THE GOLD STANDARD OF GENETICS · SPECIALIST STUDY
Separate donor origin, clonal expansion and disease before naming the finding
After allogeneic haematopoietic cell transplantation, a blood variant can belong to the donor-derived graft rather than the recipient’s constitutional genome. That single fact changes the question. The reader must connect specimen identity, lineage chimerism, variant allele fraction, time, phenotype and confirmation instead of treating every post-transplant sequence result as an inherited answer or a relapse signal.
Start with the closest route. This guide belongs to the Genetics subject guide, the Gold Standard of Education, and the How X Works learning library.
Choose a reading route
Build the biological model · chapters 1–3
Read the measurements · chapters 4–5
Work the case · chapters 6–8
Protect the interpretation · chapters 9–10
Transfer and audit · chapters 11–12
The whole problem at a glance
| Observation | What it can support | What it cannot prove alone |
|---|---|---|
| Variant in post-transplant blood | A clone is detectable in sampled haematopoietic cells | Whether the variant is donor-derived, recipient-derived or constitutional |
| High whole-blood donor chimerism | Most assayed blood DNA is donor-derived | That every lineage, tissue or variant has the same origin |
| Rising variant allele fraction | The sampled clone may be expanding | Malignancy without morphology, counts, lineage and serial context |
| Variant absent from a non-haematopoietic specimen | Evidence against recipient constitutional origin | Donor origin unless donor material or informative chimerism also supports it |
The interpretation is a chain of conditional claims. Each row removes one ambiguity but leaves others. A defensible report states which ambiguity remains.
CHAPTER 1 OF 12 · Build the biological model
1. Rebuild the genome map after transplantation
A transplant changes which genome dominates blood. The recipient still has a constitutional genome in non-haematopoietic tissues, while circulating blood cells may be largely or completely produced by donor stem cells. A laboratory result therefore needs a specimen map: whose DNA could be present, which lineage was sampled, and how complete donor engraftment was at that time.
Chimerism describes the mixture of donor and recipient haematopoietic cells. It is not a universal label attached to the whole person. Whole-blood chimerism can hide lineage differences, and a marrow aspirate can include several cell populations. Record the assay, markers, sensitivity, specimen and date before using a percentage as an origin claim.
A useful first diagram has three boxes: recipient constitutional tissue, recipient residual haematopoiesis and donor-derived haematopoiesis. Put every specimen and variant observation into one box only when evidence permits. If the evidence cannot distinguish two boxes, leave the arrow open rather than closing it with an assumption.
CHAPTER 2 OF 12 · Build the biological model
2. Define clonal haematopoiesis without turning it into a diagnosis
Clonal haematopoiesis means that a blood-cell clone carrying an acquired genetic change has expanded enough to be detected. The term describes a population pattern. It does not by itself establish myelodysplastic syndrome, leukaemia, relapse or a need for treatment. Clinical interpretation depends on counts, morphology, cytogenetics, symptoms, treatment history and the particular variant.
After transplantation, a clone can pre-exist at low level in the donor, emerge during donor-cell replication, or be selected by the recipient environment and treatment exposures. These pathways are biologically different but can produce a similar sequencing snapshot. A single result should therefore be framed as an observation plus a set of candidate explanations.
Do not import thresholds mechanically from an untransplanted population. Detection limits, specimen mixtures and selective pressures differ. State the variant, measured allele fraction, assay limit and serial behaviour, then explain which label is supported and which disease criteria have not been met.
CHAPTER 3 OF 12 · Build the biological model
3. Keep donor-derived CH, donor-engrafted CH and donor-cell neoplasia distinct
Donor-derived or donor-engrafted clonal haematopoiesis refers to a detectable clone in donor-origin blood cells without automatically implying malignancy. Donor-cell leukaemia or another donor-cell neoplasm requires a much stronger clinicopathological case. The shared word “donor” should not collapse these categories.
