Edge effects on a microplate

Edge effects on a microplate are a layout problem that masquerades as biology. The outer ring of a 96-well plate loses water faster than the inner wells, and it often sits at a slightly different temperature, because it faces the room, the lid gap and the metal of an incubator shelf. If your standards live in the middle and your treated samples live on the rim, you have compared two microclimates. The decision this page supports is practical: when you specify plates, seals and a reader, and when you draw a plate map, you decide whether an edge well is allowed to count. Format, blocking and the standard curve are set out in ELISA formats, controls and readout. This page is about the physical bias around the rim.

Who is choosing the plate, and what the rim changes

A scientist running one plate can rotate samples and notice the pattern. A buyer writing a specification has a different job. Plate flatness, well volume class, optical clarity and whether a seal actually fits the skirt are purchasing points. Evaporation and temperature are handling points. Mixing those two conversations produces a false contest between brands. A flatter plate can improve how evenly a reader illuminates the wells. It does not keep water inside an unsealed well during a two-hour incubation at 37 degrees Celsius. Ask for both, and test both.

A reader reports the liquid that is in the well when the drawer closes. If the rim evaporated during binding, the chemistry has already changed. Specify the incubation hardware and the seal in the same note as the wavelength. A method can be discussed against that note. A catalogue line for a plate is not evidence that your incubation will be uniform.

Why the outer wells are a different vessel

Water leaves a well at the air-liquid surface. Outer wells have more of their wall exposed to moving air, and the lid, if you use only a loose lid, sits above a gap that is drier at the perimeter. In a dry, air-conditioned room the gradient is obvious. In a hot laboratory the same gradient appears for a different reason: the plate is warmer, the vapour pressure is higher, and the rim still loses water first. Concentration of every dissolved reagent rises as volume falls. Capture antibody, analyte, detection antibody and enzyme conjugate all become more concentrated if they are still in the well while water leaves. Binding rates depend on concentration. A rim well can therefore develop more bound enzyme than an inner well that started identical.

The read adds a second physical effect. Colorimetric optical density depends on path length as well as on the concentration of coloured product. A well that has lost a large fraction of its volume has a shorter liquid column. Concentration pushes the signal up. A shorter path pushes it down. Which one wins depends on when the evaporation happened, whether you topped up before reading, and the substrate instructions you followed. Do not assume a single direction for every protocol. The useful habit is to measure the pattern with identical material in edge and centre wells, under the incubation you actually use.

Temperature moves with the same geography. Wells over a cold metal shelf, or beside a door that opens, equilibrate differently from wells in the middle of a stack. Enzyme substrates are timed reactions. A rim that is a degree or two warmer reaches the stop condition with a different colour even when binding was equal. Follow the substrate sheet for temperature and time. If the sheet says room temperature, record what room temperature was, because an afternoon bench is not the morning bench.

Nonspecific binding can join the pattern. A well that dried at the rim during a coating or blocking step exposes plastic that the blocker never covered evenly. Detection antibody then sticks to plastic, which is the same failure mode as a poor block, confined to the wells that dried. The primary antibody is not at fault merely because the rim is loud. Check a no-analyte well on the rim before you dilute the antibody again.

Seals, humid boxes, empty wells and the wells you refuse

Four handling classes are worth comparing before you blame a reagent lot. They are not interchangeable, and none of them is a universal cure.

An adhesive plate seal cuts the air path. It works when it truly adheres to the plastic you bought and when you do not puncture it unevenly while checking wells. A seal that is peeled and reapplied mid-incubation often leaks at the corners first, which recreates the edge pattern you were trying to remove. Confirm that the seal is optically clear if it will stay on during a read, or plan to remove it cleanly. A film the reader cannot see through is a different failure, taken up when a whole plate looks blank.

A rigid lid reduces dust and large spills. It is a weak evaporation barrier unless the fit is tight and the incubation is short. Lids also drip condensate back into random wells if the plate cools under a warmer lid. That is not an edge effect in the classic sense, but it is a position effect, and it ruins replicates in whatever wells received the drops.

A humid chamber is a closed box with a wet paper or a reservoir of water, kept with the plate so the air above the wells is closer to saturation. Many laboratories use this for long coatings in a refrigerator or for incubations that must not be sealed because a later step needs access. The chamber fails when the towel dries out, when the box is opened every few minutes, or when the plate sits on a dry rack above the water rather than in humid air. Check the towel at the end. A dry towel means the chamber was a box.

