Why Science? | Digital PCR, Reaction Partitions and Nucleic-Acid Copy Evidence

Section 1 of 36

1. Begin by dividing the sample

Digital PCR distributes a nucleic-acid mixture across many small partitions, amplifies the target and reads each partition at the endpoint as positive or negative. The fraction of positives supports an estimate of target concentration through a statistical model. It is a counting strategy built from probabilities, not a microscope that sees individual DNA molecules. Define the target sequence, material, unit and purpose before calculating. Copy number per reaction, per microlitre of extract and per millilitre of original sample are not interchangeable. Clear definitions stop dilution, extraction recovery and partition statistics from disappearing behind one impressive integer.

Contents · Next section

Section 2 of 36

2. Use amplification chemistry specifically

Primers define the target region, and probes or dyes report amplification. Sequence specificity, efficiency and side products still matter even though measurement is endpoint-based. An assay should be tested against expected variants and related sequences. Partitioning does not rescue a poorly designed primer or an inaccessible target.

Contents · Previous section · Next section

Section 3 of 36

3. Distinguish endpoint from cycle tracking

Quantitative real-time PCR follows fluorescence during cycles and often relates a threshold cycle to standards. Digital PCR classifies partitions after amplification and estimates occupancy without a conventional external calibration curve. Both depend on reaction chemistry and controls. “Digital” describes classification, not freedom from continuous measurement or uncertainty.

Contents · Previous section · Next section

Section 4 of 36

4. Create droplets or chambers reproducibly

Platforms may use droplets, wells or microfluidic chambers. Partition count and effective volume influence precision and concentration. Unequal, merged or lost partitions reduce usable information. The platform’s accepted-quality rules and volume characterisation belong in the report. Thousands of partitions do not help if their volumes are assumed incorrectly.

Contents · Previous section · Next section

Section 5 of 36

5. Model random target allocation

Target molecules are distributed among partitions approximately at random under model assumptions. Some positive partitions contain more than one target, especially at higher concentration. Counting positives alone would then undercount molecules. A Poisson correction estimates mean occupancy from the positive and negative fractions.

Contents · Previous section · Next section

Section 6 of 36

6. Read positives and negatives

After cycling, fluorescence clusters are classified relative to a threshold. Clear separation is ideal, but borderline droplets can form “rain.” Classification rules, channel compensation and controls influence the count. A plot is not naturally binary; the analyst and validated software turn continuous fluorescence into categories.

Contents · Previous section · Next section

Section 7 of 36

7. Apply the Poisson correction

If p is the positive fraction, estimated mean copies per accepted partition can be expressed as minus the natural logarithm of one minus p under a simple model. This correction accounts for multiple occupancy. It assumes independent random distribution and reliable classification. The equation is useful because its assumptions are visible and testable.

Contents · Previous section · Next section

Section 8 of 36

8. Know the partition volume

Concentration follows estimated copies per partition divided by partition volume, with dilution factors included. Small bias in effective volume becomes bias in concentration. NIST work emphasises characterising partition volume and reference materials for traceability. Manufacturer nominal volume is not automatically sufficient for the most exact claims.

Contents · Previous section · Next section

Section 9 of 36

9. Treat threshold and rain as analysis choices

A threshold too low converts noise into positives; too high hides weak true signal. Borderline partitions may reflect inhibition, damaged targets, nonspecific products or instrument variation. Define classification rules before group comparison, inspect control clusters and report excluded partitions. Quietly moving the threshold until the answer looks right introduces analyst bias.

Contents · Previous section · Next section

Section 10 of 36

10. Run no-template and negative controls

A no-template control tests reagent contamination and nonspecific amplification. A negative biological control challenges assay specificity in the relevant matrix. Rare positives in controls are especially important when the claimed target is rare. Predefined limits decide whether a run is valid, needs repetition or supports only an upper bound.

Contents · Previous section · Next section

Section 11 of 36

11. Use a positive control

A positive control shows that primers, probe, cycling and reader can detect the target. Its concentration and matrix should challenge the relevant range. A successful positive does not prove unknown extraction was effective or free of inhibition. Separate controls are needed for sample preparation and internal amplification performance.

Contents · Previous section · Next section

Section 12 of 36

12. Define the molecular measurand

Copy number may mean copies of one sequence in extracted material, copies per microlitre of reaction, or copies per unit of original sample. Fragmentation and linked targets change what is independently partitioned. State the sequence, material basis, dilution chain and reporting unit. “Absolute copies” without a measurand is incomplete.

Contents · Previous section · Next section

Section 13 of 36

13. Count accessible, amplifiable targets

Digital PCR measures targets that survive extraction, enter partitions and can be amplified by the assay. Damaged DNA, secondary structure or chemical modification may make some physical molecules inaccessible. NIST describes the importance of accessible and amplifiable targets. The number is method-defined, not a census of every molecule that once existed.