Origin evidence may come from donor testing, informative germline markers, sex-chromosome findings in appropriate donor-recipient pairs, lineage chimerism, or integration of several methods. Each method has limits. Sex mismatch is not universally available; donor testing may be unavailable or raise consent questions; and bulk chimerism may not identify the cells carrying a particular variant.
Write the conclusion in two sentences. Sentence one addresses origin: “The findings support a donor-derived clone because…” Sentence two addresses state: “The current evidence does not establish a haematologic neoplasm because…” This grammar prevents origin confidence from being mistaken for disease confidence.
CHAPTER 4 OF 12 · Read the measurements
4. Read variant allele fraction as a mixture, not a headcount
Variant allele fraction is the fraction of sequencing reads carrying the variant at an assayed position. It is influenced by zygosity, copy number, purity, lineage composition, technical bias and the proportion of cells in the clone. A 10% VAF does not simply mean 10% of all cells are abnormal.
For a heterozygous variant in a diploid region and a relatively pure cell population, a rough teaching estimate is that clone fraction may be near twice the VAF. That approximation fails with copy-number change, loss of heterozygosity, mixed lineages, low depth or assay bias. The report should name those conditions before presenting any cell-fraction inference.
Serial VAFs are most useful when specimens and assays are comparable. A rise from 4% in whole blood to 9% in enriched myeloid cells may reflect both biology and specimen change. Put assay, tissue and lineage beside every number so the timeline cannot imply a false like-for-like comparison.
CHAPTER 5 OF 12 · Read the measurements
5. Use chimerism at the resolution the question requires
Whole-blood chimerism answers a broad engraftment question. Variant origin may require lineage-specific or cell-sorted evidence. If a myeloid-associated variant is under review, a high donor percentage in whole blood is supportive but may not settle the origin of a small residual clone.
Measurement uncertainty matters near a decision boundary. Report the laboratory method, informative loci, limit of detection and whether percentages are rounded. A change from 98% to 99% donor may be analytically unimportant, while a stable 98% result can still contain enough recipient cells to produce a low-level variant signal.
Ask a counterfactual question: if the variant-bearing cells were entirely recipient-derived, would the measured recipient fraction be large enough to contain them? If not, donor origin becomes more plausible. If yes, chimerism alone cannot decide. This check turns two percentages into a falsifiable constraint.
CHAPTER 6 OF 12 · Work the case
6. Work an origin calculation without pretending it is exact
Suppose whole blood is 98% donor and a heterozygous variant has a 4% VAF in a diploid region. A rough clone-fraction estimate is 8% of nucleated cells. The measured recipient fraction is about 2%. Under those simplified assumptions, a purely recipient-derived clone cannot easily account for the signal, so donor origin becomes the leading explanation.
Now expose the assumptions. The chimerism and sequencing aliquots may differ; lineage composition may differ; copy number may not be normal; and both measurements have error. The result is therefore a constraint, not a proof. A defensible note says the numbers are inconsistent with a simple recipient-only model and identifies the confirmation needed.
Change the case to 90% donor with a 3% VAF. The rough clone fraction is 6%, which could fit within the 10% recipient component. The same calculation no longer distinguishes origin. A strong learner changes the conclusion when the feasible region changes.
Worked timeline: keep origin and behaviour separate
| Time | Finding | Provisional reading | Next check |
|---|---|---|---|
| Before transplant | Recipient blood negative at adequate depth | No detected pre-existing recipient clone | Retain assay limit and specimen details |
| Month 3 | Whole blood 98% donor; variant 4% VAF | Likely graft-associated low-level clone | Lineage chimerism and donor information if appropriate |
| Month 18 | Variant 11% VAF; counts stable | Expansion is plausible; disease not established | Repeat, examine counts and morphology |
| Month 24 | Variant 12% VAF; no diagnostic phenotype | Persistent donor-derived CH is one explanation | Clinical surveillance under specialist protocol |
The numbers are fictional. Their purpose is to show why “donor-derived” and “dangerous” are different axes.
