> For the complete documentation index, see [llms.txt](https://help.sbtinstruments.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://help.sbtinstruments.com/mpd/workflows/mpd1.md). # Identify harvest time When the cells are your product, harvest at the highest cell count. Real-time counts showed when that was, without waiting days for plates. In this work, we show how BactoBox® can be used to find the harvest time that gives the highest cell concentration, while the run is still going. When the cells are your product, you want to harvest at the highest cell concentration. Most processes find that time in an engineering run, then fix it for production. Plate counts give the answer days after the run, so they cannot tell you when to stop it. OD600 depends on cell size and on anything else that scatters light. Online signals such as pH, dissolved oxygen and off-gas do not necessarily line up with the peak. We ran an engineering run on *Staphylococcus epidermidis* with BactoBox® and plate counts side by side. ## Take-home messages * The cell count peaked at 7.2 × 10^9 cells/mL after 11.8 hours. CIZE at that point was 0.83 µm. * Plate counts gave the same harvest time, days later. * The real-time counts told us when to stop the run. * For production, set a harvest window, and check the count as you approach it. ## Study design * *S. epidermidis* in TSB, in a shake flask. * Samples about every hour from the start of deceleration phase, timed from an earlier growth curve, with longer gaps at the end of the run. * Each sample was vortexed and put on a cold rack, so the cells stopped dividing before dilution. * Three dilution series per sample for BactoBox®, and plate counts in triplicate. * A *Growth curve* measurement group in Access. The three repeats of each sample shared label and *Sampled at* time, so Access treated them as replicates. * We stopped when two points in a row changed less than 5 % per hour. The [protocol](/mpd/workflows/mpd1/protocol.md) gives every step. ## Results

S. epidermidis in TSB, measured with BactoBox® (lavender) and plate counts (yellow), both in triplicate. CIZE (cyan) is on the secondary axis. The grey box marks where most samples were taken. The magnifying glass marks the harvest time. Error bars show the standard deviation.

The count rose through deceleration phase and reached its highest mean at 11.8 hours. The next point, at 14.9 hours, was not significantly different. When two points overlap like that, pick the earlier one. Small differences in the mean most likely come from normal variation in sample preparation and measurement. Plate counts agreed. Their mean at 14.9 hours was slightly higher, but not significantly so. By 22.6 hours the cell count had fallen by about a quarter, to 5.4 × 10^9 cells/mL. A cell count falls only when cells lyse. The plate count fell further, to about 3 × 10^9 CFU/mL. More cells lost culturability than lysed. | Harvest time | Cell concentration at harvest | CIZE at harvest | | ------------ | ----------------------------- | --------------- | | 11.8 hours | 7.2 × 10^9 cells/mL | 0.83 µm | ## What it means for your process With real-time counts you end the engineering run when growth stops. Plate counts give the same answer days later, after the run is over. Production runs drift from run to run, so treat the harvest time as a window. Measure as you approach it. If the count is lower than in the engineering run, or CIZE is larger, the culture may still be growing. Extend the incubation. Some cultures level off, then grow again on a second carbon source. Before you fix the harvest time, take one more sample a few hours after the plateau to confirm it. After the peak, plate counts fall as cells lose culturability. The cell count shows only the cells that lyse. If your product is sold on CFU, [cross-check at end of fermentation](/mpd/workflows/cross-check.md) too. ## Related workflow [Track a growth curve](/mpd/mini-course/track-growth-curve.md) Questions about your own process? [Contact us](https://sbtinstruments.com/contact). --- # Agent Instructions This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com. ## Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation by asking a question. Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters: ``` GET https://help.sbtinstruments.com/mpd/workflows/mpd1.md?ask=&goal= ``` `ask` is the immediate question: it should be specific, self-contained, and written in natural language. `goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.