500-1,000 bp.

[0160] In some embodiments, the tracer DNA comprises DNA molecules having the same or substantially the same length, such as a DNA molecule having a length of about 125 bp, or about 160 bp, or about 200 bp. In some embodiments, the tracer DNA comprises DNA molecules having different lengths, such as a first DNA molecule having a length of about 125 bp, a second DNA molecule having a length of about 160 bp, and a third DNA molecule having a length of about 200 bp. In some embodiments, the DNA molecules having different lengths are used to determine size distribution of the cell-free DNA in the sample

[0161] In some embodiments, the tracer DNA comprises a target sequence, wherein the target sequence comprises a barcode positioned between a pair of primer binding sites capable of binding to a pair of primers. In some embodiments, at least part of the tracer DNA is designed based on an endogenous human SNP locus, by replacing an endogenous sequence containing the SNP locus with the barcode. During the multiplex PCR target enrichment step, the primer pair targeting the SNP locus can also amplify the portion of tracer DNA containing the barcode.

[0162] In some embodiments, the barcode is an arbitrary barcode. In some embodiments, the barcode comprises reverse complement of a corresponding endogenous genome sequence capable of being amplified by the same primer pair.

[0163] In some embodiments, the target sequence within the tracer DNA is flanked on one or both sides by endogenous genome sequences. In some embodiments, the target sequence within the tracer DNA is flanked on one or both sides by non-endogenous sequences.

[0164] In some embodiments, the tracer DNA comprises a plurality of target sequences. In some embodiments, the tracer DNA comprises a first target sequence comprising a first barcode positioned between a first pair of primer binding sites capable of binding to a first pair ofN.069.W0.01primers, and a second barcode positioned between a second pair of primer binding sites capable of binding to a second pair of primers. In some embodiments, the first and / or second target sequence is designed based on one or more endogenous human SNP loci, by replacing an endogenous sequence containing a SNP locus with a barcode. In some embodiments, the first and / or second barcode is an arbitrary barcode. In some embodiments, the first and / or second barcode comprises reverse complement of a corresponding endogenous genome sequence capable of being amplified by the first or second primer pair. In some embodiments, the first and / or second target sequence within the tracer DNA is flanked on one or both sides by endogenous genome sequences. In some embodiments, the first and / or second target sequence within the tracer DNA is flanked on one or both sides by non-endogenous sequences.

[0165] In some embodiments, the tracer DNA comprises DNA molecules having the same or substantially the same sequence. In some embodiments, the tracer DNA comprises DNA molecules having different sequences.

[0166] In some embodiments, the tracer DNA comprises a first DNA comprising a first target sequence and a second DNA comprising a second target sequence. In some embodiments, the first target sequence and second target sequence have different barcodes positioned between the same primer binding sites. In some embodiments, the first target sequence and second target sequence have different barcodes positioned between the same primer binding sites, wherein the different barcodes have the same or substantially the same lengths. In some embodiments, the first target sequence and second target sequence have different barcodes positioned between the same primer binding sites, wherein the different barcodes have different lengths. In some embodiments, the first target sequence and second target sequence are designed based on different endogenous human SNP loci, and hence comprise different primer binding sites. In some embodiments, the amount of first DNA and the amount of the second DNA are the same or substantially the same in the tracer DNA. In some embodiments, the amount of first DNA and the amount of the second DNA are different in the tracer DNA.

[0167] Determining Amount of Total cfDNA using tracer DNA

[0168] In certain embodiments, the tracer DNA can be used to improve accuracy and precision of the method described herein, help quantify over a wider input range, assess efficiency ofN.069.W0.01different steps at different size ranges, and / or calculate fragment size-distribution of input material.

[0169] Some embodiments of the present invention relate to a method of quantifying the amount of total cell-free DNA in a biological sample, comprising: a) adding a tracer DNA to extracted DNA or its derivative to obtain a mixed composition; b) performing targeted multiplex amplification on the mixed composition comprising the extracted DNA or its derivative and the tracer DNA at 100 to 20,000 different polymorphic target loci together in the same reaction volume using 100 to 20,000 different target-specific primers; c) sequencing the amplicons by high-throughput sequencing to generate sequence reads; and d) quantifying an amount of donor-derived cell-free DNA and an amount of total cell-free DNA from the sequence reads, wherein the amount of total cell-free DNA is quantified using sequence reads derived from the tracer DNA.

[0170] In some embodiments, the method comprises adding the tracer DNA to a whole blood sample before plasma extraction. In some embodiments, the method comprises adding the tracer DNA to a plasma sample after plasma extraction and before isolation of the cell-free DNA. In some embodiments, the method comprises adding the tracer DNA to a composition comprising the isolated cell-free DNA. In some embodiments, the method comprises ligating adaptors to the isolated cell-free DNA to obtain a composition comprising adaptor-ligated DNA, and adding the tracer DNA to the composition comprising adaptor-ligated DNA.

[0171] In some embodiments, the method further comprises adding a second tracer DNA before the targeted amplification. In some embodiments, the method further comprises adding a second tracer DNA after the targeted amplification.

[0172] In some embodiments, the amount of total cfDNA in the sample is estimated using the NOR of the tracer DNA (identifiable by the barcode), the NOR of sample DNA, and the known amount of the tracer DNA added to the plasma sample. In some embodiments, the ratio between the NOR of the tracer DNA and the NOR of sample DNA is used to quantify the amount of total cell-free DNA. In some embodiments, the ratio between the NOR of the barcode and the NOR of the corresponding endogenous genome sequence is used to quantify the amount of total cell-free DNA. In some embodiments, this information along with the plasma volume can also be used toN.069.W0.01calculate the amount of cfDNA per volume of plasma. In some embodiments, these can be multiplied by the percentage of donor DNA to calculate the total donor cfDNA and the donor cfDNA per volume of plasma.

[0173] Physiological Tests

[0174] In some embodiments, the method described herein further comprises quantifying a performance of the organ transplant recipient in one or more physiological tests. The physiological tests can be performed, for example, before the quantification of the amount of dd-cfDNA in the lung transplant recipient. The physiological tests can be performed can be performed, for example, after the quantification of the amount of dd-cfDNA in the organ transplant recipient. The physiological tests can be performed can be performed, for example, at about the same time as the quantification of the amount of dd-cfDNA in the organ transplant recipient. The physiological tests can be performed can be performed, for example, within 4 weeks, within 2 weeks, within 1 week, within 5 days, within 3 days, or on the same day, as the quantification of the amount of dd-cfDNA in the organ transplant recipient.

[0175] The physiological tests can be performed, for example, more than 3 months posttransplantation, more than 6 months post-transplantation, more than 9 months posttransplantation, more than 12 months post-transplantation, between 1 and 12 months posttransplantation, between 12 and 36 months post-transplantation, or between 36 and 60 months post-transplantation.

[0176] The physiological tests can comprise, for example, if the transplant organ is a lung, one or more spirometry-based pulmonary function tests (PFTs), such as forced expiratory volume in 1 second (FEV1), forced vital capacity (FVC), or forced expiratory flow 25-75 (FEF25-75%).

[0177] The physiological tests can comprise, for example, forced expiratory volume in 1 second (FEV1). FEV1 refers to the volume of air a person can forcefully exhale in the first second after taking a deep breath, essentially measuring how much air can be expelled quickly from the lungs, where a low FEV1 may suggest airflow obstruction.N.069.W0.01

[0178] The physiological tests can comprise, for example, forced vital capacity (FVC). FVC refers to the maximum amount of air a person can exhale after a full inhalation, which is a measurement of respiratory muscle function and is used to monitor respiratory health.

[0179] The physiological tests can comprise, for example, forced expiratory flow 25-75 (FEF25-75%). FEF25-75% refers to a spirometry measurement that measures the average flow rate of medium-to-small airways during a forced vital capacity (FVC) test, where a reduced FEF25-75% can be a marker of small airway obstruction.

[0180] The physiological tests can comprise, for example, a lung diffusion test, which is an evaluation of the function of alveoli and alveolar capillaries that surround them. Oxygen and carbon dioxide normally diffuse / flow through alveoli and alveolar capillaries. In some embodiments, the lung diffusion test involves inhaling a small amount of carbon monoxide that is bound to a tracer molecule such as helium, wherein exhaled carbon monoxide concentration can be compared to the inhaled concentration of carbon monoxide in order to calculate a diffusion capacity of the lungs for carbon monoxide. A concentration of exhaled carbon monoxide higher than the normal predicted value may suggest that the lungs do not efficiently absorb oxygen.

[0181] The physiological tests can comprise, for example, lung plethysmography, which is a test used to measure how much air lungs can hold. Unlike spirometry, which measures how much air one can exhale, this test measures the air inside the lungs. Lung plethysmography can be used to distinguish between obstructive diseases that prevent one from inhaling adequately and restrictive lung diseases that prevent one from exhaling adequately.

[0182] Two-Stage Model for Transplant Rejection Assessment

[0183] The present disclosure also provides a Two-Stage Model that can build on the initial assessment of transplant rejections based on dd-cfDNA by combining the dd-cfDNA based predictions with other clinical data. In another aspect, the present disclosure relates to a method for preparing a non-naturally occurring composition for assessing an organ transplant, comprising: longitudinally collecting two or more blood, plasma, serum or urine samples from an organ transplant recipient,N.069.W0.01extracting cell-free DNA (cfDNA) from the two or more blood, plasma, serum or urine samples or portions thereof from the organ transplant recipient, wherein the extracted cfDNA comprises a mixture of donor-derived cfDNA (dd-cfDNA) and recipient-derived cfDNA;preparing a sequencing library from the extracted cfDNA or its derivative and performing high-throughput sequencing on the sequencing library to obtain sequence reads;quantifying an amount of dd-cfDNA or a percentage of dd-cfDNA out of total cfDNA based on the sequence reads, wherein the organ transplant is assessed by combining the amount of dd-cfDNA or the percentage of dd-cfDNA out of total cfDNA with clinical data from the organ transplant recipient. Clinical data from the organ transplant recipient refers to herein as any data obtained about the organ transplant recipient that may be relevant to assess organ transplant rejection.

