THE GOLD STANDARD OF GENETICS · SPECIALIST STUDY
Choose the smallest test that can answer the clinical question, then state what its negative result leaves open
Gene panels, exome sequencing and genome sequencing are not a ladder from weak to perfect. Each is a measurement system with a target, capture or library method, coverage pattern, variant-calling pipeline, interpretation boundary and reporting policy. A focused panel can offer deep, well-validated coverage of a coherent phenotype. Exome sequencing broadens coding-gene analysis but may miss poorly captured exons, repeat expansions, balanced rearrangements, mitochondrial variation or deep non-coding causes. Genome sequencing reduces capture bias and can improve structural and non-coding access, yet still has coverage gaps, hard-to-map regions and an interpretation bottleneck. Diagnostic yield therefore belongs to a defined cohort, phenotype, family structure, laboratory, pipeline and date. A strong pathway matches the test to the hypothesis, checks whether the relevant region and variant class were actually detectable, confirms reportable findings, separates no finding from no genetic cause, and preserves data, consent and responsibility for purposeful reanalysis.
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 phenotype and choice · chapters 1, 2, 3
Measure what the assay can see · chapters 4, 5
Interpret yield and findings · chapters 6, 7, 8
Report and confirm · chapters 9, 10, 11
Reanalyse and govern · chapters 12, 13
The whole problem at a glance
| Test | Typical strength | Important limits |
|---|---|---|
| Focused gene panel | Deep, curated coverage for a coherent differential | Genes and variant classes outside scope |
| Exome sequencing | Broad coding-region analysis and future reanalysis | Capture gaps, non-coding and difficult structural variation |
| Genome sequencing | More even genome-wide access and structural breakpoints | Hard regions, incomplete interpretation and cost |
| Targeted orthogonal test | Repeat, methylation, CNV or biochemical question | Answers only the defined mechanism |
The clinical question, validation and actual coverage matter more than the marketing breadth of the test name.
CHAPTER 1 OF 13 · Frame the phenotype and choice
1. Start with phenotype, pedigree and the decision to be made
Record onset, systems involved, negative findings, family structure, prior tests and whether the goal is diagnosis, management, recurrence counselling or another decision. Work with one person or family, one phenotype and pedigree, one test choice, one laboratory pipeline and version, one reportable scope, one diagnostic decision and one reanalysis plan. 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, seizures and normal metabolic screening receives structured phenotype coding before test selection. Put the starting evidence, rule, decision, later observation and consequence on separate lines. This prevents a persuasive narrative from replacing a reproducible pathway.
Evidence to retain: clinical notes; phenotype terms; onset; examinations; pedigree; ancestry where relevant; prior reports; current management question Preserve provenance, dates, versions, identity checks, responsible people and intended use. If evidence is missing, label the gap instead of turning absence into reassurance.
Boundary and misconception: Sequencing cannot rescue an undefined referral, and an incomplete phenotype can make a technically broad test clinically narrow. 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 progressive hearing loss after the first report and rebuild the differential and reanalysis query. 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 and the strongest credible alternative.
- Record the next action, review date, stop rule and reopening condition.
CHAPTER 2 OF 13 · Frame the phenotype and choice
2. Choose panel, exome, genome or another method deliberately
Compare the candidate mechanisms, time, cost, coverage, variant classes, sample availability and future reinterpretation rather than ranking tests by breadth alone. Work with one person or family, one phenotype and pedigree, one test choice, one laboratory pipeline and version, one reportable scope, one diagnostic decision and one reanalysis plan. State the decision, specimen, record or person, the time boundary and the responsible owner before interpreting any result.
Worked example: A recognisable repeat-expansion disorder receives targeted testing instead of exome sequencing that is not validated for the repeat. Put the starting evidence, rule, decision, later observation and consequence on separate lines. This prevents a persuasive narrative from replacing a reproducible pathway.
Evidence to retain: differential diagnosis; genes; variant mechanisms; assay specifications; turnaround; family samples; alternatives; referral rationale Preserve provenance, dates, versions, identity checks, responsible people and intended use. If evidence is missing, label the gap instead of turning absence into reassurance.
Boundary and misconception: Genome is not automatically first or final, and a well-designed panel can outperform broader sequencing for a defined question. 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 the coherent phenotype with multiple unrelated systems and reconsider exome or genome sequencing. 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 and the strongest credible alternative.
- Record the next action, review date, stop rule and reopening condition.
