Recording instrument settings with the experiment

A result that cannot be repeated from the written voltages, the matrix, the lot and the gate hierarchy is a screenshot, not an experiment. The use this page describes is a second person, six months later, trying to acquire the same panel or the same field. The limit of the claim is that a tidy plot does not contain those settings unless you stored them. How the settings become a cytometer event is in how a flow cytometer reads a cell. How display choices affect an image is in fluorescence microscopy without the myths.

The instruments that need this habit are in the scientific instruments catalogue. The assay may sit in the molecular biology pathway. Asking whether a quoted system stores the settings in the file is a quote request question.

The cytometer row

Write one row per acquisition, not one row per project. The row names the sample, the instrument, and the software version. It lists each detector's voltage or gain. It lists the threshold and the channel that triggered it. It points to the compensation matrix and to the single-stain files, acquired at those same voltages, bright enough to match the sample. It names the gate hierarchy: time, scatter, doublet, viability, lineage, question, in the order you actually used, with each parent identified.

Antibody identity sits in the same row or in a linked table: clone, fluorophore, lot, and whether a tandem was protected from light. A new lot is a new row's worth of single stains. Event rate and dilution belong here too, because a rate above the instrument's comfort inflates doublets and makes the gate mean something else. If you changed a voltage mid-study, the later files are not "the same settings". Say so.

FCS keywords can hold much of this if the software writes them. Check once that a second program can see the stain names and the spillover. Gates often live only in the workspace. Archive the workspace. The reporting culture around those files is associated with the International Society for Advancement of Cytometry.

The microscope row

Write the objective, numerical aperture, immersion and collar position. Write the cube or the laser line and the emission band. Write exposure, gain, offset, binning and the resulting micrometres per pixel. Write pinhole and z step when you section. Write whether the lookup table in the figure is linear. A reader who has the image and not this row cannot tell bleaching and biology apart, and cannot rebuild the scale bar.

The same discipline fits other instruments that emit a file and a family name. A sequencer's run folder is useful when it contains the kit identity, the cycles and the sample sheet, and decorative when it contains only a green light in a photograph. Keep the cytometry and microscopy rows to the measurements above. Do not substitute a picture of the bench for the numbers.

Filter and illumination background for the microscope row is discussed in the MicroscopyU fluorescence section. The section does not store your exposure. You do.

What people drop, and the branch when they do

If the matrix is missing, you can recompute it only when the single stains still exist and the voltages match. If the single stains are missing, the matrix is an untested number. Say that the compensation cannot be audited. Do not invent coefficients to match a figure. If the gate parents are missing, report the events as an illustration and do not pretend the percentage is defined. If the lot is missing, the next order of the antibody is a new method, and you should bridge it with a shared sample before you pool the data.

If two operators used two voltages, split the analysis or reacquire. Averaging the medians across voltage regimes mixes scaling with biology.

Failure modes

A shared spreadsheet that records "OK" in the voltage column is not a record. Numbers are the record. A screenshot of a gated plot without the parent percentage hides how much was thrown away. A matrix emailed without the FCS files cannot be reapplied with confidence. Two microscopes with the same objective magnification and different cameras need two pixel sizes. Copying one scale bar across both is a recording failure that becomes a false distance.

After a power cut, write a new row. Lasers and LEDs do not owe you the same output when they return, and a hot room may have changed the cells while you waited. In a building where cuts are ordinary, the notebook habit matters more, not less. Do not splice the before and after into one median and one exposure series.

Bridge samples belong in the record when anything in the row changes. A new tandem lot, a replaced laser, a software upgrade or a different camera each deserve a shared sample run under the old condition and the new one before the series is pooled. If that shared sample moves by more than the usual run-to-run wobble of the assay, split the series or reacquire. Do not rescale old files quietly and leave the row looking continuous. The row is how a later reader knows an instrument event happened. Without it, the shift is indistinguishable from biology.

Safety and research limits

A complete record does not authorise the sample. Unfixed human material remains an institutional decision, with the WHO Laboratory biosafety manual and the CDC BMBL as references. This page does not approve it. Recorded research settings are not a clinical protocol. If the file contains donor identifiers, the record is also a privacy matter for the institution. Keep identifiers where that policy says they belong.

What to send with an enquiry

Ask whether voltages, threshold, spillover, pixel size, exposure and gain are written into files a second program can open. State that you will archive gates and lots beside those files. Use the scientific instruments catalogue and the quote request. Ask whether a quotation is possible. An instrument that can only show a live plot, and cannot hand you the settings, cannot support this experiment record.