Distinguish damage, mismatch, and inherited change
An abnormal base pair is not yet the same thing as a stable sequence variant in all descendants.
By the end: Trace how correction or another copying round changes the fate of a mismatch.
Identify what changed
Chemical damage alters a base or the DNA backbone. A mismatch pairs bases that do not form the expected pair. A sequence variant is a difference relative to a reference. These categories overlap, but they are not interchangeable. An A paired with C can result from an incorporation error even without an external damaging agent.
Some polymerases proofread during synthesis. Mismatch repair acts after an error escapes proofreading, while other repair pathways remove damaged bases or larger damaged segments. Repair requires a way to choose a strand or a usable template. Complementarity alone cannot reveal whether A or C in a mismatch was the intended base.
An abnormal pair and an altered chemical structure are different evidence
DNA damage is a chemical change to a molecule, whereas a sequence difference is a comparison between sequences. To separate these ideas, imagine one ordinary A:T pair and one ordinary G:C pair at the same stated position in different samples. Both pairs can have normal chemistry and pairing even though their sequences differ. Now imagine adjacent bases joined by an abnormal chemical connection. That structure can interfere with reading the template without initially replacing either base letter in a sequence listing. A sequence alphabet records base identities, but it does not describe every bond or chemical state of DNA.
A mismatch is another kind of abnormality. In the lesson's A:C example, neither letter needs to represent a chemically damaged base. The problem is that A and C do not form the expected ordinary complementary pair at that position. The A might have been copied correctly from an earlier template while C entered by mistake. Alternatively, with a different history, the C might have been the base to retain. Observing the present mismatch alone does not distinguish those histories. Pairing identifies a disagreement between strands, not a uniquely guilty base that carries an intrinsic error label.
A sequence difference need not be a mismatch
Each pair represents one site within longer antiparallel strands. The dashed connection marks the A:C disagreement. G:C pairs normally even when it differs from the stated A:T reference. No abnormal base chemistry is drawn.
A repair mechanism must therefore solve two problems that a final sequence hides. It must recognize a structure or context needing attention, and it must obtain information about what should occupy that position. An undamaged complementary strand often supplies that information when one strand has a local lesion, meaning a damaged site. For a mismatch between ordinary bases, choosing which strand to replace is essential. Otherwise a mechanism might produce normal pairing while changing the sequence that the copying process was supposed to preserve. Correct pairing and correct restoration are related goals, but they are not identical criteria.
Suppose the known earlier pair was A:T and replication produced parental A opposite new C. Replacing C with T restores the earlier pair. Replacing A with G also makes an ordinary complementary pair, but now the result is G:C. Both outcomes remove the mismatch. Only the first restores the stated earlier sequence. This contrast is why the display explicitly identifies A as parental and C as newly incorporated. That information comes from the constructed history, not from the shapes of the two letters. Real strand discrimination depends on molecular context that the one-site display does not simulate.
Proofreading and later mismatch repair use different opportunities to correct copying errors. Some replicative polymerases can remove a recently incorporated nucleotide from the growing end before continuing synthesis. Later repair can act after synthesis has moved beyond the error, removing a region and filling it again using a chosen template. Neither process requires an ability to compare the DNA against an abstract written reference genome. Their information is local and molecular. Calling these processes checking a text can be misleading if it suggests a complete external master copy against which every site is compared.
Removing a mismatch does not by itself identify the restored sequence
Retained template
Base replaced
Result
Earlier A:T restored?
Parental A
New C replaced by T
A:T
Yes
New C
Parental A replaced by G
G:C
No
The earlier A:T pair is stipulated. Both choices produce ordinary pairing, but they preserve different templates. This is a sequence comparison, not a model of repair probabilities.
Repair also needs physical completion. Removing incorrect or damaged material leaves a strand that may require new synthesis and sealing. A recognition protein alone does not rebuild the missing backbone. This connection to replication explains why polymerase and ligase can contribute to both copying and repair while performing the same underlying chemical tasks. It does not imply that every repair pathway uses identical enzymes or recognizes identical defects. The substrate, meaning the molecule an enzyme acts on, differs between pathways. The broad label DNA repair describes a group of processes rather than one universal operation that returns any altered molecule to its historical sequence.
Follow one unresolved mismatch
Suppose the original upper base was A and the newly made lower base was incorrectly C instead of T. Correcting the new strand restores A:T. If both strands instead act as templates in the next round, the A strand yields A:T while the C strand yields G:C. One daughter duplex differs from the original sequence.
