Written and maintained by CASRAI Editorial Board
Last updated
Agarose gel electrophoresis is the standard lab method for separating DNA or RNA fragments by size, using an electric current to pull nucleic acids through a porous gel matrix. Smaller fragments move faster and travel further than larger ones in a fixed run time, so after staining, the gel shows a ladder of bands that reveals fragment sizes directly. It is one of the most frequently run techniques in a molecular biology lab — used to check a PCR product, verify a restriction digest, confirm plasmid identity, or assess whether extracted DNA or RNA is intact — and it is usually the first checkpoint before a sample moves on to sequencing, cloning, or a more sensitive downstream assay. This guide covers the full workflow, the practical decisions (agarose percentage, buffer, stain) researchers get asked about most often, and a symptom-by-symptom troubleshooting reference for when a gel doesn’t come out clean.
How It Works: Separating Nucleic Acids by Size
DNA and RNA both carry a negatively charged phosphate backbone, so when a gel is placed in an electric field, nucleic acid fragments migrate toward the positive electrode (the anode). The agarose itself forms a mesh of pores when it sets, and that mesh acts as a molecular sieve: small fragments thread through the pores relatively easily and move quickly, while large fragments have to work harder to squeeze through and move more slowly. Over a fixed run time, this produces size-based separation — fragments spread out into distinct bands ordered by length, with the smallest fragments furthest from the wells and the largest fragments closest to them.
This is different from separation by charge or shape, which is roughly how proteins would behave without help — nucleic acids don’t need an added coating like SDS the way proteins do in SDS-PAGE, because the phosphate backbone already gives every fragment a charge proportional to its length, independent of base sequence.
When to Use Agarose Gel Electrophoresis (and When Not To)
Agarose gel electrophoresis is the right tool for resolving nucleic acid fragments roughly 100 base pairs to 25 kilobases in size — the range covers most PCR products, restriction digests, and plasmid preparations. For separating very small fragments (under ~100 bp, such as short oligonucleotides or certain forensic/STR fragments) or for resolving fragments that differ by only a few base pairs, polyacrylamide gel electrophoresis (PAGE) has finer resolving power because its pore size is smaller and more tunable. For proteins rather than nucleic acids, SDS-PAGE (not agarose) is the standard method, since the separation mechanism and staining/detection chemistry are different — see CASRAI’s western blot protocol guide for how SDS-PAGE fits into that workflow, and CASRAI’s western blot no-bands troubleshooting guide if the gel in question is a protein gel rather than a nucleic acid gel. Agarose is preferred over PAGE for routine nucleic acid work specifically because it is faster and simpler to cast and run, at the cost of somewhat coarser resolution.
What You Need Before You Start
A standard run requires: agarose powder, a running buffer (most commonly TAE — Tris-acetate-EDTA — or TBE, Tris-borate-EDTA), a casting tray and comb to form sample wells, a horizontal electrophoresis chamber with a power supply, a nucleic acid stain (see the stain-comparison section below for the ethidium bromide vs. safer-alternative choice), loading dye to track migration and weigh down the sample in the well, a DNA or RNA size ladder run alongside samples as a reference, and a gel imaging system (UV or blue-light transilluminator with a camera, or a dedicated gel-doc imager). TAE and TBE are not interchangeable mid-protocol: TBE has higher buffering capacity and is generally preferred for longer runs or when fragments need sharper resolution, while TAE is more common for everyday work and for fragments that will be excised from the gel afterward (TBE’s borate can interfere with some downstream enzymatic steps). For calculating buffer molarity from a stock concentrate or a solid reagent, see CASRAI’s guide to molarity and solution calculations for the lab and buffer and solution preparation guide.
Step 1: Choosing the Right Agarose Percentage
Agarose percentage controls pore size, and pore size determines which fragment sizes the gel resolves well. Vendor and protocol-repository guidance (Addgene’s published gel-electrophoresis protocol and miniPCR’s agarose-percentage guide) converge on the same general pattern, though exact cutoffs shift slightly by source — treat the ranges below as practical starting points, not fixed boundaries:
| Agarose % | Best-resolved fragment range | Typical use |
|---|---|---|
| 0.5% | ~1–30 kb | Large plasmids, genomic DNA, long PCR products |
| 0.7% | ~0.8–12 kb | Large plasmids, genomic DNA fragments |
| 1.0% (standard) | ~0.4–10 kb | Most PCR product checks and restriction digests |
| 1.5% | ~0.2–3 kb | Small-to-medium PCR products; separating close-sized fragments |
| 2.0% | ~0.1–2 kb | Small PCR products, primers, small restriction fragments |
Higher percentages give finer resolution of small fragments but run more slowly and produce a gel that is more fragile and prone to tearing; lower percentages run faster and handle large fragments better but blur the distinction between similarly sized bands. For anything smaller than roughly 100–200 bp where agarose starts to lose resolving power even at 2–3%, PAGE is the better-suited technique (see above). When in doubt, 1% is a reasonable default for a first look at an unfamiliar sample.