CHAPTER 7 OF 12 · Work the case
7. Build a serial evidence table before reading a trend
Put dates down the left and observations across the top: specimen, lineage, chimerism, VAF, blood counts, morphology, cytogenetics, symptoms and treatment changes. This prevents an attractive VAF line from becoming the whole story. A clone that rises while counts and morphology remain stable is different from one accompanied by progressive cytopenias and dysplasia.
Distinguish missing from normal. If marrow morphology was not performed, write “not assessed,” not “normal.” If a donor sample was unavailable, do not treat that absence as a negative donor test. A timeline becomes trustworthy when empty cells carry their real meaning.
Set a prospective trigger for review, such as confirmed expansion across comparable specimens plus a relevant change in counts. The trigger is a governance tool, not a universal clinical threshold. It should be chosen by the treating service and revised when the patient, assay or evidence changes.
CHAPTER 8 OF 12 · Work the case
8. Confirm tissue origin without creating a new specimen error
Recipient constitutional confirmation should use a genuinely non-haematopoietic source when possible and clinically appropriate. Buccal or saliva samples can contain blood-derived leukocytes, especially after transplantation, so they may not be independent enough. Cultured skin fibroblasts are often considered, but collection, culture and interpretation require specialist handling.
Identity controls matter as much as tissue choice. Confirm labels, collection date, chain of custody and expected donor-recipient markers. A beautiful molecular comparison cannot rescue a switched specimen. Where the donor is tested, consent, privacy and incidental findings must be addressed rather than treated as laboratory afterthoughts.
The confirmation plan should answer the minimum necessary question. If the immediate issue is whether a result is recipient constitutional, do not collect unrelated genomic information merely because sequencing is available. Proportionate testing protects both interpretability and people.
CHAPTER 9 OF 12 · Protect the interpretation
9. Separate surveillance from prediction
Some donor-derived clones persist without diagnostic disease; some expand; rare cases progress to donor-cell neoplasia. These facts support surveillance, not deterministic prediction. The individual risk depends on clone features, clinical context and evidence that remains incomplete.
A monitoring plan should state what will be measured, at what interval, with which specimen and what event prompts escalation. Repeating tests without a decision rule can create noise and anxiety. A useful plan connects each measurement to a possible action and identifies who reviews the result.
Avoid language such as “pre-leukaemia” when the evidence supports only clonal haematopoiesis. Explain uncertainty plainly: the clone is real, its origin is supported to a stated degree, and its future behaviour cannot be read from one molecular value.
CHAPTER 10 OF 12 · Protect the interpretation
10. Protect donor and recipient consent, privacy and communication
A donor-derived finding can reveal information about two people. Testing the donor may have implications for insurance, future health care, relatives and willingness to know. Services need a consent pathway that respects the recipient’s care while limiting unnecessary disclosure of donor information.
Reports should separate data available to the recipient team from data that require donor-authorised access. Communication across transplant centres must preserve specimen identity and purpose. A donor should not become a silent research subject because an unexpected recipient result raised a question.
When results are discussed with the recipient, distinguish what is known about their current health from what is inferred about the donor. Use neutral language and explain why additional testing may or may not change care. Ethical clarity is part of analytical validity because it determines what evidence can legitimately be obtained.
CHAPTER 11 OF 12 · Transfer and audit
11. Transfer the framework to a different post-transplant signal
Move from a point mutation to a donor-derived cytogenetic abnormality. The same framework still asks about specimen identity, donor-recipient mixture, serial change and phenotype. The measurement model changes: metaphase selection, clone size, assay resolution and chromosome-level events replace simple VAF assumptions.
Move again to minimal residual disease testing for the recipient’s original malignancy. Now a known recipient marker may make origin easier, but false negatives, clonal evolution and assay sensitivity become central. Do not carry the “twice the VAF” shortcut into a method that reports a different quantity.
