eduKateSG · Genetics · The Gold Standard of Education
Reverse phenotyping that preserves people, variants and uncertainty
Choose the explanation you need, or follow all twelve chapters in order.
Full chapter index · Independent practice · Genetics subject guide
A genotype-first route can reveal features that phenotype-first recruitment misses, but only when variant eligibility, recontact, measurement and interpretation remain traceable. This guide turns reverse phenotyping into a complete evidence chain rather than a dramatic story about one variant.
Continue from The Gold Standard of Genetics within The Gold Standard of Education: learning that transfers into life and the How X Works learning library.
Full chapter index
See the mechanism · Chapters 1–2
Build the worked model · Chapters 3–5
Test what changes · Chapters 6–8
Check evidence and limits · Chapters 9–10
Practise and continue · Chapters 11–12
Traditional clinical genetics often begins with features and asks which genetic change might explain them. A genotype-first study identifies people carrying a specified variant or variant class, then measures their characteristics systematically. State whether the study estimates penetrance, describes a spectrum, compares variant classes or tests a mechanism. Name the source population, age range, ancestry and observation window. A biobank carrier group is not automatically representative of every carrier.
Evidence check: preserve the observation before the interpretation. Record target claim and source population. Identify its source, unit, time window and uncertainty. Then state the inference, one credible alternative, and the additional observation that would distinguish them. A fluent explanation is not yet evidence until another reader can inspect the same trail.
Changed-context transfer: repeat this chapter’s decision after changing one consequential condition while holding the others fixed. Calculate or predict what should change, name the threshold or mechanism, and compare it with the original case. State the range tested. If the conclusion reverses, explain why; if it survives, say what remained invariant.
Limit and misconception check: do not turn a relative comparison into an absolute guarantee, treat an unmeasured condition as absent, or let one convenient indicator stand for the full mechanism. Name the boundary where this chapter’s method stops, the tempting overclaim it prevents, and the safer sentence the available evidence actually supports.
Practical decision record: write one sentence for what is known, one for what remains uncertain, and one for the next safe action. Include the evidence owner, the latest useful decision time, the review date, and the condition that would stop or reverse the action. This turns a technical result into a choice that can be checked later.
CHAPTER 2 OF 12 · See the mechanism
2. Separate reverse phenotyping, recall and record review
Reverse phenotyping is an interpretive direction: genotype information prompts a deliberate search for features. Recall by genotype is a prospective recruitment design. Retrospective record review is different and limited to what care systems captured. A record may miss subtle sensory, developmental or biochemical features. Prospective recall can measure them consistently, yet participation can depend on health, access and willingness. Do not treat an unrecorded feature as absent.
Evidence check: preserve the observation before the interpretation. Record reverse phenotyping, recall and record review separately. Identify its source, unit, time window and uncertainty. Then state the inference, one credible alternative, and the additional observation that would distinguish them. A fluent explanation is not yet evidence until another reader can inspect the same trail.
Changed-context transfer: repeat this chapter’s decision after changing one consequential condition while holding the others fixed. Calculate or predict what should change, name the threshold or mechanism, and compare it with the original case. State the range tested. If the conclusion reverses, explain why; if it survives, say what remained invariant.
Limit and misconception check: do not turn a relative comparison into an absolute guarantee, treat an unmeasured condition as absent, or let one convenient indicator stand for the full mechanism. Name the boundary where this chapter’s method stops, the tempting overclaim it prevents, and the safer sentence the available evidence actually supports.
Practical decision record: write one sentence for what is known, one for what remains uncertain, and one for the next safe action. Include the evidence owner, the latest useful decision time, the review date, and the condition that would stop or reverse the action. This turns a technical result into a choice that can be checked later.
