Debate Round Table

Leukemia (2004) 18, 909–911. doi:10.1038/sj.leu.2403325 Published online 4 March 2004

Interactive ChimerTrack software facilitates computation, visual displays and long-term tracking of chimeric status based on STRs

Copies of the copyrighted application are available without charge from the authors for hospital and university based laboratories.

D Kristt1, J Stein2, I Yaniv2 and T Klein1

  1. 1Immunogenetics and Histocompatibility/Tissue Typing Laboratory, Rabin Medical Center, Petach Tikvah, Israel
  2. 2Bone Marrow Transplantation Unit, Schneider Children's Medical Center, Petach Tikvah, Israel

Correspondence: D Kristt, Laboratory of Histocompatibility and Immunogenetics, Rabin Medical Center, Petach Tikvah, Israel. Fax: +972 3 937 6733; E-mail: dkristt@clalit.org.il

Received 19 August 2003; Accepted 12 January 2004; Published online 4 March 2004.

Chimerism testing has become an important service provided by the post-transplantation monitoring laboratory. This is because chimeric status serves as a useful indication of the cellular dynamics involved in engraftment, or relapse of a malignant recipient cell population, particularly following stem cell transplantation. Previous reports in this Debate Round-Table on chimerism testing1,2,3,4,5 reflect the current interest in quantitatively estimating chimeric status based on an analysis of microsatellite markers, or short tandem repeats (STRs). The principal advantage of this approach is its high sensitivity coupled with the potential quantitative character of the results. The latter affords a most useful capability to follow a patient's chimeric status over time, that is, long-term trends.

Despite sophisticated technological approaches for analyzing DNA, the final phase of the process – when % chimerism is calculated – usually relies on manual transcription of the sequencer results followed by manual computation. Moreover, although STR data are well suited for long-term profiling of chimeric status in a patient, a manual retrospective graphic analysis comparing the present to previous samples is not a feasible undertaking for every patient sample.

Fortunately, such computational and data analytic tasks are typical of the capabilities of current computer technology. Platforms are available that can interactively facilitate computation while providing the capabilities for immediate data analysis within the current sample, and across different samples. For this reason it has been possible to develop an application within the Microsoft® Excel® platform, called ChimerTrack©. The software operates in Excel and can be used in many of the laboratories whose papers have appeared in this Debate-Roundtable.1,2,3,4,5 The purpose of this communication is to describe the features and functionality options of the algorithm.

At the outset it should be emphasized that the software was created in response to our particular laboratory's needs. Most of our patients are HLA matched and receive allogeneic hematopoetic stem cell transplants. We do not routinely enrich or fractionate our cell samples prior to DNA extraction from peripheral blood. Chimeric status evaluation is based on 10 STR markers (plus amelogenin) amplified using a multiplex PCR kit, with fluorescence detection, optimized for forensic DNA analysis, viz, ABI AmpliflSTR SGM+®. PCR products are sequenced with an ABI 3100 Genetic Analyzer®, and these data are analyzed with the ABI Genescan® program, which produces an estimate of the quantity of DNA at the STR alleles. The ChimerTrack utility, which normally is used for Genescan data, runs on Excel 2000 and Excel 97 and will accommodate STR analysis results obtained from any software utilizing the Microsoft platform. Other laboratories have developed their own macros for import and computation of sequencer data,4 and a commercial product is also available by SystemLink® as part of a larger laboratory management software package.

Top

Description of ChimerTrack

This utility application was constructed within a single Excel spreadsheet ('worksheet'). As a result, it can be visualized in its entirety on one screen, and operated by merely scrolling from one site in the worksheet to another. For printing convenience, the worksheet is divided into six parts ('pages'). However, only the first page is routinely printed and issued as the report, with the remainder stored on the computer for further reference or research in the laboratory. It is organized as follows:

  • Page 1Report page (Figure 1), where numeric and graphic displays of % chimerism are automatically displayed.
  • Page 2 – Data Summary page, where the computational results and characteristics of each informative locus are displayed in tabular and graphic forms for the present and previous samples on the same patient.
  • Pages 3, 4 and 5Electropherogram Display page (see below)
  • Page 6Data Record page, where STR data from Genescan is pasted and where laboratory-relevant information is recorded, such as system features, kit and DNA quality. The formulas for computation of % chimerism, average and standard deviation are embedded here. The results are automatically transferred to useful locations throughout the worksheet for numeric, tabular and graphic displays.