[0184] The dd-cfDNA measurements can be a current measurement of dd-cfDNA or a percentage of dd-cfDNA out of total cfDNA, or at one or more time points after the transplant procedure. The dd-cfDNA measurements can also be made longitudinally from before the transplant procedure, at the time of the transplantation, and at one or more time points after the transplant procedure as described in more detail elsewhere herein. In some embodiments, the method further comprises determining longitudinal dd-cfDNA changes based on the amount of dd-cfDNA or the percentage of dd-cfDNA out of total cfDNA. The longitudinal changes in the amount of dd-cfDNA or the percentage of dd-cfDNA out of total cfDNA are referred to herein as longitudinal dd-cfDNA changes.

[0185] Based on dd-cfDNA measurement, transplant recipients first undergo routine dd-cfDNA testing, which provides a two-threshold call (2TC) identifying low-risk and high-risk samples. The 2TC assay determines high-risk if at least one measurement of dd-cfDNA is above a predefined threshold within a certain time window. The time windows for this assay can be for example, within 3 months post-transplantation, within 6 months post-transplantation, within 9 months post- transplantation, within 12 months post-transplantation, between 1 and 60 months post- transplantation, between 1 and 36 months post- transplantation, between 1 and 12 months post- transplantation, between 12 and 36 months post- transplantation, or between 36 and 60 months post-transplantation.N.069.W0.01

[0186] A designation of high-risk by the 2TC assay can be further analyzed by using a Two-Stage model disclosed herein. The Two-Stage model is an hierarchical model combining the dd-cfDNA measurements with clinical data to further assess an organ transplant status and discriminate between different subtypes of organ transplant rejections. The Two-Stage model includes a first stage caller and a second stage caller. The first stage caller combines dd-cfDNA measurement information with additional clinical and immunologic features and generates a refined rejection risk score, classifying each sample as either low risk or high risk for rejection. For samples that are assigned to be high-risk in both the 2TC model and the first stage caller, a second stage caller is applied to differentiate the likely rejection subtypes. The second stage caller produces risk scores for ABMR and TCMR and assigns a label (TCMR, AB MR, or type no-call) for eligible high-risk samples. This hierarchical architecture allows the Two-Stage model to preserve established low-risk calls while adding both enhanced rejection detection and rejection subtype prediction for patients at elevated risk.

[0187] In some embodiments, the Two-Stage model can be a machine learning (ML) model such as a temporal random forest (TeRF) model. Both the first and second stage callers can be TeRF models. The first stage caller can be a rejection detection model trained on patient data, using a temporal random forest (TeRF) classifier to distinguish rejection from non-rejection. Its inputs can include dd-cfDNA measurements and other clinical data from the 2TC based routine assay (including quality metrics like donor fraction estimate and donor quantification score (DQS)). The clinical data can be any information relevant to assessment of organ transplants, such as age, ethnicity, prior rejection history, serum creatinine, and donor specific antibody (DSA) status. The first stage caller can output a calibrated risk score, a binary high-risk / low-risk label, a rejection threshold, and feature importance rankings. The second stage caller can include two additional TeRF models: one model for ABMR versus no-ABMR and a separate model for TCMR versus no-TCMR. The second stage caller outputs calibrated risk scores for ABMR and TCMR, a final categorical label (TCMR, ABMR, or type no-call), and feature importance values for each subtype.

[0188] In some embodiments, the method further comprises determining a TCMR risk score for T cell-mediated rejection (TCMR) rejection, an ABMR risk score for antibody-mediated rejection (ABMR) rejection, and / or a Mixed risk score for mixed rejection based on combiningN.069.W0.01(i) the amount of dd-cfDNA or the percentage of dd-cfDNA out of total cfDNA or longitudinal dd-cfDNA changes with (ii) clinical data from the organ transplant recipient. In some embodiments, the second stage caller comprises combining longitudinal dd-cfDNA changes with clinical data. In some embodiments, the first stage caller comprises combining the amount of dd-cfDNA or the percentage of dd-cfDNA out of total cfDNA with prior transplant rejection history.

[0189] In some embodiments, the second stage caller comprises combining longitudinal dd-cfDNA changes with demographic data and a plurality of organ function markers to determine a TCMR risk score for T cell-mediated rejection (TCMR) rejection.

[0190] In some embodiments, the second stage caller comprises combining the amount of dd-cfDNA or the percentage of dd-cfDNA out of total cfDNA with prior AB MR history to determine an AB MR risk score.

[0191] In some embodiments, the amount of dd-cfDNA or the percentage of dd-cfDNA out of total cfDNA is combined with clinical data to determine the likelihood of a positive outcome of the organ transplant. In some embodiments, the method further comprises determining T cell-mediated rejection (TCMR) rejection, antibody-mediated rejection (AB MR) rejection, and / or mixed rejection based on combining the amount of dd-cfDNA or the percentage of dd-cfDNA out of total cfDNA with clinical data.

[0192] Clinical data used to help predict transplant rejection can include patient and donor demographics, transplant history and immunologic risk factors (HLA match, PRA / DSA, prior sensitization), immunosuppression regimen and adherence, vital signs and clinical symptoms, routine labs and organ-function markers (e.g., creatinine, LFTs), drug levels (e.g„ tacrolimus), immunology / serology testing (DSA, complement), biopsy and histopathology findings, imaging results (ultrasound, Doppler, CT / MRI), infection and inflammatory markers, and molecular or omics biomarkers (gene-expression profiles). In some embodiments, the clinical data comprises prior transplant rejection history, demographic data, a plurality of immunological risk factors, vital signs, a plurality of organ function markers, imaging results, inflammatory markers, a gene expression profile indicative of transplant rejection, a plurality of serology test markers, and / or a plurality of biomarkers indicative of transplant rejection. In some embodiments, the demographic data comprises age or ethnicity. In some embodiments, the plurality of serology test markersN.069.W0.01comprises donor-specific antibodies (DSA). Tn some embodiments, the plurality of organ function markers comprises an amount of serum creatinine.

[0193] The data obtained from the amount of dd-cfDNA or the percentage of dd-cfDNA out of total cfDNA can be combined with clinical data for assessing or monitoring organ transplant rejection status, including determining rejection subtypes like TCMR and AB MR and Mixed rejection.

[0194] In some embodiments, a machine learning model is used to combine data obtained from the amount of dd-cfDNA or the percentage of dd-cfDNA out of total cfDNA with the clinical data. In some embodiments, combining data obtained from the amount of dd-cfDNA or the percentage of dd-cfDNA out of total cfDNA with the clinical data is performed by using any model available, such as statistical model or mathematical model or system. Such models can include for example multivariable regression models, rule-based scoring systems, threshold based decision trees, or Bayesian models. The machine learning models used for combining data in this disclosure can include tree-based ensemble models (Gradient Boosting Machines (GBM), Regularized regression models, logistic regression, support vector machines (SVM), neuronal network based approaches, probabilistic and Bayesian ML models, k-nearest neighbor (k-NN), or combinations of these approaches or stacking or ensembles.

[0195] In some embodiments, the machine learning model provides (i) a risk score of transplant rejection, and / or (ii) a TCMR risk score for T cell-mediated rejection (TCMR) rejection, an AB MR risk score for antibody-mediated rejection (ABMR) rejection, and / or a Mixed risk score for mixed rejection.

[0196] In some embodiments, the machine learning model comprises a temporal random forest model. In some embodiments a statistical, mathematical, or machine learning model disclosed herein could be used for the first stage caller or the second stage caller.

[0197] In some embodiments, the first stage caller and the second stage caller are temporal random forest models.

[0198] Treatment of Transplant rejectionN.069.W0.01

[0199] In some embodiments, the method described herein further comprises administering a treatment to an organ (e.g. kidney) transplant recipient suffering transplant rejection condition such as AB MR relapse. In some embodiments, the treatment comprises administering the anti-CD38 antibody or antigen-binding fragment thereof (e.g., daratumumab, felzartamab, or isatuximab) subcutaneously to an organ (e.g. kidney) transplant recipient suffering transplant rejection condition such as AB MR relapse. In some embodiments, the treatment comprises administering the anti-CD38 antibody or antigen-binding fragment thereof (e.g., daratumumab, felzartamab, or isatuximab) intravenously.

[0200] Abbreviations: AR: acute rejection, AB MR: antibody-mediated rejection, BxNeg: biopsy-negative, BPAR: biopsy-proven acute rejection, dd-cfDNA donor-derived cell-free DNA, DSA: donor- specific antibodies, DTW: dynamic time warping, IQR: interquartile range, KTR: kidney transplant recipients, MVI: microvascular inflammation, sABMR: suspected AB MR, TCMR: T-cell mediated rejection.Working Examples

[0201] Working Example 1: k-Medoids clustering with dynamic time warping of longitudinal dd-cfDNA testing

[0202] Summary: The presently disclosed approach for longitudinal donor-derived cfDNA (dd-cfDNA was based on unsupervised clustering and dynamic time warping distances. Dynamic time warping is a measure for assessing similarity between two time series that accounts for temporal differences by stretching or compressing segments of each series to achieve alignment with minimal cost. This was done by tracing a warping path that minimizes the cumulative distance between corresponding points. Moreover, the collected samples can be categorized into time periods, missing values can be masked and dropped. The presently disclosed method can involve a two-step approach to clustering: an initial step to define clusters based on subjects with complete data, and a second step to fit subjects with incomplete data to these clusters.

[0203] Method: Dynamic Time Warping Analysis of dd-cfDNA testingN.069.W0.01

[0204] For each organ transplant recipient having measured dd-cfDNA, the dd-cfDNA data was grouped into buckets corresponding to the closest assigned time point to the sample collection date. In other words, samples taken within 7 days prior to biopsy would be identified as the index sample (or first sample in the time series), samples between 7 and 21 days post-rejection diagnosis would be identified as the 2 week sample, etc. If a subject did not receive a draw within the specified time window, the value for that time point would be considered to be null / unknown. Using these bucketed sample sets, a time series of data for each subject was created using the values at each of the time points and null values to represent the unknown / missing data. Then to effectively perform an unsupervised clustering analysis, a pairwise distance matrix of values was created as an input to the clustering algorithm. Given the variations in timing and sampling frequency of the subjects in this analysis. Dynamic Time Warping (DTW) was selected as the distance measure in order to provide a method of comparing time series of unequal length and with temporal variation. The time series were limited to subjects with at least the index sample and one other sample in the time frame. Given limitations in the algorithm for handling null values, the null data was masked by removing any missing values and creating arrays of the known sample values for each subject.

