Flow Files and Templates
Working across many files, grouping them into conditions, and reusing a gating strategy you've already built.
One Gating Strategy, Many Files
An analysis has a single set of gates that applies to every file in it. Gates are matched to each file by channel name, not by channel position, so a panel acquired with the detectors in a different order still lines up.
When staining varies enough between tubes that one boundary won't do, you can move a gate on a single file without moving it everywhere. With more than one file loaded, Customize for this file on the gate's ⋮ menu unlinks that gate on the file you're viewing, and dragging it from then on adjusts only that copy. The gate draws with a dashed outline and its label reads · this file, so a count that differs from the other tubes reads as your decision rather than a fault. Reset to shared puts it back on the shared boundary and discards the tailored one.
Files exist in three states, and the difference matters when you have a lot of them:
- Attached to the analysis, so they're part of it
- In workspace, meaning loaded into memory and ready to plot
- Staged, meaning attached but not currently loaded
Manage (from the ⚙ on the Files section) is where you attach files from your datasets, upload new ones, and choose which are loaded. The header keeps a running count of each.
Click a file in the sidebar to make it the active one, or press Shift + ← and Shift + → to walk through them.
Tip: Nothing caps how much data an analysis can cover. A staged file costs no memory while it waits, so a study far larger than your machine's RAM still works: load what you're looking at and leave the rest staged. Conspecta brings up to 200 files into the workspace at a time, moves the ones you haven't touched in a while out to browser storage, and reads them back when you return to them. Reopening a big analysis restores what fits in 200 files or 5 GB and tells you how many it left staged.
Conditions
A condition labels what a file is: treated or untreated, day 0 or day 7, patient or control. Create them with + New condition in Manage files and assign each file to one.
Conditions do real work rather than just tidying the list. They band the aggregate statistics table, they're the default way the compare view groups plots, and they're what makes a significance test possible, since a test needs to know which files belong to which arm.
A file belongs to at most one condition, and the rest sit under Unassigned.
Comparing Across Files
There are three ways to look at more than one file, and they answer different questions.
Compare all files (C, or the grid button on a plot) takes one plot and draws it for every loaded file at once. Group the grid by condition, tube name, experiment, cytometer, acquisition date, or a segment of the filename, then group again within that. Display settings control whether gates, population statistics, axis labels, and filenames appear on each panel. File names come off the panels as soon as the arrangement names them for you, either a row and column grid or a grouping with one file under each label, and the panels take that room back for the plot. Turn File names back on under Display if you want both, and the choice sticks while you regroup. A grid holds one file per cell, so replicates and files missing one of the two values stay out of it, and the count next to the grouping controls tells you how many. Panels size themselves to fill the window; step them bigger when you want to read one closely, or back to Fit to see the whole set again. This is the visual check: does the gate sit right on all of them.
Compare in the Files menu applies your gate tree across a set of files you pick and lands on a Batch Statistics table of counts and percentages. It needs at least one gate and two files with data loaded. Files where a gate couldn't be placed are reported individually rather than silently skipped.
Statistics in the same menu opens Aggregate Statistics, which is the quantitative comparison. Rows are files, columns are populations, and cells are percentages of parent. A quadrant gets a column for itself and one for each of its four corners, so a double-positive frequency compares across files like any other population. With conditions assigned it does more:
- Each condition's band ends with a mean ± standard deviation row
- A cell more than two standard deviations from its band's mean is flagged, which is how an outlier file announces itself
- With exactly two named conditions and at least two files in each, a Welch's t-test row appears, with p-values corrected for multiple comparisons and significance marked
- Below the table, a strip of per-gate dot plots shows every file as a point, ordered so the most significant gate comes first
From there, Export CSV or Send comparison to a Data Table.
Gating Templates
A template is a gating strategy saved for reuse: its gates, the plots they sit on, and its compensation matrix.
Templates in the sidebar opens the library, and everything you do with templates happens there. Search it, or work from the recently used group at the top. Save current gating as a template is pinned above the list once you have at least one gate drawn, so saving and applying live in the same place. Name it, describe it, and Conspecta shows exactly what it's capturing.
A template can also come from another program. Drop a FlowJo .wsp workspace or a Gating-ML file (.xml or .gml) anywhere on the library, or use Import gating at the bottom of it, and Conspecta reads the file in your browser, shows you the gating it found, and names any population it couldn't rebuild. Confirm, and each sample in the file becomes its own template, carrying the axis scales and compensation matrix it was drawn with. Nothing here is checked against data. The counts on screen are the ones the source file recorded. To have Conspecta run the strategy on your own files and compare population by population before you keep it, use Import gating on the dataset holding those files instead.
The list already knows what fits. A template drawn on a color your open files never recorded is flagged on its row with the missing channel, and a row with no flag applies cleanly. The filter icon at the right end of the search box narrows the list: to templates that fit your open files, to approved ones, or to the ones you made, and the choices combine. Opening a template shows what it contains and who made it, across three tabs:
- Details shows its gates, its plots, and whether it fits the files you have open. The description sits at the top, and you can write or change it here at any time, whether or not one was set when the template was saved. The ⋮ menu holds two ways out: Export as Gating-ML writes the template as an ISAC-standard
.xmlfile that FlowJo, Cytobank, and the RCytoMLstack read, and Export as FlowJo .wsp writes a FlowJo 10 workspace instead, for collaborators who live in FlowJo. Either way, gates the format can't state are left out and named on screen with the reason. - Connections maps what the template is connected to. Every analysis built from it appears here, and clicking through to one opens it. Underneath the map, the same analyses are listed with the template version each was built on, so one still pinned to an older version is marked behind latest. From any item on the map you can re-center on it and keep going, which is how you get from a template to the files an analysis read.
- Discussion holds comments and sign-off, which is how a lab agrees on a standard panel.
Applying One
Before it applies anything, Conspecta checks the template against your files and tells you what will happen. Each gate is reported as fine, blocked because its channel is missing, or dropped because its parent was. Missing channels are listed per file, across every file in the analysis, not just the active one. Gates that can't be placed are left out and named.
Then you choose:
- Keep both adds the template's gates alongside what you already have. This is the default.
- Replace existing gates deletes your current gates and everything gated beneath them first.
Replace is destructive, and a template carrying compensation overwrites the session's either way. Undo still works until the analysis saves, but don't count on it. If you're unsure, keep both and delete what you don't want.
Templates are shared with everyone on the project, and there's no private one. Members with view-only access can open the library, read the discussion, and export a template. Applying, saving, and editing stay with members who can edit. You can manage the library without opening an analysis, from Templates in the Analysis tab's header.
When a Template Moves On
Applying a template records which version you used. If someone improves the template later, the analysis details tell you that you're built from an older version and offer to re-apply the latest. It's advice, not an automatic update: your gates stay exactly where they are until you decide.
Quality Control
Quality Control, under Analysis → Cell Analysis, looks for instability during acquisition: a clog, a bubble, a pressure drift. It bins the run by time and flags bins that sit outside your threshold, defaulting to 100 bins and 3 standard deviations.
Review the flagged regions, deselect any you're happy with, and apply. A green dot marks the filter as active.
Applying a time gate re-parents your whole tree. Conspecta creates a gate named "QC Time" covering the good time windows and moves every top-level gate and plot underneath it, so all your existing populations are now filtered by it. That's the point, but it's a big structural change. Removing the filter later clears the result without deleting the gates it made.
Quality control needs the file to carry a TIME channel. Without one there's nothing to bin.