THE GOLD STANDARD OF GENETICS · SPECIALIST STUDY
Treat a microarray finding as a bounded copy-number measurement across a defined interval, not as a complete chromosome picture or a diagnosis by size alone
Chromosomal microarray analysis can detect losses and gains across the genome at a finer scale than conventional chromosome analysis. That reach creates a common trap: a technically precise interval can look self-explanatory even when its clinical meaning is not. Every result depends on platform design, probe density, genome build, quality controls, reporting thresholds and whether the array measures only dosage or also single-nucleotide polymorphism patterns. A deletion and a duplication over the same coordinates are not mirror images; genes and regions may be sensitive to loss, gain, both or neither. Size matters, but gene content, dosage mechanism, population evidence, case evidence, inheritance and phenotype matter more. The array also has blind spots. Balanced rearrangements, many sequence variants, repeat expansions, small events between probes and low-level mosaicism may escape detection. Responsible interpretation therefore separates analytical detection from evidence-based classification and then separates classification from the question of whether the finding explains this person’s presentation. It confirms technically or through family studies when the decision requires it, reports uncertain findings without manufacturing certainty, and directs unresolved cases to the next appropriate method.
This guide belongs to the Genetics subject root, the Gold Standard of Education, and the How X Works hub. It begins with a concrete decision, then tests whether the method still holds when the context changes.
Choose a reading route
Frame the question · chapters 1, 2, 3
Validate the array result · chapters 4, 5
Interpret the interval · chapters 6, 7, 8
Connect result to person · chapters 9, 10, 11
Report and transfer · chapters 12, 13
The whole problem at a glance
| Question | Evidence needed | Unsafe shortcut |
|---|---|---|
| What did the array measure? | Platform, probes, build and validated scope | Genome-wide means every variant |
| Is the call reliable? | Quality, probe pattern and confirmation rule | Software output is proof |
| Is the interval dosage sensitive? | Curated gene and region evidence | Large always means pathogenic |
| Does it explain the phenotype? | Independent classification plus clinical correlation | Pathogenic means causal for every feature |
Classification and patient-level explanation are related but separate decisions.
CHAPTER 1 OF 13 · Frame the question
1. Define indication and test capability
State whether the question concerns developmental difference, congenital anomalies, pregnancy, cancer, infertility or another indication, and whether a constitutional or neoplastic array is intended. Begin with one person or pregnancy, one clinical indication, one validated array platform, one copy-number interval, one genome build, one classification decision, one confirmation plan and one review date. State the decision, specimen, record or person, the time boundary and the responsible owner before interpreting any result.
Worked example: A child with developmental delay receives a constitutional SNP array; a tumour copy-number question is not silently answered by the same assumptions. Put the starting evidence, applicable rule, intermediate decision, later observation and consequence on separate lines so that a fluent story cannot replace a reproducible pathway.
Evidence to retain: referral; phenotype; age; sample; constitutional or tumour scope; prior tests; consent; intended decision Preserve provenance, dates, versions, identity checks, responsible people and intended use. If a necessary item is missing, label the gap and its effect instead of converting absence into reassurance.
Boundary and misconception: Microarray is a method family, not one universal test, and performance claims do not transfer automatically between constitutional, prenatal and tumour applications. Test the strongest credible alternative explanation and name the new evidence that would change the conclusion. The pass condition is a bounded, reviewable claim, not the removal of uncertainty.
Evidence discipline: record the observation, version or time boundary that supports this chapter’s conclusion. Preserve the rejected alternative as well as the selected action. That makes a later review able to distinguish a changed world from an undocumented change of mind.
Changed-context transfer: Keep the sample fixed but change the question from constitutional to neoplastic and rebuild validation. The method transfers only if the decision question and measurement meaning survive the change. If either changes, rebuild the comparison rather than copying the earlier threshold or weight.
Chapter 1 check · what should survive an audit
- Name the exact decision, unit, time boundary and responsible authority.