[0205] For each subject, the DTW distance between that subject’s time series data and the time series data of all other subjects meeting the minimum sampling criteria was computed. These distances were arranged into a pairwise distance matrix for input into the unsupervised machine learning algorithm.

[0206] To perform unsupervised clustering of the dd-cfDNA time series data, the k-Medoids clustering algorithm was selected to reduce the impact of outliers on the clusters. Given the presence of subjects with incomplete data (defined as any missing samples within the defined time windows), the initial clusters were generated based on the subjects that had dd-cfDNA testing data available at every time point. The pairwise distance matrix was limited to only these subjects and input into a UMedoids clustering algorithm (Partitioning Around Medoids) to generate initial clusters. The cluster number was optimized at 4 given a review of cluster counts from 2-10 and the resultant silhouette scores and visual cluster separation.N.069.W0.01

[0207] FIG. 1 A is a visualization of an example of a clustering of DTW distances obtained from kidney transplant study described in Working Example 2 below. The clustering was based solely on the dd-cfDNA (%) values and not informed by any other data. Using the output clusters of dd-cfDNA trends, shape-based trends were identified based on the median line of the identified clusters. These trends were confirmed based on a manual review of the individual patient dd-cfDNA trajectories after the initial rejection diagnosis. The abstracted trends are identified in FIG. IB.

[0208] Four different types of trajectories of dd-cfDNA post rejection diagnosis were identified and assigned names that are descriptive of their general characteristics. LOW - dd-cfDNA that begins with no elevation and does not rise throughout the monitoring period; DROP - dd-cfDNA that starts elevated at the time of rejection but falls and stays low throughout the 8 week monitoring period; MID - “middling” dd-cfDNA that initially lowers like the DROP group but begins to rise again towards or above the clinically validated 1% threshold; HIGH - dd-cfDNA does not come down from the initial elevation (a slight dip, but then the dd-cfDNA rises again).

[0209] The abstraction of trends is based on individual variability and the presence of the overall trends in subjects with different magnitudes of dd-cfDNA levels, hence the absence of a numeric y-axis in the schematic diagram. The abstracted trends were used to classify each of the subjects based on the most closely related trend group, with respect to only the dd-cfDNA (%) trajectory and without consideration of rejection type or other clinical variables.

[0210] Results

[0211] Using the composite endpoints defined in the above section, the outcome group was determined based on the presence of any of the negative outcomes by the one year follow-up visit. A Sankey diagram showing the assignment of subjects to the different cluster groups and the resultant outcomes is shown in FIG.2.Working Example 2: Longitudinal Monitoring of the Response to Treatment of Kidney Allograft Rejection with Donor-derived Cell-free DNA

[0212] This working example provided an approach for longitudinal dd-cfDNA testing that take into account variations in individual baselines and sampling frequency as well as disease-relatedN.069.W0.01fluctuations. This working example provided an unsupervised machine learning approach for longitudinal dd-cfDNA testing, based on dynamic time warping distances, which enabled differentiation of dd-cfDNA trajectories after rejection diagnosis into groups that correlated with outcomes.

[0213] Brief Summan’ of Methods: 79 kidney transplant recipients with biopsy-proven acute rejection (BPAR) were included. dd-cfDNA tests were performed at the time of rejection diagnosis and every 2-week for 8 weeks thereafter. Dynamic time warping distances between the dd-cfDNA time series data were calculated in a pairwise manner. A distance matrix of dynamic time warping distances was created and input into a K-medoids clustering algorithm. Initial clusters were defined based on subjects with dd-cfDNA data at each time point. The number of clusters specified to the algorithm was optimized based on a combination of silhouette score and visual separation of the clusters as shown in FIG. 1.

[0214] After defining initial clusters, subjects with less complete data were fit to the existing clusters. Cluster assignments were then reviewed for accuracy based on dd-cfDNA data. The optimal number of clusters was chosen as 4, based on a silhouette score of 0.64 for subjects with complete dd-cfDNA data and clear visual separation between the clusters. Fitting the subjects with incomplete data to the existing clusters resulted in a median silhouette score of 0.58, retaining visual separation. The clusters were then abstracted into trends based on the apparent visual patterns. These trends showed significant association (p=3.18x10-7) with the clinical outcomes of the patients, including improvement vs. no improvement in kidney dysfunction, subsequent rejection episode, and presence of DSA.

[0215] Brief summary of results and conclusions: The optimal number of clusters was chosen as 4, based on a silhouette score of 0.64 for subjects with complete dd-cfDNA data and clear visual separation between the clusters (FIG. 1). Fitting the subjects with incomplete data to the existing clusters resulted in a median silhouette score of 0.58, retaining visual separation. The clusters were then abstracted into trends based on the apparent visual patterns. These trends showed significant association (p=3.18x10-7) with the clinical outcomes of the patients, including improvement vs. no improvement in kidney dysfunction, subsequent rejection episode, and presence of DSA. An unsupervised machine learning approach with K-medoids clustering andN.069.W0.01dynamic time warping presents a novel approach to characterize longitudinal trends in dd-cfDNA. Further, it revealed clinically meaningful kinetic patterns that were not detectable when applying conventional Euclidean distance as further explained below. This approach may also be applicable to other longitudinal biomarkers in organ transplantation.

[0216] Introduction

[0217] The primary goal of kidney transplantation is to improve the quality and life expectancy for patients with end stage kidney disease. A major impediment to successful transplantation is persistent alloreactivity leading to graft dysfunction and graft loss. See Hariharan et al., Long-Term Survival after Kidney Transplantation, N Engl J Med; 385(8):729-743 (2021). Acute rejection (AR) remains the leading cause of death-censored graft loss. See Meier-Kriesche et al., Lack of improvement in renal allograft survival despite a marked decrease in acute rejection rates over the most recent era, Am J Transplant;4(3):378-83 (2004); Sellares et al., Understanding the causes of kidney transplant failure: the dominant role of antibody-mediated rejection and nonadherence, Am J Transplant; 12(2):388-99 (2012); and Bbhmig et al., The therapeutic challenge of late antibody-mediated kidney allograft rejection, Transpl Int.;32(8):775-788 (2019).

[0218] Despite the plethora of treatments for transplant rejection available, as many as 32% to 37% of T-cell mediated rejection (TCMR) episodes, and 36% to 51% of antibody-mediated rejection (ABMR) episodes remain unresolved after treatment. See Landsberg et al.. Follow-up biopsies identify high rates of persistent rejection in pediatric kidney transplant recipients after treatment of T cell-mediated rejection, Pediatr Transplant'. 28(l):el4617 (2024); Pineiro et al., Influence of Persistent Inflammation in Follow-Up Biopsies After Antibody-Mediated Rejection in Kidney Transplantation, Front Med (Lausanne); 8:761919 (2021) and Parajuli et al., The Trend of Serum Creatinine Does Not Predict Follow-Up Biopsy Findings Among Kidney Transplant Recipients With Antibody-Mediated Rejection, Transplant Direct; 9(6):el489 (2023). Furthermore, among those who demonstrate recovery, recurrent AR episodes and their molecular sequelae have been shown to negatively impact long-term graft survival. See Alasfar et al., Current Therapies in Kidney Transplant Rejection, J Clin Med.;12(15) (2023). As such, there isN.069.W0.01a need to better understand and monitor the molecular factors related to AR that contribute to long-term graft dysfunction and failure.

[0219] Currently, there is a lack of consensus on the optimal methods for monitoring the resolution of rejection after treatment. There is also substantial heterogeneity regarding the definition of rejection resolution and “successful” treatment. See Koshy et al., European Survey on Clinical Practice of Detecting and Treating T-Cell Mediated Kidney Transplant Rejection, Transpl Int.;37: 12283 (2024); Sood et al., Kidney allograft rejection: Diagnosis and treatment practices in USA- A UNOS survey, Clin Transplant, 35(4) :el4225 (2021); Leblanc et al„ Practice Patterns in the Treatment and Monitoring of Acute T Cell-Mediated Kidney Graft Rejection in Canada, Can J Kidney Health Dis.; 5:2054358117753616 (2018). Evaluation of allograft function by the return of serum creatinine levels to baseline can have a delayed response and be insensitive to the type of injury, while biopsy is not ideal for routine monitoring due to its invasive nature, risks of morbidity and considerable interobserver and sampling variability. Additionally, there is often discordance between functional and histological responses, which suggests that histopathology is not entirely reflective of allograft status. In a series of 163 patients with TCMR, 68% of patients with no clinical response exhibited complete histological resolution, whereas 14% with a complete clinical response exhibited only a partial or no histologic response. See Aziz et al., How Should Acute T-cell Mediated Rejection of Kidney Transplan ts Be Treated: Importance of Follow-up Biopsy, Transplant Direct, 8(4):el 305 (2022).

[0220] Over time, inadequately treated TCMR leads to severe inflammation, epithelial dedifferentiation, tubulitis, interstitial fibrosis and tubular atrophy, and nephron loss. ABMR remains especially recalcitrant to treatment, with low rates of complete remission. Djamali et al., Diagnosis and management of antibody-mediated rejection: current status and novel approaches, Am J Transplant; 14(2):255-71 (2014). In one series of 90 patients with ABMR treated with plasma exchange, intravenous immune globulin (IVIG), and rituximab, 71% had persistent microvascular inflammation (MVI) on follow-up biopsy and 19% had persistent tubulitis on follow-up biopsy, both of which are associated with inferior graft survival. Pineiro et al., Influence of Persistent Inflammation in Follow-Up Biopsies After Antibody-Mediated Rejection in Kidney Transplantation, Front Med (Lausanne)', 8:761919 (2021). Persistent ABMR may cause interstitial fibrosis, tubular atrophy, reduced graft function, and result in chronicN.069.W0.01ABMR, and ultimately graft loss. Betjes et al., Causes of Kidney Graft Failure in a Cohort of Recipients With a Very’ Long-Time Follow-Up After Transplantation. Front Med (Lausanne);9:842419 (2022) and Pineiro et al., Influence of Persistent Inflammation in Follow-Up Biopsies After Antibody-Mediated Rejection in Kidney Transplantation, Front Med (Lausanne)', 8:761919 (2021). Consequently, there is a critical need to identify adequate biomarkers that can distinguish between those with complete resolution of their rejection from those with persistent alloreactivity and continued low-level (smoldering, subclinical) rejection. Improved immune monitoring would enable the use of appropriate and timely therapies following treatment of initial rejection episodes to better identify and control ongoing alloreactivity with the goal of enhancing overall allograft outcomes.

