Tech Transfer Mantra #2

Process Development is Responsible for the Process

The Value of Debug plans

One of the Mantras of that Tech transfer philosophy is that, during the start-up phase, “Process Development is responsible for the process” - Mantra #2. That mantra has many obligations associated with it, including the need to build the right scaled-down models, the importance of understanding the process deeply, etc. And, it is critical that process development lead the troubleshooting of the process should the process not perform as expected in the scaled up run. And it has to be fast, as time is money!

I learned over and over again that a critical tool facilitating the speed of this troubleshooting is the Debug Plan. This is a plan that allows the process engineers to learn quickly and efficiently during start-up. It requires significant pre-thinking to focus troubleshooting effort without having to do another full-scale run and have it fail at considerable expense. As an example, for every fermentation, the PD engineer should always be able to rapidly assess whether the root cause of a problem is the media, the inoculum, or the bioreactor operation (This is something I learned very early on from Dr. Lee Lynd at the Thayer School of Engineering. If I had a nickel for every time I’ve recited that nugget …).

A story: as a young process engineer, I had crafted some Debug plan activities for a fermentation we were scaling up to a CMO, including the vial culture thaw and initial incubation. We were transferring a protein production process to a new facility, and wanted to understand every step, even something as simple as that first shake flask step.

Just one day into the startup, when that very first shake flask culture grew poorly, I was surprised but ecstatic that I had and executed that Debug plan. I had the data to be able to (somewhat nervously … gulp!) engage the plant manager and say “I think there’s something wrong with your incubator.”

Now I know what you’re thinking – that doesn’t seem particularly likely – and the plant manager thought the same.  And sure, he was an imposing figure, so this was not something I asserted lightly. But the debug plan data was clear, and sure enough, we did a follow up test (after having attached a strip chart recorder to the incubator – so 20th century!) and we confirmed that for some unknown reason, the incubator-shaker was periodically losing temperature & agitation control. A different incubator-shaker was used for the next run and away we went.

The point is this: the debug plan was easy to implement, and though it required additional resources to execute, it quickly focused our attention, and it allowed us to find even unusual points of failure. That’s the real value in my opinion – many fewer meetings after failures with people speculating, potentially getting caught up in group think, doing many different types of experiments because of the lack of data, etc. It afforded efficient problem rectification.

I’ve been associated with the opposite too – I’m aware of a technology transfers in which shortcuts were taken in the Debug plan. Sometimes that seems like the only approach – you only have so many resources. But in some cases, it’s the result of complacency. I’ve seen process steps underperform in N=1 scale-up runs without sufficient debug plan data to know what happened, leading to an entire year of uncertainty about how to fix the problem - not where you want to be.

I have many more stories like this one, some of which we covered here[2], and many of which end in the same way: by using these tried and true techniques, you’ll be up and running in the plant at full capacity (or close to it) much faster in your start-up campaign.

Please reach out if you feel like you’d like some support in crafting and assessing your technology transfer or scale-up/scale-down plan.

[1] Hill PW, Benjamin K, Bhattacharjee B, Garcia F, Leng J, Liu C-L, Murarka A, Pitera D, Porcel, EMR, da Silva I, Kraft C, “Clean manufacturing powered by biology: how Amyris has deployed technology and aims to do it better”, Journal of Industrial Microbiology & Biotechnology, 07 October 2020,  https://doi.org/10.1007/s10295-020-02314-3