- Preserve the supporting evidence, controls and strongest credible alternative.
- Record the next action, review date, stop rule and reopening condition.
CHAPTER 2 OF 13 · Frame the question
2. Identify platform, build and reporting thresholds
Record array type, manufacturer or validated design, genome build, probe density, effective resolution, calling software and laboratory thresholds. Begin with one person or pregnancy, one clinical indication, one validated array platform, one copy-number interval, one genome build, one classification decision, one confirmation plan and one review date. State the decision, specimen, record or person, the time boundary and the responsible owner before interpreting any result.
Worked example: Two laboratories report different breakpoints for the same event because probe layouts differ; the intervals are compared on one build before declaring disagreement. Put the starting evidence, applicable rule, intermediate decision, later observation and consequence on separate lines so that a fluent story cannot replace a reproducible pathway.
Evidence to retain: platform; probe design; build; software; filters; resolution; size threshold; gene threshold; report version Preserve provenance, dates, versions, identity checks, responsible people and intended use. If a necessary item is missing, label the gap and its effect instead of converting absence into reassurance.
Boundary and misconception: Nominal probe spacing is not uniform clinical resolution, and build conversion can shift or split coordinates. Test the strongest credible alternative explanation and name the new evidence that would change the conclusion. The pass condition is a bounded, reviewable claim, not the removal of uncertainty.
Evidence discipline: record the observation, version or time boundary that supports this chapter’s conclusion. Preserve the rejected alternative as well as the selected action. That makes a later review able to distinguish a changed world from an undocumented change of mind.
Changed-context transfer: Reduce probe density across the interval and widen the boundary uncertainty. The method transfers only if the decision question and measurement meaning survive the change. If either changes, rebuild the comparison rather than copying the earlier threshold or weight.
Chapter 2 check · what should survive an audit
- Name the exact decision, unit, time boundary and responsible authority.
- Preserve the supporting evidence, controls and strongest credible alternative.
- Record the next action, review date, stop rule and reopening condition.
CHAPTER 3 OF 13 · Frame the question
3. Protect specimen identity and cellular context
Trace collection, extraction, sex-chromosome expectation, contamination, transplant or mosaic context and sample quality before interpreting a call. Begin with one person or pregnancy, one clinical indication, one validated array platform, one copy-number interval, one genome build, one classification decision, one confirmation plan and one review date. State the decision, specimen, record or person, the time boundary and the responsible owner before interpreting any result.
Worked example: A low-amplitude blood finding after stem-cell transplant is paused until donor and recipient contributions are understood. Put the starting evidence, applicable rule, intermediate decision, later observation and consequence on separate lines so that a fluent story cannot replace a reproducible pathway.
Evidence to retain: sample ID; tissue; collection; extraction; quality; contamination; transfusion; transplant; cell mixture; identity checks Preserve provenance, dates, versions, identity checks, responsible people and intended use. If a necessary item is missing, label the gap and its effect instead of converting absence into reassurance.
Boundary and misconception: A technically valid signal may describe the wrong cellular population, while blood may not represent every mosaic tissue. Test the strongest credible alternative explanation and name the new evidence that would change the conclusion. The pass condition is a bounded, reviewable claim, not the removal of uncertainty.
Evidence discipline: record the observation, version or time boundary that supports this chapter’s conclusion. Preserve the rejected alternative as well as the selected action. That makes a later review able to distinguish a changed world from an undocumented change of mind.
Changed-context transfer: Add a donor-derived cell population and redesign confirmation. The method transfers only if the decision question and measurement meaning survive the change. If either changes, rebuild the comparison rather than copying the earlier threshold or weight.
Chapter 3 check · what should survive an audit
- Name the exact decision, unit, time boundary and responsible authority.
- Preserve the supporting evidence, controls and strongest credible alternative.
- Record the next action, review date, stop rule and reopening condition.