Transfer succeeds when the questions remain stable but the evidence model changes. Name the stable questions, then replace the method-specific assumptions. That is safer than reusing a familiar threshold in a new assay.
CHAPTER 12 OF 12 · Transfer and audit
12. Write an auditable interpretation record
End with five lines: specimen and date; assay and limit; origin evidence; disease-state evidence; and the next discriminating check. Add the alternatives considered and why they were not selected. This compact record lets another team reconstruct the reasoning without guessing what “likely donor-derived” meant.
Audit the wording against the evidence table. Every adjective should have a source: “rising” requires comparable serial values; “donor-derived” requires origin evidence; “clonal haematopoiesis” requires an acquired clone without unsupported disease language. Remove any conclusion whose evidence cell is empty.
The gold standard is not certainty where certainty is impossible. It is a transparent boundary between observation, inference and action, maintained across laboratory, clinical and ethical decisions.
Build a variant-provenance dossier
A provenance dossier begins before interpretation. Record the original recipient diagnosis, pre-transplant molecular profile, donor relationship, conditioning regimen, graft source, transplant date, engraftment history and any prior chimerism studies. Add the exact laboratory identifiers for every specimen used in the new comparison. These facts establish which genomes and selective pressures could plausibly contribute to the result.
For the variant itself, preserve transcript, genome build, genomic coordinate, nucleotide and protein notation, depth, alternate reads, quality filters, assay panel version and reportable range. A label such as DNMT3A p.R882H is not enough to compare results if laboratories used different transcripts, filters or limits. Harmonise the representation before drawing a trend.
Link each molecular result to a specimen map. A marrow aspirate, peripheral blood sample, sorted myeloid fraction and cultured skin fibroblast do not answer the same question. Note cellular enrichment, collection timing, transfusions where relevant and the possibility of donor-derived leukocyte contamination. The map prevents apparently independent confirmations from repeating the same biological source.
Finish with an evidence-status column: observed, derived, reported by another laboratory, or assumed. This classification matters when a case crosses institutions. A copied statement that the donor was negative may mean formal sequencing, targeted testing, a clinical history, or only that no disorder was known. The dossier should never treat those as equivalent.
Distinguish technical movement from clonal movement
Before interpreting a VAF change, compare analytical conditions. Depth, unique-molecule counting, strand balance, read position, mapping context and the caller’s filters can change the estimate, especially at low levels. Review raw quality metrics or a validated laboratory summary rather than comparing rounded percentages from two reports.
Estimate an uncertainty interval or at least the plausible counting variation. With 500 informative reads, 20 alternate reads produce 4%, but another draw from the same underlying proportion will not reproduce exactly. A change from 4% to 5% may be less compelling than the printed decimals suggest. Technical replicates or orthogonal assays can help near a consequential boundary.
Biological comparability is separate. Whole blood after granulocyte recovery may have a different lineage mixture from whole blood during lymphocyte predominance. A stable clone restricted to one lineage can appear to expand or contract as counts change. Pair VAF with differential counts, lineage chimerism and treatment context whenever those data are material.
A defensible trend statement names the comparable interval: “The variant increased across three peripheral-blood specimens tested by the same assay while myeloid counts were broadly stable.” If those conditions are absent, use narrower wording such as “the reported VAFs differ” and state what would distinguish analytical from biological change.
Plan donor testing as a governed clinical question
Donor testing should not begin with the broadest available panel. Define the question first: confirm origin of one variant, evaluate a suspected donor-derived disorder, or inform care of the donor. Each purpose changes consent, test scope, result ownership and who is responsible for follow-up. A recipient-care need does not automatically authorise unrestricted analysis of donor DNA.
Identify the contact pathway before the sample is requested. Transplant registries, collection centres and clinical teams may hold different responsibilities. Avoid informal contact that bypasses prior consent or reveals recipient information unnecessarily. Document which facts must be shared to explain the request and which can remain confidential.
