ToolsConductScience tool
Polymerase ComparisonFree in-browser calculator

PCR Fidelity Estimator.

Predict the fraction of error-free PCR product for any polymerase, amplicon length, and cycle count. Compare Q5, Phusion, KAPA HiFi, Platinum SuperFi II, Pfu, OneTaq, and Taq side-by-side. Runs entirely in your browser.

PrivateData stays in your browser
LiveNo sign-up required
Open methodFormula shown on this page
CitableMethods and citation included

Calculator

Results update in place

Try it out

Load example pcr-fidelity data to see the full workflow

bp
cycles

Fidelity Estimate

% correct after 30 cycles
98.8%
Error rate
5.30e-7 /bp
Cycles ≥ 95% fidelity
30
Cycles ≥ 99% fidelity
25

Expected molecules after 30 cycles

Total molecules
1.07e+13
Error-free molecules
1.06e+13
Error-containing molecules
1.27e+11

Fidelity across cycles

Comparison across preset polymerases (30 cycles, 1500 bp)

PolymeraseError rateFinal % correctCycles ≥ 95%Cycles ≥ 99%
KAPA HiFi2.8e-799.4%3030
Platinum SuperFi II3.0e-799.3%3030
Phusion High-Fidelity4.4e-799.0%3030
Q5 High-Fidelity5.3e-798.8%3025
Pfu1.3e-697.1%3010
OneTaq Hot Start2.0e-563.5%30
Taq2.9e-552.3%20

When to use

  • Decide how many PCR cycles to run before error accumulation becomes a concern
  • Compare high-fidelity polymerases (Q5, Phusion, KAPA HiFi, Platinum SuperFi II) for a cloning or NGS job
  • Estimate the ceiling for error-free product on long or GC-extreme amplicons
  • Plug in a lab-measured error rate via the Custom option to estimate fidelity for an in-house enzyme
  • Teach the cycle-by-cycle compounding of PCR errors in a methods or reagents class

Do not use for

  • Estimating variant allele frequencies from deep sequencing — the tool does not model sampling noise or strand bias
  • Predicting chimera / recombination rates on mixed-template PCR (e.g. metagenomics)
  • Fidelity of reverse transcription — RT error rates are separate and much higher
  • Error-correcting chemistries like rolling-circle or duplex consensus sequencing, which are modeled differently

Halve e·L per cycle, not multiply

The intuitive "fraction error-free drops by e·L per cycle" is wrong by a factor of 2. Only the half of the population that is newly synthesized each cycle is exposed to error — the other half is inherited. This is why the closed form is (1 − e·L/2)ⁿ.

More cycles is not more data

Past ~25–30 cycles most polymerases are already in plateau and fidelity is dropping. If you need more product, increase starting template copies rather than adding cycles.

Long amplicons amplify the fidelity penalty

Doubling amplicon length doubles the per-cycle error probability. A 4 kb amplicon at 30 cycles with Q5 is still ~98% error-free, but with Taq it drops below 10%. Budget accordingly.

Manufacturer error rates are best-case

Published error rates are measured on optimized buffers and templates. Real-world rates are typically 1.5–3× higher on difficult templates. Treat this calculator as an upper bound on fidelity.

Use starting copies to set sequencing depth

If you plug in starting copies, the tool reports expected error-free molecule counts at the end of the run. This is the ceiling for unique error-free reads, and a useful sanity check for deep sequencing experiments where you need a minimum clean template count.

1

Method

Cycle-by-cycle recursion Xₙ₊₁ = Xₙ·(1 − e·L/2) in closed form: P_correct(n) = (1 − e·L/2)ⁿ. Polymerase presets drawn from Barnes (PNAS 1994), Cline et al. (NAR 1996), and manufacturer product documentation (NEB, Roche/KAPA, Thermo Fisher). Does not model chimera formation, GC bias, or polymerase stalling.

2

Published

Published 2026-04-07. Calculations support planning and documentation; verify procurement decisions against manufacturer specifications or institutional SOPs.

3

How to cite

How to Cite

ConductScience PCR Fidelity Estimator. ConductScience, Inc. 2026. Available at: https://conductscience.com/tools/pcr-fidelity-calculator

Barnes WM. PCR amplification of up to 35-kb DNA with high fidelity and high yield from lambda bacteriophage templates. Proc Natl Acad Sci USA. 1994;91(6):2216-2220. doi:10.1073/pnas.91.6.2216

Cline J, Braman JC, Hogrefe HH. PCR fidelity of Pfu DNA polymerase and other thermostable DNA polymerases. Nucleic Acids Res. 1996;24(18):3546-3551. doi:10.1093/nar/24.18.3546

Sun F. The polymerase chain reaction and branching processes. J Comput Biol. 1995;2(1):63-86. doi:10.1089/cmb.1995.2.63

What is PCR fidelity?