Freeze genome build, transcript, variant representation, annotation release, classification date, frequency rule and quality filters. Decide whether eligibility uses one named variant, predicted loss-of-function variants, a domain or a curated set. Preserve exclusions. Create a flow from sequenced population to eligible, contactable, invited, enrolled and assessed carriers. Report available comparisons with nonparticipants. Separate independent observations from relatives and define how relatedness enters analysis.
Evidence check: preserve the observation before the interpretation. Record variant manifest and participant flow. Identify its source, unit, time window and uncertainty. Then state the inference, one credible alternative, and the additional observation that would distinguish them. A fluent explanation is not yet evidence until another reader can inspect the same trail.
Changed-context transfer: repeat this chapter’s decision after changing one consequential condition while holding the others fixed. Calculate or predict what should change, name the threshold or mechanism, and compare it with the original case. State the range tested. If the conclusion reverses, explain why; if it survives, say what remained invariant.
Limit and misconception check: do not turn a relative comparison into an absolute guarantee, treat an unmeasured condition as absent, or let one convenient indicator stand for the full mechanism. Name the boundary where this chapter’s method stops, the tempting overclaim it prevents, and the safer sentence the available evidence actually supports.
Practical decision record: write one sentence for what is known, one for what remains uncertain, and one for the next safe action. Include the evidence owner, the latest useful decision time, the review date, and the condition that would stop or reverse the action. This turns a technical result into a choice that can be checked later.
A fictional GENE-Q cohort contains protein-truncating, critical-domain missense, other missense and synonymous comparison variants. Twenty-four, thirty, forty and sixty participants complete a standardised assessment. If a neurological feature appears in 15 of 24 truncating carriers and 12 of 30 critical-domain carriers, the gradient is suggestive but imprecise. Age, family structure and ascertainment may explain part of it. Report denominators and missing assessments for every feature.
| Variant class | Participants | Feature present | Interpretation |
|---|---|---|---|
| Protein-truncating | 24 | 15 | Suggestive high proportion |
| Critical-domain missense | 30 | 12 | Intermediate, uncertain |
| Other missense | 40 | 7 | Lower, heterogeneous |
| Synonymous comparison | 60 | 4 | Process-matched comparison |
Evidence check: preserve the observation before the interpretation. Record every count, denominator and interval. Identify its source, unit, time window and uncertainty. Then state the inference, one credible alternative, and the additional observation that would distinguish them. A fluent explanation is not yet evidence until another reader can inspect the same trail.
Changed-context transfer: repeat this chapter’s decision after changing one consequential condition while holding the others fixed. Calculate or predict what should change, name the threshold or mechanism, and compare it with the original case. State the range tested. If the conclusion reverses, explain why; if it survives, say what remained invariant.
Limit and misconception check: do not turn a relative comparison into an absolute guarantee, treat an unmeasured condition as absent, or let one convenient indicator stand for the full mechanism. Name the boundary where this chapter’s method stops, the tempting overclaim it prevents, and the safer sentence the available evidence actually supports.
Practical decision record: write one sentence for what is known, one for what remains uncertain, and one for the next safe action. Include the evidence owner, the latest useful decision time, the review date, and the condition that would stop or reverse the action. This turns a technical result into a choice that can be checked later.
CHAPTER 5 OF 12 · Build the worked model
5. Define an allelic series without forcing one
An allelic series is an ordered relationship between variant classes and effect. The order may follow dosage, residual function or domain disruption. Test the proposed order against an unconstrained model. Inspect influential variants because one founder variant can create a false class effect. Functional evidence may help classify variants, but assay activity and human phenotype are different outcomes. If class effects cross by phenotype, report a multidimensional pattern rather than one severity ladder.
Evidence check: preserve the observation before the interpretation. Record mechanistic ordering and influential variants. Identify its source, unit, time window and uncertainty. Then state the inference, one credible alternative, and the additional observation that would distinguish them. A fluent explanation is not yet evidence until another reader can inspect the same trail.
