Printer Friendly

Ultrasensitive Detection of Chimerism by Single-Molecule Molecular Inversion Probe Capture and High-Throughput Sequencing of Copy Number Deletion Polymorphisms.

The detection of genomic chimerism, [greater than or equal to] 2 cell populations from differing genetic origins within an individual, has diverse clinical applications including identifying fetal microchimerism in autoimmune diseases (1), graft injury and transplant rejection (2, 3), and the clinical management of allogeneic hematopoietic stem cell transplantation (HSCT) [2] patients. In HSCT, identification of an increasing proportion of host cells in the transplanted marrow is a harbinger of poor outcomes, potentially indicating the expansion of residual malignant cells and subsequent disease relapse (4-7) or stem cell graft rejection (8). Ultrasensitive detection of chimerism in this and other clinical contexts could improve patient risk stratification and provide opportunities for earlier therapeutic intervention.

Early methods for distinguishing chimeric cells involved Y-chromosome fluorescence in situ hybridization and was restricted to sex-mismatched cases (9). Later approaches exploited inherent genetic variation on autosomes, making them more generally applicable. The most common clinical methods type polymorphic short tandem repeat markers, which have a limit of sensitivity approaching 1% (10). We (11) and subsequently others (12, 13) more recently developed methods to interrogate deletion copy number variants, a subclass of copy number variation representing genomic regions tens to hundreds of thousands of base pairs long that are present in some individuals but absent in others (14-17). The first application of these markers used microscopic examination of cells subjected to fluorescence in situ hybridization for informative loci (11), whereas later, real-time PCR (12) and digital droplet PCR (ddPCR) (13) assays offered advantages in target multiplexing and sensitivity. Nevertheless, even ddPCR chimerism assays have finite sensitivities (1 in 2000 cells, or 0.05%) (13) and lesser limits of quantification (accuracy to 0.5% chimerism) (13), while interrogating only a limited number of loci because of restrictions with primer multiplexing and requiring specialized instrumentation (18).

Here, we demonstrate single-molecule molecular inversion probe (smMIP) capture (19, 20) coupled with high-throughput sequencing to identify and quantify deletion copy number variants for the purposes of chimerism analysis (Fig. 1). smMIP technology couples highly multiplexed targeted sequencing with molecular counting through unique molecular identifiers (UMIDs) (21), degenerate oligonucleotide barcodes that mark sequence reads derived from a common progenitor (19). Quantification of deletion copy number variants is achieved by counting sequence reads derived from smMIPs targeted to approximately 150-bp regions within those larger markers after UMID-mediated error correction, which eliminates bias from PCR duplicates (19). As a clinical application, we demonstrate the potential of this approach by identifying chimerism in the context of post-allogeneic transplantation for hematologic malignancy (6), irrespective of disease lineage.

Materials and Methods


This project was approved by the University of Washington Human Subjects Division and conducted in accordance with the Declaration of Helsinki. Residual clinical bone marrow samples (Table 1) were obtained retrospectively and deidentified according to University of Washington Institutional Review Board guidelines. This represents a convenience sample of HSCT patients, in which all patients with confirmed molecular or flow cytometry diagnoses were selected if sufficient residual DNA (at least 500 ng) was available for analysis. Clinical minimal residual disease (MRD) testing was performed by the University of Washington Hematopathology Laboratory as described (22-24), and clinical chimerism testing was done by the Seattle Cancer Care Alliance Immunogenetics Laboratory using the PowerPlex 16 Human Identity Kit (Promega).

Deidentified residual peripheral blood samples from randomly selected, healthy blood bank donors were similarly extracted. DNA from Centre d'Etude du Polymorphism Humain (CEPH) family 1463 (see Fig. 1 in the Data Supplement that accompanies the online version of this article at issue6) was purchased from Coriell Biorepository.


smMIPs were designed against 40 deletion copy number variants found in high frequencies across populations (13, 25) and supplemented with 3 additional polymorphic loci (UGT2B17, chr2:52751531-52782353, and chr8:39251030-39372444) [3] (11) and 2 sex-linked loci (F9 and SRY). Design of probes optimized for uniqueness of binding sites and guanine--cytosine content was performed against hg19 using the program MIPgen (26) as previously described (27) to tile across each region. Targeted loci were divided into equally sized quadrants, and a single smMIP was selected from each to identify 4 probes per locus. Salt-purified probes were synthesized by IDT.

Probes were 5;-phosphorylated and used in capture reactions as elsewhere (27). Briefly, 500 ng of DNA was hybridized with probes in the presence of DNA polymerase and DNA ligase, exonuclease treated, and PCR amplified to generate sequencing libraries. Sequencing used an Illumina Miseq or NextSeq500 with 300 cycles and dual indexing. Initially, smMIPs were pooled at equimolar concentrations and used to analyze 8 normal blood donors. The average relative abundance of UMIDs from each probe was determined for patients yielding grossly positive read counts. Probes were empirically rebalanced to improve uniformity and remove those with poor performance. The final panel of 159 smMIP probes interrogated 42 loci (see Table 1 in the online Data Supplement). Uniformity of the panel was evaluated using smMIP capture of 8 donors and plotting relative abundance of UMIDs for each smMIP (Fig. 2A).