Step 2: Casting the Gel
Weigh out agarose powder and dissolve it in the chosen running buffer at the target percentage (for example, 1 g of agarose in 100 mL of buffer for a 1% gel), then heat the mixture — typically in a microwave, swirling at intervals — until the agarose fully dissolves and the solution turns clear. Let the molten agarose cool to roughly 50–60°C (cool enough not to warp a plastic casting tray or melt a comb, but well before it starts to set) before adding stain, if the stain is being cast directly into the gel rather than applied after the run. Pour the gel into a casting tray with a comb positioned to form sample wells, and let it solidify fully at room temperature, typically 20–30 minutes, before removing the comb carefully and slowly to avoid tearing the wells — a common, avoidable cause of ragged or leaking wells later in the run (see troubleshooting below).
Step 3: Preparing and Loading Samples
Mix each sample with loading dye before loading — loading dye serves two purposes: it adds density so the sample sinks into the well instead of diffusing into the buffer, and it contains a visible tracking dye that migrates through the gel at a roughly predictable rate, giving a rough visual cue for how far the run has progressed without needing to stop and check under UV or blue light repeatedly. Load a DNA or RNA ladder into at least one well on the gel — without a ladder run alongside the samples, there is no way to estimate fragment size from the finished gel, only relative comparisons between sample lanes. Ladders are sold in different size ranges (for example, a 100 bp ladder for small-fragment work versus a 1 kb ladder for larger fragments), so pick one whose range brackets the fragment sizes actually expected in the samples, and load it at the concentration the manufacturer recommends — an underloaded ladder is one of the most common causes of a “faint bands” complaint that isn’t actually about the samples at all. Load samples carefully and consistently; uneven loading volumes across lanes make band-intensity comparisons between samples unreliable.
Step 4: Running the Gel
Submerge the cast gel in an electrophoresis chamber filled with the same running buffer used to cast it — mixing TAE-cast gels with TBE running buffer, or vice versa, changes the effective ionic strength and produces inconsistent migration. Apply a constant voltage, commonly in the range of about 5–10 volts per centimeter of gel length (in absolute terms, this often works out to roughly 80–150 V on a standard mini-gel rig, consistent with Addgene’s published protocol), and run until the tracking dye has migrated an appropriate distance for the fragment sizes being resolved — typically until the dye front is about 75–80% of the way down the gel. A full run commonly takes 30 minutes to a couple of hours depending on gel percentage, voltage, and fragment size. Running too fast (high voltage) generates heat that can distort bands, especially on longer runs; running too slow wastes time without improving resolution meaningfully once the dye has migrated far enough. As a rule of thumb, a lower voltage over a longer time gives sharper bands than a rushed high-voltage run.
Step 5: Staining and Visualization
Nucleic acids in an unstained gel are invisible to the eye, so a fluorescent dye that intercalates into (or otherwise binds) DNA or RNA is needed to visualize bands under UV or blue light. Stain can be added directly into the molten agarose before casting, into the running buffer, or applied to the gel as a post-run soak — each has trade-offs in sensitivity, background, and how much extra handling time it adds. Which specific dye to use, and the safety and disposal considerations that come with that choice, is significant enough to warrant its own section below.
Step 6: Reading and Interpreting Results
Image the gel under UV or blue-light transillumination and compare each sample band’s migration distance to the ladder to estimate fragment size — the closer a sample band lines up with a specific ladder band, the closer its size is to that reference. Because migration distance is roughly proportional to the log of fragment length rather than to length itself, size differences compress at the large-fragment end of the gel (near the wells) and expand at the small-fragment end — which is part of why very large fragments are harder to distinguish from each other on a standard gel. A single sharp band at the expected size is the typical sign of a clean, successful reaction (a specific PCR product, a completed restriction digest). Multiple bands can mean multiple real products (partial digestion, alternative splice products, non-specific PCR amplification) or contamination, and distinguishing between those explanations usually requires additional evidence beyond the gel itself, such as sequencing or a repeat reaction with adjusted conditions. A diffuse smear rather than a distinct band is a common sign of degraded or sheared nucleic acid, especially for RNA, which is far more prone to degradation than DNA due to the ubiquity of RNases; a smear in an RNA gel is often the first indicator that a sample needs to be re-extracted rather than run forward into a downstream application.
