What drives the DNA
Every nucleotide in DNA carries one negative charge on its phosphate, so in an electric field DNA moves toward the positive electrode (the red one; remember "run to red"). In free solution both the charge and the friction grow in step with length, so a 100 bp fragment and a 10,000 bp fragment would move at the same speed. The gel is what separates them: agarose sets into a mesh of pores, and longer molecules are held back more as they snake through it.
What decides where a band ends up
- Fragment length. Over the useful range of a gel, the distance moved falls roughly in a straight line with the logarithm of the length. Doubling the size moves a band back by about the same distance anywhere in that range.
- Agarose percentage. A denser gel has smaller pores. Roughly: 0.7% for 5 to 20 kb, 1% for 0.5 to 10 kb, 1.5% for 0.2 to 3 kb, 2 to 3% for 50 bp to 1 kb.
- Field strength. Speed is proportional to the voltage per centimeter between the electrodes, not the voltage itself. 5 to 8 V/cm is usual; a mini gel in a tank with electrodes 15 cm apart runs at about 100 V.
- Shape. Supercoiled plasmid is compact and runs ahead of linear DNA of the same length; nicked (open circle) plasmid drags and runs behind; RNA and single strands fold up and run unpredictably.
- Buffer and dye. TAE separates large fragments a little better, TBE small ones. Ethidium bromide slows DNA slightly, and loading dyes run with fragments of a size that depends on the gel: bromophenol blue with about 300 to 500 bp in 1% agarose, xylene cyanol with about 3 to 4 kb.
Sizing a band from the ladder
A ladder is a mix of fragments of known size run in its own lane. Measure the distance from the well to each ladder band, plot log10(size) against distance, and fit a straight line: the standard curve. Read an unknown band's size off the line.
Worked example: in a 1% gel, the 3,000 bp ladder band moved 2.6 cm and the 1,000 bp band 4.0 cm. The line through them has slope (log10 1,000 − log10 3,000) ÷ (4.0 − 2.6) = −0.477 ÷ 1.4 = −0.341 per cm. An unknown band at 3.3 cm is at log10(size) = log10 3,000 − 0.341 × 0.7 = 3.477 − 0.239 = 3.238, so its size is 103.238 ≈ 1,730 bp. Use the two ladder bands either side of the unknown, or a line fitted through several, and do not read off past the last ladder band.
Reading common results
- Uncut plasmid: two or three bands, supercoiled lowest, nicked highest, sometimes linear between. None of them sits at the plasmid's true size on a linear ladder.
- Linearized plasmid (one cut): one band at the full length, a good check that a single enzyme cut.
- Double digest: fragments that should add up to the plasmid size. If they do not, a small fragment may have run off the gel or be too faint to see, or two fragments of similar size may sit in one band.
- PCR product: one sharp band at the expected size. A smear means too much template or too many cycles; extra bands, primers binding in the wrong places; a fuzzy band near the bottom, primer dimers.
Using the simulation
In the Gel Electrophoresis Simulator lanes 1 to 6 hold a ladder, an uncut plasmid, the plasmid cut once and twice, a PCR product and sizes you type. The run starts at once. Stop when the bromophenol blue front is about three quarters of the way down, then read the sizes under the picture, each estimated from a line fitted to the ladder bands still on the gel. Things to try:
- Set lane 6 to 100, 150, 200 and run a 1% gel: the three run close together, just ahead of the dye front. Rerun at 2.5% and they separate.
- Set lane 6 to 20000, 30000, 48000 in 1% agarose: they pile up in one band, the limiting mobility.
- Turn the voltage to 250 V: the run is fast, but the bands blur and the middle lanes run ahead of the edges (smiling) as the gel heats.
- Let the run go on: the small fragments and the dye run off the end and are lost.
What the model assumes
- An empirical mobility formula, μ = μ0 ÷ (1 + (L ÷ L*)0.9) with L* = 2,000 bp × (1% ÷ T)1.4, chosen to give realistic positions for 0.5 to 3% agarose in TAE. It is a fit, not a theory of how DNA moves through pores, and its numbers are typical rather than exact for any brand of agarose.
- A limiting mobility for very long DNA, below which no fragment slows; in 1% agarose it is reached at about 20 kb.
- Fixed shape factors for plasmids: supercoiled DNA runs like linear DNA 0.6 times its length, nicked like 1.8 times. In reality the factors change with the gel percentage, the voltage, the amount of supercoiling and whether ethidium bromide is in the gel.
- Equal amounts of each fragment in a digest (equimolar), so band brightness is proportional to length, and 300 ng of DNA per sample lane. A band needs about 2 ng to show.
- A uniform field and temperature across the gel, 15 cm between the electrodes and a buffer resistance of 1.6 kΩ, except for a simple heating penalty above 10 V/cm.
- Band width from diffusion plus the width of the well; overloading, salt in the sample and uneven wells, which all distort real bands, are left out.
Edge cases
- Fragments outside the ladder get a question mark: the line is extrapolated beyond its data and can be badly off.
- Two fragments of nearly the same size run as one band, twice as bright. Only the brightness gives it away; a longer run or a denser gel can split them.
- Very small fragments (under about 100 bp in 1% agarose) run with the dye front and are hidden by it, and very faint bands below about 2 ng never show.
- A fragment longer than the plasmid in a double digest is impossible; if the cut distance is set larger than the plasmid allows, the simulation clamps it.
- Running past the end: shortly after the dye leaves the gel the run stops, and the bands that have left cannot be sized.
Where the model stops being right
- Above about 20 to 50 kb DNA moves by reptation, snaking end first, and all long fragments run together. Separating them needs pulsed-field gel electrophoresis, which the model cannot describe.
- Polyacrylamide gels separate fragments differing by a single base, as in sequencing; agarose cannot, and the formula does not apply to them.
- Proteins are separated on SDS-PAGE gels, by a different mechanism (SDS coats them with a charge proportional to length); use the SDS-PAGE Gel Calculator for those.
- Single-stranded DNA and RNA fold into shapes that change their speed; they need denaturing gels to run by size.
- Real gels are not uniform: uneven heating, a sloping gel, salt or too much DNA in a lane all bend and smear bands in ways a reader learns to spot and a model does not show.
Calculators
Pour the gel with the Agarose Gel Calculator (and its guide, How to Make an Agarose Gel), size bands from your own measurements with the Gel Band Size Estimator, and predict the fragments of a digest with the Restriction Map.