CHAPTER 4 OF 13 · Validate the array result
4. Review raw probe pattern and quality metrics
Inspect log-ratio and allele-frequency patterns, probe count, segmentation, waviness, noise and genome-wide quality rather than accepting a software label alone. Begin with one person or pregnancy, one clinical indication, one validated array platform, one copy-number interval, one genome build, one classification decision, one confirmation plan and one review date. State the decision, specimen, record or person, the time boundary and the responsible owner before interpreting any result.
Worked example: A small deletion call rests on two poorly performing probes and disappears under a validated reanalysis, so it is not reported as a clinical event. Put the starting evidence, applicable rule, intermediate decision, later observation and consequence on separate lines so that a fluent story cannot replace a reproducible pathway.
Evidence to retain: raw plots; probes; BAF; LRR; QC thresholds; segmentation; analyst review; repeat rule; control performance Preserve provenance, dates, versions, identity checks, responsible people and intended use. If a necessary item is missing, label the gap and its effect instead of converting absence into reassurance.
Boundary and misconception: Algorithms can oversegment noise or merge adjacent events, and a clean summary table does not replace raw-pattern review. Test the strongest credible alternative explanation and name the new evidence that would change the conclusion. The pass condition is a bounded, reviewable claim, not the removal of uncertainty.
Evidence discipline: record the observation, version or time boundary that supports this chapter’s conclusion. Preserve the rejected alternative as well as the selected action. That makes a later review able to distinguish a changed world from an undocumented change of mind.
Changed-context transfer: Hold coordinates constant but fail a genome-wide noise metric and suspend interpretation. The method transfers only if the decision question and measurement meaning survive the change. If either changes, rebuild the comparison rather than copying the earlier threshold or weight.
Chapter 4 check · what should survive an audit
- Name the exact decision, unit, time boundary and responsible authority.
- Preserve the supporting evidence, controls and strongest credible alternative.
- Record the next action, review date, stop rule and reopening condition.
CHAPTER 5 OF 13 · Validate the array result
5. Define interval and variant structure cautiously
Report minimum and maximum plausible boundaries, copy state, mosaic evidence, affected genes and whether a complex or discontinuous pattern is possible. Begin with one person or pregnancy, one clinical indication, one validated array platform, one copy-number interval, one genome build, one classification decision, one confirmation plan and one review date. State the decision, specimen, record or person, the time boundary and the responsible owner before interpreting any result.
Worked example: A gain with an internal normal segment is described as a complex pattern rather than one simple tandem duplication. Put the starting evidence, applicable rule, intermediate decision, later observation and consequence on separate lines so that a fluent story cannot replace a reproducible pathway.
Evidence to retain: coordinates; build; copy state; probes; breakpoint uncertainty; mosaic fraction estimate; complex pattern; genes; orientation limits Preserve provenance, dates, versions, identity checks, responsible people and intended use. If a necessary item is missing, label the gap and its effect instead of converting absence into reassurance.
Boundary and misconception: Microarray usually cannot determine orientation, insertion site or balanced structure, so duplication does not mean tandem duplication. Test the strongest credible alternative explanation and name the new evidence that would change the conclusion. The pass condition is a bounded, reviewable claim, not the removal of uncertainty.
Evidence discipline: record the observation, version or time boundary that supports this chapter’s conclusion. Preserve the rejected alternative as well as the selected action. That makes a later review able to distinguish a changed world from an undocumented change of mind.
Changed-context transfer: Keep dosage change fixed but alter insertion site and show what the array cannot resolve. The method transfers only if the decision question and measurement meaning survive the change. If either changes, rebuild the comparison rather than copying the earlier threshold or weight.
Chapter 5 check · what should survive an audit
- Name the exact decision, unit, time boundary and responsible authority.
- Preserve the supporting evidence, controls and strongest credible alternative.
- Record the next action, review date, stop rule and reopening condition.
