Run a Protocol

A protocol you wrote in the Notebook is a recipe. A run is one execution of it: the day you actually did it, on those samples, with those lot numbers, and whatever happened along the way. Runs are what turn a protocol from a document into a record.

Starting a Run

Start run sits at the top of any protocol page. You can also start one from a sample or a sample group with Run protocol, which links those samples to the run as inputs and drops you back on the sample when you're done. That's the right entry point when the question is "what did we do to this sample", rather than "let's do this procedure".

If you already have a run of this protocol open, or abandoned one in the last two weeks, Conspecta asks before starting a fresh one. You can Resume where you left off, reopen an abandoned run, look at the full history, or start new anyway. This is deliberate: half-finished runs are usually a tab you closed, not a decision you made.

Starting a run also nudges a protocol still marked Draft to Work in Progress, since something you're executing isn't a draft anymore.

Working Through It

A run replaces the editor with a stripped-down view built for a bench: one step at a time, big targets, and no way to accidentally edit the recipe.

The run works from a frozen copy of the protocol. Edits someone makes to the protocol page while you're running it don't change what's in front of you. Whatever you followed is what gets recorded.

The header carries how many steps you've checked, its run number, and when you started; once the run stops it says Paused or Completed, and a finished run shows how long it took. Pause run holds any timers that were counting when you step away. Resuming restarts only the ones it paused.

The Run Card

Above the steps sits a card with the context for this particular run.

Running on lists the samples this run is being performed on, and you can add more at any point.

Batch is a multiplier for parallel replicates. Set it to 3 and every reagent volume and every capture field scales with it, with the original shown struck through next to the scaled figure so you can still see what the recipe said.

Materials rolls up everything the run will consume, so you can check you have it before you start rather than halfway through.

Steps

Tick a step to mark it done. It dims, drops its capture box, and keeps a Done badge and a count of what you captured, so the list stays short and the next thing to do is always at the top. A chevron reopens any step you want to look at again.

An open step can hold more than instructions:

• A timer you start, pause, and reset from the step itself. It keeps counting while you work anywhere else in Conspecta, rides along at the bottom of the screen with a tap that takes you back to that step, and chimes wherever you are when it's up. When it reaches zero the run writes it down, so months later the record shows the incubation ran its length and when it ended, alongside the moment you ticked the step
• An instrument chip that jumps to that instrument's booking calendar
• An image analysis chip that launches the author's saved settings and records the result against this run
• Materials with live stock, showing what's on the shelf and which lot to use, defaulting to the one expiring first while letting you pick a different one
• Choose path, on a branching step, where picking a case reveals only the steps that belong to it

Checking a step off draws its materials out of Inventory for real. The stock numbers on the step update as you go, so the last person to run the protocol doesn't leave the next one guessing.

Capture

Each step's capture box collects the fields the protocol's author asked for, plus two that are always there: Notes, for what you observed, and Deviation, for anything you did differently. Those two are the ones that matter six months later. Values save when you move off the field.

A multi-day protocol shows day headings so you know where the natural stopping points are.

At the Bench

A run is meant to be carried. Open it on a phone and you get the same steps, the same timers and the same capture boxes, so what happened gets written down where it happened.

A step timer belongs to the whole app rather than to the screen you started it on. Once one is counting, a bar sits at the bottom of every screen you move to, naming the timer finishing soonest and counting the rest. Tapping it takes you back to that step of that run, and the timer chimes wherever you are when it is up.

Every countdown you have going is also listed under Active timers in Timers, whether you dialled it there or started it on a step. A run's step timer carries the name of the protocol underneath it, and that name is a link back into the run.

Timers is also where a sequence runs, which is a preset holding several steps: a stain that blocks, stains, washes three times and counterstains, without you dialling each wait as it comes. The card names the run, then the step it is on, fills in a tick for each step behind it, and says how much of the whole run is left. Conspecta at the Bench covers building one.

What the Run Produced

Below the last step, a section records the run's outputs.

Consumed lists the materials that were drawn down, filled in automatically as you checked steps.

Produced is where you record what came out. New sample creates one and links it to the run, and Link existing sample points at one you already registered. Either way the new sample's Connections tab shows this run upstream of it, so its provenance is traceable without anyone writing it down.

Image analysis collects any analysis launched from a step.

Finishing

At the end of the list a checklist tells you where the run stands: how many steps are checked, whether capture fields are filled, whether outputs are recorded, whether deviations are logged. Anything unfinished links straight to the step in question.

Finish run asks to confirm, and says plainly how many steps are still unchecked, because sometimes the honest record is that you skipped two. Finishing locks the run and files it in the protocol's run history.

You're then offered a choice: mark it Completed, or Needs review and pick who should look at it. Review is suggested when the run logged deviations or produced samples, but nothing is blocked either way, and you can decide later.

If you're the only person on the project there's nobody to ask, so Flag for review marks the run for you to check yourself later. The flag sits on the run, so it's waiting for you whenever you come back to it.

The run header of a finished run carries a review picker, holding No review needed, Needs review, In review, and Reviewed. Use it to move the run along as the check happens, or to clear the flag once you've looked.

A finished run can be reopened if you realize you missed something, and a paused one picks up where it stopped.

Scheduling a Multi-Day Protocol

A protocol with day breaks can be laid onto the calendar before you start. Use Schedule run on the protocol page, or Schedule days in the run's menu once it's underway.

The dialog box has the protocol's days on the left, each showing its step count, rough duration, and any instrument it needs, and a live calendar on the right. Pick a day, then click a time to place it. Clicking an hour slot sizes the block from the recipe's estimate, dragging lets you set the length yourself, and clicking a date in Month view schedules the day without committing to a time. Placed blocks can be dragged and resized afterwards.

Switch the calendar between My calendar and Team calendar to see what else is booked before you commit.

Instrument bookings come along. If a day uses an instrument, that instrument's existing bookings and downtime are drawn right in the grid, hours it can't be booked are refused as start times, and placing a day at a time offers to book the instrument for you. The offer is always a question you answer, and skipping it leaves the day scheduled without a booking.

Splitting is available too. A day with several steps that hasn't been placed yet offers Split into sessions, so a long day can become a morning and an afternoon on different dates.

Each scheduled day becomes a task under the run, which is why it appears on the Planner calendar and in the table but never as a card on the board: it's part of a run, not a separate piece of work. Removing a placement sends its task to Trash rather than silently clearing its date, since by then it may carry notes or an assignee.

History and Discussion

A run has its own activity panel, holding comments, review requests and decisions, and a complete audit trail: who started it, every step checked and unchecked, every step timer that ran out and how long it ran, every captured value before and after it was edited, every status change, and every change to the batch multiplier.

Run history in the protocol page's menu lists every run of that protocol, which is where you go to compare the day it worked with the day it didn't.