Changed-context transfer: repeat this chapter’s decision after changing one consequential condition while holding the others fixed. Calculate or predict what should change, name the threshold or mechanism, and compare it with the original case. State the range tested. If the conclusion reverses, explain why; if it survives, say what remained invariant.
Limit and misconception check: do not turn a relative comparison into an absolute guarantee, treat an unmeasured condition as absent, or let one convenient indicator stand for the full mechanism. Name the boundary where this chapter’s method stops, the tempting overclaim it prevents, and the safer sentence the available evidence actually supports.
Practical decision record: write one sentence for what is known, one for what remains uncertain, and one for the next safe action. Include the evidence owner, the latest useful decision time, the review date, and the condition that would stop or reverse the action. This turns a technical result into a choice that can be checked later.
Penetrance is the proportion of carriers meeting a stated phenotype by a specified age under a defined observation process. It is not independent of population and definition. Use time-to-event methods when age matters. Variable expressivity concerns onset, severity, organs and course among affected carriers. A carrier with no current feature may develop it later, have a subclinical form or remain unaffected. Record measurement sensitivity and follow-up.
Evidence check: preserve the observation before the interpretation. Record age, penetrance and expression definitions. Identify its source, unit, time window and uncertainty. Then state the inference, one credible alternative, and the additional observation that would distinguish them. A fluent explanation is not yet evidence until another reader can inspect the same trail.
Changed-context transfer: repeat this chapter’s decision after changing one consequential condition while holding the others fixed. Calculate or predict what should change, name the threshold or mechanism, and compare it with the original case. State the range tested. If the conclusion reverses, explain why; if it survives, say what remained invariant.
Limit and misconception check: do not turn a relative comparison into an absolute guarantee, treat an unmeasured condition as absent, or let one convenient indicator stand for the full mechanism. Name the boundary where this chapter’s method stops, the tempting overclaim it prevents, and the safer sentence the available evidence actually supports.
Practical decision record: write one sentence for what is known, one for what remains uncertain, and one for the next safe action. Include the evidence owner, the latest useful decision time, the review date, and the condition that would stop or reverse the action. This turns a technical result into a choice that can be checked later.
Select assessments from a prespecified mechanism and participant-benefit review. Use validated instruments, calibrated measurements and blinded assessors where feasible. Include a comparison group recruited through the same process. If carriers receive detailed examinations while controls contribute routine records, detection opportunity is unequal. Explain which findings may be returned, who confirms them and what support follows. Incidental findings and family implications require governance before invitations begin.
Evidence check: preserve the observation before the interpretation. Record matched measurement opportunity and participant burden. Identify its source, unit, time window and uncertainty. Then state the inference, one credible alternative, and the additional observation that would distinguish them. A fluent explanation is not yet evidence until another reader can inspect the same trail.
Changed-context transfer: repeat this chapter’s decision after changing one consequential condition while holding the others fixed. Calculate or predict what should change, name the threshold or mechanism, and compare it with the original case. State the range tested. If the conclusion reverses, explain why; if it survives, say what remained invariant.
Limit and misconception check: do not turn a relative comparison into an absolute guarantee, treat an unmeasured condition as absent, or let one convenient indicator stand for the full mechanism. Name the boundary where this chapter’s method stops, the tempting overclaim it prevents, and the safer sentence the available evidence actually supports.
Practical decision record: write one sentence for what is known, one for what remains uncertain, and one for the next safe action. Include the evidence owner, the latest useful decision time, the review date, and the condition that would stop or reverse the action. This turns a technical result into a choice that can be checked later.
A molecular assay may test transcript abundance, protein localisation, binding or pathway activity. Use controls distinguishing complete loss, partial loss, altered function and artefact. Replicate batches and state limits of the biological system. Concordance between a functional gradient and a phenotype gradient strengthens a mechanism but does not prove it. Discordance may indicate an inadequate assay, heterogeneous phenotype or modifiers. Predefine which results would weaken the proposed series.