Worked copy-number interpretation
| Observation | Possible meaning | Needed check | Do not infer |
|---|---|---|---|
| 650 kb deletion | Loss of material in tested interval | Build, probes, genes and HI evidence | Size alone proves disease |
| Inherited duplication | May be benign or variably expressive | Parent phenotype, penetrance and TS evidence | Inherited means harmless |
| Low-amplitude shift | Possible mosaic CNV or artefact | Cell fraction, QC and orthogonal method | Exact mosaic percentage |
| Normal array | No reportable CNV in validated scope | Blind spots and next test | Genome is normal |
Copy-number gains and losses require their own evidence pathways.
CHAPTER 6 OF 13 · Interpret the interval
6. Evaluate dosage sensitivity for loss and gain separately
Use curated haploinsufficiency and triplosensitivity evidence for genes and regions, considering disease mechanism rather than assuming any included gene is causal. Begin with one person or pregnancy, one clinical indication, one validated array platform, one copy-number interval, one genome build, one classification decision, one confirmation plan and one review date. State the decision, specimen, record or person, the time boundary and the responsible owner before interpreting any result.
Worked example: A deletion fully removes a gene with established haploinsufficiency, while the reciprocal duplication lacks triplosensitivity evidence and is classified separately. Put the starting evidence, applicable rule, intermediate decision, later observation and consequence on separate lines so that a fluent story cannot replace a reproducible pathway.
Evidence to retain: ClinGen curation; gene mechanism; region evidence; transcript; overlap; inheritance; literature date; conflicts Preserve provenance, dates, versions, identity checks, responsible people and intended use. If a necessary item is missing, label the gap and its effect instead of converting absence into reassurance.
Boundary and misconception: Loss and gain are not opposites of equal meaning, and association with a gene does not prove dosage is the mechanism. Test the strongest credible alternative explanation and name the new evidence that would change the conclusion. The pass condition is a bounded, reviewable claim, not the removal of uncertainty.
Evidence discipline: record the observation, version or time boundary that supports this chapter’s conclusion. Preserve the rejected alternative as well as the selected action. That makes a later review able to distinguish a changed world from an undocumented change of mind.
Changed-context transfer: Reverse the event from deletion to duplication and rebuild the evidence score. The method transfers only if the decision question and measurement meaning survive the change. If either changes, rebuild the comparison rather than copying the earlier threshold or weight.
Chapter 6 check · what should survive an audit
- Name the exact decision, unit, time boundary and responsible authority.
- Preserve the supporting evidence, controls and strongest credible alternative.
- Record the next action, review date, stop rule and reopening condition.
CHAPTER 7 OF 13 · Interpret the interval
7. Apply an evidence framework without double counting
Score genomic content, established regions, case evidence, population frequency and segregation under a documented CNV framework, keeping dependent evidence visible. Begin with one person or pregnancy, one clinical indication, one validated array platform, one copy-number interval, one genome build, one classification decision, one confirmation plan and one review date. State the decision, specimen, record or person, the time boundary and the responsible owner before interpreting any result.
Worked example: Several publications describe the same family; the cases are not counted as independent support three times. Put the starting evidence, applicable rule, intermediate decision, later observation and consequence on separate lines so that a fluent story cannot replace a reproducible pathway.
Evidence to retain: framework version; evidence categories; points; case overlap; controls; frequency; segregation; dependencies; reviewer Preserve provenance, dates, versions, identity checks, responsible people and intended use. If a necessary item is missing, label the gap and its effect instead of converting absence into reassurance.
Boundary and misconception: A numeric score aids transparency but cannot rescue weak, duplicated or misapplied evidence. Test the strongest credible alternative explanation and name the new evidence that would change the conclusion. The pass condition is a bounded, reviewable claim, not the removal of uncertainty.
Evidence discipline: record the observation, version or time boundary that supports this chapter’s conclusion. Preserve the rejected alternative as well as the selected action. That makes a later review able to distinguish a changed world from an undocumented change of mind.
Changed-context transfer: Remove duplicated cases and recalculate classification. The method transfers only if the decision question and measurement meaning survive the change. If either changes, rebuild the comparison rather than copying the earlier threshold or weight.