DNA Stains Compared: Performance and Safety
Not all nucleic acid stains carry the same safety and disposal burden, and this is a genuine research-administration and EHS concern, not just a bench-chemistry choice.
| Stain | Sensitivity | Hazard profile | Disposal |
|---|---|---|---|
| Ethidium bromide (EtBr) | High; long-standing default | Intercalates into DNA and is classified as a mutagen; requires dedicated handling (gloves, contained work area, dedicated waste stream) | Almost universally treated as hazardous chemical waste under institutional EHS programs — not disposed of down the drain or in regular trash |
| SYBR Safe | Comparable to EtBr for most routine work; excites well under blue light | Marketed and generally treated as substantially less mutagenic than EtBr | Some institutions permit disposal via standard aqueous waste at low concentrations, but this varies by institution — confirm locally rather than assuming |
| GelRed / GelGreen | High sensitivity, often marketed as comparable to or exceeding EtBr | Marketed as non-mutagenic and membrane-impermeable (reducing cellular uptake) | Same caveat — institutional policy varies and should be confirmed, not assumed, before treating waste as non-hazardous |
Ethidium bromide has historically been the most common stain because it is inexpensive and highly sensitive, but it is a mutagen and requires careful handling and disposal as hazardous waste under most institutional environmental health and safety programs. Safer alternatives — SYBR Safe, GelRed, and similar dyes — are now widely used in place of ethidium bromide because they are designed to be less mutagenic and, in some cases, compatible with standard (non-hazardous) waste disposal. That said, “designed to be less hazardous” is not the same as “exempt from EHS oversight” at every institution: waste-disposal rules vary by dye, concentration, and institution, so confirm the specific policy with institutional EHS rather than assuming a given dye is automatically exempt from hazardous-waste handling. Whichever stain is chosen, UV transillumination itself also carries a real handling consideration — prolonged direct UV exposure to skin and eyes is a hazard independent of the stain, which is part of why many labs have moved to blue-light transilluminators (safer for both the user and, since UV can nick DNA, for downstream applications like cloning that require intact excised fragments).
Extracting and Purifying DNA from a Gel
When a band needs to be recovered for downstream use — cloning, sequencing, or as a template for another reaction — the standard approach is to excise the band with a clean scalpel under the dimmest, shortest visualization exposure practical (a blue-light or long-wave UV transilluminator with a UV-blocking face shield, since short-wave UV can nick DNA and reduce cloning efficiency), then purify it away from the agarose. Most workflows use a silica-membrane spin-column kit: the excised gel slice is dissolved in a chaotropic binding buffer, the solution is passed through a silica membrane that binds DNA under high-salt conditions, contaminants are washed away, and purified DNA is eluted in a small volume of low-salt buffer or water. The result is a workable but not perfectly pure preparation — sufficient for most cloning, restriction digestion, or PCR-template uses, but not a substitute for the higher-purity preparations some sensitive downstream applications (certain library-prep chemistries, for instance) require.
When a Gel Isn’t the Right Assay: Quantification Limits
A stained gel gives a useful but rough estimate of both size and relative quantity — band brightness scales loosely with the amount of nucleic acid present, but reading precise concentration off a gel image is not reliable, and a gel says nothing about sample purity in the way a spectrophotometric ratio does. When precise quantification actually matters — for library prep, for normalizing samples across a large batch, or for any application with a strict input-mass requirement — a fluorometric assay (for example, a Qubit-type fluorometer, which measures dye binding specifically to double- or single-stranded nucleic acid rather than to any UV-absorbing contaminant) gives a far more accurate concentration than either a gel or a UV-absorbance reading. Where fragment-size distribution needs to be characterized precisely rather than just checked (RNA integrity number, library fragment-size distribution before sequencing), a microfluidic capillary electrophoresis instrument such as an Agilent Bioanalyzer or TapeStation is the appropriate tool — it resolves and quantifies size distributions with far finer precision than a stained agarose gel, at the cost of specialized, more expensive instrumentation. The practical rule: use a gel to confirm a reaction worked and fragments are roughly the expected size; use a fluorometer or capillary electrophoresis system when a number needs to go into a calculation or a QC report.
