Transformation Efficiency Calculator — cfu per µg of DNA

Background.

This transformation efficiency calculator turns a colony count into colony-forming units per microgram of DNA, correcting for the fraction of the transformation you actually spread on the plate. That correction is the part people get wrong, and it scales the answer directly: plating a tenth of your recovery means the plate shows a tenth of the transformants, so the count has to be divided by 0.1 before it means anything.

The measurement has a fixed shape. You put a known mass of DNA — usually an intact supercoiled control plasmid such as pUC19, at picogram amounts — into competent cells, heat-shock or electroporate them, add recovery medium, and plate some fraction of the result on selective agar. The efficiency is the total number of transformants across the whole recovery volume, divided by the DNA mass in micrograms. Hanahan's 1983 paper in the Journal of Molecular Biology is where that definition and the optimised chemical method come from; it reports 10⁶ to 10⁸ cfu per µg, and puts it in memorable terms: under the best conditions about one plasmid molecule in every 400 produces a transformed cell. Every efficiency band this calculator prints is attributed to that paper by name and year, because a band without a source is a number someone invented.

The calculator also runs backwards, which is often more useful than the forward direction. Enter a known efficiency from your competent-cell datasheet and it tells you how many colonies to expect, which is how you find out before the experiment that your control plate is going to be a lawn. At 10⁸ cfu/µg with 0.1 ng of DNA and a ten percent plating, you get about a thousand colonies on one plate — uncountable, and a wasted control. The third mode goes the other way and tells you how much DNA to use for a target colony count.

Two caveats belong next to the number rather than in a footnote. First, an efficiency is a property of one batch of competent cells on one day, measured against an intact control plasmid. It is not a constant of the strain and it is not transferable between preparations. Second, and more practically: Addgene records that transformation efficiencies are approximately ten-fold lower for a ligation of insert into vector than for an intact control plasmid. Comparing the colony count from your ligation against the manufacturer's control-plasmid figure and concluding the cells are bad is the commonest misuse of this number.

The arithmetic also assumes you are in the linear regime, where one colony corresponds to one transforming molecule and efficiency does not depend on how much DNA you added. That holds at picogram amounts and breaks down above roughly one to ten nanograms per transformation, where the cells saturate and the apparent efficiency falls. That is precisely why control assays use tiny amounts of DNA, and why a high-DNA transformation cannot be used to characterise a cell batch.

What is transformation efficiency calculator?

Transformation efficiency measures how readily a batch of competent bacterial cells takes up plasmid DNA. It is expressed as colony-forming units per microgram of DNA — the number of transformants you would get if you had used exactly one microgram, extrapolated from however much you actually used.

The underlying biology dates to Mandel and Higa's 1970 discovery that calcium ions allow E. coli to take up bacteriophage DNA, and to Cohen, Chang and Hsu's 1972 demonstration that plasmids could be introduced the same way — the experiment that made molecular cloning possible. Hanahan systematically optimised the chemical method in 1983, working through the effects of magnesium in the growth medium and of manganese, calcium, rubidium, DMSO, dithiothreitol and hexamine cobalt in the competence buffer, and reporting efficiencies of 10⁶ to 10⁸ cfu per µg.

The number is used for two things. As quality control, you transform a known picogram amount of a control plasmid and check that the batch of cells performs the way it should before committing a precious ligation to it. As diagnosis, when a cloning experiment yields nothing, the control tells you whether the cells or the DNA were at fault. Both uses depend on the plating fraction being handled correctly, because the raw colony count is meaningless without knowing what proportion of the transformation it represents.

What the number is not: a fixed property of a strain, a prediction of how a ligation will behave, or anything with a biological meaning beyond the assay. It is a measured, method-dependent value tied to one preparation of cells, one DNA sample and one day's technique.

How to use this calculator.

1. 01Run a control transformation with a known picogram amount of an intact plasmid — 0.1 ng of pUC19 is the usual choice. Do not use a ligation for this; ligations transform about ten-fold worse and will not characterise the cells.
2. 02Record the total recovery volume: cells plus the SOC or LB you added after the heat shock. 50 µL of cells plus 950 µL of SOC is a recovery volume of 1000 µL.
3. 03If you diluted before plating, record that separately. Taking 10 µL of the recovery up to 1 mL is an extra dilution factor of 100. Leave it at 1 if you plated straight from the recovery.
4. 04Enter the volume you actually spread on the plate, and the colony count after overnight growth.
5. 05Read the efficiency, and check the fraction plated shown beside it. If that percentage is not what you expected, the rest of the answer is wrong — it scales the result directly.
6. 06Check the plate was countable. Below about 30 colonies the Poisson error alone exceeds 18 percent; above about 300 the colonies merge and you under-count. Neither gives a trustworthy efficiency.
7. 07Quote the result to at most two significant figures. A count of 200 colonies carries a Poisson error of about 14, so the third digit is not real.
8. 08Use the 'colonies to expect' mode before your next control to choose a plating volume that lands inside the countable window instead of producing a lawn.

The formula.

The calculation has three steps and one trap.

Step one is the plated fraction: f = (volume plated ÷ recovery volume) ÷ extra dilution factor. Plating 100 µL of a 1000 µL recovery with no further dilution gives f = 0.1, or ten percent. Plating 30 µL after taking 10 µL of a 300 µL recovery up to 1 mL gives f = (30 ÷ 300) ÷ 100 = 0.001, which is the same thing as the two-step form (10 ÷ 300) × (30 ÷ 1000). The single extra-dilution field exists so both workflows can be expressed without arithmetic on the reader's part.