Contents · Previous section · Next section

Section 14 of 36

14. Control extraction and inhibition

Sample extraction can lose nucleic acid or co-purify inhibitors. Dilution can reduce inhibition but changes concentration and uncertainty. Spike controls and multiple dilutions can test recovery and inhibition. A clean amplification plot does not prove quantitative extraction, so the result must state whether it refers to extract or original material.

Contents · Previous section · Next section

Section 15 of 36

15. Treat RNA as an extra conversion

For RNA targets, reverse transcription creates complementary DNA before digital PCR. Enzyme efficiency and priming add variation, so the assay may be called digital RT-PCR. Controls should separate RNA, reverse-transcription and amplification effects. Copy estimates of cDNA are not automatically exact counts of original RNA molecules.

Contents · Previous section · Next section

Section 16 of 36

16. Practise with an invented partition table

These fictional values support statistical reasoning only; they are not assay thresholds.

Both too few and too many positives can weaken precision.

Contents · Previous section · Next section

Section 17 of 36

17. Calculate concentration transparently

Report accepted partitions, positive fraction, occupancy estimate, partition volume and all dilution factors. Carry units through the calculation. Software may display copies per microlitre, but the analyst must know whether that refers to reaction mixture or original specimen. Preserve raw counts so the number can be recalculated.

Contents · Previous section · Next section

Section 18 of 36

18. Link precision to partition count

More independent accepted partitions can narrow sampling uncertainty, especially at suitable occupancy. Precision also depends on classification, volume and preparation. Repeating wells may increase total partitions, but wells from one extraction remain technical replicates. Confidence intervals should reflect the statistical model and other uncertainty sources. Pooling partitions can improve counting precision only when wells are comparable and the pooling rule is justified. Independent extractions address a different question: whether sampling and preparation reproduce. Show both levels instead of collapsing every partition into one enormous sample size. Statistical power comes from the experimental unit, while partition count refines measurement of that unit.

Contents · Previous section · Next section

Section 19 of 36

19. Stay inside the dynamic range

At very low occupancy, a few background positives dominate interpretation. At very high occupancy, almost every partition is positive and the number of unseen multiple targets grows. Dilution brings samples into an informative window. Digital PCR is not infinitely sensitive or unlimited simply because the readout is partitioned.

Contents · Previous section · Next section

Section 20 of 36

20. Control false positives and contamination

Highly sensitive assays can detect carryover from previous amplifications, environmental DNA or reagent impurities. Physical workflow separation, clean controls and predefined reporting rules matter. One positive partition may be a true rare target or contamination. Replication and independent preparation help distinguish those possibilities.

Contents · Previous section · Next section

Section 21 of 36

21. Use multiplex ratios carefully

Different fluorescent channels can measure targets together, supporting copy-number ratios or linked-target studies. Spectral overlap, competition and unequal assay performance can bias ratios. Compensation and single-target controls are needed. A ratio may reduce some volume uncertainty but does not remove extraction, accessibility or classification effects.

Contents · Previous section · Next section

Section 22 of 36

22. Challenge the claim “absolute means exact”

Digital PCR can avoid a conventional calibration curve and support direct counting models, but it still has uncertainty from partition volume, random allocation, thresholds, accessible targets, dilution and controls. “Absolute” is often used to contrast calibration strategies, not to claim perfect truth. Report traceability and uncertainty instead of the adjective alone.

Contents · Previous section · Next section

Section 23 of 36

23. Challenge the claim “one positive proves a rare variant”

One positive partition is compatible with a rare target, but also with contamination, nonspecific amplification or misclassification. Evidence becomes stronger through sequence-specific probes, clean controls, independent extraction, repeat partitions and orthogonal confirmation. Rare-event claims demand stricter attention to background because every count carries more weight.

Contents · Previous section · Next section

Section 24 of 36

24. Compare digital PCR with qPCR

qPCR tracks amplification curves and commonly uses standards or relative quantification; digital PCR uses endpoint partition occupancy. qPCR can be efficient across many samples, while digital PCR can be useful for low-level differences and reference measurement. Neither is universally superior. The sample, range, throughput and claim determine the method.

Contents · Previous section · Next section

Section 25 of 36

25. Did You Know? Negative partitions carry information

A negative partition may look like “nothing happened,” yet the fraction of negatives anchors the Poisson calculation. When almost no negatives remain, concentration becomes hard to estimate precisely. Absence at the tiny-partition level is therefore part of the count. Science often learns from structured zeros as well as bright positives.

Contents · Previous section · Next section

Section 26 of 36

26. Did You Know? Too many positives can be a problem

A plate full of positive partitions feels convincing, but it hides how many targets shared each compartment. Dilution can improve information by restoring a mixture of positives and negatives. More visible signal is not always more quantitative evidence. The best occupancy is chosen for the estimation problem.

