A request lands and the queue gets sorted
Discovery handed over an anti-CD20 panel: twenty candidates to express as full-length IgG for characterization, with the lead three jumped to the front. For the production lead, the morning starts at the request queue: what's ready to run, and what's still waiting on a plasmid or a decision.
Each request arrives carrying its context, so triage is reading what's in front of you, not chasing the requester for the details. The team lead works down the queue: the requests ready to go get separated from the ones still incomplete, the three priority constructs get peeled off for single-construct runs, and the rest of the panel stays grouped for a subsequent batch. When a request is for a protein the lab already has on hand, triage fills it straight from inventory instead of sending it to the bench, the stored aliquots and their amounts surfaced right there. Anything headed to an outside vendor is turned into an Excel order sheet based on the scientists' requests. What isn't moving forward gets cancelled with a reason that stays on the record, so the queue shows the current state of the work.
The construct moves through expression
The three priority constructs go one at a time; the panel of twenty runs as a single parallel batch. The work is the same shape either way, and it plays out over weeks at the bench. The difference is whether Affinity is following one construct or a whole plate.
A single construct moves through expression a phase at a time, and Affinity records each phase as the bench finishes it: the transfection with its cell line and volume, the purification with its method and the yield and concentration recovered, the characterization with its SEC purity and endotoxin, and the last step placing the finished protein into an inventory location. A task advances only as the data for it lands, so the queue is always a true read on where the construct is: waiting to start, in transfection, through purification, or complete. The plate-based panel of twenty runs that same arc in parallel: a single transfection step records the cell line and volume across the batch, the wells move through together, and the stragglers and failures get marked off as the batch thins out.
However the construct got made, the production data lands where the work happened, not in a record on someone's drive that the rest of the program can't see.
Every lot gets reviewed before release
Twenty expressed lots came out of HTE. Some have great yields and clean SEC traces; some have low yields or visible aggregation. The next step is reviewing each lot against the program's criteria before releasing anything to downstream characterization.
The Lot Review page puts each lot's production record next to its quality numbers, so the release call is made on the whole picture in one place. The review-status field on each lot accepts Flagged or Do Not Use values when a problem is caught, which are then displayed to scientists in Affinity, so a low-purity Flagged lot doesn't accidentally end up in a characterization assay without the reviewer's reasoning visible.
The released lot lands in inventory with its history attached
Lot review is done. The lots that passed are now first-class in inventory, ready to be pulled when the next assay batch starts.
The protein's page in the bioregistry carries the full production lineage: the construct, the cell line it was expressed in, transfection information (date, volume), the yield and concentration recovered, the SEC purity and endotoxin readings, any review-status flags applied, the request the lot originated from, the project it belongs to, and the freezer location of every aliquot. When the next scientist pulls a vial six weeks later and asks what's the production history of this lot — the answer is on one page, not a hunt across spreadsheets and emails.