Documentation, Reproducibility, and Image Integrity
Because a gel image is often the primary evidence that an experiment worked, treat the image itself as a data record, not just a bench check. Capture and retain the raw, unedited image file (not just a cropped screenshot), and record the run conditions alongside it — gel percentage, buffer, voltage, run time, ladder used, and sample identities — in a lab notebook or electronic lab notebook (ELN) entry, so the result can be reproduced or re-examined later without guesswork. This matters beyond the individual experiment: published gel and blot images are subject to the same image-integrity expectations as any other figure, and organizations such as the Committee on Publication Ethics (COPE) treat undisclosed splicing, duplication, or other digital manipulation of gel and blot images as a form of image integrity misconduct — a recognized, documented category of post-publication correction and retraction in the literature. If a figure needs to combine lanes from more than one gel or exposure, that must be disclosed explicitly (for example, with a clear dividing line and a note in the figure legend), never presented as a single continuous run. Keeping the original, unedited image on file is the simplest safeguard against this kind of dispute arising later.
Troubleshooting Agarose Gels by Symptom
Most agarose gel problems trace back to one of a small number of causes — heat, a buffer mismatch, sample quality, or a loading/orientation mistake. Work through the symptom that matches what’s actually on the gel.
No Bands at All
Check gel orientation first — DNA migrates toward the positive electrode (anode), so a gel run backward (leads reversed) shows nothing in the expected lanes, with sample instead having migrated out of the well toward the negative electrode or not migrated meaningfully at all. If orientation is correct, confirm stain was actually present in or applied to the gel (an omitted staining step is a surprisingly common cause of a genuinely blank image), and confirm the sample itself contains enough nucleic acid to detect — a failed PCR reaction or a very dilute extraction can produce a gel with nothing to see because there is genuinely nothing there.
Faint or Weak Bands
Distinguish this from “no bands”: a faint band means something is present but under-detected. Common causes: too little sample loaded, a low-yield reaction (dilute PCR product or a low-efficiency digest), insufficient stain concentration or exposure time, or an imaging system set to too low a sensitivity/exposure. If the ladder itself is also faint, check ladder loading volume and concentration against the manufacturer’s recommendation before concluding anything about the sample lanes.
Smeared Bands
A smear instead of a sharp band most often indicates degraded or sheared nucleic acid — RNA is especially prone to this because RNases are extremely common and hard to fully eliminate from a lab environment, so a smeared RNA gel is often a re-extraction problem rather than a gel-running problem. Overloading a well (too much sample forced through too little space) and running the gel too hot (too high a voltage for too long, generating heat that distorts the matrix) can both also produce smearing even with an intact sample, so rule those out before assuming degradation.
Distorted or “Smiling” Bands
Bands that curve upward at the outer lanes relative to the center (“smiling”) are almost always a heat or buffer problem: uneven heat distribution across the gel during the run, most often from running at too high a voltage, an inadequate buffer volume covering the gel, or an inconsistent buffer level between the two electrode chambers. Lowering the voltage and confirming the buffer fully and evenly covers the gel resolves most cases.
Bands at an Unexpected Size
A single band that runs at a different size than expected can indicate non-specific PCR amplification (wrong or degenerate primer binding), incomplete or off-target restriction digestion, or genuine biological variation (an unexpected splice variant, insertion, or deletion) — treat this as a real result worth investigating rather than a running error to explain away. Confirm the ladder was read correctly first (an easy source of apparent size error is simply misreading which ladder band corresponds to which size), then consider sequencing or a repeat reaction with adjusted conditions to confirm the finding.
Extra or Unexpected Bands
Additional bands beyond the expected single product usually point to one of: primer-dimer formation (very small, faint extra bands near the dye front in a PCR gel), incomplete restriction digestion (a partial-digest band alongside the fully cut product), non-specific PCR amplification, or cross-contamination between samples or from a shared reagent. Include a no-template (negative) control on the gel routinely — a band appearing in a no-template lane is the clearest evidence of contamination rather than a genuine result.
Wells Tearing, Leaking, or Sample Floating Out
Torn or leaking wells are usually a casting problem: the comb was removed too early (before the gel fully solidified), removed too quickly or at an angle, or the gel percentage is high enough (2%+) that the set gel is naturally more fragile and needs gentler handling. Sample floating up out of the well rather than settling usually means too little loading dye/glycerol was used relative to sample volume, so the sample isn’t dense enough to sink and stay put — or the well itself was damaged during comb removal and no longer holds liquid cleanly. Recasting with a slower, more careful comb removal after full solidification, and confirming loading dye ratio, resolves most of these cases.