Chapter 7 check · what should survive an audit
- Name the exact decision, unit, time boundary and responsible authority.
- Preserve the supporting evidence, controls and strongest credible alternative.
- Record the next action, review date, stop rule and reopening condition.
CHAPTER 8 OF 13 · Interpret the interval
8. Separate classification from phenotype correlation
Classify the CNV using evidence about the variant, then ask separately whether its known effects fit the person’s findings, penetrance and alternative diagnoses. Begin with one person or pregnancy, one clinical indication, one validated array platform, one copy-number interval, one genome build, one classification decision, one confirmation plan and one review date. State the decision, specimen, record or person, the time boundary and the responsible owner before interpreting any result.
Worked example: A pathogenic deletion is real but does not explain a new symptom outside its established spectrum; the symptom remains open. Put the starting evidence, applicable rule, intermediate decision, later observation and consequence on separate lines so that a fluent story cannot replace a reproducible pathway.
Evidence to retain: classification; phenotype ontology; onset; penetrance; expressivity; known spectrum; alternatives; second diagnoses Preserve provenance, dates, versions, identity checks, responsible people and intended use. If a necessary item is missing, label the gap and its effect instead of converting absence into reassurance.
Boundary and misconception: Pathogenic does not mean every feature is caused by the CNV, while poor phenotype fit should not automatically downgrade a well-established CNV. Test the strongest credible alternative explanation and name the new evidence that would change the conclusion. The pass condition is a bounded, reviewable claim, not the removal of uncertainty.
Evidence discipline: record the observation, version or time boundary that supports this chapter’s conclusion. Preserve the rejected alternative as well as the selected action. That makes a later review able to distinguish a changed world from an undocumented change of mind.
Changed-context transfer: Keep classification fixed but change the referral phenotype and revise only correlation. The method transfers only if the decision question and measurement meaning survive the change. If either changes, rebuild the comparison rather than copying the earlier threshold or weight.
Chapter 8 check · what should survive an audit
- Name the exact decision, unit, time boundary and responsible authority.
- Preserve the supporting evidence, controls and strongest credible alternative.
- Record the next action, review date, stop rule and reopening condition.
CHAPTER 9 OF 13 · Connect result to person
9. Use inheritance and family studies proportionately
Test parents or relatives when inheritance can clarify de novo status, segregation, reproductive risk or the structure of a complex finding, with consent for family implications. Begin with one person or pregnancy, one clinical indication, one validated array platform, one copy-number interval, one genome build, one classification decision, one confirmation plan and one review date. State the decision, specimen, record or person, the time boundary and the responsible owner before interpreting any result.
Worked example: A seemingly de novo deletion is confirmed after parentage and sensitive parental testing rather than inferred from family history alone. Put the starting evidence, applicable rule, intermediate decision, later observation and consequence on separate lines so that a fluent story cannot replace a reproducible pathway.
Evidence to retain: pedigree; parental samples; identity; method sensitivity; phenotype; de novo confirmation; consent; unexpected relationships Preserve provenance, dates, versions, identity checks, responsible people and intended use. If a necessary item is missing, label the gap and its effect instead of converting absence into reassurance.
Boundary and misconception: Inherited does not equal benign because penetrance may be reduced, and apparently de novo does not prove causality. Test the strongest credible alternative explanation and name the new evidence that would change the conclusion. The pass condition is a bounded, reviewable claim, not the removal of uncertainty.
Evidence discipline: record the observation, version or time boundary that supports this chapter’s conclusion. Preserve the rejected alternative as well as the selected action. That makes a later review able to distinguish a changed world from an undocumented change of mind.
Changed-context transfer: Add an unaffected carrier parent and update evidence without erasing penetrance uncertainty. The method transfers only if the decision question and measurement meaning survive the change. If either changes, rebuild the comparison rather than copying the earlier threshold or weight.
Chapter 9 check · what should survive an audit
- Name the exact decision, unit, time boundary and responsible authority.