Step two scales the plate back up: total transformants = colonies ÷ f. Step three divides by the DNA mass in micrograms: efficiency = total transformants ÷ (ng ÷ 1000).

The trap is that f divides twice over, so it is the term most likely to be silently wrong. If you believe you plated ten percent when you really plated one percent, your efficiency is off by a factor of ten in the direction that flatters you. That is why the fraction plated is reported as its own output.

Work through the loaded default. Two hundred colonies from a 0.1 ng transformation, 1000 µL recovered, 100 µL plated, no extra dilution. The plated fraction is 100 ÷ 1000 = 0.1, so ten percent. Total transformants = 200 ÷ 0.1 = 2000 cfu. The DNA mass is 0.1 ÷ 1000 = 0.0001 µg. The efficiency is 2000 ÷ 0.0001 = 2.0 × 10⁷ cfu per µg — inside the 10⁶ to 10⁸ range Hanahan reported, and typical of well-made homemade chemically competent cells.

Now the inverse, which is the worked example on this page. Suppose your datasheet says 1 × 10⁸ cfu per µg and you plan the same 0.1 ng, 1000 µL, 100 µL setup. Total transformants = 10⁸ × 0.0001 = 10,000 cfu, and the plate would receive ten percent of that: 1000 colonies. That is a lawn, not a countable plate, which is exactly why high-efficiency protocols dilute before plating.

The directions are worth stating explicitly because two of them read backwards at first. More DNA for the same colony count means a lower efficiency, because efficiency is per microgram. Plating less of the same transformation for the same colony count means a higher efficiency, because you found the same number of transformants in a smaller sample. An extra dilution multiplies the efficiency by exactly that factor for the same colony count, for the same reason.

Rounding: every step runs in arbitrary-precision decimal arithmetic, and rounding happens once, at the end, to ten decimal places. The plated fraction is never rounded before it is used as a divisor — rounding a value of 0.001 first would visibly distort the answer. Report your own result to two significant figures.

Invalid inputs are refused rather than fudged. A zero recovery volume is a singularity in the plated fraction and is rejected, as is a zero or negative plated volume or extra dilution factor. Plating more than the recovery volume is rejected with a message pointing you at the extra-dilution field. A DNA mass of zero is rejected in the efficiency mode, because an efficiency per zero DNA is undefined rather than infinite. A plate with zero colonies is accepted, but the interpretation says clearly that this means the efficiency is below the detection limit of that plating — not that it is zero. One plate cannot demonstrate zero.

A worked example.

Frequently asked questions.

References& sources.

1. [1]Hanahan, D. (1983). 'Studies on transformation of Escherichia coli with plasmids.' Journal of Molecular Biology 166(4): 557–580. The paper that defines transformation efficiency as transformants per microgram of plasmid DNA and systematically optimises the chemical method — elevated Mg²⁺ in the growth medium, and Mn²⁺, Ca²⁺, Rb⁺ or K⁺, DMSO, dithiothreitol and hexamine cobalt(III) in the competence buffer. Reports 10⁶ to 10⁸ cfu µg⁻¹ for the optimised method, with the framing that under the best conditions about one plasmid molecule in every 400 produces a transformed cell. Every efficiency band printed by this calculator is attributed to this paper. Paywalled; listed bibliographically. Verified 2026-07-29.
2. [2]Addgene — 'Protocol: Bacterial Transformation.' Source of the DNA-amount guidance quoted on this page — 'Mix 1 - 5 μl of DNA (usually 10 pg - 100 ng) into 20-50 μL of competent cells' — and of the caveat that 'Transformation efficiencies will be approximately 10-fold lower for ligation of inserts to vectors than for an intact control plasmid', which is the basis of the warning against comparing a ligation plate with a control-plasmid figure. Non-profit plasmid repository, free. Retrieved 2026-07-29.
3. [3]New England Biolabs — FAQ, 'How should I calculate the transformation efficiency?' Independent second authority consulted specifically to cross-check the plating-fraction arithmetic, which Hanahan's paper does not spell out for a modern recovery-and-plate workflow. States TE = Colonies / µg / Dilution, with the note that the dilution term exists because you rarely plate the entire transformation mixture, and gives a worked example — 150 colonies from 100 pg of pUC19, 300 µL recovery, 10 µL diluted to 1 mL, 30 µL plated — yielding 1.5 × 10⁹ cfu/µg. This calculator reproduces that example exactly, and it is asserted as a test. Cited WITHOUT a link: neb.com returns HTTP 403 to automated retrieval, so the FAQ was read through a search-engine excerpt rather than fetched and verified directly, and its arithmetic rather than its prose is what was checked.
4. [4]Mandel, M. & Higa, A. (1970). 'Calcium-dependent bacteriophage DNA infection.' Journal of Molecular Biology 53(1): 159–162. doi:10.1016/0022-2836(70)90051-3. The discovery that calcium ions render Escherichia coli competent to take up DNA — the origin of the chemical transformation this calculator quantifies. Paywalled; listed bibliographically.
5. [5]Cohen, S. N., Chang, A. C. Y. & Hsu, L. (1972). 'Nonchromosomal antibiotic resistance in bacteria: genetic transformation of Escherichia coli by R-factor DNA.' Proceedings of the National Academy of Sciences USA 69(8): 2110–2114. doi:10.1073/pnas.69.8.2110. The first demonstration that plasmid DNA could be introduced into E. coli by CaCl₂ treatment and select for its resistance marker — the experiment that made plasmid cloning, and therefore this measurement, possible. Open access via PubMed Central.