Evidence check: preserve the observation before the interpretation. Record functional controls and human triangulation. Identify its source, unit, time window and uncertainty. Then state the inference, one credible alternative, and the additional observation that would distinguish them. A fluent explanation is not yet evidence until another reader can inspect the same trail.
Changed-context transfer: repeat this chapter’s decision after changing one consequential condition while holding the others fixed. Calculate or predict what should change, name the threshold or mechanism, and compare it with the original case. State the range tested. If the conclusion reverses, explain why; if it survives, say what remained invariant.
Limit and misconception check: do not turn a relative comparison into an absolute guarantee, treat an unmeasured condition as absent, or let one convenient indicator stand for the full mechanism. Name the boundary where this chapter’s method stops, the tempting overclaim it prevents, and the safer sentence the available evidence actually supports.
Practical decision record: write one sentence for what is known, one for what remains uncertain, and one for the next safe action. Include the evidence owner, the latest useful decision time, the review date, and the condition that would stop or reverse the action. This turns a technical result into a choice that can be checked later.
CHAPTER 9 OF 12 · Check evidence and limits
9. Test selection, classification and family dependence
Perturb one element at a time: pathogenicity threshold, transcript, domain boundary, age restriction, phenotype definition and family-clustering method. Compare participants with eligible nonparticipants using permissible variables. Run leave-one-family and leave-one-variant-out analyses. Use a negative-control feature to reveal healthcare-use or reporting differences. A synonymous comparison can help, but synonymous variants are not universally neutral and recruitment still must be matched.
| Check | Changed | Held fixed | Question |
|---|---|---|---|
| Leave one family out | Related cluster | Phenotype rule | Family-driven signal? |
| Leave one variant out | Variant composition | Class definition | Founder effect? |
| Age restriction | Risk window | Variant set | Age-dependent penetrance? |
| Blinded reassessment | Measurement process | Participants | Observer influence? |
Evidence check: preserve the observation before the interpretation. Record classification, family and selection sensitivity. Identify its source, unit, time window and uncertainty. Then state the inference, one credible alternative, and the additional observation that would distinguish them. A fluent explanation is not yet evidence until another reader can inspect the same trail.
Changed-context transfer: repeat this chapter’s decision after changing one consequential condition while holding the others fixed. Calculate or predict what should change, name the threshold or mechanism, and compare it with the original case. State the range tested. If the conclusion reverses, explain why; if it survives, say what remained invariant.
Limit and misconception check: do not turn a relative comparison into an absolute guarantee, treat an unmeasured condition as absent, or let one convenient indicator stand for the full mechanism. Name the boundary where this chapter’s method stops, the tempting overclaim it prevents, and the safer sentence the available evidence actually supports.
Practical decision record: write one sentence for what is known, one for what remains uncertain, and one for the next safe action. Include the evidence owner, the latest useful decision time, the review date, and the condition that would stop or reverse the action. This turns a technical result into a choice that can be checked later.
Recontact based on genomic information may reveal risk participants did not expect. Check consent, recontact permission, clinical confirmation, counselling and privacy protections. Research findings are not automatically clinical diagnoses. Small carrier groups create unstable estimates and re-identification risk. Aggregate or suppress when necessary. Association does not mean every carrier will develop a condition, and no detected phenotype is not proof of harmlessness. Bound each claim to population and assessment window.
Evidence check: preserve the observation before the interpretation. Record consent, privacy and clinical-validation limits. Identify its source, unit, time window and uncertainty. Then state the inference, one credible alternative, and the additional observation that would distinguish them. A fluent explanation is not yet evidence until another reader can inspect the same trail.
Changed-context transfer: repeat this chapter’s decision after changing one consequential condition while holding the others fixed. Calculate or predict what should change, name the threshold or mechanism, and compare it with the original case. State the range tested. If the conclusion reverses, explain why; if it survives, say what remained invariant.