To model the false-positive rate for loci absent in a specimen, we performed capture and high-depth sequencing of DNA from 5 normal blood donors. The expected relative abundance of UMIDs per smMIP was estimated from the total UMID count per specimen and the expected abundance of UMID counts per smMIP as determined from 8 individuals typed during the rebalancing process (Fig. 2B). smMIPs yielding <5% expected abundance were considered false-positive findings. The false-positive rate was calculated for each smMIP by dividing observed UMID counts by the expected. False-positive rates for each smMIP were integrated across specimens fit to a [beta] distribution (27) given their mean and SD. For probes with no false-positive findings observed, we assumed a rate equal to the average of measured false-positive events per individual divided by the sum of expected UMID counts (4.61 X 10 6).


Data analysis was double-blinded. Demultiplexed reads were preprocessed with barcodecop ( nhoffman/barcodecop) to remove artefactual barcode misassignments (28), then processed and mapped to hg19 as described (27), except that we discarded self-assembled reads < 140 bp, those with mapping quality <50, and those mapping with < 100-bp matched sequence. Mapped reads were grouped by UMID to collapse PCR duplicates and quantify unique capture events.

The expected UMID count from each smMIP was estimated as for the error model, and P values per smMIP were calculated under the error model based on observed relative abundances.

When genotypes of minor chimeric populations were known (i.e., when pretransplant material was available), informative loci were identified as those present in the minor chimeric population but absent from the major population (<0.1 of the expected abundance in the donor unless otherwise noted).

For cases when genotypes of minor populations were not available, informative loci were identified as those at <0.1 of expected abundance in the donor unless otherwise noted, with P [less than or equal to] 0.05 required to consider an event analytically valid.

The probability that a chimeric species had been detected was expressed as the cumulative probability of P values for individual smMIPs meeting the above criteria. Relative abundance of the species was estimated as the measured divided by expected UMIDs for each smMIP, averaged across informative probes.

Sequence data are accessible from the National Center for Biotechnology Information Sequence Read Archive (accession SRP124971). Source code for the analysis pipeline is available from ssh://git@bitbucket. org/uwlabmed/smmips_analysis.git.



To detect polymorphic genomic regions that could serve as markers of chimerism, we designed an smMIP probe set targeting 40 common autosomal deletion copy number variants (11, 25) and 2 sex-linked loci (see Table 1 in the online Data Supplement).

To characterize the false-positivity rate of smMIP capture, manifesting as sequence reads matching deletion copy number loci that are homozygously deleted, we sequenced genomic DNA from 5 individuals to high depth. In all, 1.78 x [10.sup.7] [+ or -] 2.2 x [10.sup.6] [mean (SD)] reads per individual were generated, with 4.7 x [10.sup.4] [+ or -] 1.4 x 104 uniquely tagged smMIP reads per specimen. The presence or absence of deletion copy number loci was apparent by comparing the number of reads bearing unique UMIDs from each smMIP with the expected performance (Fig. 2A). Individual probes showed consistent performance across experiments. Most smMIPs had either reads falling within a factor of 2 from the predicted read counts (Fig. 2B), reflecting homozygous presence (2 alleles) and heterozygous states (1 allele), or undetectable or near-undetectable read counts, representing homozygous deletions (null alleles). The false-positive rate of the assay was estimated by quantifying the number of nonzero reads present in the lowest mode.

Aggregating results, we found 389 of the 795 (49%) individual smMIPs corresponded to regions of homozygous copy number loss. Remarkably, only 9 of the 389 (2.3%) probes produced artifactual sequence reads in homozygously deleted regions. We quantified 27 false-positive reads bearing unique UMIDs, out of 2.4 x [10.sup.6] total UMIDs, representing an effective error rate of 1 in 85000 events. Although this error is nonlimiting, we subsequently assessed the significance of positive reads using an empiric error model trained on these data.


The informativity of interrogating various numbers of deletion copy number loci has been previously modeled (11, 12) or experimentally evaluated (12, 13). We empirically determined the informativity of our panel for 8 unrelated donors and for 2 parents and 10 siblings of a CEPH reference family to model transplantation occurring between first-degree relatives (see Tables 2 and 3 in the online Data Supplement). We analyzed only autosomal markers, providing a conservative estimate of performance outside of sex-mismatched transplants. Informativity was evaluated with respect to 2 metrics.

First, we quantitated the homozygous deleted loci per individual, which could potentially serve as informative markers in chimerism analyses (Fig. 3A). For the randomly selected cohort, an average of 20.5 [+ or -] 2.9 homozygous deleted loci were observed per person out of the 40 total loci. Slightly lower values were observed for first-degree relatives from the CEPH family (14.4 [+ or -] 2.1).

Second, we examined all possible donor-recipient combinations within cohorts to quantify the informative loci (loci absent from the donor but present in the recipient) that could be functionally used in each pairing (Fig. 3B). In all, 9.6 [+ or -] 2.7 informative loci were seen for simulated donor-recipient combinations (n = 56 permutations) among randomly selected individuals. This decreased to 4.9 [+ or -] 2.1 loci for first-degree relatives (n = 132 permutations); nevertheless, only 2 pairings (1.5%) yielded no informative loci.

Although smMIPs targeted to different regions of a given copy number deletion locus were typically concordant within an individual, specific probes were variably present or absent within these larger domains. Eighteen of 304 autosomal loci (5.9%) typed for the 8 unrelated individuals showed biological discordance among independent smMIPs, suggesting localized regions of presence or absence within loci 10, 13, 18, 23, 24, and UGT2B17. Consequently, if individual smMIP probes, rather than their consensus across larger regions, are considered, diversity among individuals will be higher than reported here.