[0221] Novel technologies that measure the molecular processes of injury that underlie rejection and graft dysfunction hold promise as effective tools to assess response to therapy with complete resolution of rejection. One such biomarker, donor-derived cell-free DNA (dd-cfDNA), reflects the degree of molecular injury of the renal allograft, and has been shown to be more strongly associated with molecular rejection than histologic rejection. Halloran et al., The Trifecta Study: Comparing Plasma Levels of Donor-derived Cell-Free DNA with the Molecular Phenotype of Kidney Transplant Biopsies, J Am Soc Nephrol, (2022). Currently used to assess the risk of ongoing rejection, dd-cfDNA may be a promising tool for monitoring the response to treatment following allograft rejection. Recent evidence that dd-cfDNA levels increase 2 and 5 months before histologic diagnosis of TCMR and ABMR, respectively, demonstrates that dd-cfDNA is dynamic and sensitive to molecular injury. Bromberg et al., Elevation of Donor-derived Cell-free DNA Before Biopsy-proven Rejection in Kidney Transplant, Transplantation', 108(9): 1994-2004 (2024).

[0222] While dd-cfDNA levels have shown excellent performance in detecting ongoing injury and rejection, studies focused on monitoring dd-cfDNA levels during and after rejection treatment are limited. Findings from these studies generally indicate that dd-cfDNA levels decrease during and after treatment. Kanzow et al., Graft-derived cell-free DNA as an early organ integrity biomarker after transplantation of a marginal HELLP syndrome donor liver, Transplantation; 98(5):e43-5(2014); Hidestrand et al., Highly sensitive noninvasive cardiac transplant rejection monitoring using targeted quantification of donor-specific cell-freeN.069.W0.01deoxyribonucleic acid, J Am Coll Cardiol.;63(12):1224-1226 (2014); Beck et al., Donor-Derived Cell-Free DNA Is a Novel Universal Biomarker for Allograft Rejection in Solid Organ Transplantation, Transplant Proc.; 47(8):2400-3 (2015); Hinojosa et al., Donor-derived Cell-free DNA May Confirm Real-time Response to Treatment of Acute Rejection in Renal Transplant Recipients, Transplantation;103(4):e61(2019); Shen et al., Prognostic value of the donor-derived cell-free DNA assay in acute renal rejection therapy: A prospective cohort study, Clin Transplant.; 34(10):el4053 (2020); Wolf-Doty et al., Dynamic Response of Donor-Derived Cell-Free DNA Following Treatment of Acute Rejection in Kidney Allografts, Kidney360;2(4):729-736 (2021); Steggerda et al., Use of a donor-derived cell-free DNA assay to monitor treatment response in pediatric renal transplant recipients with allograft rejection, Pediatr Transplant;26(4):el4258 (2022); Osmanodja et al,, Donor-Derived Cell-Free DNA for Kidney Allograft Surveillance after Conversion to Belatacept: Prospective Pilot Study. J Clin Med:, 12(6) (2023); Osmanodja et al., Donor-Derived Cell-Free DNA as a Companion Biomarker for AMR Treatment With Daratumumab: Case Series, Transpl Int:, 37:13213 (2024). However, data on longitudinal dd-cfDNA responses and associations with outcomes following treatment are lacking. In this study, we assessed the role of dd-cfDNA levels in monitoring response to therapy for rejection by identifying longitudinal dd-cfDNA trends during treatment of AR and assessing their relationships with clinical outcomes at one- year post-rejection diagnosis.

[0223] Methods

[0224] Study Design: Patients were included in the analysis if they had a renal biopsy diagnosis and a dd-cfDNA result within the 7 days before the biopsy. Patients were considered eligible for enrollment if they: 1) were 18 years of age or older at the time of signing the informed consent form; 2) had at least one renal allograft transplant; 3) were scheduled to undergo a kidney biopsy; 4) were able to read, understand, and provide written informed consent; and 5) were willing and able to comply with the study-related procedures. Patients were excluded if they had received a non-kidney allograft, received a kidney from an identical twin, or were receiving dialysis or had active cancer at the time of enrollment. Patients were monitored and provided care in accordance with the local standard of care. At the time of index biopsy, clinical and demographic data related to recipient characteristics, donor characteristics, transplant procedure, clinical examination, and blood and urine laboratory tests were collected (FIG. 3).N.069.W0.01

[0225] Patients with biopsy-proven acute rejection (BPAR) at the index biopsy were asked to provide at least four additional blood samples for dd-cfDNA testing at weeks 2, 4, 6, and 8. At follow-up visits, transplant assessment comprising clinical examination, laboratory data, immunosuppressive treatment, and dd-cfDNA testing were conducted. Results of the dd-cfDNA testing were not shared with either the patients or the attending physicians and consequently were not used in patient management or decision-making. Follow-up data were collected from all patients at one-year post-enrollment, with outcomes determined using kidney function tests (serum creatinine, estimated glomerular filtration rate, and proteinuria) and the presence of subsequent BPAR, death-censored graft loss, and detection of donor- or HLA-specific antibodies (DSA). Health Insurance Portability and Accountability Act (HIPAA) and General Data Protection Regulation (GDPR) guidelines were followed for US and EU sites, respectively. Informed consent was obtained from all 584 subjects. This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guidelines. The study was performed in full adherence to the Declaration of Helsinki. The clinical and research activities reported are consistent with the Principles of the Declaration of Istanbul outlined in the Declaration of Istanbul on Organ Trafficking and Transplant Tourism.

[0226] dd-cfDNA Testing: All blood samples collected for dd-cfDNA testing using the Prospera™ test (Natera, Inc., Austin, TX) were drawn in two 10 mL Streck Cell-Free DNA BCT tubes and shipped to the processing laboratory. Extracted cfDNA was amplified using massively-multiplexed PCR targeting 13,926 single nucleotide polymorphisms (SNPs) selected to maximize the number of informative SNPs across ethnicities, followed by next-generation sequencing of the resulting amplicons on the Illumina NextSeq 500 with a minimum of 8 million reads per sample. See Altug et al., Analytical Validation of a Single -nucleotide Polymorphism-based Donor-derived Cell-free DNA Assay for Detecting Rejection in Kidney Transplant Patients, Transplantation; 103(12):2657-2665 (2019). The samples were processed according to a standard operating protocol used in the Clinical Laboratory Improvement Amendments laboratory responsible for running the Prospera tests.

[0227] Allograft pathology classification protocol: Allograft pathology reports consisting of histology, indirect immunofluorescence, and electron microscopy were retrospectively reviewed at a central pathology site, focusing on classification and standardization based on the updatedN.069.W0.01BANFF 2019 and 2022 Kidney Meeting Report criteria. Cases with fewer than six glomeruli were excluded due to inadequate sampling, while cases with 6-10 glomeruli and no mediumsized vessels were included with limitations, and their positive findings were recorded. All pathology reports were categorized into three main diagnostic groups: Acute Rejection (AR), Biopsy Negative (BxNeg), and Other Injury. The AR group was further classified into acute antibody-mediated rejection (AB MR), T-cell mediated rejection (TCMR), and mixed rejection, as diagnosed according to the updated BANFF 2019 classification criteria. BxNeg cases were characterized by a lack of significant findings or interstitial fibrosis and tubular atrophy without evidence of active inflammation or injury. ‘Other Injury’ was a broad category that included borderline inflammation, various infection types, and recurrent disease (Table 1). In accordance with the 2022 BANFF commentary, cases demonstrating threshold microvascular inflammation (g+ptc score greater than 2) in the absence of C4d and DSA detection (MVI, DSA- and C4d-), or subthreshold microvascular inflammation with positive DSA (probable AB MR) were grouped into a new category called suspected AB MR (sABMR). Biopsies that demonstrated features of both AB MR and TCMR were classified as AR under the mixed rejection category.Table 1N.069.W0.01

[0228] Definition of the composite endpoint at 1 year post-rejection: A composite endpoint, termed “Negative Outcome”, was defined as one or more of the following 4 components: 1) the presence of death-censored graft loss, 2) the development of a subsequent BPAR episode. 3) the presence of persistently elevated donor specific antibodies (DSA) levels at the time of biopsy, or detection of de novo DSA during the follow-up period, and 4) lack of resolution of renal dysfunction, defined as either a lack of improvement of in eGFR (at least 5% from the baseline measurement at the time of BPAR), or a proteinuria / creatinine ratio >0.5 mg / mg. A “Positive Outcome” was defined by the absence of all 4 events in the composite endpoint. eGFR was calculated using the 2021 CKD-EPI formula without race. Inker et al.. Am J Kidney Dis.;78(5):736-749 (2021).