A well-paired G:C duplex does not announce that it was once an error. Its history is no longer visible from pairing alone. This is why preventing fixation of an error and recognizing a stable difference are distinct problems. Our one-site model ignores cell death, additional repair, segregation details, and selection.
Copying a mismatch preserves both competing templates
Track the A:C mismatch as a duplex, not as a single ambiguous letter. The upper strand contains the parental A, and the lower strand contains the newly incorporated C. When the strands separate for another ordinary copying round, polymerase encounters an A template on one side and a C template on the other. The A directs addition of T. The C directs addition of G. There is no step in this path where the original C is converted into G. C persists in one daughter as a template, paired with a newly synthesized G on the opposite strand.
The orientation can be checked over a longer invented interval. Let the parental upper strand be 5′-GAT-3′ and the new lower strand be 3′-CCA-5′. The middle position is the A:C mismatch. Copying the upper strand produces a lower 3′-CTA-5′ partner, restoring the original pair at the middle site. Copying the lower strand produces an upper 5′-GGT-3′ partner. The two daughter upper sequences are therefore GAT and GGT. The flanking positions remain the same in this example. Their inclusion makes clear that ordinary copying makes complete complementary partners rather than exchanging the two central bases between daughters.
Retain C as C while synthesizing its G partner
The boxed middle A and C track the original disagreeing bases through copying. Daughter 1 retains the upper GAT strand; daughter 2 retains the lower CCA strand. All rows have explicit ends, and the flanking positions are schematic context rather than a real gene.
The new G:C pair is stable in a limited sequence sense. It can guide ordinary complementary copying in later rounds without regenerating the original mismatch. Stable does not mean chemically indestructible, permanently expressed, or guaranteed to persist in a growing population. It means that at this site both strands now specify the same complementary sequence state. Mismatch recognition no longer has the original A:C disagreement to detect. Recovering the earlier A:T state would need some additional information or another sequence-changing event. Pairing alone cannot distinguish the altered daughter from a G:C duplex with a much older history.
Compare this with correction before the next copying round. If repair removes the new C-containing interval and uses the retained A-containing strand to fill it, the duplex becomes A:T. Ordinary copying of that corrected duplex gives two A:T daughters. The difference between the paths therefore appears before the next generation of products is counted. One path removes the disagreement while retaining the stated parental information. The other gives each disagreeing strand an opportunity to act as a template. Timing matters because the same mismatch can lead to different product sets depending on whether correction precedes copying.
The one-site result is one original-sequence daughter and one altered-sequence daughter, not a claim that half of all future cells must carry the change. A duplex is not a whole organism, and a molecular product count does not specify how cells survive, divide, or are sampled. Even if both daughters enter separate cells, those cells need not contribute equally to later samples. The example also omits other chromosome copies and any later repair or sequence change. Its exact conclusion concerns the two duplexes produced by the stated copying event. A population prediction would require additional assumptions beyond complementary base pairing.
Correction before copying changes the two-daughter product set
Path from parental A:new C
Duplex before copying
Products after one copying round
Correct the new strand first
A:T
A:T and A:T
Leave mismatch unresolved
A:C
A:T and G:C
Counterfactual: replace parental A
G:C
G:C and G:C
Only the last step in each row is an ordinary copying round. These exact duplex counts do not predict the frequencies of cells in a later population.
This distinction also clarifies the word fixed. Here an error becomes fixed in one duplex when copying produces a normally paired altered sequence. In population genetics, fixation can mean that a variant reaches all copies at a locus in a population. Those are different uses of the same word. The A:C example supports the first molecular description and says nothing by itself about the second. To avoid ambiguity, describe the G:C result as a stable paired sequence change in one daughter. Then state separately whether any claim concerns later cells, sampled DNA molecules, or reproductive descendants of an organism.
Work through the mechanism
One mismatch, two possible histories
Initial A:C mismatch
Parental
5′A3′
New, mismatched
3′C5′
Outcome
Restored duplex
5′A3′
Correct partner
3′T5′
New C was replaced by T using the stated parental template.
A was the parental base; C was the newly incorporated error. Positions and molecular sizes are schematic. The original and the chosen outcome remain visible.
Path from A:C
Result
What is retained
Repair new C to T
A:T
Original pair
Copy both without repair
A:T and G:C
One original and one altered pair
One copying event is not a prediction of descendant population frequencies.