We next evaluated the ability of smMIP capture to identify minor cell populations within a dominant background. To simulate transplants between first-degree relatives, we generated synthetic mixtures of DNA from a HapMap trio: NA12878 (child), NA12891 (father), and NA12892 (mother). DNA from each parent was separately mixed into the child at 50%, 25%, 1%, 0.1%, and 0.01% relative abundance. Duplicate preparations of each dilution were analyzed to estimate assay variability, yielding 4 replicates. Genotyping members of the trio predicted 9 informative loci (31 smMIPs) for spike-ins using NA12891, and 5 (21 smMIPs) for spike-ins using NA12892. Specimens were sequenced to provide 4.7 x [10.sup.5] [+ or -] 1.1 x [10.sup.5] uniquely tagged smMIP reads per sample. The significance ofpositive signal from each smMIP probe was calculated using an empiric, site-specific error model, and that of the overall result estimated as the cumulative probability of those events.

Linearity was achieved over 4 orders of magnitude (Fig. 4A; [R.sup.2] > 0.89); however, the estimated relative abundance of minor chimeric populations was consistently overestimated (range, approximately 3-14-fold). Variability among replicates inversely correlated with abundance of the minor population (% CV at 50% relative abundance = 3.5%; 25% = 5.5%; 1% = 31%; 0.1% = 50%; 0.01% = 70%). All smMIPs predicted as informative were detected for minor populations at 1% abundance, although smaller subsets were found as relative abundance decreased (Fig. 4B). Statistical significance of individual informative probes was strong across specimens (average P at [greater than or equal to]1% relative abundance = <[10.sup.-10] CI; 0.1% = 0.002; 0.01% = 0.03), and the cumulative P values remained significant across replicates to 0.01% relative abundance (range P = 0.05-0.0003). These results suggest that the limit of detection lies below the 1 in 10 000 threshold tested in this study.

We next evaluated our ability to identify minor populations without previous knowledge of informative loci. For dilutions [less than or equal to]10%, we quantified the abundance of smMIPs yielding measurable sequence reads relative to their predicted read counts, while attributing any high-level signal (>20% of expected UMID counts) to the major population. This automated analysis correctly identified informative smMIPs at each of the 3 relative abundances in all replicates, perfectly recapitulating manual analysis using previous knowledge of the specimen genotypes.


Mixed chimerism of [greater than or equal to] 3 genotypes may occur in pooled cord blood transplants, often from unrelated donors, or because of repeat transplantation (29). To assess the ability of smMIPs to resolve complex mixtures derived from multiple individuals, we manufactured and typed 2 different synthetic DNA mixtures of 4 randomly selected and presumably unrelated individuals each, combined in relative abundances of 88.9%, 10%, 1%, and 0.1%. Genotyping the individuals comprising those mixtures indicated an average of 11.14 (range, 6-18) informative smMIPs that uniquely identified each individual, whereas only 1 donor (the individual at 10% relative abundance in the second mixture) could not be uniquely distinguished from the 3 others and was subsequently excluded from analysis. smMIP-based quantification of the informative markers for each uniquely identifiable individual approximated their expected relative contributions (Fig. 5).


To provide a proof-of-principle assessment of assay performance in clinical practice, we applied the panel to a collection of 28 specimens derived from 15 patients (Table 1) having undergone allogeneic HSCT for acute myeloid leukemia (n = 9), B lymphoblastic leukemia (n = 3), or T lymphoblastic leukemia (n = 3), evaluated at various points before and after transplant. Two patients (2 and 4) received transplants from sex-mismatched siblings. Results of smMIP capture for detecting residual host cells, which may be of either neoplastic or nonneoplastic origin (6, 30, 31), were correlated with conventional diagnosis of MRD using flow cytometry (32, 33), RT-PCR for BCR-ABL gene fusion (34), FLT3 allelic ratio testing (35), NPM1 deep sequencing (24), and/or conventional chimerism analysis (10) when clinically available. Pretransplant specimens were analyzed against the error model to identify artifactual sequence reads of sufficient abundance to be ascribed to biological chimerism, allowing assessment of false-positivity rates in practice.

Although only 11 of the 22 posttransplantation specimens had measurable MRD by conventional clinical testing, all 22 specimens tested positive for host-derived cells by smMIPs, even at time-points far from the date of original transplantation. In cases for which conventional approaches detected MRD and smMIPs identified recipient cells, there was a positive correlation between the quantitated cells ([R.sup.2] = 0.68). However, there was considerable variability surrounding the quantification of patient-derived cells by smMIPs, especially in cases when they were of low abundance, for which the SD of quantitated chimerism approached or exceeded the average estimate obtained across individual smMIPs. In general, error-modeled P values for abundance of individual smMIPs diminished with decreasing MRD, while remaining below our cutoff of P [less than or equal to] 0.05 for analytical validity. Even so, the cumulative P value of each case, integrating the significance of each informative smMIP, was uniformly high (P value range <6.8 x [10.sup.-108] to 2.95 x [10.sup.-14]). Conventional chimerism analysis was available for 11 cases and proved less sensitive for detecting recipient-derived cells than flow cytometry, which identified MRD in 3 of the 7 cases testing negative for chimerism, or smMIP analysis, which identified residual patient-derived cells in all 7 of those cases.

A single false-positive smMIP, quantified at significantly higher levels than predicted by the error model (1.99%; P < 1 x [10.sup.-100]), was detected across the 6 pretransplant cases, when chimerism is constitutionally absent (Patient 3). Yet, this technical artifact was a solitary aberrant event: The number of informative smMIPs in specimens legitimately harboring chimerism was considerably higher (17.4 probes; range, 3-116).