Figure 1.
Figure 1 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

Typical Report page of ChimerTrack sent to the clinical transplantation team. All the computational and graphic results appear automatically here, once the donor and recipient allelic data, from Genescan, have been copied and pasted into the Data Record page.

Full figure and legend (159K)

Additionally, it should be recalled that since this is an Excel application, the format, column headings, and many other features, can be easily modified to suit the laboratory's or clinical service's particular requirements.

ChimerTrack performs 13 automatic functions such as selection, computation, transfer, graphic display and updating. These provide the basis for its principal capabilities. Nonetheless, all automatic features can be overridden by merely typing over an automatically generated entry in any cell. Some of the features we find most useful are listed below.

  • Automatic computation of % chimerism
  • Automatic displays of % chimerism on the single Report page as:
    • Numeric display of average and standard deviation for current sample (3–7 loci)
    • Bar graph of previous examinations compared to current sample
  • Single-operation, computer-assisted transfer of Genescan data into ChimerTrack, for individual alleles. The utility selects out the peak area and size from these data for use in its computational algorithm
  • Duplicate occurrence of tables and/or graphs, and critical information throughout worksheet, enables instantaneous updating and confirmation of the operator's results, while minimizing need to scroll through worksheet
  • Provides space for pasting electropherogram displays from Genescan. This is a convenience feature for the operator that eliminates the necessity of moving between programs (i.e. ChimerTrack and Genescan), and provides a convenient paperless archival record, independent of Genescan, which we find useful for clinical consultations, research and preparing lectures

We would like to underscore that ChimerTrack is an 'interactive' utility that does not usurp the decision-making phase of the analysis. It is merely a tool for computation and display. Consequently, the process begins with an examination of the electropherogram display in Genescan, for selecting donor and recipient alleles at each informative locus. The operator must import the data for each locus individually, not as batch data, to insure evaluation of each relevant allele. There are specific, labeled and color-coded domains in ChimerTrack for each of the four possible STR alleles at a locus to accurately guide this task. The actual process of selecting informative loci is beyond the scope of this short correspondence, and has been dealt with in several previous papers.1,6 Nonetheless, a short description of our criteria is provided below, in so far as it bears on the design of our algorithm.

Top

Description of algorithm

The functional core of the utility is the algorithm that automatically calculates the ratio of component DNAs in a mixture based on data derived from an analysis of STRs. It can be equally used to compute percent donor or recipient chimerism. As recommended by a number of workers we base our analysis of STRs on the following formula:1,5,6

Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

where X is the "peak area" of either a donor or recipient allele; and Y is the opposite.

Since there are workers who recommend using other algorithms,1 a few comments regarding our rationale for choosing this particular formula as the basis for ChimerTrack may be appropriate.

As others have noted,7 the multiplex ABI PCR kits often result in considerable variation in interallelic amplification efficiency of 5–15% at a locus. Our experience is similar, and illustrations from the Round-Table show comparable effects.1 This fact has a significant influence on the accuracy of any quantitative algorithm based on a ratio of allelic DNAs at a particular locus. Although, we intend to make this the subject of a future communication, in practical terms, the inaccuracies will be most pronounced, and not directly assessable, in cases of a shared allele. The relevant allelic patterns in loci with a shared allele are:

Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

where R1 is the smaller of the two recipient alleles, and D1 is the smaller of the two donor alleles at a locus in a patient with a mixed chimerism.

Our preliminary data suggest that where there is one shared allele, and the remaining distinct donor and recipient alleles are within 15 bp of each other the computational bias resulting from imbalanced amplification may be minimized. In this case, the algorithm automatically assigns a zero to R2 and D2 contained in the shared allele. We are evaluating a number of other variables that influence quantitative estimates of chimerism. If these are found to produce systematic errors (linear or exponential) then a mathematical correction will be formulated and incorporated into the algorithm. In the interim we are using these restrictive criteria for the selection of loci 'informative for quantification', which has provided at least three loci informative for quantification in each sample. Further, all quantified loci in the sample show little interlocus variability (mean+5%), as expected from a biologically reliable assay. Similar results have been reported previously.5 Thus, our computational algorithm, reflecting these criteria, is mathematically justified and biologically appropriate.