Agarose Gel Electrophoresis vs. Related Techniques
Agarose gel electrophoresis is sometimes confused with related separation methods that serve different purposes. Polyacrylamide gel electrophoresis (PAGE) uses a synthetic polymer matrix instead of agarose and offers finer resolution, making it the better choice for very small nucleic acid fragments or for protein separation (as SDS-PAGE, the first stage of a western blot). Capillary electrophoresis runs the separation inside a thin capillary tube rather than a flat gel slab, giving higher resolution and enabling automated, quantitative sizing (used in DNA sequencing and fragment-length analysis such as microsatellite genotyping), at the cost of specialized instrumentation that a standard gel rig doesn’t require. Choosing among these comes down to what the experiment needs: agarose for routine, fast, low-cost checks of nucleic acid fragments in the common size range; PAGE for finer resolution or protein separation; capillary electrophoresis for automated, high-precision sizing.
Frequently Asked Questions
What is agarose gel electrophoresis?
It is a laboratory technique that separates DNA or RNA fragments by size by pulling them through a porous agarose gel matrix using an electric current — smaller fragments move faster and travel further than larger fragments in a given run time, producing a pattern of bands that reveals fragment sizes when compared against a size ladder.
Why does DNA migrate toward the positive electrode?
DNA’s phosphate backbone carries a negative charge along its entire length, so in an electric field the molecule is pulled toward the positive electrode (the anode). This charge is roughly proportional to fragment length, which is part of why migration through the gel’s pores — not raw charge alone — ends up being the dominant factor in size-based separation.
What’s the difference between TAE and TBE buffer?
Both are Tris-based buffers used to run and sometimes cast agarose gels. TBE (Tris-borate-EDTA) has higher buffering capacity, making it better suited to longer runs and sharper resolution, while TAE (Tris-acetate-EDTA) is more common for everyday, shorter runs and for samples that will be excised from the gel afterward, since TBE’s borate can interfere with some downstream enzymatic reactions.
How long does an agarose gel run take?
Typically 30 minutes to a couple of hours, depending on gel percentage, applied voltage, and how far the fragments need to migrate to separate clearly — a quick check of a PCR product on a 1% gel at a moderate voltage often takes 30–45 minutes, while resolving closely sized fragments on a higher-percentage gel can take longer.
Why is my DNA or RNA smearing instead of forming a clean band?
A smear instead of a sharp band most often indicates degraded or sheared nucleic acid — RNA is especially prone to this because RNases are extremely common and hard to fully eliminate from a lab environment. Overloading the well or running the gel too hot (too high a voltage for too long) can also produce smearing even with intact sample.
Why are my gel bands “smiling” instead of running straight?
Smiling bands are almost always a heat-distribution or buffer problem — most commonly too high a voltage, uneven buffer coverage over the gel, or an inconsistent buffer level between the electrode chambers. Lowering the voltage and confirming even buffer coverage resolves most cases.
Is ethidium bromide still used for staining agarose gels?
It is still used in some labs because it is inexpensive and sensitive, but many labs have shifted to alternatives such as SYBR Safe or GelRed, which are designed to be less mutagenic and, depending on institutional policy, may allow simpler waste disposal. Whichever stain is used, always follow the specific institutional EHS guidance for handling and disposal rather than assuming a given dye is exempt from hazardous-waste rules.
Can a gel image be cropped or spliced together for a publication figure?
Only with explicit disclosure. Combining lanes from different gels or exposures into a single figure without a clear dividing line and a note in the legend is treated as an image-integrity problem by publication-ethics bodies such as COPE, and undisclosed gel or blot splicing is a documented, recognized cause of post-publication corrections and retractions. Keep the original, unedited image on file.
Why did my sample float out of the well instead of staying put?
This usually means the sample wasn’t dense enough to sink and stay in the well — check that enough loading dye (which contains glycerol or a similar density agent) was mixed in relative to sample volume, and check that the well itself wasn’t torn or damaged during comb removal.
For the concentration and dilution math behind preparing running buffer or diluting a stock reagent to a working molarity, see CASRAI’s guide to molarity and solution calculations for the lab, and for the serial-dilution technique often used to prepare a dilution series of a sample or standard, see CASRAI’s serial dilution guide. For a protein-gel counterpart to this workflow, see CASRAI’s western blot protocol guide and its dedicated no-bands troubleshooting decision tree.