In this pilot work, we have adapted smMIP technology for the purposes of detecting genomic chimerism based on targeting polymorphic copy number deletion loci (11). By examining neutral markers of genomic variation that distinguish individuals, rather than phenotypic or genetic lesions specific to a disease, this approach enables universal diagnosis of chimerism without reliance on or expectation of a stereotyped disease marker. Because copy number deletion loci are either present or absent in a given individual, this binary signal provides analytically sensitive and background-free measurement of chimeric populations (11, 12). Use of smMIP capture to assay such loci provides high specificity at 2 levels. First, hybridization of an smMIP to its target is sequence-specific, so that off-target events are not favored. Second, detection of informative loci is accomplished through massively parallel sequencing, which provides sequence information for each capture event and allows them to be accurately mapped within the genome, further reducing the measurement of stochastic off-target captures. These features offer advantages over next-generation sequencing approaches that identify chimerism based on single-nucleotide polymorphisms (36, 37), which are subject to sequencing errors that limit their sensitivity (7, 37).

The application of smMIPs to chimerism detection provides significant advantages over conventional and advanced diagnostic methods reported previously. First, smMIP capture and error correction enables ultrasensitive detection of chimeric populations. Here, we demonstrated limits of detection of at least 1 in 10000 chimeric cells, exceeding the reported sensitivity of ddPCR-based assays by a factor of 5 (13). Given minor variation in the performance of individual probes (Fig. 2A), however, the assay's precise sensitivity will vary somewhat based on the genotype examined, which could potentially be mitigated by further rebalancing (19). Quantification of minor cell populations was precise, with linearity achieved over 4 orders of magnitude. Although the relative abundance of minor cell populations was consistently overestimated in our linearity studies (Fig. 3B), bias in clinical specimens compared with measured MRD or chimerism was less predictable (Table 1). This finding is likely genotype-specific, reflecting differences in copy number of informative loci from analyzed specimens and those estimated by our error model, which is an amalgam of normal individuals spanning various copy states. Computational modeling of these compositional data or empiric correction factors matched to patient genotypes may overcome quantitative bias in future iterations; nevertheless, the positive or negative detection of informative loci remains robust. Second, smMIPs are readily multiplexable, enabling simultaneous interrogation of hundreds or even thousands of different copy number deletion loci (19), far more than can be practically examined by multiplexed PCR. This provides robustness in interrogating each informative locus multiple times independently, buffering against stochastic loss of signal, particularly for low-abundance chimeric populations. We also found that polymorphic subregions occur within larger genomic regions (e.g., UGT2B17), providing increased genomic diversity through redundant testing. Although our analysis shows that 40 autosomal loci can distinguish even among first-degree relatives with reliability (Fig. 3B), the ability to accommodate much larger numbers of sites may prove advantageous for specific, closely related donors or for mixed chimerism arising from sequential or pooled bone marrow transplants. Third, the quantitative nature of smMIP data enables computational inference of the genotypes of major and multiple minor cell populations without the need for independently genotyping donor and recipient material, as has proven necessary by both conventional short tandem repeat (10, 12) and PCR-based copy number deletion assays (12, 13), and eliminates the need for calibration by standard curve (12). Fourth, smMIP capture is cost-effective and readily scalable to multiple patient specimens (38), facilitating high-throughput testing in clinical practice.

While showing concordance with flow cytometry and other conventional diagnostics for detection of MRD, analysis of clinical specimens identified low levels of host-derived cells in all 22 posttransplant specimens tested, even those testing negative by clinical assays. The presence of very low-level mixed chimerism following HSCT is not necessarily uncommon (6) and is compatible with previous work by our group (24, 27) and others (7, 31, 39) using other ultrasensitive techniques. As our approach does not identify the lineage of chimeric cell populations, this finding may reflect both ongoing immune surveillance and suppression of abnormal cells (30) and the persistence of normal recipient hematopoietic, stromal, or nonhematopoietic cells after transplant (6, 30, 31). The rate of change in a chimeric population over time (6, 24, 27) or the quantification of chimerism over a critical threshold (7, 31) may consequently serve as a more robust marker of hematopoietic disease relapse in the posttransplant state than the absolute detection of extremely low-abundance chimerism. Alternatively, testing of peripheral blood could improve the specificity for detecting abnormal cells, at the likely expense of sensitivity (40).

Chimerism testing by smMIP capture can scale linearly to greater levels of sensitivity if increased quantities of input DNA and sequencing coverage are used (27), allowing enhanced performance to be readily achieved. This approach holds promise for many other applications requiring the sensitive detection of genomically distinct minor cell populations, including fetal microchimerism in autoimmune diseases (1) and solid organ graft injury and transplant rejection (2, 3). Future work will seek to establish the utility of smMIP capture in monitoring a variety of disease states, as well as in larger cohorts of HSCT patients as a potential surrogate for MRD detection.

Author Contributions: All authors confirmed they have contributed to the intellectual content of this paper and have met the following 3 requirements: (a) significant contributions to the conception and design, acquisition of data, or analysis and interpretation of data; (b) drafting or revising the article for intellectual content; and (c) final approval of the published article.

Authors' Disclosures or Potential Conflicts of Interest: Upon manuscript submission, all authors completed the author disclosure form. Disclosures and/or potential conflicts of interest:

Employment or Leadership: None declared.

Consultant or Advisory Role: None declared.