Use the repair display to test strand identity rather than just final letters
Start with Repair before copying selected. Read the initial panel as a stated historical claim, the upper A belongs to the parental strand and the lower C entered during the recent copying event. The outcome panel retains A and supplies T as its complementary partner. The visible result is one restored duplex, not two products of another replication round. The display compares a correction outcome with a copying outcome, so equal numbers of panels would not mean equal numbers of biological events. Count the molecules described by the selected path rather than treating screen layout as a population measurement.
Switch to Copy without repair and look for the retained C in the second daughter. The new G appears above it, while C remains below. This is the identity check that a letter-only reading can miss. Writing C becomes G would describe a substitution within one strand, not the complementary copying shown. Instead, the two daughters contain different retained templates, one with A and one with C. Their new partners are T and G respectively. Both daughters pair normally, so pairing status no longer identifies which one preserves the sequence that existed before the incorporation error.
The two controls compare different histories from the same mismatch
Display choice
Retained information
Shown endpoint
Not calculated
Repair before copying
Parental A supplies template
One restored A:T duplex
Repair time or success rate
Copy without repair
A and C each supply a template
A:T plus G:C
Later cell frequencies
Selecting an option replaces the displayed outcome; it does not advance a simulation. The repair option stops at one corrected duplex, while the copying option shows two daughters.
The display contains a choice between two stipulated histories, not a measurement of how often either occurs. Repeatedly selecting the copying option does not accumulate generations or mutate the initial panel. Each choice restates the same starting mismatch and replaces the shown outcome. Likewise, choosing repair does not show an enzyme moving along the chain or establish the time required for correction. This matters when interpreting a simple interactive model. Its controls can expose a logical contrast between mechanisms without containing the variables needed to estimate rates, probabilities, or the proportion of lesions repaired in a living cell.
Consider a counterfactual outside the control choices. If a mechanism instead removed the A-containing region and used C as its template, it could make a G:C duplex before another replication round. After ordinary copying, both daughters would carry G:C at that site. This counterfactual demonstrates why detecting a mismatch is not enough to guarantee restoration. It is not a third measured outcome supplied by the interface, nor a claim about its frequency. The comparison changes only which strand supplies the sequence information. The normal base-pairing rules remain the same, and they give a different result when the retained template changes.
A reference sequence adds a different kind of information. If a listed reference contains A:T while the sampled duplex contains G:C, their difference can be recorded precisely. That does not prove which state was ancestral in the sampled lineage, which state is common in another population, or which state best supports cellular function. The reference is a coordinate and comparison standard, not a chemically privileged DNA molecule. In this lesson the earlier A:T state is known because the example specifies it. Without that stipulation, the current paired duplex and a reference alone would establish a difference rather than a complete history.
A current G:C pair does not disclose its history
The left path is a counterfactual strand-choice example, not an extra interface option. The right path starts with an existing G:C pair. Their common endpoint cannot tell which history occurred, and comparison with an A:T reference does not resolve that ambiguity.
The display's omissions help locate the next questions rather than answer them. It does not show the chemical signals that distinguish newly made DNA, how a repair complex finds the affected region, or whether the cell pauses division. Those mechanisms vary with organism and pathway. The model's contribution is narrower and testable on paper. Given the parental A and new C, retaining A during correction yields A:T, whereas copying both strands without correction yields A:T and G:C. Any proposed verbal account of these controls should preserve the same strand identities and should not append population frequencies that the model never calculated.
Separate cell lineage from organismal inheritance
A somatic mutation occurs in a body-cell lineage. It can pass to daughter cells without being passed to children of the organism. A germline change can be transmitted through reproductive cells, but transmission still depends on which gamete contributes and subsequent development.
A change need not be harmful or beneficial in every setting. Its effect depends on location, expression, molecular function, genetic background, and environment. A repair defect can alter the rate or spectrum of future variants without making every existing molecule defective in the same way.
Lineage and function require evidence beyond a corrected base pair
A cell lineage is a set of cells related by division. A stable sequence change can pass along that lineage when a cell copies the altered duplex and its descendants retain the copies. This is a form of inheritance between cells. It is not automatically inheritance between organisms. In a multicellular organism, somatic lineages make body tissues, while reproductive lineages can contribute genetic material to offspring. Finding a variant in several related body cells supports persistence through those cell divisions. It does not establish that the variant entered a reproductive cell or contributed to an offspring's genome.