Contents · Previous section · Next section

Section 27 of 36

27. Preserve the complete digital-PCR audit trail

Record specimen, extraction, nucleic-acid quality, target sequence, primers and probes, reagent lots, partition platform, accepted counts, volume basis, cycling, fluorescence channels, thresholds, rain rules, controls, dilution, inhibition tests, reverse transcription if used, replicate structure, raw files, calculation, uncertainty, software and analysis version. MIQE-style transparency makes results reusable. Keep fluorescence plots and per-well counts, not only the final concentration. State which wells were excluded, why, and whether the rule was set before viewing groups. Document how partition volume was obtained and how uncertainty from dilution and volume entered the interval. Link every reported unit back to the original specimen. These records allow another analyst to move a defensible threshold, recalculate occupancy and see whether the biological conclusion persists. A digital result is strongest when its apparently simple integer remains open to statistical and laboratory inspection.

Contents · Previous section · Next section

Section 28 of 36

28. Write a claim–evidence–limit paragraph

Try: “Across two independent extracts, target occupancy fell within the validated range, no-template and negative controls met acceptance rules, and dilution-corrected estimates agreed. The assay measured 1,240 accessible target copies per millilitre with stated uncertainty. This does not count inaccessible molecules, establish organism viability or prove clinical significance.”

Contents · Previous section · Next section

Section 29 of 36

29. Connect Biology and Mathematics

Biology supplies DNA, RNA, sequence variation and amplification. Chemistry supplies enzymes, probes and buffers. Physics supplies fluorescence and microfluidics; Mathematics supplies probability, logarithms and uncertainty. Singapore’s 2026 O-Level Biology syllabus builds genetic and practical reasoning, while digital PCR shows how many tiny yes/no observations become a quantitative biological claim.

Contents · Previous section · Next section

Section 30 of 36

30. Learn safely with simulated partitions

Students can classify provided clusters, calculate occupancy and test how dilution changes precision using safe datasets. Real work may involve clinical, environmental or genetically modified material, contamination control and specialist waste. Do not collect unknown samples or improvise amplification at home. Use simulations and supervised approved laboratories.

Contents · Previous section · Next section

Section 31 of 36

31. Make science tuition earn its place

Good science tuition connects amplification to probability and asks what each negative partition contributes. Learners can progress from Primary Science fair tests and PSLE Science graphs to Secondary Science, O-Level Science and STEM reasoning about DNA, logarithms, controls and uncertainty. The method becomes understandable without turning it into a recipe.

Contents · Previous section · Next section

Section 32 of 36

32. Use the topic for school choices

Verify official descriptions of genetics, biotechnology, computing and laboratory safety when comparing schools or enrichment. Strong foundations can use simulated droplets and open data without owning a digital-PCR system. Do not infer admissions advantage, scholarships, placements or careers from an equipment list. Look for mathematical reasoning and responsible biosafety.

Contents · Previous section · Next section

Section 33 of 36

33. See the career ecosystem

Partition-based nucleic-acid measurement appears in genomics, biotechnology, reference laboratories, environmental monitoring, food testing and biomedical research. Roles include assay design, metrology, bioinformatics, quality assurance and regulation. Qualifications and clinical authorisations vary. Current official institution, standards and employer sources should guide pathway decisions.

Contents · Previous section · Next section

Section 34 of 36

34. Use a jars-of-beads analogy—with limits

Imagine randomly distributing invisible beads among many jars, then asking only whether each jar contains at least one. The mix of yes and no supports an estimate even though some jars hold several beads. The analogy captures occupancy and Poisson correction but misses amplification efficiency, thresholds, target accessibility and partition volume.

Contents · Previous section · Next section

Section 35 of 36

35. Ask what digital PCR cannot tell you alone

Digital PCR can estimate accessible target copies for a defined assay and material. It may not reveal full sequence, organism viability, cell location, protein expression, causal mechanism or clinical meaning. It also cannot recover targets lost during sampling or extraction. Pair it with sequencing, culture, imaging or clinical evidence when those questions matter. A precise copy estimate can still describe a biased specimen. Ask whether collection, extraction, independent replication or an orthogonal sequence test is the largest remaining uncertainty, then design the next step around that boundary.

Contents · Previous section · Next section

Section 36 of 36

36. Keep the sample-to-partition-to-claim chain visible

Begin with a defined molecular measurand, collect representative material, validate extraction and assay specificity, partition reproducibly, run positive and negative controls, classify fluorescence by predefined rules, stay within occupancy range, apply the stated Poisson and volume model, repeat independent samples and preserve raw counts. Then distinguish amplifiable target copies from physical molecules and biological meaning. Digital PCR becomes rigorous when every reported copy remains connected to a partition, probability model, control and uncertainty. A strong report should expose accepted partition counts, positive fractions, threshold plots, rain handling, partition-volume basis, dilution factors and confidence intervals rather than relying on one software output. It should distinguish repeated wells from independent extractions and explain how background positives affect the detection claim. When nearly every partition is positive, dilution is an information-improving step rather than an admission of failure. When only one or two are positive, an independently prepared repeat and orthogonal confirmation are proportionate safeguards. These practices let another analyst reproduce the calculation, challenge the model assumptions and discover whether the apparent biological difference survives a different defensible classification rule.

Contents · Previous section · Continue to the Science Learning Hub