Stock Ownership: None declared.

Honoraria: None declared.

Research Funding: S.J. Salipante, grant CA192980 from the National Cancer Institute. The University of Washington Department of Laboratory Medicine.

Expert Testimony: None declared.

Patents: A. Waalkes, D. Wu, S.J. Salipante, provisional patent 62625167.

Role of Sponsor: The funding organizations played no role in the design of study, choice of enrolled patients, review and interpretation of data, or final approval of manuscript.

Acknowledgments: The authors thank B. Wood and Y. Zhou for helpful discussions, and staff from the UW Hematopathology laboratory for technical help.


(1.) Boddy AM, Fortunato A, Wilson Sayres M, Aktipis A. Fetal microchimerism and maternal health: a review and evolutionary analysis of cooperation and conflict beyond the womb. BioEssays News Rev Mol Cell Dev Biol 2015;37:1106-18.

(2.) Gadi VK, Nelson JL, Boespflug ND, Guthrie KA, Kuhr CS. Soluble donor DNA concentrations in recipient serum correlate with pancreas-kidney rejection. Clin Chem 2006;52:379-82.

(3.) Beck J, Bierau S, Balzer S, Andag R, Kanzow P, Schmitz J, et al. Digital droplet PCR for rapid quantification of donor DNA in the circulation of transplant recipients as a potential universal biomarker of graft injury. Clin Chem 2013;59:1732-41.

(4.) Walter RB, Gooley TA, Wood BL, Milano F, Fang M, Sorror ML, et al. Impact of pretransplantation minimal residual disease, as detected by multiparametric flow cytometry, on outcome of myeloablative hematopoietic cell transplantation for acute myeloid leukemia. J Clin Oncol 2011;29:1190 -7.

(5.) Walter RB, Buckley SA, Pagel JM, Wood BL, Storer BE, Sandmaier BM, et al. Significance of minimal residual disease before myeloablative allogeneic hematopoietic cell transplantation for AML in first and second complete remission. Blood 2013;122:1813-21.

(6.) Khan F, Agarwal A, Agrawal S. Significance of chimerism in hematopoietic stem cell transplantation: new variations on an old theme. Bone Marrow Transplant 2004; 34:1-12.

(7.) Debeljak M, Mocci E, Morrison MC, Pallavajjalla A, Beierl K, Amiel M, et al. Haplotype counting for sensitive chimerism testing: potential for early leukemia relapse detection. J Mol Diagn 2017;19:427-36.

(8.) Masmas TN, Petersen SL, Madsen HO, Ryder LP, Kornblit B, Svejgaard A, et al. Graft rejection after hematopoietic cell transplantation with nonmyeloablative conditioning. Am J Hematol 2008;83:563-9.

(9.) Przepiorka D, Gonzales-Chambers R, Winkelstein A, Rosenfeld C, Shadduck RK. Chimerism studies using in situ hybridization for the Y chromosome after T cell-depleted bone marrow transplantation. Bone Marrow Transplant 1990;5:253-7.

(10.) Odriozola A, Riancho JA, Colorado M, Zarrabeitia MT. Evaluation of the sensitivity of two recently developed STR multiplexes for the analysis of chimerism after haematopoietic stem cell transplantation. Int J Immunogenet 2013;40:88-92.

(11.) Wu D, Vu Q, Nguyen A, Stone JR, Stubbs H, Kuhlmann G, et al. In situ genetic analysis of cellular chimerism. Nat Med 2009;15:215-9.

(12.) Bruno DL, Ganesamoorthy D, Thorne NP, Ling L, Bahlo M, Forrest S, et al. Use of copy number deletion polymorphisms to assess DNA chimerism. Clin Chem 2014; 60:1105-14.

(13.) Whitlam JB, Ling L, Swain M, Harrington T, Mirochnik O, Brooks I, et al. Use of ubiquitous, highly heterozygous copy number variants and digital droplet polymerase chain reaction to monitor chimerism after allogeneic haematopoietic stem cell transplantation. Exp Hematol 2017;49:39 - 47.e5.

(14.) Conrad DF, Andrews TD, Carter NP, Hurles ME, Pritchard JK. A high-resolution survey of deletion polymorphism in the human genome. Nat Genet 2006;38: 75-81.

(15.) McCarroll SA, Hadnott TN, Perry GH, Sabeti PC, Zody MC, Barrett JC, et al.Commondeletion polymorphisms in the human genome. Nat Genet 2006;38:86-92.

(16.) Sebat J, Lakshmi B, Troge J, Alexander J, Young J, Lundin P, et al. Large-scale copy number polymorphism in the human genome. Science 2004;305:525- 8.

(17.) Hinds DA, Kloek AP, Jen M, Chen X, Frazer KA. Common deletions and SNPs are in linkage disequilibrium in the human genome. Nat Genet 2006;38:82-5.

(18.) Wertheim GBW, Bagg A. NPM1 for MRD? Droplet like it's hot! J Mol Diagn 2017;19:498-501.

(19.) Hiatt JB, Pritchard CC, Salipante SJ, O'Roak BJ, Shendure J. Single molecule molecular inversion probes for targeted, high-accuracy detection of low-frequency variation. Genome Res 2013;23:843-54.

(20.) Eijkelenboom A, Kamping E, Kastner-van Raaij A, Hendriks-Cornelissen S, Neveling K, Kuiper R, et al. Reliable next-generation sequencing of formalin-fixed, paraffin-embedded tissue using single molecule tags. J Mol Diagn 2016;18:851-63.