[0229] Identification of dd-cfDNA trends: dd-cfDNA samples were assigned to the closest time point from the index biopsy date: 2, 4, 6, and 8 weeks, and must have been drawn within + / - 7 days of the time point. An unsupervised machine learning approach was taken to identify trends, in which dd-cfDNA results at each timepoint for all patients were clustered using the Partitioning Around Medoids algorithm. See Kaufman et al.. Partitioning Around Medoids (Program PAM). In: Kaufman L, Rousseeuw PJ, eds. Kinding Groups in Data: 1990:68-125, Wiley Series in Probability and Statistics. Dynamic time warping (DTW), which assesses the distance of the dd-cfDNA values over time between subjects with at least 2 dd-cfDNA samples (of which one was matched with the index biopsy) was used to generate a pairwise distance matrix, defined as a dictionary of the DTW distance values between any two given subjects in the dataset. Missing data points were masked and removed from the arrays for the distance calculation. Clustering was first performed using the DTW distances between only patients with dd-cfDNA data at eachN.069.W0.01of the 5 time points. The appropriate number of clusters was selected based on a balance of silhouette scores and visual cluster separation. After the initial clustering was performed, the remaining patients (those with fewer than 5 samples) were fit to the existing clusters. The cluster assignments for each individual generated by the machine learning algorithm were manually reviewed and reclassified as appropriate, without reference to clinical characteristics. All analyses were performed in python version 3.10.9.

[0230] Statistical analysis: Continuous variables were reported as mean and standard deviation or median and interquartile range as appropriate. The distributions were compared with the Mann Whitney test. Paired dd-cfDNA tests were compared over time using the Wilcoxon signed-rank text. Categorical variables were compared with Chi-square tests, and odds ratios were calculated and compared using Fisher’s exact test. P-values were corrected for multiple comparisons as appropriate using false discovery rate corrections (Benjamini-Hochberg).

[0231] Results

[0232] Demographics and clinical characteristics: A total of 584 kidney transplant recipients were enrolled, of which 96 were excluded for reasons including no dd-cfDNA test result prior to index biopsy (n=4), dd-cfDNA blood draw within 14 days of transplant (n=14), no index biopsy was performed (n=51 ), or the index biopsy sample was insufficient or unable to be diagnosed (n=27). The remaining 488 subjects with contemporaneous dd-cfDNA test and biopsy results had a median age of 51.4 years (IQR: 39.9-61.4), were 58% male, 57% White and 10% (n=49) had a prior kidney transplant (Table 2). Among this cohort, the median body mass index was 29 (kg / m2) (IQR: 25.2-33.6), and the median time from transplant to index biopsy was 330 days (IQR: 110-856). The primary indications for transplantation were glomerulonephritis (18.4%; 90 / 488), hypertension (18.7%; 91 / 488), diabetes mellitus (18.2%; 89 / 488), and polycystic kidney disease (10.0%; 49 / 488). The cohort was similar to national data obtained from the Scientific Registry of Transplant Recipients for KTR from 2018 to 2024 regarding age, sex, BMI, race, ethnicity, primary indication for transplantation, and retransplant status. Lentine et al., OPTN / SRTR 2022 Annual Data Report: Kidney, Am J Transplant; 24(2S 1): S 19-S 118 (2024).Table 2N.069.W0.01<><<N.069.W0.01<

[0233] Among the 488 index biopsies, 80.3% (n=392) were considered non-rejection (biopsy negative: n=204; Other Injury: n=188), and 19.7% (n=96) were classified as AR (ABMR: n=39, TCMR: n=48, and Mixed: n=9) (FIG.4). A total of 70.1% (n=342) of biopsies were for cause, and 29.9% (n=146) were for surveillance.

[0234] dd-cfDNA characteristics among different pathology groups: A total of 722 dd-cfDNA samples were included in the analysis, of which 392 samples were from subjects in the non-AR cohort (one sample per patient), and 330 were from subjects in the AR cohort (median 4 samples per patient). At the time of the index biopsy, the median dd-cfDNA fractions among all pathology groups were significantly elevated (TCMR [n=48; 0.78%, IQR: 0.46-2.0%], ABMR [n=39; 2.6%. IQR: 1.5-3.6%], Mixed [n=9; 3.2%. IQR: 2.1-6.4%], and Other Injury [n=188; 0.42% IQR: 0.18-1.0%]) as compared to biopsy negative (n=204; 0.23%, IQR: 0.10-0.52%) (p- values all <0.0001), (FIG. 5; and see FIG. 6 for dd-cfDNA values for sub-categories in the Other Injury category.

[0235] At the time of rejection diagnosis, the median dd-cfDNA% was 1.73% (IQR:0.58-3.31%) for all rejection cases, which decreased to 0.51% at week 2 post-diagnosis (IQR:0.19-1.04%), and remained low (0.63%, IQR:0.24-1.86%) at weeks 4-8 (FIG.7). When stratified by rejection type, the median dd-cfDNA% among TCMR cases was significantly decreased at weeks 2-8 post-diagnosis (0.78% at T=0 vs. 0.33%, 0.38%, 0.38%, 0.27% at weeks 2, 4, 6, 8; p<0.001 at all time points), while the median dd-cfDNA% among ABMR and mixed rejection samples rebounded at weeks 4-8, after an initial decrease at 2 weeks post-diagnosis (2.57% at T=0 vs. 0.65%, 1.76%, 1.21%, 2.27% at weeks 2, 4, 6, 8) (FIG. 8).

[0236] dd-cfDNA trends post-rejection diagnosis and during treatment: Longitudinal dd-cfDNA test results were classified into four distinct dd-cfDNA trends within subjects experiencing rejection: those with consistently low dd-cfDNA (LOW), those with initially high but steeplyN.069.W0.01dropping dd-cfDNA levels that stayed low (DROP), those with medium levels of dd-cfDNA that initially decreased and slightly increased again (MID), and those with high dd-cfDNA levels that initially decreased but increased to a high level (HIGH) (FIG. 9).

[0237] Of the 96 AR cases, 68.8% (n=66) had follow-up data and were assigned to one of the dd-cfDNA trend groups based on the individuals’ longitudinal dd-cfDNA data (TCMR: n=37; ABMR: n=24; Mixed rejection: n=5; FIG. 9). In total, 13 cases were classified in the LOW and 12 in the DROP groups, 17 in the MID group, and 24 in the HIGH group (FIG. 2). ABMR, TCMR, and mixed rejection cases were analyzed separately in non-responder and responder categories defined by dd-cfDNA trends; comparable results were found (FIG. 10).

[0238] Associations between dd-cfDNA trends in the two months following BPAR and clinical outcomes at one year: Each rejection case was categorized into one of two endpoints based on outcomes and kidney function at the end of the 1-year follow-up period: (1) Negative Outcome, defined as the occurrence of death censored graft loss, subsequent rejection, presence of DSA and / or no improvement in kidney function or (2) Positive Outcome, defined as resolving kidney dysfunction and the absence of any of the conditions mentioned above. The relationships between the dd-cfDNA trend groups and endpoints were assessed (FIG. 2). In total, 75.8% (50 / 66) of KTR experienced a Negative Outcome, including death-censored graft loss (n=5), future rejection (n=l 0), presence of DSA (n=14), and lack of resolving kidney dysfunction (n=45). Most cases in the MID and HIGH dd-cfDNA trend groups experienced Negative Outcomes (17 / 17 and 23 / 24, respectively), suggesting that the cases in these groups did not adequately respond to treatment (“non-responders”). The majority of the cases in the LOW and DROP dd-cfDNA trend groups (8 / 13 and 7 / 12, respectively) experienced a Positive Outcome, suggesting that these cases mostly responded to treatment (“responders”).

[0239] The odds of exhibiting resolving kidney dysfunction independent of other outcomes were 13x higher among patients in the responder category compared to the non-responder category (OR: 12.8, p=2.15xl0-5). In contrast, the odds of experiencing a Negative Outcome were 60x higher among patients in the non-responder category, compared to the responder category (OR: 60.0, p=3.18xl0’7)).N.069.W0.01

[0240] We also compared using Euclidean distance and DTW to analyze the dd-cfDNA longitudinal trends from the kidney transplant recipients. FIG. 11 schematically illustrates how Euclidean distance and DTW distance differ when applied to the dd-cfDNA trajectories of these kidney transplant recipients. Under a Euclidean framework, each subject’s dd-cfDNA value at a given nominal time point must be compared directly to the value at the same nominal time point in another subject. In contrast, the DTW-based comparison depicted in FIG. 11 allows the time axis of the trajectories to be warped, so that comparable phases of the dd-cfDNA response after rejection diagnosis are aligned despite irregular or asynchronous sampling. This illustrates why Euclidean distance is effectively limited to complete, uniformly sampled time series, whereas DTW can robustly capture similarity in real-world longitudinal biomarker data.

[0241] FIG. 12 provides a visualization of the pairwise dynamic time warping distance matrix used as input for the k-Medoids clustering. Each cell in the clustermap represents the DTW distance between the longitudinal dd-cfDNA trajectories of a pair of subjects in the cohort, with warmer colors indicating smaller distances and therefore more similar kinetics, and cooler colors indicating larger distances and less similar kinetics. The accompanying dendrograms illustrate how subjects group together based on these DTW distances and demonstrate the presence of distinct clusters of patients with similar dd-cfDNA temporal patterns.

[0242] Discussion

[0243] Incomplete resolution of kidney allograft rejection substantially impacts long-term graft outcomes. This presents a clear unmet need for a non-invasive molecular test that can assess treatment responses in the period immediately after rejection diagnosis. The study provided herein is the first to use longitudinal dd-cfDNA trends to assess rejection resolution during treatment in KTR after BPAR, with the findings showing a clear association between dd-cfDNA trends and outcomes. Herein, we identified a cluster classification model for stratifying transplant recipients based on dd-cfDNA dynamics in the 8 weeks after diagnosis of AR. The four clusters are named LOW, DROP, MID, and HIGH. We found the odds of resolving kidney dysfunction to be 13x higher among patients in the LOW and DROP groups compared to the MID and HIGH groups. At the same time, the odds of experiencing a negative outcome were 60x higher in the MID and HIGH groups, as compared to the LOW and DROP groups. TheseN.069.W0.01findings suggest those in the LOW and DROP groups were responsive to treatment (“responders”), while those in the MID and HIGH groups had an inadequate response to treatment (“non-responders”). The use of dd-cfDNA trends indicative of response to treatment may enable risk stratification for long-term outcomes and can guide the need for alternate or additional treatments.