- Preserve the supporting evidence, controls and strongest credible alternative.
- Record the next action, review date, stop rule and reopening condition.
Microarray interpretation gate
| Gate | Pass evidence | Remain open when |
|---|---|---|
| Analytical | Identity, QC and call boundaries are sound | Probe pattern or contamination is unresolved |
| Dosage | HI or TS evidence and content are reviewed | Only interval size is known |
| Correlation | Phenotype and inheritance are tested independently | Classification is made to fit referral |
| Confirmation | The next decision has adequate support | A high-stakes call remains technically uncertain |
Reanalysis can change interpretation while the measured interval stays the same.
CHAPTER 10 OF 13 · Connect result to person
10. Recognise array blind spots and choose the next method
State inability to reliably detect balanced rearrangements, most sequence variants, some repeats, small events and mosaicism below validated range, then match the residual question to another test. Begin with one person or pregnancy, one clinical indication, one validated array platform, one copy-number interval, one genome build, one classification decision, one confirmation plan and one review date. State the decision, specimen, record or person, the time boundary and the responsible owner before interpreting any result.
Worked example: A normal array with a strong monogenic phenotype leads to sequencing, while a suspected balanced translocation leads to cytogenetic or genome-structural testing. Put the starting evidence, applicable rule, intermediate decision, later observation and consequence on separate lines so that a fluent story cannot replace a reproducible pathway.
Evidence to retain: validated limitations; phenotype; residual hypothesis; sequencing; karyotype; FISH; long-read or genome option; decision owner Preserve provenance, dates, versions, identity checks, responsible people and intended use. If a necessary item is missing, label the gap and its effect instead of converting absence into reassurance.
Boundary and misconception: A normal array is not a normal genome, and adding more analysis to the same data cannot reveal signals the platform never measured. Test the strongest credible alternative explanation and name the new evidence that would change the conclusion. The pass condition is a bounded, reviewable claim, not the removal of uncertainty.
Evidence discipline: record the observation, version or time boundary that supports this chapter’s conclusion. Preserve the rejected alternative as well as the selected action. That makes a later review able to distinguish a changed world from an undocumented change of mind.
Changed-context transfer: Change the residual hypothesis from SNV to balanced rearrangement and select a different method. The method transfers only if the decision question and measurement meaning survive the change. If either changes, rebuild the comparison rather than copying the earlier threshold or weight.
Chapter 10 check · what should survive an audit
- Name the exact decision, unit, time boundary and responsible authority.
- Preserve the supporting evidence, controls and strongest credible alternative.
- Record the next action, review date, stop rule and reopening condition.
CHAPTER 11 OF 13 · Connect result to person
11. Confirm high-stakes and ambiguous findings
Use FISH, qPCR, MLPA, sequencing, karyotype, parental testing or another validated method according to event size, structure, mosaicism and decision. Begin with one person or pregnancy, one clinical indication, one validated array platform, one copy-number interval, one genome build, one classification decision, one confirmation plan and one review date. State the decision, specimen, record or person, the time boundary and the responsible owner before interpreting any result.
Worked example: A prenatal mosaic gain is confirmed on an independent preparation with a method that can test the suspected cell fraction before counselling. Put the starting evidence, applicable rule, intermediate decision, later observation and consequence on separate lines so that a fluent story cannot replace a reproducible pathway.
Evidence to retain: confirmation purpose; method; independent extraction; sensitivity; locus; structure; tissue; concordance; discrepancy plan Preserve provenance, dates, versions, identity checks, responsible people and intended use. If a necessary item is missing, label the gap and its effect instead of converting absence into reassurance.
Boundary and misconception: Repeating the same pipeline is not necessarily orthogonal, and confirmation method must be capable of testing the asserted structure. Test the strongest credible alternative explanation and name the new evidence that would change the conclusion. The pass condition is a bounded, reviewable claim, not the removal of uncertainty.