These considerations raise one further point. With numerous sources of technical and biological variability in sensitivity and accuracy, it currently seems reasonable to view STR analysis with the SGM+ kit as producing semiquantitative results; ABI itself concurs with this assessment (N Oldroyd, personal communication). There are two immediate implications of this conclusion. First, we routinely note that % chimerism represents an estimate only. Second, we customarily use other more sensitive methods to detect very low levels of recipient DNA, as in relapse. A corollary of this tentative conclusion is that long-term profiling of chimeric status may be the most clinically useful application of chimerism analysis based on STRs. It is this application, principally, that motivated us to develop the graphic-display features of ChimerTrack.

In conclusion, we describe a new software application that utilizes STR analysis data for paper-less computation of % chimerism and rapid generation of a laboratory report. The utility eliminates transcription errors, and produces graphic, numeric and tabular displays that facilitate clear visualization of the central tendency and range of variation in the current specimen. Yet, ChimerTrack's unique capability for conveniently facilitating rapid assessment of long-term trends in chimeric status may be its most singular and useful contribution. This feature has provided an important and novel perspective on a patient's chimeric status that has aided our transplantation teams in their clinical decision-making. In daily practical terms – if the operator is experienced with Excel – ChimerTrack is convenient and fun to use. It represents an efficient, time-saving tool for the laboratory task of chimerism monitoring.

A by-product of this software approach is the setting of operational standards for the final phases of chimerism testing: quantification and reporting. ChimerTrack implicitly represents an initial statement of standards for these processes. Since this Excel application is easily modifiable, it can serve as a convenient framework for further dialogue leading to the formulation of a broad consensus regarding standards for quantification and reporting. This, in turn, would facilitate sharing of technological and clinical experiences between centers throughout the world. For further information interested parties may contact the corresponding author DK.

Top

References

  1. Fernandez-Aviles F, Urbano-Ispizua A, Aymerich M, Colomer D, Rovira M, Marinez C et al. Serial quantification of lymphoid and myeloid mixed chimerism using multiplex PCR amplification of short tandem repeat-markers predicts graft rejection and relapse respectively, after allogeneic transplantation of CD34+ selected cells from peripheral blood. Leukemia 2003; 17: 613–620. | Article | PubMed | ISI | ChemPort |
  2. Schraml E, Daxberger H, Watzinger F, Lion T. Quantitative analysis of chimerism after allogeneic stem cell transplantation by PCR amplification of microsatellite markers and capillary electrophoresis with fluorescence detection: the Vienna experience. Leukemia 2003; 17: 224–227. | Article | PubMed | ISI | ChemPort |
  3. Acquaviva C, Duval M, Mirebeau D, Bertin R, Cavé H. Quantitative analysis of chimerism after allogeneic stem cell transplantation by PCR amplification of microsatellite markers and capillary electrophoresis with fluorescence detection: the Paris-Robert Debré experience. Leukemia 2003; 17: 224–227. | Article | PubMed |
  4. Kreyenberg H, Holle W, Mohrle S, Niethammer D, Bader P. Quantitative analysis of chimerism after allogeneic stem cell transplantation by PCR amplification of microsatellite markers and capillary electrophoresis with fluorescence detection: the Tuebingen experience. Leukemia 2003; 17: 237–240. | Article | PubMed | ISI | ChemPort |
  5. Koehl U, Beck O, Seifried E, Klingebiel T, Schwabe D, Seidle C. Quantitative analysis of chimerism after allogeneic stem cell transplantation by PCR amplification of microsatellite markers and capillary electrophoresis with fluorescence detection: the Frankfurt experience. Leukemia 2003; 17: 232–236. | Article | PubMed | ISI | ChemPort |
  6. Senitzer D, Gaidulis L. Short tandem repeat analysis of engraftment in allogeneic stem cell transplantation. ASHI Quart 2001; 25: 49–53.
  7. Butler JM. Commonly Used Short Tandem Repeat Markers, in Forensic DNA Typing. San Diego: Academic Press, 2001, pp 53–79.