[0244] A closer examination of these two categories provides valuable insights. Most responders (60%) went on to have positive outcomes at the end of 1 year, a rate considerably higher than that of the full cohort (24.2%). In contrast, non-responders had high dd-cfDNA levels that may have initially dropped but then rebounded, reflecting their inadequate response to treatment. Indeed, only 1 of 41 non-responders experienced a positive outcome. These KTR may have benefited from a more effective treatment. The dd-cfDNA kinetics of these groups reflect the known poor prognosis with graft loss among patients diagnosed with ABMR, which comprised 51% of these two groups. See Schinstock et al., Recommended Treatment for Antibody-mediated Rejection After Kidney Transplantation: The 2019 Expert Consensus From the Transplantion Society Working Group, Transplantation; 104(5):911-922 (2020). Fortunately, promising new treatments for ABMR, such as anti-CD38 monoclonal antibodies, have shown efficacy in ameliorating ABMR. In a recent study of 22 patients randomized to treatment with the CD38 monoclonal antibody felzartamab or placebo, 9 of 11 patients in the interventional group showed resolution of morphological ABMR compared to only 2 of 10 patients in the placebo group. Notably, the interventional group had lower dd-cfDNA levels after treatment, reflective of the rejection resolution. Mayer et al.. A Randomized Phase 2 Trial of Felzartamab in Antibody-Mediated Rejection, N Engl J Med.; 391(2): 122-132 (2024). With these therapies progressing toward approval, the use of dd-cfDNA levels to monitor treatment efficacy in ABMR patients will become increasingly important.

[0245] A prior report showed that dd-cfDNA levels measured pre-diagnosis was associated with outcomes in patients with low-grade rejection-like activity. Stites et al., High levels of dd-cfDNA identify patients with TCMR 1A and borderline allograft rejection at elevated risk of graft injury, Am J Transplant; 20(9):2491-2498 (2020). Specifically, 42 patients with borderline / TCMR1A and elevated dd-cfDNA levels demonstrated a higher drop in eGFR, and higher rates of de novo DSA detection and future or persistent rejection, as compared to a similar cohort of 37 patientsN.069.W0.01with lower dd-cfDNA levels. Our data suggest that longitudinal post-diagnosis dd-cfDNA trends are more discerning of future outcomes compared to a single dd-cfDNA value at the time of biopsy. While previous studies have shown a statistically significant reduction in dd-cfDNA levels in the first few weeks after treatment for rejection was initiated, Benning et al., Donor-Derived Cell-Free DNA (dd-cfDNA) in Kidney Transplant Recipients With Indication Biopsy-Results of a Prospective Single-Center Trial, Transpl Int.; 36:11899 (2023), Wolf-Doty et al.. 2021, Shen et al., 2020, and Hidestrand et al., 2014, we are not aware of any investigation into the use of post-treatment dd-cfDNA testing to stratify patients by efficacy of treatment or future outcomes. Patients in the DROP group exemplify the utility of post-diagnosis dd-cfDNA trends; if they had been risk-stratified based on their dd-cfDNA results at the time of biopsy, they would have been characterized as more likely to have poor outcomes (non-responders) due to their initially high dd-cfDNA, despite their relatively high rate of positive outcomes.1 246 / Demonstrating that post-rejection dd-cfDNA trends reflect outcomes one year after diagnosis suggests that these dd-cfDNA trends can provide valuable insights into future allograft monitoring and function. This could not only help physicians select the optimal therapeutic plan for each patient, but also minimize unnecessary interventions, reduce side effects, and improve long-term graft survival. This may also be helpful after the treatment period in terms of stratifying patients by risk for closer surveillance, especially in patients further from transplant receiving care from a general nephrology clinic. Of note, we found a lack of correlation between serum creatinine and dd-cfDNA levels, suggesting that serum creatinine levels would not follow similar trends. Additionally, we note that these trends are agnostic of the underlying cause of rejection, e.g., non-compliance, and thus should apply to all patients. Looking forward, dd-cfDNA-guided patient management strategies could be key in advancing precision medicine for transplant patients.

[0247] The main limitation of this study was that a post-treatment biopsy was not required, limiting our ability to confirm the near-term impact of treatment histologically. We note that the limited number of post-treatment biopsies in the study (n=2 in the 8 weeks post-biopsy), and low adoption rate of this practice among transplant centers, specifically demonstrates the clinical need for a biomarker such as dd-cfDNA to monitor rejection resolution. Moreover, since we have shown an association between the dd-cfDNA trends and longer-term kidney function andN.069.W0.01outcomes one year later, the lack of biopsy, which is only a proxy for outcomes, is less important.

[0248] In conclusion, these data show four distinct clusters of dd-cfDNA trends observed in the immediate post-rejection period that were statistically associated with kidney function and outcomes at one year. These dd-cfDNA trends could allow physicians to optimize treatment of patients with BPAR in a personalized manner and potentially improve their long-term prognosis.

[0249] Working Example 3: Hierarchical Machine Learning Framework for Predicting Kidney Rejection Subtypes Using Longitudinal dd-cfDNA and Clinical Features

[0250] This working example describes a hierarchical machine learning approach for noninvasively detecting kidney allograft rejection and differentiating rejection subtypes using longitudinal donor-derived cell-free DNA (dd-cfDNA) testing in combination with clinical and immunologic variables. A Two- Stage temporal random forest model was developed to (i) identify biopsy-proven rejection and (ii) classify rejection into T cell-mediated rejection (TCMR) or antibody-mediated rejection (ABMR). This example demonstrates that integrating longitudinal dd-cfDNA with routine clinical data allows accurate prediction of rejection subtypes and provides a scalable extension of existing dd-cfDNA-based transplant surveillance.

[0251] FIG. 13 schematically illustrates how the hierarchical model operates in general as it can be applied to any organ transplant rejection assessment. Transplant recipients first undergo routine dd-cfDNA testing, which provides a two-threshold call (2TC call) identifying low-risk and high-risk samples. The samples identified as high risk are then further analyzed in the Two-Stage model. The first stage model combines dd-cfDNA measurement information with additional clinical and immunologic features and generates a refined rejection risk score, classifying each sample as either low risk or high risk for rejection. For samples that are assigned to be high-risk in both the 2TC model and the first stage caller (Stage 1), a second stage caller (Stage 2) is applied to differentiate the likely rejection subtypes. Stage 2 produces calibrated risk scores for ABMR and TCMR and assigns a final label (TCMR, ABMR, or type no-call) for eligible high-risk samples. This hierarchical architecture allows the model to preserve established low-risk calls while adding both enhanced rejection detection and rejection subtype prediction for patients at elevated risk.N.069.W0.01

[0252] FIG. 14 further illustrates the Two-Stage model architecture. Stage 1 is a rejection detection model trained on all available draws in the cohort, using a temporal random forest (TeRF) classifier to distinguish rejection from non-rejection. Its inputs include dd-cfDNA fraction and quality metrics from the 2TC based routine assay (such as donor fraction estimate and data quality score), along with clinical features such as age, ethnicity, prior rejection history, serum creatinine, and donor specific antibody (DSA) status. Stage 1 outputs a calibrated risk score, a binary high-risk / low-risk label, a rejection threshold, and feature importance rankings. Stage 2 comprises two additional TeRF models: one model for ABMR versus no-ABMR and a separate model for TCMR versus no-TCMR. Stage 2 outputs calibrated risk scores for ABMR and TCMR, a final categorical label (TCMR, ABMR, or type no-call), and feature importance values for each subtype. This model is designed to extend existing dd-cfDNA testing by adding a two stage machine learning framework for both detection and subtype classification of rejection, using longitudinal data and clinical context.

[0253] Methods

[0254] Study cohort and monitoring: Data were obtained from an analysis of a cohort of 1,371 kidney transplant recipients enrolled at 37 centers in the United States as part of a multicenter prospective study. Subjects were monitored for the first 19 months after transplantation with serial dd-cfDNA testing described herein. During the monitoring period, a total of 11,243 dd-cfDNA tests were performed, with a median of 8 tests per patient (interquartile range [IQR]: 7-9). Clinical and demographic variables, including age, ethnicity, body weight, serum creatinine, donor-specific antibodies (DSA), and rejection history, were collected according to standard clinical practice.

[0255] Biopsy and rejection classification: During the observation period. 85 patients experienced at least one biopsy-proven rejection event. Biopsy findings were classified according to the Banff criteria, as described above, into T cell-mediated rejection (TCMR), antibody-mediated rejection (ABMR), mixed rejection, and non-rejection categories. Among the rejection episodes included in this analysis, 57 had at least one TCMR occurrence, 47 had at least one ABMR occurrence and 15 had mixed rejection.

[0256] Feature construction: For each dd-cfDNA test, a feature set was constructed incorporating both the current biomarker values and longitudinal information from prior tests. The followingN.069.W0.01features were considered: dd-cfDNA fraction (%) measured by the dd-cfDNA assay and further metrics were obtained, such as donor fraction estimate (DFE) and donor quantification score (DQS); summary measures of longitudinal dd-cfDNA dynamics over recent months (for example, changes or trends in dd-cfDNA levels over time); demographic variables (including age and ethnicity); serum creatinine; DSA history; and rejection history (prior biopsy-proven rejection, its subtype, and time since rejection). Where applicable, additional clinical variables such as body weight were included to reflect overall patient status. Longitudinal features were derived from repeated measurements within a predefined look-back window prior to the time of interest, similar to the time-windowed approaches described in working example 2, to capture temporal patterns in dd-cfDNA and related variables.

[0257] Two-Stage hierarchical model: A two stage temporal random forest framework (Stage 1 is a first stage caller, and Stage 2 is a second stage caller as explained above) was developed to model rejection risk and subtype, as illustrated in FIGS. 13, 14, and 15. In Stage 1. the objective was to classify each dd-cfDNA test as indicative of rejection (biopsy-proven acute rejection) versus nonrejection. Stage 1 was applied across the entire cohort of 1,371 kidney transplant recipients and used all available dd-cfDNA tests and associated clinical features. In Stage 2, the objective was to classify the subtype of rejection (TCMR versus AB MR) among subjects at elevated risk.