(21.) Kinde I, Wu J, Papadopoulos N, Kinzler KW, Vogelstein B. Detection and quantification of rare mutations with massively parallel sequencing. Proc Natl Acad Sci U S A 2011;108:9530-5.

(22.) Wood BL. Flow cytometric monitoring of residual disease in acute leukemia. Methods Mol Biol 2013;999: 123-36.

(23.) Zhou Y, Wood BL. Methods of detection of measurable residual disease in AML. Curr Hematol Malig Rep 2017; 12:557-67.

(24.) Salipante SJ, Fromm JR, Shendure J, Wood BL, Wu D. Detection of minimal residual disease in NPM1-mutated acute myeloid leukemia by next-generation sequencing. Mod Pathol 2014;27:1438-46.

(25.) Sudmant PH, Rausch T, Gardner EJ, Handsaker RE, Abyzov A, Huddleston J, et al. An integrated map of structural variation in 2,504 human genomes. Nature 2015;526:75-81.

(26.) Boyle EA, O'Roak BJ, Martin BK, Kumar A, Shendure J. MIPgen: optimized modeling and design of molecular inversion probes for targeted resequencing. Bioinformatics 2014;30:2670-2.

(27.) Waalkes A, Penewit K, Wood BL, Wu D, Salipante SJ. Ultrasensitive detection of acute myeloid leukemia minimal residual disease using single molecule molecular inversion probes. Haematologica 2017;102: 1549-57.

(28.) Wright ES, Vetsigian KH. Quality filtering of Illumina index reads mitigates sample cross-talk. BMC Genomics 2016;17:876.

(29.) Ende N, Lu S, Alcid MG, Chen R, Mack R. Pooled umbilical cord blood as a possible universal donor for marrow reconstitution and use in nuclear accidents. Life Sci 2001;69:1531-9.

(30.) Paietta E. Minimal residual disease in acute myeloid leukemia: coming of age. Hematol Am Soc Hematol Educ Program 2012;2012:35-42.

(31.) Jimenez-Velasco A, Barrios M, Roman-Gomez J, Navarro G, Buno I, Castillejo JA, et al. Reliable quantification of hematopoietic chimerism after allogeneic transplantation for acute leukemia using amplification by real-time PCR of null alleles and insertion/deletion polymorphisms. Leukemia 2005;19:336-43.

(32.) Miyazaki T, Fujita H, Fujimaki K, Hosoyama T, Watanabe R, Tachibana T, et al. Clinical significance of minimal residual disease detected by multidimensional flow cytometry: serial monitoring after allogeneic stem cell transplantation for acute leukemia. Leuk Res 2012;36: 998-1003.

(33.) Hourigan CS, Karp JE. Minimal residual disease in acute myeloid leukaemia. Nat Rev Clin Oncol 2013;10: 460-71.

(34.) Luu MH, Press RD. BCR-ABL PCR testing in chronic myelogenous leukemia: molecular diagnosis for targeted cancer therapy and monitoring. Expert Rev Mol Diagn 2013;13:749-62.

(35.) Levis M. FLT3 mutations in acute myeloid leukemia: what is the best approach in 2013? Hematol Am Soc Hematol Educ Program 2013;2013:220-6.

(36.) Kim J, Hwang IS, Kim HS, Joo DJ, Hong KR, Choi JR. Bone marrow chimerism detection using next generation sequencing based on single nucleotide polymorphisms following liver transplantation: comparison with short tandem repeat-PCR. Ann Lab Med 2016;36: 82-4.

(37.) Aloisio M, Licastro D, Caenazzo L, Torboli V, D'Eustacchio A, Severini GM, et al. A technical application of quantitative next generation sequencing for chimerism evaluation. Mol Med Rep 2016;14: 2967-74.

(38.) Neveling K, Mensenkamp AR, Derks R, Kwint M, Ouchene H, Steehouwer M, et al. BRCA testing by single-molecule molecular inversion probes. Clin Chem 2017;63:503-12.

(39.) Uy GL, Duncavage EJ, Chang GS, Jacoby MA, Miller CA, Shao J, et al. Dynamic changes in the clonal structure of MDS and AML in response to epigenetic therapy. Leukemia 2017;31:872-81.

(40.) Zeijlemaker W, Kelder A, Oussoren-Brockhoff YJM, Scholten WJ, Snel AN, Veldhuizen D, et al. Peripheral blood minimal residual disease may replace bone marrow minimal residual disease as an immunophenotypic biomarker for impending relapse in acute myeloid leukemia. Leukemia 2016;30:708-15.

David Wu, [1] Adam Waalkes, [1] Kelsi Penewit, [1] and Stephen J. Salipante [1] *

[1] Department of Laboratory Medicine, University of Washington, Seattle, WA.

* Address correspondence to this author at: University of Washington, Box 357110,1959 NE Pacific St., Seattle WA, 98195. Fax 206-598-6189; e-mail

Received November 22, 2017; accepted February 21, 2018.

Previously published online at DOI: 10.1373/clinchem.2017.284737

[2] Nonstandard abbreviations: HSCT, hematopoietic stem cell transplantation; ddPCR, digital droplet PCR; smMIP, single-molecule molecular inversion probe; UMID, unique molecular identifier; MRD, minimal residual disease; CEPH, Centre d'Etude du Polymorphism Humain.

Caption: Fig. 1. Schematic of smMIP-based chimerism analysis.