CHAPTER 3 OF 13 · Frame the phenotype and choice
3. Define the reportable scope and consent
Explain primary findings, secondary findings, carrier findings, uncertain variants, data storage, family implications and the possibility of recontact. Work with one person or family, one phenotype and pedigree, one test choice, one laboratory pipeline and version, one reportable scope, one diagnostic decision and one reanalysis plan. State the decision, specimen, record or person, the time boundary and the responsible owner before interpreting any result.
Worked example: Parents choose whether to receive a defined list of secondary findings separately from the child’s diagnostic question. Put the starting evidence, rule, decision, later observation and consequence on separate lines. This prevents a persuasive narrative from replacing a reproducible pathway.
Evidence to retain: consent version; decision maker; capacity; primary scope; secondary findings policy; VUS policy; data sharing; withdrawal; contact preferences Preserve provenance, dates, versions, identity checks, responsible people and intended use. If evidence is missing, label the gap instead of turning absence into reassurance.
Boundary and misconception: Consent to sequencing is not consent to every future use or every category the platform could technically reveal. 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 from an adult with capacity to a child and rebuild best-interest and future-autonomy safeguards. 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 and the strongest credible alternative.
- Record the next action, review date, stop rule and reopening condition.
CHAPTER 4 OF 13 · Measure what the assay can see
4. Verify sample identity, quality and family relationships
Track collection, identity, contamination, DNA quality and whether stated biological relationships are consistent with analysis while planning sensitive findings. Work with one person or family, one phenotype and pedigree, one test choice, one laboratory pipeline and version, one reportable scope, one diagnostic decision and one reanalysis plan. State the decision, specimen, record or person, the time boundary and the responsible owner before interpreting any result.
Worked example: A trio analysis detects a relationship inconsistency; the laboratory follows its consented policy instead of disclosing it casually in the diagnostic report. Put the starting evidence, rule, decision, later observation and consequence on separate lines. This prevents a persuasive narrative from replacing a reproducible pathway.
Evidence to retain: identifiers; collection; chain of custody; DNA metrics; contamination; sex-chromosome checks; relationship checks; disclosure policy; deviations Preserve provenance, dates, versions, identity checks, responsible people and intended use. If evidence is missing, label the gap instead of turning absence into reassurance.
Boundary and misconception: Trio sequencing improves interpretation only if identities and relationships are reliable and unexpected information is handled responsibly. 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 one parent from the trio and state which de novo and phase conclusions weaken. 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 and the strongest credible alternative.
- Record the next action, review date, stop rule and reopening condition.
CHAPTER 5 OF 13 · Measure what the assay can see
5. Read coverage at the region that matters
Inspect depth, breadth, mapping quality and laboratory thresholds for candidate exons, splice regions, mitochondrial sequence or other targets rather than relying on mean coverage. Work with one person or family, one phenotype and pedigree, one test choice, one laboratory pipeline and version, one reportable scope, one diagnostic decision and one reanalysis plan. State the decision, specimen, record or person, the time boundary and the responsible owner before interpreting any result.
Worked example: The average exome depth is high, but one exon in the leading gene has no reliable reads and requires fill-in testing. Put the starting evidence, rule, decision, later observation and consequence on separate lines. This prevents a persuasive narrative from replacing a reproducible pathway.
Evidence to retain: per-base coverage; low-coverage list; mapping quality; GC content; paralogous regions; fill-in method; acceptance criteria; report statement Preserve provenance, dates, versions, identity checks, responsible people and intended use. If evidence is missing, label the gap instead of turning absence into reassurance.
Boundary and misconception: A high mean can coexist with a clinically decisive blind spot. Covered is not the same as callable. 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 mean depth fixed but move the candidate variant into a pseudogene-rich exon. 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 and the strongest credible alternative.
- Record the next action, review date, stop rule and reopening condition.
Worked negative-result review
| Observation | Residual explanation | Next evidence | Do not claim |
|---|---|---|---|
| Panel negative | Gene absent or region below threshold | Scope and exon-level coverage review | Condition is not genetic |
| Exome negative | Non-coding, structural, repeat or mosaic cause | Phenotype-led method review | All genes were normal |
| Genome negative | Hard region or unknown gene-disease relationship | Coverage, pipeline and reanalysis plan | Whole genome means complete |
| VUS reported | Evidence is insufficient for causal use | Segregation and future classification review | Uncertain means almost positive |
A negative report is bounded by the assay and knowledge available on its report date.