Filling unused wells with the incubation buffer changes the microclimate of the working wells next to them. An empty well is a dry neighbour. This is cheap and often effective, and it is still a test, not a theorem. Fill with the same buffer, not with a random liquid that wicks under a seal. Do not fill with substrate or with a solution that will be aspirated into a washer and contaminate the next plate.

The fourth choice is to abandon the outer ring for anything you will interpret. Columns 1 and 12 and rows A and H become buffer or nothing. You lose wells. You buy a map that does not ask the rim to impersonate the centre. For a small study this is often the most honest specification. For a screen that needs every well, you then owe yourself seals, humidity and a uniformity plate.

A systematic short dispense on one side of the plate looks like an edge effect and is a hand effect. Accurate micropipetting is the volume check before you redesign the incubator.

A workflow that branches when the rim disagrees

Coat, block, bind and detect as the format requires, with times taken from the reagent instructions rather than from memory. On the first plate of a new incubation length, place the same control material in at least four inner wells and four edge wells. Read them before you interpret any treatment.

If edge and centre agree within the scatter you already accept for replicates, keep the map and record the seal and the room conditions that produced the agreement. If the edge is consistently higher or lower, do not average it into the centre and call the mean a biological replicate. Branch instead. Repeat once with a seal or a humid chamber. If the pattern shrinks, adopt that handling and re-run the samples. If the pattern remains, leave the outer ring out of the analysis and say so in the plate record. A third branch is optical. If the disagreement appears only after substrate, and the volumes visibly differ, the binding step may have been fine and the read volume was not. Top up only if the substrate protocol allows a fixed final volume, and record that you did.

Washer residual volume can fake the same picture. Outer wells sometimes retain more or less wash buffer when the aspirator head is misaligned. Realign before you change the primary antibody dilution.

protocols.io is a place to see plate maps written down before loading. Promega protocols illustrate enzyme timing. The bottle you opened still sets the clock.

What changes between rim and centre

Use a uniformity plate, not a treated experiment, to fill a table like this. The direction of the signal change is something you observe. It is not a constant you can copy from another laboratory's room.

When the pattern is the finding

A single outlier on the corner can be a bubble or a short pipette. A whole column that marches upward from the centre is a position effect until proven otherwise. Do not delete edge wells quietly after you have seen the treatment labels. Pre-commit: either the uniformity plate passed and edges stay in, or edges were never part of the design.

Stacked plates add a vertical version of the same problem. The top plate faces the door and the bottom plate faces the shelf. Record which plate was where if you must stack. A stopped plate left open on a dry bench still evaporates before a free reader. Read promptly, and note the delay.

Safety, and the limit of a research plate

Plate seals and stop solutions need the hazard notes on their bottles. Acid stop reagents are corrosive. Some antibody diluents contain preservatives that are toxic if misused. Human or animal specimens follow the biosafety rules of your institution. The WHO Laboratory Biosafety Manual is a public reference for that decision. It does not turn a research plate into a diagnostic test, and a uniform rim does not either. You may claim a comparison among wells that shared a climate. You may not claim a clinical concentration because the edge happened to look tidy.

NIST is a public reminder to write a uniformity difference as an observation with a method, not as a property of a brand. Citing that habit is not a claim that your plate is a national reference.

Dry conditioned air, a hot bench, and a reader queue

Air conditioning that makes a laboratory comfortable also dries the air above an open plate. A humid city outside the building does not humidify the well. If you move a plate from a refrigerator coating step onto a sunny windowsill to warm up, the rim warms first and the centre lags, which is an edge effect in temperature before any evaporation shows. Keep incubations in the place the protocol names.

Power cuts matter when a 37 degree Celsius incubation stops halfway and the room is hotter or cooler than the cabinet. The plate does not pause its chemistry evenly. Record the interruption, and do not pool that plate with an uninterrupted run. If the reader queue is long because several groups share one instrument, develop only when a slot is actually free, or use a stop reagent the substrate sheet allows so the colour waits. An open developed plate in a dry room will start a new edge effect while it waits.

What to put on a plate and reader enquiry

State the well format, the read mode you need, whether incubations are long enough to need a seal or a humid chamber, and whether the outer ring must be usable. Name the sample matrix and the assay format so a plate binding capacity is discussed against a real coat, not against a slogan. The reagents and chemicals catalogue is a starting list for plates, buffers and related reagent classes. The molecular biology pathway places the assay among other bench methods. Use the quote request to ask whether a quotation is possible and to describe the incubation conditions you will actually run. A method can be discussed from that description. Ask, in the reply you want, for well volume class, seal compatibility and any flatness note the manufacturer publishes. Do not treat a quotation as a statement that the rim will behave.