Evidence discipline: record the observation, version or time boundary that supports this chapter’s conclusion. Preserve the rejected alternative as well as the selected action. That makes a later review able to distinguish a changed world from an undocumented change of mind.
Changed-context transfer: Replace independent confirmation with a second software caller and identify the remaining uncertainty. The method transfers only if the decision question and measurement meaning survive the change. If either changes, rebuild the comparison rather than copying the earlier threshold or weight.
Chapter 11 check · what should survive an audit
- Name the exact decision, unit, time boundary and responsible authority.
- Preserve the supporting evidence, controls and strongest credible alternative.
- Record the next action, review date, stop rule and reopening condition.
CHAPTER 12 OF 13 · Report and transfer
12. Report uncertainty, secondary findings and limits
Present coordinates, build, copy state, classification, evidence, inheritance, phenotype correlation, confirmation and test limits in language that supports action without deterministic claims. Begin with one person or pregnancy, one clinical indication, one validated array platform, one copy-number interval, one genome build, one classification decision, one confirmation plan and one review date. State the decision, specimen, record or person, the time boundary and the responsible owner before interpreting any result.
Worked example: A VUS is described with the evidence gap and review trigger, not converted into a diagnosis or ignored as meaningless. Put the starting evidence, applicable rule, intermediate decision, later observation and consequence on separate lines so that a fluent story cannot replace a reproducible pathway.
Evidence to retain: HGVS or ISCN notation; build; interval; genes; classification; evidence; correlation; limitations; recommendations; sign-off Preserve provenance, dates, versions, identity checks, responsible people and intended use. If a necessary item is missing, label the gap and its effect instead of converting absence into reassurance.
Boundary and misconception: A report can be technically complete yet misleading if uncertain findings are framed as answers or important blind spots are buried. Test the strongest credible alternative explanation and name the new evidence that would change the conclusion. The pass condition is a bounded, reviewable claim, not the removal of uncertainty.
Evidence discipline: record the observation, version or time boundary that supports this chapter’s conclusion. Preserve the rejected alternative as well as the selected action. That makes a later review able to distinguish a changed world from an undocumented change of mind.
Changed-context transfer: Keep the classification but rewrite the report for a family making a reproductive decision. The method transfers only if the decision question and measurement meaning survive the change. If either changes, rebuild the comparison rather than copying the earlier threshold or weight.
Chapter 12 check · what should survive an audit
- Name the exact decision, unit, time boundary and responsible authority.
- Preserve the supporting evidence, controls and strongest credible alternative.
- Record the next action, review date, stop rule and reopening condition.
CHAPTER 13 OF 13 · Report and transfer
13. Preserve data for reanalysis and transfer
Store raw files, QC, platform, build, interpretation evidence and consent; set triggers for new dosage curation, phenotype change or improved methods. Begin with one person or pregnancy, one clinical indication, one validated array platform, one copy-number interval, one genome build, one classification decision, one confirmation plan and one review date. State the decision, specimen, record or person, the time boundary and the responsible owner before interpreting any result.
Worked example: A CNV VUS is reclassified after a gene gains sufficient dosage-sensitivity evidence, with the original measurement remaining traceable. Put the starting evidence, applicable rule, intermediate decision, later observation and consequence on separate lines so that a fluent story cannot replace a reproducible pathway.
Evidence to retain: raw data; array design; QC; software; build; evidence snapshot; databases; report versions; recontact policy; review date Preserve provenance, dates, versions, identity checks, responsible people and intended use. If a necessary item is missing, label the gap and its effect instead of converting absence into reassurance.
Boundary and misconception: Reanalysis changes interpretation, not historical facts, and build or platform changes require explicit bridging. Test the strongest credible alternative explanation and name the new evidence that would change the conclusion. The pass condition is a bounded, reviewable claim, not the removal of uncertainty.
Evidence discipline: record the observation, version or time boundary that supports this chapter’s conclusion. Preserve the rejected alternative as well as the selected action. That makes a later review able to distinguish a changed world from an undocumented change of mind.