CHAPTER 6 OF 13 · Interpret yield and findings
6. Match validation to variant classes
Confirm which single-nucleotide, indel, copy-number, structural, repeat, mosaic, mitochondrial and non-coding variants the laboratory has validated and at what limits. Work with one person or family, one phenotype and pedigree, one test choice, one laboratory pipeline and version, one reportable scope, one diagnostic decision and one reanalysis plan. State the decision, specimen, record or person, the time boundary and the responsible owner before interpreting any result.
Worked example: A genome pipeline reliably calls large deletions but not a small repeat expansion, so a negative genome result leaves that mechanism open. Put the starting evidence, rule, decision, later observation and consequence on separate lines. This prevents a persuasive narrative from replacing a reproducible pathway.
Evidence to retain: validation samples; sensitivity; precision; size range; allele fraction; genomic contexts; orthogonal methods; software version Preserve provenance, dates, versions, identity checks, responsible people and intended use. If evidence is missing, label the gap instead of turning absence into reassurance.
Boundary and misconception: One dataset does not imply one validated test for every variant class that might be visible in raw reads. 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: Lower the mosaic fraction below the validated limit and revise the negative claim. 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 and the strongest credible alternative.
- Record the next action, review date, stop rule and reopening condition.
CHAPTER 7 OF 13 · Interpret yield and findings
7. Use gene-disease validity before variant classification
Ask whether the gene is convincingly related to the phenotype and inheritance pattern before investing causal meaning in a rare variant. Work with one person or family, one phenotype and pedigree, one test choice, one laboratory pipeline and version, one reportable scope, one diagnostic decision and one reanalysis plan. State the decision, specimen, record or person, the time boundary and the responsible owner before interpreting any result.
Worked example: A predicted damaging variant in a disputed gene remains research-level even though several algorithms call it deleterious. Put the starting evidence, rule, decision, later observation and consequence on separate lines. This prevents a persuasive narrative from replacing a reproducible pathway.
Evidence to retain: gene-disease curation; inheritance; disease mechanism; case evidence; segregation; functional evidence; contradictory studies; curation date Preserve provenance, dates, versions, identity checks, responsible people and intended use. If evidence is missing, label the gap instead of turning absence into reassurance.
Boundary and misconception: A strong-looking variant cannot create a valid gene-disease relationship by itself. 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 variant but change the phenotype to one outside the established disease spectrum. 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 and the strongest credible alternative.
- Record the next action, review date, stop rule and reopening condition.
CHAPTER 8 OF 13 · Interpret yield and findings
8. Classify variants with independent evidence
Combine population frequency, computational, functional, segregation, de novo, allelic and case evidence without double counting correlated support. Work with one person or family, one phenotype and pedigree, one test choice, one laboratory pipeline and version, one reportable scope, one diagnostic decision and one reanalysis plan. State the decision, specimen, record or person, the time boundary and the responsible owner before interpreting any result.
Worked example: A de novo missense variant receives weight only after parentage, phenotype fit and gene mechanism are checked. Put the starting evidence, rule, decision, later observation and consequence on separate lines. This prevents a persuasive narrative from replacing a reproducible pathway.
Evidence to retain: reference build; transcript; nomenclature; databases and dates; evidence codes; conflicts; analyst; approval; classification version Preserve provenance, dates, versions, identity checks, responsible people and intended use. If evidence is missing, label the gap instead of turning absence into reassurance.
Boundary and misconception: Rare does not mean pathogenic, and multiple tools trained on similar data are not independent witnesses. 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: Introduce a high-quality benign functional assay and update the classification transparently. 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 and the strongest credible alternative.
- Record the next action, review date, stop rule and reopening condition.
CHAPTER 9 OF 13 · Report and confirm
9. Interpret diagnostic yield in its denominator
Report yield for a defined cohort, test, phenotype, family design, inclusion criteria and follow-up period; separate definitive diagnoses, probable findings and VUS. Work with one person or family, one phenotype and pedigree, one test choice, one laboratory pipeline and version, one reportable scope, one diagnostic decision and one reanalysis plan. State the decision, specimen, record or person, the time boundary and the responsible owner before interpreting any result.
Worked example: A laboratory’s 35 percent yield in deeply phenotyped trios is not promised to a person with a different indication and singleton testing. Put the starting evidence, rule, decision, later observation and consequence on separate lines. This prevents a persuasive narrative from replacing a reproducible pathway.
Evidence to retain: cohort; referral criteria; phenotype; trio rate; method; result categories; denominator; follow-up; missing outcomes; date Preserve provenance, dates, versions, identity checks, responsible people and intended use. If evidence is missing, label the gap instead of turning absence into reassurance.
