The mutation rate calculator above does two related jobs. In its first mode it converts mutant and viable colony counts, with their plating volumes and dilution factors, into a mutant frequency — the proportion of the population carrying the selected mutation. In its second mode it estimates a mutation rate per cell per generation from a fluctuation experiment, using the p0 null-class method, which infers the number of mutational events from how many parallel cultures produced no mutants at all.
Arb Digital publishes this as a laboratory arithmetic and teaching tool. It handles the counting maths that sits between a stack of plates and a number in a notebook, and it is explicit about the limits of each method rather than presenting one figure as the answer. It is not a substitute for the statistical treatment a published estimate requires.
What This Mutation Rate Calculator Does
In frequency mode, each plate count is converted to a concentration by dividing by the volume plated and multiplying by the dilution factor. The mutant concentration divided by the viable concentration gives the mutant frequency, which is reported both as a decimal and as mutants per hundred million cells — the form results are most often quoted in. A rough per-generation figure divides the frequency by the number of generations, and the article below explains carefully why that figure should be treated as an order-of-magnitude sanity check rather than a rate.
In p0 mode, the tool takes the number of parallel cultures, how many of them yielded no mutant colonies, and the final population size of a single culture. The proportion of blank cultures is the null class, and the mutation rate follows from it directly. This is the simplest of the fluctuation-analysis estimators and the only one that can be worked by hand.
Boundary worth stating: this page counts mutants. If you are working out the dilutions that produced the plates, use the serial dilution calculator; for converting a measured density to a target one, the cell density dilution calculator; and for the growth arithmetic behind the number of generations, the cell doubling time calculator.
How to Use It
- Choose the mode that matches the experiment you actually ran. One pair of plates gives a frequency; only a set of independent parallel cultures can give a rate.
- Enter the selective plate count with the volume plated and the dilution factor. Mutant plating is usually done undiluted, so the dilution factor is often 1.
- Enter the non-selective plate count the same way. This one is almost always heavily diluted, and getting the exponent wrong here is the single most common source of a wildly wrong answer.
- For p0, enter the number of cultures, how many were blank, and the cells per culture. The tool will warn you when the blank proportion falls outside the range where the method behaves.
- Report the method with the number. A figure quoted without saying whether it is a frequency or a rate, and by which estimator, is not interpretable.
The Formulas and How They're Calculated
Concentration = colonies ÷ volume plated × dilution factor. With 23 mutant colonies from 0.1 mL of undiluted culture, the mutant concentration is 23 ÷ 0.1 × 1 = 230 per mL. With 148 colonies from 0.1 mL of a millionfold dilution, the viable concentration is 148 ÷ 0.1 × 1,000,000 = 1.48 × 109 per mL.
Mutant frequency = mutants ÷ viable cells. That is 230 ÷ 1.48 × 109 = 1.554 × 10−7, or about 15.5 mutants per hundred million cells.
The p0 method: if mutational events occur independently, the number per culture follows a Poisson distribution, so the proportion of cultures with none is p0 = e−m, where m is the mean number of mutational events per culture. Rearranged, m = −ln(p0), and the mutation rate is μ = m ÷ N, with N the number of cells per culture. With 12 of 40 cultures blank, p0 is 0.30, so m is −ln(0.30) = 1.204, and against 1.2 × 108 cells per culture the rate is 1.00 × 10−8 per cell per generation.
Why Frequency Is Not Rate
This is the distinction the whole subject turns on. Mutation frequency is the share of the population that is mutant at the moment you plate it. Mutation rate is the probability that a cell acquires the mutation during one division. They differ because a mutation that happens early in the growth of a culture is inherited by every descendant of that cell, producing a large clone of mutants from a single event. A mutation that happens in the last division before plating produces one mutant. Both are one event; they contribute wildly differently to the frequency.
That is exactly the phenomenon Luria and Delbrück exploited. Across replicate cultures, the number of mutants varies far more than random sampling could explain — a few cultures carry enormous numbers, most carry a handful — and the shape of that distribution encodes the rate. As Rosche and Foster put it in Determining Mutation Rates in Bacterial Populations, a properly determined rate is more accurate and more reproducible than a mutant frequency, because it is not hostage to when in the growth curve the mutations happened to occur.
Where the Per-Generation Estimate Breaks Down
Dividing a frequency by a number of generations is a common shortcut and it is a poor estimator. It assumes mutants accumulate linearly with divisions, which is precisely what clonal expansion prevents. In practice it can be off by an order of magnitude in either direction, and its error is not symmetric: because the distribution of mutant numbers has a long tail, a single culture that happened to catch an early mutation drags the estimate up dramatically.