Changed-context transfer: Move the interval to a new genome build and verify every boundary and annotation before comparison. The method transfers only if the decision question and measurement meaning survive the change. If either changes, rebuild the comparison rather than copying the earlier threshold or weight.
Chapter 13 check · what should survive an audit
- Name the exact decision, unit, time boundary and responsible authority.
- Preserve the supporting evidence, controls and strongest credible alternative.
- Record the next action, review date, stop rule and reopening condition.
Verification lab: work the decision twice
Interpret a fictional 650 kb deletion by separating analytical call, dosage evidence, classification and patient-level correlation.
A duplication is inherited from an apparently unaffected parent. Build a penetrance-aware family study without declaring benignity by inheritance alone.
A normal microarray follows a strong phenotype. Match residual hypotheses to sequencing, cytogenetic or structural methods.
Final review: distinguish copy state, interval, dosage sensitivity, classification, correlation and confirmation. If any pair is collapsed, the pathway fails.
Limits and misconceptions
Array resolution is uneven
Effective detection depends on probe design, quality, event type and reporting thresholds.
Dosage direction matters
Loss and gain require separate haploinsufficiency and triplosensitivity evidence.
Pathogenic is not total explanation
A real CNV may explain some, all or none of a particular person’s features.
Normal is method-bounded
Balanced changes, sequence variants, repeats and low-level mosaicism may remain unseen.
Questions readers ask
What does chromosomal microarray detect?
- It detects validated copy-number losses and gains and, on SNP arrays, certain allele-pattern changes across the genome.
Is a larger CNV always more serious?
- No. Gene content, dosage sensitivity, inheritance, penetrance and phenotype matter more than size alone.
Can a normal array rule out a genetic condition?
- No. Many sequence, repeat, balanced and low-level mosaic variants are outside its scope.
Why test parents?
- Inheritance can clarify de novo status, segregation, reproductive implications and sometimes event structure.
Must every CNV be confirmed?
- Confirmation depends on platform validation, event confidence and the decision; high-stakes or ambiguous findings often need independent support.
Can a VUS be used as a diagnosis?
- It should not be treated as definitive; the report should preserve the evidence gap and review plan.
Practise, verify and continue
Build a CNV worksheet with indication, sample, platform, build, QC, interval uncertainty, copy state, probes, dosage evidence, evidence score, population and case data, inheritance, phenotype correlation, blind spots, confirmation and reanalysis trigger.
Apply the method to a structural variant from genome sequencing. Preserve dosage evidence and phenotype correlation, then add breakpoint, orientation and sequence-level validation that microarray cannot provide.
Practical evidence check
Before implementation, write a one-page evidence map with the decision, unit of analysis, version or time boundary, inputs, outcomes, uncertainty, authority and stop rule. Attach one worked case that passes and one that fails. A reviewer should be able to reproduce the reasoning without relying on the memory of the person who built the system.
Run one changed-context rehearsal before approval. Alter a mechanism rather than a cosmetic detail, then identify which calculation, safeguard or governance gate must change. Record any conclusion that does not transfer. This short rehearsal is often the fastest way to expose a hidden assumption before it becomes an operational dependency.
Related eduKate reading
Genome sequencing, panels and coverage gaps
Variant classification and evidence weight
Gene–disease validity and evidence strength
Prenatal diagnostic testing and sampling limits
Long-read sequencing and structural-variant confirmation
Authoritative sources and scope
ACMG and ClinGen constitutional CNV interpretation standards
Full ACMG and ClinGen CNV scoring framework
ACMG chromosomal microarray technical standard, 2021 revision
ClinGen dosage-sensitivity curation
These sources support the laboratory, interpretation and counselling distinctions described here. They do not replace patient-specific evaluation, informed consent, validated laboratory methods, local regulation or advice from a qualified genetics professional. All worked cases are fictional. A result must be interpreted with phenotype, family history, assay limits and current evidence; it must not be used alone to diagnose or change care.