Boundary and misconception: Diagnostic yield is a group statistic, not a personal probability, and it changes as knowledge and pipelines change. 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 from trio to singleton testing and explain why published yield may not transfer. 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 and the strongest credible alternative.
- Record the next action, review date, stop rule and reopening condition.
Sequencing interpretation gate
| Gate | Pass evidence | Remain open when |
|---|---|---|
| Question | Phenotype, pedigree and decision are explicit | Testing is an undirected search |
| Detection | Relevant regions and variant classes are validated | Coverage summary hides critical gaps |
| Interpretation | Gene validity and variant evidence are current | A weak association drives diagnosis |
| Follow-up | Confirmation, communication and reanalysis are owned | Data are archived without a plan |
Reanalysis and retesting are different actions and need separate triggers.
CHAPTER 10 OF 13 · Report and confirm
10. Confirm findings and connect them to the person
Use orthogonal confirmation when required, verify sample identity and phase, review phenotype and avoid changing care until the result meets clinical standards. Work with one person or family, one phenotype and pedigree, one test choice, one laboratory pipeline and version, one reportable scope, one diagnostic decision and one reanalysis plan. State the decision, specimen, record or person, the time boundary and the responsible owner before interpreting any result.
Worked example: A likely pathogenic deletion is confirmed by an independent method and shown to be de novo before recurrence counselling is finalised. Put the starting evidence, rule, decision, later observation and consequence on separate lines. This prevents a persuasive narrative from replacing a reproducible pathway.
Evidence to retain: raw evidence; confirmation; second sample; phase; segregation; phenotype review; multidisciplinary sign-off; updated report Preserve provenance, dates, versions, identity checks, responsible people and intended use. If evidence is missing, label the gap instead of turning absence into reassurance.
Boundary and misconception: A plausible sequence finding is not a diagnosis until technical, genetic and clinical coherence are established. 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 confirmation and state which decisions must remain provisional. 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 and the strongest credible alternative.
- Record the next action, review date, stop rule and reopening condition.
CHAPTER 11 OF 13 · Report and confirm
11. Write a bounded negative or uncertain report
State tested scope, important coverage and variant-class limits, unresolved candidates, residual hypotheses and the circumstances that should trigger more testing. Work with one person or family, one phenotype and pedigree, one test choice, one laboratory pipeline and version, one reportable scope, one diagnostic decision and one reanalysis plan. State the decision, specimen, record or person, the time boundary and the responsible owner before interpreting any result.
Worked example: The report names a low-coverage exon in the leading gene and recommends targeted analysis rather than saying no pathogenic variant was found without qualification. Put the starting evidence, rule, decision, later observation and consequence on separate lines. This prevents a persuasive narrative from replacing a reproducible pathway.
Evidence to retain: assay and pipeline; reference build; reportable regions; low coverage; excluded classes; VUS policy; residual differential; next actions Preserve provenance, dates, versions, identity checks, responsible people and intended use. If evidence is missing, label the gap instead of turning absence into reassurance.
Boundary and misconception: No reportable variant is not proof of no genetic condition, and a VUS should not drive predictive testing or irreversible care by itself. 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: A new clinical feature strongly supports a mechanism the assay did not test; revise the next action without rewriting the old result. 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 and the strongest credible alternative.
- Record the next action, review date, stop rule and reopening condition.
CHAPTER 12 OF 13 · Reanalyse and govern
12. Distinguish reanalysis from retesting
Reanalysis revisits existing data with updated phenotype, databases or pipelines; retesting generates new data when technology, chemistry or specimen choice can answer a previously inaccessible question. Work with one person or family, one phenotype and pedigree, one test choice, one laboratory pipeline and version, one reportable scope, one diagnostic decision and one reanalysis plan. State the decision, specimen, record or person, the time boundary and the responsible owner before interpreting any result.
Worked example: An old exome is reanalysed for new gene-disease knowledge, but a suspected repeat expansion requires a new targeted assay rather than another annotation pass. Put the starting evidence, rule, decision, later observation and consequence on separate lines. This prevents a persuasive narrative from replacing a reproducible pathway.
Evidence to retain: original files; data quality; pipeline version; phenotype update; knowledge changes; new method; time elapsed; consent; owner; trigger Preserve provenance, dates, versions, identity checks, responsible people and intended use. If evidence is missing, label the gap instead of turning absence into reassurance.