[0258] Stage 1 - Rejection detection: Stage 1, or the first stage caller, used a temporal random forest classifier to predict biopsy-proven acute rejection at or near the time of a dd-cfDNA draw. The model incorporated dd-cfDNA measurements and rejection history. The primary outcome was a binary label indicating low or high risk of biopsy-confirmed rejection. For training and evaluation, dd-cfDNA draws were aligned to the nearest biopsy date where applicable, using a predefined time window, and non-rejection draws were defined based on absence of biopsy-proven rejection and clinical stability in the corresponding time frame. Model performance was assessed using three-fold cross-validation across the entire cohort. For comparison to an existing noninvasive approach, Stage 1 performance was evaluated alongside the established two-threshold caller, which interprets dd-cfDNA fraction using predefined cutoffs. A two-threshold call (2TC) classifies high risk if one call of dd-cfDNA are made above a predefined threshold value over a pre-defined time window as described elsewhere herein.N.069.W0.01

[0259] Stage 2 - Rejection subtype classification. Stage 2 (or the second stage caller) was designed to operate on a subset of tests considered high risk, focusing on predicting the underlying rejection subtype in patients with determined to be high risk by the two-threshold caller (2TC) and the Stage 1 prediction in the Two-Stage model. To construct the Stage 2 test dataset, dd-cfDNA tests were selected from patients with high-risk 2TC results (as determined by the clinically validated thresholds) who also had at least two dd-cfDNA draws in the prior six months. The Stage 2 model has two temporal random forest classifiers: one trained to discriminate AB MR versus no-ABMR and another trained to discriminate TCMR versus no-TCMR, using features including longitudinal dd-cfDNA dynamics, age, serum creatinine, DSA status and history, and prior rejection subtype. For each eligible high-risk sample, Stage 2 produced calibrated risk scores for ABMR and TCMR and a final categorical label (ABMR, TCMR, or type no-call), as depicted schematically in FIG.15.

[0260] Model training and evaluation: Both Stage 1 and Stage 2 models were trained using standard random forest implementations. Each model was evaluated using three-fold cross-validation to estimate generalization performance. For Stage 1 (any rejection versus non-rejection), sensitivity and specificity were calculated and compared to those of the two-threshold (2TC) caller. For Stage 2 (TCMR versus ABMR), classification performance was summarized using area under the receiver operating characteristic curve (AUC) for each subtype. To further characterize model behavior, a confusion matrix was generated comparing Stage 2 predicted subtypes with biopsyconfirmed diagnoses among high-risk Prospera tests, and feature importance scores were analyzed to identify the most influential predictors in each stage.

[0261] Results

[0262] Cohort characteristics: The interim analysis included 1,371 kidney transplant recipients with a median age of 55.4 years and a median body weight of 81.3 kg. The cohort was 56.1% Caucasian. Over the 19-month monitoring period, 11,243 dd-cfDNA tests were performed, with a median of 8 tests per patient (IQR: 7-9). Eighty-five patients experienced at least one biopsyconfirmed rejection event, classified as >=1 TCMR (n=57), => 1 ABMR (n=47), or >= mixed rejection (n=15).N.069.W0.01

[0263] Stage 1 - Rejection detection performance: Across the entire cohort, the Stage 1 temporal random forest model achieved a sensitivity of 0.93 (95% confidence interval [CI]: 0.84-1.00) and a specificity of 0.96 (95% CI: 0.95-0.96) for detecting biopsy-proven rejection. In comparison, the two-threshold algorithm (2TC) alone demonstrated a sensitivity of 0.89 (95% CI: 0.85-0.93) and a specificity of 0.94 (95% CI: 0.91-0.98). The improvement in performance observed with the Stage 1 model is consistent with the added contribution of longitudinal dd-cfDNA trends and rejection history beyond single time point dd-cfDNA measurements.

[0264] Stage 2 - Rejection subtype classification performance: Stage 2 classification was evaluated in patients with high-risk Prospera results who had at least two dd-cfDNA tests in the preceding six months. Within this subset, the temporal random forest model achieved robust discrimination between TCMR and ABMR. For TCMR, the model achieved an AUC of 0.89 (95% CI: 0.78-1.00), with a discrimination score of 4.18, indicating that cases classified as high probability for TCMR were approximately four times more likely to represent true TCMR than would be expected by chance. For ABMR, the model achieved an AUC of 0.96 (95% CI: 0.95-0.98), with a discrimination score of 11.75, indicating strong enrichment for true ABMR among cases predicted as high probability ABMR. See Table 3 and FIG. 16.

[0265] Table 3. Rejection subtype classification performance.

[0266] A confusion matrix (Table 4 and FIG. 17) comparing Stage 2 subtype predictions to biopsy-confirmed diagnoses for high-risk tests demonstrated accurate differentiation between TCMR and ABMR, with a low rate of misclassification between the two subtypes. Mixed rejection cases were infrequent in this subset and, where misclassified, tended to be assigned to the predominant immunologic phenotype (TCMR or ABMR) present in the biopsy.N.069.W0.01

[0267] Table 4. Confusion matrix of Stage 2 subtype predictions versus clinical truth for high risk Prospera results.

[0268] Feature importance analysis indicated that Stage 1 detection was primarily driven by the current dd-cfDNA fraction, measures of recent longitudinal dd-cfDNA behavior, and prior rejection history. In contrast, TCMR classification relied mostly on longitudinal dd-cfDNA dynamics, age, prior TCMR rejection history, and serum creatinine and ABMR classification was driven largely by prior ABMR history, rejection recency, and current dd-cfDNA fraction.

[0269] This working example demonstrated that a hierarchical machine learning framework enabled noninvasive detection of kidney allograft rejection and differentiation of rejection subtypes by integrating longitudinal dd-cfDNA measurements with routinely collected clinical and immunologic features. By first identifying patients at high risk of rejection and then estimating whether the rejection is more likely TCMR or ABMR, this two-stage approach can provide earlier and more precise risk stratification than use of dd-cfDNA thresholds alone. The framework represents a practical and scalable extension of existing dd-cfDNA-based surveillance in kidney transplantation.

Claims

N.069.W0.01CLAIMS WHAT IS CLAIMED IS:

1. A method for preparing a non-naturally occurring composition for assessing an organ transplant, comprising:longitudinally collecting two or more blood, plasma, serum or urine samples from an organ transplant recipient;extracting cell-free DNA (cfDNA) from the two or more blood, plasma, serum or urine samples or portions thereof from the organ transplant recipient, wherein the extracted cfDNA comprises a mixture of donor-derived cfDNA (dd-cfDNA) and recipient-derived cfDNA;preparing a sequencing library from the extracted cfDNA or its derivative and performing high-throughput sequencing on the sequencing library to obtain sequence reads;quantifying an amount of dd-cfDNA or a percentage of dd-cfDNA out of total cfDNA based on the sequence reads; andassessing the organ transplant by generating time series data of the amount of dd-cfDNA or the percentage of dd-cfDNA out of total cfDNA, wherein the time series data of the amount of dd-cfDNA or the percentage of dd-cfDNA out of total cfDNA are input for a machine learning (ML) model, wherein the ML model improves upon the use of a computerized system for determining likely outcomes by assigning the time series data for the organ transplant recipient to a cluster classification generated by the ML model from training data, and wherein the cluster classification provides a likelihood of a positive outcome of the organ transplant.

2. The method of claim 1, wherein a first sample of the two or more blood, plasma, serum or urine samples is collected prior to a scheduled biopsy or diagnostic test of the transplant recipient, and a second sample of the two or more blood, plasma, serum or urine samples is collected after the scheduled biopsy or diagnostic test.

3. The method of claim 1 or claim 2, wherein the organ transplant recipient suffers from a transplant rejection condition, wherein a first sample of the two or more blood, plasma, serum or urine samples is collected prior to starting a treatment for the transplant rejection condition, and a second sample of the two or more blood, plasma, serum or urine samples is collected after starting the treatment for the transplant rejection condition.N.069.W0.

014. The method of any one of claims 1 -3, wherein the percentage of dd-cfDNA out of total cfDNA for each of the two or more longitudinally collected blood, plasma, serum or urine samples are categorized into two or more time periods.

5. The method of claim 4, wherein the two or more time periods comprise a first time period within 7 days prior to (i) a scheduled biopsy or diagnostic test for transplant rejection, or (ii) starting the treatment for the transplant rejection condition.

6. The method of claim 4 or 5, wherein a pairwise distance matrix of the percentage of dd-cfDNA out of total cfDNA values at each of the two or more time periods is generated.

7. The method of claim 6, wherein the pairwise distance matrix is the input for the ML model.

8. The method of claim 7, wherein the ML model is an unsupervised clustering algorithm.

9. The method of any one of claims 6-8, wherein pairwise distance matrix comprises a distance measure, and wherein the distance measure is dynamic time warping (DTW).

10. The method of any of the preceding claims, wherein the cluster classification was generated from time series data of a population of transplant recipients as training data.

11. The method of claim 10, wherein the cluster classification for the organ transplant recipient is performed by comparing the DTW of the time series data of the organ transplant recipient to the DTW of the time series data of the population of transplant recipients11. The method of any of the preceding claims, wherein the cluster classification indicates a likelihood of a positive outcome or a negative outcome of the organ transplant or a treatment for transplant rejection.

12. The method of claim 8, wherein the unsupervised clustering algorithm is a k-Medoids clustering algorithm.

13. The method of claim 12, wherein a cluster number is based on a resultant silhouette score and a visual cluster separation.N.069.W0.0114. The method of claim 13, wherein the cluster number is 2- 10, preferably wherein the cluster number is 4.

15. The method of claim 14, wherein the cluster number is 4, and the clusters comprise i) low, ii) drop, iii) mid, and iv) high, wherein low and drop are indicative of a positive outcome, and wherein mid and high are indicative of a negative outcome.

16. The method of claim 14 or claim 15, wherein i) low indicates that the percentage of dd-cfDNA out of total cfDNA is below 1% in the first sample and remains below 1% thereafter, ii) drop indicates that the percentage of dd-cfDNA out of total cfDNA is above 1% in the first sample, drops below 1% within a time period, and remains below 1% thereafter, ii) mid indicates that the percentage of dd-cfDNA out of total cfDNA is above 1% in the first sample, drops below 1% within a time period, and increases to above 1% after a second time period, and iv) high indicates that the percentage of dd-cfDNA out of total cfDNA is above 1% in the first sample and remains above 1% thereafter.