PCR fidelity is the fraction of amplified molecules that are perfect copies of the original template. It is set by two things: the polymerase’s per-base error rate and the number of bases it has to copy each cycle (the amplicon length). Errors compound geometrically across cycles, so even a small per-base error rate can dominate the final population after enough doublings.

This tool estimates that fraction and its behavior across cycles. Use it to decide how many cycles to run, which polymerase to pick for a given cloning/NGS/variant-calling job, and what your realistic ceiling for error-free product is.

The cycle-by-cycle recursion

Let Xₙ be the expected fraction of error-free molecules after n cycles. Each cycle every molecule is kept AND copied once. A newly synthesized molecule is error-free iff its parent was AND the copy itself was clean (probability 1 − e·L). Therefore:

X_{n+1} = (1/2)·Xₙ + (1/2)·Xₙ·(1 − e·L) = Xₙ · (1 − e·L / 2)

Closed form: P_correct(n) = (1 − e·L / 2)ⁿ

This matches the first-order approximation used by NEB’s PCR Fidelity Estimator and by Sun (1995). It assumes independent errors and ideal amplification.

Choosing a polymerase

Cloning / site-directed mutagenesis: use Q5, Phusion, KAPA HiFi, or Platinum SuperFi II. These have error rates ~10⁻⁷ to 10⁻⁶ and produce cloning-grade product at 30–35 cycles.
NGS library prep: KAPA HiFi is the standard choice for amplicon-based libraries due to its very low error rate and GC-tolerant buffer.
Long amplicons (>5 kb): Q5 or Platinum SuperFi II are the most reliable. Fidelity alone does not guarantee yield on long templates — processivity matters just as much.
Colony PCR, genotyping, diagnostic work: OneTaq or Taq is usually fine. You do not need proofreading to confirm the presence or absence of an amplicon.
Avoid Taq for anything downstream of Sanger or NGS: the raw error rate (~3×10⁻⁵) makes every 1 kb amplicon carry at least one substitution in most molecules after 25–30 cycles.

Frequently asked

What math does this calculator use?
It uses the standard cycle-by-cycle recursion for PCR fidelity. Each cycle, every existing molecule is preserved and a new copy is made with probability (1 − e·L) of being error-free, where e is the polymerase per-base error rate and L is the amplicon length. The expected fraction of error-free molecules after n cycles is (1 − e·L/2)^n. This is the same model used by NEB’s PCR Fidelity Estimator and is the first-order approximation from Sun (1995).
Where do the default error rates come from?
The presets are drawn from published manufacturer literature and primary references: Q5 and Phusion from NEB product sheets, KAPA HiFi from the Roche/KAPA datasheet, Platinum SuperFi II from Thermo Fisher, Pfu from Cline et al. (NAR 1996), and Taq from Barnes (PNAS 1994). If you have a lab-measured error rate, switch to the Custom option and enter it.
Why is fidelity higher than I expected?
Because only half of the population is newly synthesized on any given cycle — the other half are parent molecules inherited from the previous cycle. This is why the effective error compounding per cycle is e·L/2, not e·L. For typical high-fidelity polymerases (e ~ 10⁻⁶) at 1 kb and 30 cycles, you should see ≥99% error-free product.
Why does fidelity drop so fast for Taq?
Taq’s per-base error rate is ~3×10⁻⁵ — roughly 50–100× higher than a modern high-fidelity enzyme. For a 1 kb amplicon that is ~1.4% loss per cycle, and 30 cycles compounds it to 30–50% loss depending on the exact rate. For any cloning, NGS, or sequencing-grade work, use a proofreading enzyme (Q5, Phusion, KAPA HiFi, or Platinum SuperFi II).
What does "cycles ≥ 95% fidelity" mean?
It is the largest number of cycles at which the expected fraction of error-free molecules is still above 95%. Running more cycles than this means more than 5% of your product will carry at least one substitution. This is a useful ceiling for cloning, single-molecule sequencing, or variant discovery.
Does this account for recombination or amplification bias?
No. The model assumes every molecule duplicates in every cycle (ideal exponential amplification) and that errors are independent. It does not model chimera formation, GC-content amplification bias, or polymerase stalling. For long (≥10 kb) amplicons or GC-extreme templates, use the estimate as an upper bound on achievable fidelity, not a prediction.
Does my data leave the browser?
No. All computation runs locally in JavaScript. Nothing is sent to any server. You can verify by disconnecting from the internet and reloading the tool.
How should I cite this tool?
See the “How to cite” section on this page for a ready-to-paste methods statement and citation. The underlying math follows Sun (1995) and manufacturer-published error rates (Barnes 1994; Cline et al. 1996; NEB product documentation).
325
Free tools
1,200+
Institutions
100%
Client-side
0
Uploads required