Limit and misconception check: do not turn a relative comparison into an absolute guarantee, treat an unmeasured condition as absent, or let one convenient indicator stand for the full mechanism. Name the boundary where this chapter’s method stops, the tempting overclaim it prevents, and the safer sentence the available evidence actually supports.
Practical decision record: write one sentence for what is known, one for what remains uncertain, and one for the next safe action. Include the evidence owner, the latest useful decision time, the review date, and the condition that would stop or reverse the action. This turns a technical result into a choice that can be checked later.
CHAPTER 11 OF 12 · Practise and continue
11. Practise transfer with a founder variant and older cohort
First discover that ten truncating carriers share one founder variant and two families. Recalculate after leaving out each family and the founder. If the gradient weakens, design independent recruitment. Then move the study to an older biobank. Apparent penetrance may rise because carriers had more time, or fall because severe cases were less likely to enrol or survive. Write both explanations and identify observations that separate them before comparing raw percentages.
Evidence check: preserve the observation before the interpretation. Record founder-variant and older-cohort contexts. Identify its source, unit, time window and uncertainty. Then state the inference, one credible alternative, and the additional observation that would distinguish them. A fluent explanation is not yet evidence until another reader can inspect the same trail.
Changed-context transfer: repeat this chapter’s decision after changing one consequential condition while holding the others fixed. Calculate or predict what should change, name the threshold or mechanism, and compare it with the original case. State the range tested. If the conclusion reverses, explain why; if it survives, say what remained invariant.
Limit and misconception check: do not turn a relative comparison into an absolute guarantee, treat an unmeasured condition as absent, or let one convenient indicator stand for the full mechanism. Name the boundary where this chapter’s method stops, the tempting overclaim it prevents, and the safer sentence the available evidence actually supports.
Practical decision record: write one sentence for what is known, one for what remains uncertain, and one for the next safe action. Include the evidence owner, the latest useful decision time, the review date, and the condition that would stop or reverse the action. This turns a technical result into a choice that can be checked later.
CHAPTER 12 OF 12 · Practise and continue
12. Continue through sources and the Genetics route
Read the NIH-led paper Genotype first: clinical genomics research through a reverse phenotyping approach, the NHGRI genotype-first explanation, and the allelic-series rare-variant association paper. Continue to phenotype harmonisation, cross-cohort replication, and rare-variant meta-analysis.
Evidence check: preserve the observation before the interpretation. Record the narrowest reproducible genotype-first claim. Identify its source, unit, time window and uncertainty. Then state the inference, one credible alternative, and the additional observation that would distinguish them. A fluent explanation is not yet evidence until another reader can inspect the same trail.
Changed-context transfer: repeat this chapter’s decision after changing one consequential condition while holding the others fixed. Calculate or predict what should change, name the threshold or mechanism, and compare it with the original case. State the range tested. If the conclusion reverses, explain why; if it survives, say what remained invariant.
Limit and misconception check: do not turn a relative comparison into an absolute guarantee, treat an unmeasured condition as absent, or let one convenient indicator stand for the full mechanism. Name the boundary where this chapter’s method stops, the tempting overclaim it prevents, and the safer sentence the available evidence actually supports.
Practical decision record: write one sentence for what is known, one for what remains uncertain, and one for the next safe action. Include the evidence owner, the latest useful decision time, the review date, and the condition that would stop or reverse the action. This turns a technical result into a choice that can be checked later.
Questions readers ask
Is genotype-first research genetic screening?
No. Screening offers testing to identify risk in a defined population. Genotype-first research begins with existing genomic information for a research question.
Does a pathogenic label determine penetrance?
No. Penetrance still depends on age, definition, population, modifiers and observation.
Can one carrier prove a new feature?
One case can generate a hypothesis, but repeated independent observations and mechanism are needed before generalising.
Why include apparently unaffected carriers?
They reveal incomplete penetrance and subtle expression when absence is measured rather than assumed.