Prepare for three outcomes: variant detected, variant not detected, or result uninformative. A negative result can reflect tissue, timing, assay sensitivity or a clone that emerged after donation. An uninformative result should not be rewritten as reassurance. Decide in advance how each outcome affects the recipient interpretation and the donor’s access to counselling.
Where donor testing is not available or proportionate, say so. Integrate chimerism, serial behaviour and recipient non-haematopoietic testing, and retain residual uncertainty. Ethical limits on evidence do not justify pretending the evidence exists; they justify a carefully bounded conclusion.
Use a multi-hypothesis evidence matrix
List competing hypotheses before selecting one: recipient constitutional variant, recipient somatic clone, donor constitutional variant, donor clonal haematopoiesis, donor-cell neoplasm, sequencing artefact and specimen error. Add or remove hypotheses to fit the case. This step prevents the first plausible explanation from monopolising later evidence.
For each observation, mark whether it supports, weakens or does not distinguish the hypotheses. High donor chimerism may support donor-origin explanations but not distinguish donor constitutional variation from donor clonal haematopoiesis. Progressive cytopenias may increase concern for disease but do not determine origin. The matrix preserves these two axes.
Weight evidence by reliability and independence. Donor testing and variant-linked cell sorting may be highly informative, while two bulk specimens from the same biological compartment may be correlated. Do not count a laboratory report, a copied clinical note and a multidisciplinary summary as three independent observations if they trace to one test.
End with the next test’s expected information gain. Choose the feasible test most likely to separate the leading hypotheses, not the one that merely produces more data. If two hypotheses would lead to the same current care, monitoring may be more proportionate than invasive confirmation.
Write risk communication that preserves uncertainty
Lead with the finding that is secure. For example: a blood-cell clone carrying a specified variant is detectable after transplantation. Then describe the supported origin, the evidence for or against a current neoplasm, and the monitoring plan. This order prevents a rare severe outcome from becoming the headline before its probability is known.
Use absolute statements only for absolute evidence. “This is not inherited by the recipient” may be appropriate after strong non-haematopoietic evidence; “the blood result alone does not establish an inherited variant” is safer before confirmation. Explain that donor-derived haematopoietic cells do not replace the recipient’s reproductive cells, while acknowledging any separate constitutional question.
Avoid vague reassurance. Saying “nothing to worry about” can undermine trust if surveillance continues. A clearer explanation is that no current diagnostic evidence of malignancy has been established, the clone has a specified behaviour, and the team will review named clinical or laboratory changes. Concrete boundaries reduce both alarm and false certainty.
Document the conversation and the terms used. Recipients may later read reports from several teams that use “clone,” “mutation,” “CH” and “leukaemia” differently. A shared glossary and written summary help prevent terminology drift from becoming clinical misinformation.
Verification laboratory: review a complete fictional case
Case file: a 52-year-old recipient is two years after an unrelated-donor transplant. Pre-transplant sequencing did not report the new variant. Current peripheral blood shows 99% donor chimerism and DNMT3A p.R882H at 7% VAF; counts are stable, marrow has not been repeated, and no donor sample is currently available. The facts are fictional and designed to exercise the method.
Observation layer: the variant is reproducibly detected in post-transplant blood by a validated assay. Origin layer: 99% bulk donor chimerism makes a donor-associated clone likely, and a simple recipient-only model struggles to explain a rough 14% clone fraction. Disease layer: stable counts and absent current morphology do not establish a neoplasm, but “not assessed” must not be written as “normal.”
Quality review asks whether chimerism and sequencing used comparable aliquots, whether the region is diploid, whether the VAF is free from mapping artefact, and whether lineage composition could change the estimate. It also retrieves the pre-transplant assay’s coverage and limit rather than treating an unreported variant as proven absence.
The leading hypothesis is donor-derived clonal haematopoiesis. Alternatives remain: an unusual recipient-derived clone inconsistent with the bulk mixture, a donor constitutional variant present in haematopoietic cells, or a technical or identity error. Donor-cell neoplasia is a monitored possibility, not a current conclusion.