Boundary and misconception: Reanalysis cannot recover sequence that was never captured well, and retesting is not always necessary when interpretation alone changed. 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 knowledge fixed but introduce a new long-read method and decide whether new data are justified. 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 and the strongest credible alternative.
- Record the next action, review date, stop rule and reopening condition.
CHAPTER 13 OF 13 · Reanalyse and govern
13. Govern storage, recontact and report corrections
Define how long data and samples are retained, who can request reanalysis, how classifications and reports are corrected and how reasonable recontact is attempted. Work with one person or family, one phenotype and pedigree, one test choice, one laboratory pipeline and version, one reportable scope, one diagnostic decision and one reanalysis plan. State the decision, specimen, record or person, the time boundary and the responsible owner before interpreting any result.
Worked example: A variant is downgraded after strong new evidence; the laboratory issues a dated amended report and follows the recorded clinical contact route. Put the starting evidence, rule, decision, later observation and consequence on separate lines. This prevents a persuasive narrative from replacing a reproducible pathway.
Evidence to retain: retention policy; file formats; consent; request route; review interval; classification alerts; contact details; amendments; acknowledgements Preserve provenance, dates, versions, identity checks, responsible people and intended use. If evidence is missing, label the gap instead of turning absence into reassurance.
Boundary and misconception: A promise to reinterpret forever is unrealistic, but silence about responsibility leaves clinically important change without an owner. 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: The patient changes health system and contact details; test whether the record remains usable and responsibility is clear. 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 and the strongest credible alternative.
- Record the next action, review date, stop rule and reopening condition.
Verification lab: work the decision twice
Compare a 50-gene panel, exome and genome for a coherent cardiomyopathy phenotype. Mark which regions and variant classes each can actually assess.
A negative exome has 98 percent average coverage but zero reliable reads in the leading gene’s second exon. Rewrite the report and next action.
A laboratory advertises 40 percent diagnostic yield. Reconstruct the cohort, denominator, trio rate and result categories before applying it to one family.
Final review: distinguish data generation, variant calling, gene validity, variant classification, clinical diagnosis and reanalysis. If one word such as sequencing represents all six, the pathway is incomplete.
Limits and misconceptions
Genome does not mean complete
Short-read genome sequencing still has hard-to-map regions, coverage gaps and incompletely interpretable non-coding variation.
Coverage is local
Mean depth can hide a decisive exon or locus that was not reliably callable.
Yield belongs to a cohort
It is not a guaranteed personal probability and depends on phenotype, family design, pipeline and date.
Reanalysis cannot create missing data
A new method or specimen may require retesting rather than reinterpretation.
Questions readers ask
Is a gene panel less advanced than genome sequencing?
- No. A focused panel may provide deeper, well-validated coverage for a coherent question, while genome sequencing has broader but still incomplete reach.
What does coverage mean?
- It describes how reliably sequence was observed and callable at specific regions; an average alone is not enough.
Can exome sequencing detect structural variants?
- Some laboratories validate selected copy-number or structural calls, but capability varies and must be stated.
What is diagnostic yield?
- The proportion of a defined tested cohort receiving a result category under a stated method and time period.
Why might a negative result change later?
- New gene-disease relationships, variant evidence, phenotype information and pipelines can make existing data more informative.
When is retesting better than reanalysis?
- When the prior data lack relevant regions, variant classes, quality or specimen information that a newer method can provide.
Practise, verify and continue
Build a sequencing evidence map with phenotype, pedigree, decision, test choice, reportable scope, sample, coverage, validated variant classes, gene validity, variant evidence, confirmation, bounded negative statement and reanalysis or retesting trigger.
Apply the method to tumour sequencing. Preserve coverage, variant-class validation and versioning, then rebuild the unit, because tumour purity, clonality and somatic-germline separation change the interpretation.
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
Long-read sequencing, repeat expansions and structural variants
Variant classification, conflicting assertions and reanalysis
Gene–disease validity, disputed genes and report correction
Epigenetic testing and methylation signatures
Authoritative sources and scope
GeneReviews: Genetic Testing, Current Approaches
ACMG evidence-based guideline for exome and genome sequencing in paediatric congenital anomalies and developmental disorders
ACMG points to consider for reevaluation and reanalysis of genomic test results
ACMG technical standard for chromosomal microarray analysis
These sources support the testing distinctions and evidence practices described here. They do not replace patient-specific genetic counselling, reproductive counselling, diagnostic evaluation, laboratory validation, consent, local regulation or a clinician’s interpretation. All worked cases are fictional. An educational example, research result or database entry must not be used alone to diagnose, treat or change care.