(A), Deletion copy number variants are naturally polymorphic among different individuals, wherein some loci are deleted (pale regions) heterozygously or homozygously, or present on both copies of a chromosome. (B), smMIP capture of genomic regions containing common deletion copy number variants followed by high-throughput sequencing allows quantitative detection of DNA corresponding to these loci. (C), Loci that are present in 1 individual but homozygously deleted in another (colored bars) can be used as informative markers for identifying and quantifying the relative abundance of each cell population. Loci that are present at any copy state in both individuals (gray bars) are not informative.

Caption: Fig. 2. Relative capture efficiencies of smMIPs and measured abundance of targeted loci.

(A), Relative abundance of unique capture events recovered per smMIP (arbitrarily sorted along xaxis) for 8 different control specimens, after eliminating probes directed against deletion copy number loci that were absent in an individual. (B), Bean plot displaying the difference between observed and expected smMIP capture counts over the expected smMIP capture count for 5 germline specimens. Expected values estimate the relative abundance of smMIP capture events for a specimen when all targeted loci are present. A value of 0 indicates agreement between observed and expected counts. Positive values indicate greater numbers of observed events than expected events. Negative numbers indicate fewer events than expected to a minimum value of-1, which is consistent with homozygous copy number loss. Black horizontal lines indicate the weight of individual data points, and gray polygons represent the estimated density of the data. Note that the modes of this distribution are centered around 0 and -1, corresponding to sequence reads being present at expected abundance or fully absent in typed specimens.

[3] Human genes: UGT2B17, UDP glucuronosyltransferase family 2 member B17; F9, coagulation factor IX; SRY sex determining region Y; BCR, BCR, RhoGEF and GTPase activating protein; ABL, ABL proto-oncogene 1, non-receptor tyrosine kinase; FLT3, fms related tyrosine kinase 3; NPM1, nucleophosmin 1.

Caption: Fig. 3. Informativity analysis for autosomal loci.

(A), The number of autosomal copy number deletion loci that are homozygous deleted per individual in cohorts of unrelated and related individuals. (B), The number of informative autosomal loci (those that are absent from a donor but present in a recipient) for all possible pairwise comparisons within cohorts of unrelated and related individuals. Gray bars indicate average values for the indicated population, and whiskers show SD.

Caption: Fig. 4. Performance characteristics of smMIP-based chimerism analysis.

(A), Quantitation of minor chimeric cell populations in synthetic control specimens derived from related individuals over a span of relative abundances. Blue data series indicates measured abundance of chimerism, and dotted line indicates theoretical perfect relationship between measured and known abundance. (B), Fraction of the total number of informative probes that were recovered at each relative abundance. Error bars in both panels represent SD of 4 independently prepared replicates.

Caption: Fig. 5. Deconvolution of mixed chimeric specimens.

Two mixtures comprising 4 individuals each are presented, with the relative abundance of each component individual indicated along the x axis. The relative abundance of each individual is estimated as the count of unique capture events divided by the expected UMIDs averaged across each of the smMIPs that uniquely identify that individual. Data from individual probes are represented on the y axis, with gray bars indicating average values for the indicated population and whiskers showing SD. The cumulative probability of the result for each population was significant at P <4 x [10.sup.-8]. The individual present at 10% relative abundance in mixture 2 could not be analyzed because of the lack of loci whose presence uniquely identified that individual from others in the pool.
Table 1. Patient characteristics and testing results.

Patient               Specimen
ID        Diagnosis     type         Transplant status

1          AML (b)       BM      Pretransplant

1            AML         PB      156 days posttransplant

1            AML         BM      179 days posttransplant

2            AML         BM      Pretransplant

2            AML         BM      505 days posttransplant

2            AML         BM      582 days posttransplant

2            AML         BM      659 days posttransplant

2            AML         BM      785 days posttransplant

3            AML         BM      Pretransplant

3            AML         BM      28 days posttransplant

3            AML         BM      177 days posttransplant

3            AML         BM      182 days posttransplant

3            AML         BM      414 days posttransplant

4            AML         BM      Pretransplant

4            AML         BM      28 days posttransplant

5            AML         BM      132 days posttransplant

6            AML         BM      Pretransplant

6            AML         BM      28 days posttransplant

7            AML         BM      141 days posttransplant

8            AML         BM      286 days posttransplant

9            AML         BM      Pretransplant

9            AML         BM      28 days posttransplant

10           TLL         BM      70 days posttransplant

11           TLL         BM      365 days posttransplant

12           TLL         BM      334 days posttransplant

13           BLL         BM      358 days posttransplant

14           BLL         BM      364 days posttransplant

15           BLL         PB      365 days posttransplant

Patient         Clinical MRD            Clinical chimerism
ID              quantitation               analysis (a)


1         2.9% abnormal cells        0% recipient (flow sorted
            by flow cytometry (c)      CD3+ I cells and
                                       CD33+ myeloid
                                       fractions tested)

1         60% abnormal cells by
            flow cytometry


2         1.5% abnormal cells        1% recipient
            by flow cytometry

2         Negative by NPM1           0% recipient (flow sorted
            deep sequencing (d)        CD3+ T ceils and
                                       CD33+ myeloid
                                       fractions tested)

2         Negative by NPM1
            deep sequencing

2         Negative by NPM1
            deep sequencing


3         Negative by flow

3         21% abnormal cells by      0% recipient in flow
            flow cytometry             sorted CD3+ T cells;
                                       26% recipient in flow
                                       sorted CD33+
                                       myeloid cells