The next discriminating evidence could include comparable serial blood testing, lineage-specific chimerism, clinical review of counts and morphology, and a governed donor-testing discussion. Recipient non-haematopoietic testing addresses constitutional status but does not alone distinguish donor constitutional variation from donor-acquired CH.
A provisional report might state: “A reproducible DNMT3A variant is present in post-transplant blood. Given near-complete donor chimerism and the measured allele fraction, the findings support a donor-derived haematopoietic clone. Current available evidence does not establish a donor-cell neoplasm; serial clinicopathological correlation is advised.”
Now change one fact: the variant is present at 49% VAF in blood and the donor is confirmed heterozygous in a non-haematopoietic specimen. The interpretation shifts toward donor constitutional carriage rather than a small acquired clone. The biological and counselling questions change even though donor origin remains.
Change another fact: the recipient had the same variant at 2% before transplant and now has mixed chimerism with 20% recipient cells. Recipient-clone persistence or expansion becomes feasible. The exercise shows why no title, gene name or threshold can replace the full specimen-and-time model.
Before sign-off: the minimum evidence checklist
Confirm that every sample has a verified identity, tissue description, collection date and relationship to transplantation. Confirm that comparison results use compatible variant notation and genome build. Resolve clerical ambiguity before adding biological interpretation.
Confirm that VAF, depth, assay sensitivity, copy-number context and lineage composition have been reviewed. Mark which trend comparisons are genuinely like for like. Explain any approximation used to translate VAF into clone fraction.
Confirm that chimerism evidence is matched to the relevant date, specimen and lineage. State whether it constrains origin or proves it. Do not let a rounded percentage carry more precision than the method supports.
Confirm that constitutional, somatic, donor-derived and disease-state questions are answered separately. Retain competing hypotheses, negative evidence and unperformed tests. State what new evidence would change the conclusion.
Questions readers ask
Does full donor chimerism mean every blood variant came from the donor?
It makes donor origin more plausible, but resolution, lineage, assay sensitivity and specimen mixture still matter. Variant-specific evidence may be needed.
Can a donor-derived clone be inherited by the recipient’s children?
No. Donor-derived haematopoietic cells do not replace the recipient’s reproductive cells. Constitutional inheritance questions require recipient germline evidence.
Does a rising VAF prove leukaemia?
No. It can support clonal expansion, but diagnosis requires clinical, morphological and other laboratory evidence.
Why not use saliva for germline confirmation?
Saliva and buccal specimens can contain donor-derived blood cells after transplantation. Suitability depends on the question and specialist laboratory practice.
Practise, verify and continue
A recipient has 97% donor whole-blood chimerism and a new heterozygous variant at 6% VAF, with stable counts and no marrow examination. Build the feasible origin argument, name four assumptions, and write one sentence about origin plus one sentence about disease state.
Then change the chimerism to 88% donor and move the sequencing specimen from whole blood to enriched myeloid cells. Recalculate what can be excluded and explain why the same conclusion cannot simply be copied.
Related eduKate reading
Clonal haematopoiesis, blood-only false signals and germline confirmation
Mosaicism, variant allele fractions and tissue-specific detection
Reverse phenotyping, allelic series and genotype-first recall
Phenotype harmonisation and cross-biobank case definitions
Authoritative sources and scope
Gondek and colleagues on donor-cell leukaemia arising from clonal haematopoiesis after transplantation
Nevejan and colleagues on malignant progression of donor-engrafted clonal haematopoiesis
Review of donor-cell leukaemia after allogeneic transplantation
Study of recipient clonal haematopoiesis and donor-derived findings after transplantation
The sources support the biological distinctions, reported pathways and need for integrated interpretation. The numerical timeline is a fictional teaching case. This article is educational and does not replace transplant-haematology, genetics or laboratory advice.