Longitudinal flow cytometry studies introduce a problem that single-timepoint experiments do not have: the gating strategy must remain consistent across weeks or months of data collection, often with different analysts running different timepoints. When it does not, population shifts that appear to reflect biology may actually reflect a gate that moved two pixels to the left between week three and week four. This article describes the practices we use to prevent that.

Lock the gating template at timepoint one

Before the week-one data is delivered to the client, a senior analyst reviews and approves the gating hierarchy. The approved workspace is saved as a locked template with a version number and a date. Every subsequent timepoint is gated by applying this template to the new FCS files, not by drawing gates from scratch. The template is only modified if there is a documented biological or technical reason, and any modification is versioned separately.

Use the same instrument configuration for every timepoint

Instrument configuration changes, including laser power adjustments, detector filter changes, or software updates, can shift population positions in scatter plots even when the biology has not changed. For longitudinal studies, we freeze the instrument configuration at the start of the study and document it in the study record. If a configuration change is unavoidable, we run a bridging experiment with samples from the previous timepoint to quantify the effect.

Detecting drift before it affects the data

We compare the position of stable reference populations (typically CD45+ leukocytes or a bead population) across timepoints as a drift check. If the reference population shifts by more than five percent of the axis range between timepoints, we investigate before releasing the data. This check takes approximately fifteen minutes per timepoint and has caught three gating drift events in the past two years.

Documenting deviations

Any deviation from the locked template, whether a gate adjustment, a re-run, or a configuration change, is documented in the study record with a date, a reason, and the analyst's name. This documentation is delivered to the client alongside the final dataset. It allows the client to account for the deviation in their statistical analysis and to report it accurately in a methods section.

What to do when a timepoint has to be re-run

Re-runs happen. Samples degrade, instruments have unexpected downtime, or a compensation error is identified after acquisition. When a timepoint is re-run, we use the locked template from timepoint one and run a side-by-side comparison of the original and re-run data where original data is available. The comparison is included in the data delivery. We do not simply replace the original file without documentation.

If you are planning a longitudinal study and want to discuss gating strategy before the first timepoint, a panel design consultation is the right starting point. We can review the experimental design and agree the template locking procedure before any samples are run.