17. The method of any of the preceding claims, wherein the positive outcome is improved or restored organ function; and / or wherein the positive outcome comprises absence of death censored graft loss, subsequent rejection, and / or donor- specific antibodies (DSA).

18. The method of claim 11, wherein the negative outcome is an occurrence of death censored graft loss, a subsequent rejection, and / or a presence of donor-specific antibodies (DSA).

19. The method of any of the preceding claims, wherein the organ transplant is from a human.

20. The method of any of claims 1-18, wherein the organ transplant is a xenotransplant.

21. The method of claim 20, wherein the organ transplant is from a pig. a primate, a baboon, a cow, or a dog, preferably from a pig.

22. The method of any of the preceding claims, wherein preparing a sequencing library comprises performing universal amplification on the extracted cell-free DNA or its derivative.N.069.W0.0123. The method of any of the preceding claims, wherein preparing a sequencing library comprises performing targeted enrichment on the extracted cell-free DNA or its derivative to enrich a plurality of polymorphic target loci, preferably a plurality of SNP loci.

24. The method of any of the preceding claims, wherein the organ transplant is one or more organs selected from the group consisting of heart, kidney, liver, lung, pancreas, and intestine.

25. The method of any of the preceding claims, wherein the organ transplant is kidney.

26. The method of any of the preceding claims, the transplant recipient suffers from an antibody-mediated allograft rejection as classified by a histologic method or molecular methods, optionally wherein the histologic method is the Banff 2019 Classification.

27. The method of any of the preceding claims, wherein the two or more blood, plasma, serum or urine samples are collected over a period of 8 weeks.

28. The method of any of the preceding claims, wherein longitudinally collecting two or more blood, plasma, serum or urine samples from an organ transplant recipient comprises collecting at a first time period comprising 7 days prior to a scheduled biopsy or diagnostic test of a transplant rejection, or starting the treatment for the transplant rejection condition, and a second time period within 3 weeks after the scheduled biopsy or diagnostic test, or starting the treatment for the transplant rejection condition.

29. The method of claim 28, further comprising a third time period that is after the second time period.

30. The method of any of the preceding claims, wherein the quantifying step comprises: a) adding a tracer DNA to extracted DNA or its derivative to obtain a mixed composition; b) performing targeted multiplex amplification on the mixed composition comprising the extracted DNA or its derivative and the tracer DNA to amplify 100 to 20,000 different polymorphic target loci together in the same reaction volume using 100 to 20,000 different target- specific primers;c) sequencing the amplicons by high-throughput sequencing to generate sequence reads; andN.069.W0.01d) quantifying the amount of donor-derived ceil- free DNA and the amount of total cell - free DNA from the sequence reads, wherein the amount of total cell-free DNA is quantified using sequence reads derived from the tracer DNA.

31. The method of any of the preceding claims, wherein the organ transplant recipient has a kidney transplant and has biopsy-proven acute rejection (BPAR).

32. The method of claim 31, wherein the two or more samples are collected within one week prior to diagnosis of BPAR, and at weeks 2, 4, 6, and 8 after the BPAR diagnosis.

33. The method of claim 32, wherein the organ transplant recipient received immunosuppressive treatment after the BPAR diagnosis.

34. The method of any of claims 3-33. wherein the treatment for the transplant rejection condition comprises anti-CD38 monoclonal antibodies.

35. The method of any of claims 1-34, wherein the time series data input for the ML model is generated from the percentage of dd-cfDNA out of total cfDNA.

36. The method of any of claims 1-34, wherein the time series data input for the ML model is generated from the amount of dd-cfDNA normalized to a reference.

37. The method of any of claims 1-36, wherein the amount of dd-cfDNA or the percentage of dd-cfDNA out of total cfDNA is combined with clinical data to determine the likelihood of a positive outcome of the organ transplant.

38. The method of claim 37, further comprising determining T cell-mediated rejection (TCMR) rejection, antibody-mediated rejection (ABMR) rejection, and / or mixed rejection based on combining the amount of dd-cfDNA or the percentage of dd-cfDNA out of total cfDNA with clinical data.

39. The method of any of clams 37-38, further comprising using a Two-Stage model comprising a first stage caller and a second stage caller, wherein the first stage caller provides a risk score of transplant rejection, and if the risk score from the first stage caller is above a threshold value, then the second stage caller provides a TCMR risk score for T cell-N.069.W0.01mediated rejection (TCMR) rejection, an AB MR risk score for antibody-mediated rejection (ABMR) rejection, and / or a Mixed risk score for mixed rejection.

40. The method of claim 39, wherein the first stage caller comprises combining the amount of dd-cfDNA or the percentage of dd-cfDNA out of total cfDNA or longitudinal changes of the amount of dd-cfDNA or the percentage of dd-cfDNA out of total cfDNA with clinical data.

41. The method of claim 39 or 40, wherein the second stage caller comprises combining longitudinal changes of the amount of dd-cfDNA or the percentage of dd-cfDNA out of total cfDNA with clinical data.

42. The method of any of clams 37-41, wherein the clinical data comprises prior transplant rejection history, demographic data, a plurality of immunological risk factors, vital signs, a plurality of organ function markers, imaging results, inflammatory markers, a gene expression profile indicative of transplant rejection, a plurality of serology test markers, and / or a plurality of biomarkers indicative of transplant rejection.

43. The method of claim 42, wherein the demographic data comprises age or ethnicity.

44. The method of claim 42, wherein the plurality of serology test markers comprises donorspecific antibodies (DSA).

45. The method of claim 42, wherein the plurality of organ function markers comprises an amount of serum creatinine.

46. The method of any of clams 37-45, wherein a machine learning model is used to combine data obtained from the amount of dd-cfDNA or the percentage of dd-cfDNA out of total cfDNA with the clinical data.

47. The method of claim 46, wherein the machine learning model provides (i) a risk score of transplant rejection, and / or (ii) a TCMR risk score for T cell-mediated rejection (TCMR) rejection, an ABMR risk score for antibody-mediated rejection (ABMR) rejection, and / or a Mixed risk score for mixed rejection.N.069.W0.0148. The method of claims 46 or 47, wherein the machine learning model comprises a temporal random forest model.

49. The method of claims 39 to 48, wherein the first stage caller and the second stage caller are temporal random forest models.

50. A method for preparing a non-naturally occurring composition for assessing an organ transplant, comprising:longitudinally collecting two or more blood, plasma, serum or urine samples from an organ transplant recipient,extracting cell-free DNA (cfDNA) from the two or more blood, plasma, serum or urine samples or portions thereof from the organ transplant recipient, wherein the extracted cfDNA comprises a mixture of donor-derived cfDNA (dd-cfDNA) and recipient-derived cfDNA;preparing a sequencing library from the extracted cfDNA or its derivative and performing high-throughput sequencing on the sequencing library to obtain sequence reads;quantifying an amount of dd-cfDNA or a percentage of dd-cfDNA out of total cfDNA based on the sequence reads, wherein the organ transplant is assessed by combining the amount of dd-cfDNA or the percentage of dd-cfDNA out of total cfDNA with clinical data from the organ transplant recipient.

51. The method of claim 50, determining longitudinal dd-cfDNA changes based on the amount of dd-cfDNA or the percentage of dd-cfDNA out of total cfDNA.

52. The method of claims 50 or 51, further determining a TCMR risk score for T cell-mediated rejection (TCMR) rejection, an AB MR risk score for antibody-mediated rejection (AB MR) rejection, and / or a Mixed risk score for mixed rejection based on combining (i) the amount of dd-cfDNA or the percentage of dd-cfDNA out of total cfDNA or longitudinal dd-cfDNA changes with (ii) clinical data from the organ transplant recipient.

53. The method of any of claims 50 to 52, further comprising using a Two-Stage model comprising a first stage caller and a second stage caller, wherein the first stage caller provides a risk score of transplant rejection, and if the risk score from the first stage caller is above a threshold value, then the second stage caller provides a TCMR risk score for T cell-mediatedN.069.W0.01rejection (TCMR) rejection, an ABMR risk score for antibody-mediated rejection (AB MR) rejection, and / or a Mixed risk score for mixed rejection.

54. The method of claim 53, wherein the first stage caller comprises combining the amount of dd-cfDNA or the percentage of dd-cfDNA out of total cfDNA with clinical data.

55. The method of claims 53 or 54, wherein the second stage caller comprises combining longitudinal dd-cfDNA changes with clinical data.

56. The method of any of clams 50-55, wherein the clinical data comprises prior transplant rejection history, demographic data, a plurality of immunological risk factors, vital signs, a plurality of organ function markers, imaging results, inflammatory markers, a gene expression profile indicative of transplant rejection, a plurality of serology test markers, and / or a plurality of biomarkers indicative of transplant rejection.

57. The method of claim 56, wherein the demographic data comprises age and / or ethnicity.

58. The method of claim 56, wherein the plurality of serology test markers comprises donorspecific antibodies (DSA).

59. The method of claim 56, wherein the plurality of organ function markers comprises an amount of serum creatinine.

60. The method of any one of claims 53 to 59, wherein the first stage caller comprises combining the amount of dd-cfDNA or the percentage of dd-cfDNA out of total cfDNA with prior transplant rejection history.

61. The method of any one of claims 52 to 60, wherein the second stage caller comprises combining longitudinal dd-cfDNA changes with demographic data and a plurality of organ function markers to determine a TCMR risk score for T cell-mediated rejection (TCMR) rejection.

62. The method of any one of claims 52 to 60, wherein the second stage caller comprises combining the amount of dd-cfDNA or the percentage of dd-cfDNA out of total cfDNA with prior ABMR history to determine and ABMR risk score.N.069.W0.0163. The method of claim 50, wherein combining the amount of dd-cfDNA or the percentage of dd-cfDNA out of total cfDNA with clinical data from the organ transplant recipient is performed by using a statistical model.

64. The method of claims 38 or 52, wherein determining TCMR rejection, ABMR rejection, and / or mixed rejection is performed simultaneously.

Images