This tool reports the figure because it is a useful order-of-magnitude sanity check and because readers expect to see it, not because it is defensible in a publication. Where the number matters, the answer is a fluctuation experiment analysed with a proper estimator. The Ma–Sandri–Sarkar maximum-likelihood method is the standard, and Patricia Foster's Methods for Determining Spontaneous Mutation Rates sets out seven approaches with their assumptions and failure modes.
When the p0 Method Is Usable
The null-class method is elegant because it uses only one piece of information — how many cultures were blank — and therefore needs no colony counting on the cultures that were not. Its weakness is that it throws away everything else. If almost every culture has mutants, p0 approaches zero and the logarithm becomes unstable; if almost none do, the estimate rests on very few informative cultures. Published guidance puts the usable window at roughly m between 0.3 and 2.3, which corresponds to somewhere between about ten and about seventy-four percent of cultures being blank.
Falling outside that window is not a disaster, it is a design signal: adjust the inoculum size or the final population so that a workable proportion of cultures come up empty, and run the experiment again. The tool flags the condition rather than silently returning a number that looks fine.
Counting Errors That Dominate the Result
The dilution factor on the viable count is the big one. A factor of 105 entered where 106 belonged shifts the frequency by a full order of magnitude, and the result still looks entirely plausible — which is what makes it dangerous. Read the exponent back off the tube label rather than from memory.
Plate counts outside the countable range are the second. A selective plate with three colonies carries enormous Poisson noise, and a non-selective plate with six hundred is undercounted by crowding. Both errors propagate directly into the frequency. And because mutants are rare, the selective plates are almost always the ones with too few colonies, which is why replicate plating matters more on that side. For the underlying log maths, our logarithm calculator is useful, and the standard deviation calculator handles the spread across replicates.
Arb Digital publishes hundreds of free, no-signup tools covering molecular biology, laboratory maths, statistics and everyday arithmetic — no accounts, no stored data.
Browse All Free Tools Contact Arb DigitalCommon Mistakes to Avoid
- Calling a frequency a rate. They are different quantities with different units and different meanings, and the substitution invalidates any comparison.
- Entering the wrong dilution exponent on the non-selective plate, which moves the answer by powers of ten while looking perfectly reasonable.
- Using the total cells across all cultures in the p0 method instead of the population of a single culture.
- Running a fluctuation experiment with cultures that are not independent, for example by splitting one grown culture, which destroys the whole basis of the analysis.
- Reporting a rate from a single pair of plates. One culture cannot distinguish an early mutation from a late one.
Related Free Tools From Arb Digital
For growth arithmetic, the cell doubling time calculator converts two counts and an interval into a doubling time and a number of generations. For the plating work, use the serial dilution calculator and the cell density dilution calculator. For quantitative PCR, the qPCR efficiency calculator scores a standard curve and the DNA copy number calculator converts mass to copies. For inheritance problems there is the Punnett square calculator. Browse the full free online tools hub for more.
Frequently Asked Questions
Divide the mutant cell concentration by the viable cell concentration, with both derived from colony counts, plating volumes and dilution factors. For example 230 mutants per mL against 1.48 billion viable cells per mL gives a frequency of about 1.55 in 10 million.
Frequency is the proportion of a population that is mutant at one moment. Rate is the probability a cell acquires the mutation per division. An early mutation creates a large clone, inflating the frequency without changing the rate, which is why the two can differ substantially.
If mutational events are Poisson distributed, the proportion of parallel cultures containing no mutants equals e raised to minus m, where m is the mean number of events per culture. So m equals minus the natural log of that proportion, and the rate is m divided by the cells per culture.
When too few or too many cultures come up blank. Published guidance puts the usable window at roughly m between 0.3 and 2.3, meaning about ten to seventy-four percent of cultures should contain no mutants. Outside that range the estimate becomes unstable.
Only as an order-of-magnitude check. The shortcut assumes mutants accumulate linearly with divisions, which clonal expansion prevents, so it can be wrong by a factor of ten. A published rate needs a fluctuation experiment and a proper estimator.
Because the timing of the mutational event within each culture's growth varies. A mutation in an early division is inherited by a huge number of descendants; one in a late division is not. That variance is the signal fluctuation analysis is built on.
Enough that the blank proportion is estimated reliably, which in practice usually means tens of independent cultures rather than a handful. The exact number depends on the expected rate and the estimator being used, and should be settled at the design stage.
This calculator performs standard laboratory arithmetic on values you supply and is provided for education and planning. Published mutation rate estimates require an appropriately designed fluctuation experiment and a statistical treatment suited to the data; the choice of estimator and the interpretation of any result rest with the investigator.