3         44% abnormal cells by      84% recipient
            flow cytometry

3         26% abnormal cells by
            flow cytometry


4         2.5% suspicious cells      0% recipient (flow sorted
            by flow cytometry,         CD3+T cells, CD33+
            PL 13-11D positive at      myeloid, and CD56+
            5% allelic ratio by        NKcell fractions
            fragment analysis          tested)

5         14% abnormal cells by
            flow cytometry


6         16.5% abnormal cells       79% recipient
            by flow cytometry

7         0.02% abnormal cells
            by flow cytometry

8         Negative by flow           0% recipient


9         Negative by flow           0% recipient (flow sorted
            cytometry                  CD3+ T cells)

10        Negative by flow

11        Negative by flow

12        Negative by flow

13        Negative by flow

14        Negative by flow           0% recipient

15        0.015% abnormal cells      0% recipient
            by flow cytometry,
            real-time PCR
            positive at 0.03%
            ratio to BCR-ABL1 to
            ABL1 Cepheid

          Chimerism quantitation
          by smMIP (average per      Number
Patient   smMIP [+ or -] SD, if    informative
ID            applicable), %         smMIPs

1                   0                   0

1               3.20(2.37)             40

1              74.44(39.38)            36

2                   0                   0

2              2.01 (2.22)             14

2              0.27 (0.213)            11

2               0.92(0.52)             13

2               0.30(0.17)             11

3                  1.99                 1

3              1.07 (0.44)             13

3              67.96(28.89)            13

3             98.03 (26.24)            13

3             91.54 (43.37)            13

4                   0                   0

4               7.55(1.05)              8

5            17.00 (3.71) (e)        116 (e)

6                   0                   0

6              7.71 (1.81)              7

7               0.95(1.7)              31

8               3.05(2.02)              4

9                   0                   0

9              0.18(0.084)             19

10             0.10(0.039)              7

11             0.20(0.077)             16

12              0.23(0.13)             24

13              0.76(0.13)              3

14              1.70(2.93)             27

15              0.62(0.28)             42

             P value per smMIP             Cumulative
                (average per               P value of
Patient    smMIP [+ or -] SD, if          diagnosis by
ID              applicable)                  smMIP

1                    NA                        NA

1            5.36 x [10.sup.-4]        <1 x [10.sup.-100]
            (3.1 x [10.sup.-4]3)

1            <1 x [10.sup.-100]        <1 x [10.sup.-100]

2                    NA                        NA

2            7.12 x [10.sup.-6]        <1 x [10.sup.-100]
            (1.80 x [10.sup.-5])

2               0.078(0.19)            <1 x [10.sup.-100]

2               0.018(0.061)           <1 x [10.sup.-100]

2               0.012(0.030)           <1 x [10.sup.-100]

3            <1 x [10.sup.-100]        <1 x [10.sup.-100]

3              0.0021 (0.011)          <1 x [10.sup.-100]

3            <1 x [10.sup.-100]        <1 x [10.sup.-100]

3            <1 x [10.sup.-100]        <1 x [10.sup.-100]

3            <1 x [10.sup.-100]        <1 x [10.sup.-100]

4                    NA                        NA

4            <1 x [10.sup.-100]        <1 x [10.sup.-100]

5          <1 x [10.sup.-100] (e)    <1 x [10.sup.-100] (e)

6                    NA                        NA

6            5.18 x [10.sup.-4]        <1 x [10.sup.-100]
            (1.37 x [10.sup.-3])

7              0.0052(0.012)           <1 x [10.sup.-100]

8            <1 x [10.sup.-100]        <1 x [10.sup.-100]

9                    NA                        NA

9               0.12(0.018)            <1 x [10.sup.-100]

10              0.020(0.014)          2.95 x [10.sup.-14]

11             0.0045(0.0082)          1.8 x [10.sup.-63]

12            0.0051 (0.0068)         6.8 x [10.sup.-108]

13        2.11 x [10.sup.-10](0.0)     9.4 x [10.sup.-30]

14             0.0006(0.0023)          <1 x [10.sup.-100]

15             0.0016(0.0058)          <1 x [10.sup.-100]

(a) Unsorted specimen tested unless otherwise specified; all samples
tested clinically using Promega PowerPlex assay with reported
sensitivity 1% to 5%.

(b) AML, acute myeloid leukemia; TLL, T lymphoblastic leukemia; BLL, B
lymphoblastic leukemia; BM, bone marrow; PB, peripheral blood; NA, not

(c) Flow cytometry sensitivity is approximately 0.1% to 0.01% of
total leukocytes following erythroid lysis, depending on

(d) Maximal sensitivity for NPM1 deep sequencing is 0.007%.

(e) Threshold for identifying informative probes raised to 0.2
relative abundance because of high quantitation of MRD by conventional
COPYRIGHT 2018 American Association for Clinical Chemistry, Inc.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2018 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:Molecular Diagnostics and Genetics
Author:Wu, David; Waalkes, Adam; Penewit, Kelsi; Salipante, Stephen J.
Publication:Clinical Chemistry
Date:Jun 1, 2018
Previous Article:Neurogranin as Cerebrospinal Fluid Biomarker for Alzheimer Disease: An Assay Comparison Study.
Next Article:Accurate Pan-Cancer Molecular Diagnosis of Microsatellite Instability by Single-Molecule Molecular Inversion Probe Capture and High-Throughput...

Terms of use | Privacy policy | Copyright © 2021 Farlex, Inc. | Feedback | For webmasters |