Robert “Bob” Hariri (B.H.), MD, PhD, founder and CEO of Celularity in Warren, NJ, talks with Mike May, PhD, (M.M.) about developing your bioprocess with a focus on collecting and analyzing all the critical process data.
M.M.: In today’s bioprocessing environment, how do you use data?
B.H.: First and foremost, by developing your bioprocess with a focus on collecting and analyzing all of the data attendant to that process, you can go back and audit what you’ve done in a high resolution manner. You should also realize that the data is going to include some very subtle and nuanced things. In manufacturing a cell product, for instance, you’re going to want to dive deep into the starting material analytics—everything from genomics to transcriptomics and secretomics, etc. If you generate this huge dataset and then analyze it, you’ll be able to figure out what subtle differences between, for example, the starting material genome or the starting material transcriptome actually impacts and influences what you get in the end product.
M.M.: Can you do this now in bioprocessing?
B.H.: There’s no doubt that the investment in collecting that data is going to be valuable, but there still aren’t any intelligence tools yet developed that set you up to properly analyze this data. So, it means you have to either partner with the right AI and machine-learning groups or you’re going to have to ultimately incur the expense of developing that capability yourself.
M.M.: What makes that data so important?
B.H.: Once you have to scale and produce a product—hundreds of thousands to millions of doses of product a year—if you haven’t figured out the impact of subtle differences in the starting material and subtle differences in the behavior of the product to manufacturing, if you haven’t monitored all that, you’re not going to be able to know what at the beginning predicts the quality of the product at the end.
M.M.: If you could make one suggestion for speeding up analysis in bioprocessing, what would it be?
B.H.: If early on you enter into partnerships—with academic researchers, intelligence-tool developers—you may find that the partnership is mutually beneficial. The people developing these intelligence tools, what’s the thing they’re looking for most? Data. So, if you give them access to data and tell them what the objective is, and then they design the tool, that tool is very often translatable to multiple customers. But they do need that starting data to be able to do anything. That’s a great place to work together.