Even within a somatic lineage, the fraction of sampled cells carrying a change is not determined by the original mismatch arithmetic. A change that occurs early in a branching lineage can be present in several descendant branches. A change arising later can be confined to one branch. Some branches may divide more, disappear, or be absent from the sampled tissue. A diagram showing eight final cells with a marked branch is therefore a stipulated history, not an expected fraction for every tissue. Interpreting a real sample would also require knowing how many DNA copies each cell contributes and which cells the sample contains.
Persistence in a body-cell branch is not transmission to an offspring
Each circle is one cell in an invented branching somatic history. Equal drawing size does not imply equal eventual contribution to a tissue sample. No reproductive lineage or offspring is shown or inferred.
Sequence persistence and functional effect are separate questions too. A variant outside a protein-coding interval might alter a regulatory interaction, or might have no detectable effect under the conditions examined. A variant within a coding interval might alter the resulting protein, but that possibility requires the additional steps of expression and translation developed later in the course. The repair model deliberately stops before those steps. It predicts base pairs, not protein amount or cellular activity. A normal-looking pair can have a sequence-dependent effect, while a repaired site can restore sequence without guaranteeing that every other molecule in the cell is undamaged.
A repair defect changes a process rather than specifying one universal outcome at every site. If a pathway recognizes certain kinds of damage poorly, the resulting changes may depend on which lesions arise and which alternative processes act on them. It does not follow that every DNA molecule must contain the same defect or that every previously present sequence difference resulted from that pathway's failure. This is a causal distinction between changing the conditions under which future variants arise and identifying the history of a particular observed variant. The second claim needs site-specific or lineage evidence beyond the first.
What observation would distinguish a current mismatch from a stable paired variant? Information about the bases on both complementary strands at the same molecular site would address pairing status. Comparing an ordinary single sequence listing to a reference instead addresses sequence difference. A reading that collapses both strands into one reported letter may not preserve the evidence needed to reconstruct an A:C intermediate. This is an information limit, not a claim that every assay has identical weaknesses. A defensible interpretation first asks what molecular feature the measurement retained, then asks whether the proposed conclusion requires information that the readout discarded.
Different evidence supports different levels of claim
Evidence available
What it addresses
Still unresolved
A and C opposite at one site
Current mismatch
Which base should be retained without history
G:C differs from reference A:T
Current paired sequence difference
Ancestral state and function
Variant in related body cells
Somatic lineage persistence
Transmission to an offspring
Same sequence in two samples
Sequence identity at measured positions
Molecular state and activity
These are information requirements rather than experimental instructions. Knowing the paired bases, a sequence difference, or a somatic lineage addresses distinct questions; none alone establishes a functional effect.
The strongest conclusion available from the constructed case is therefore a conditional molecular history. Given parental A and new C, correction directed at the new strand restores A:T; ordinary copying without correction produces one A:T duplex and one G:C duplex. Additional observations are needed to establish later lineage distribution, transmission to an offspring, or functional effects. These are not interchangeable uncertainties. Each concerns a different step, from a molecular copying product to a branching cell history or an expressed phenotype. Keeping those steps separate lets a precise sequence prediction remain useful without turning it into an unsupported claim about whole organisms.
Exercises
Can pairing identify the original?
You observe a correctly paired G:C where a reference has A:T. Can mismatch repair identify the older pair just from the current duplex?
Hint
Ask whether there is a mismatch now.
Worked answer
Reference pair
Upper
5′A3′
Lower
3′T5′
Current pair, no mismatch
Upper
5′G3′
Lower
3′C5′
Current evidence
Conclusion
G paired with C
No mismatch at that site
Reference A:T
A sequence difference
No historical template
Original state not recoverable from pairing alone
Reference sequence and repair target are different concepts.
No. G:C is correctly paired. A reference comparison establishes a difference, not an intrinsic mismatch. Recovering a previous sequence would require other information or a suitable template, and even then the reference is not automatically the biologically preferred allele.
Which descendants can carry it?
A variant appears in a skin-cell lineage and is found in several descendant skin cells. Does this establish transmission to an offspring?
Hint
Distinguish mitotic descendants from reproductive descendants.
Worked answer
Lineage
Supported by the observation?
Descendant skin cells
Yes
Reproductive cells
Not established
Offspring
Not established
The word inherited needs a specified lineage.
No. It establishes persistence in a somatic lineage. Transmission to offspring would require involvement of a reproductive lineage and contribution to the offspring. A skin sample does not establish either.
Sources and further reading
Selected textbook passages support the mechanisms below. The prose and diagrams are original, and examples are constructed rather than measured. These references are not a full literature review or independent specialist validation. Optional links are not runtime dependencies.