252: How to Use Media Supplements to Tailor Biosimilar Glycan Quality to Your Reference Product in Two Rounds

Are you still using one-factor-at-a-time experiments for biosimilar development, losing months, missing interactions, and risking costly dead-ends?In this episode, David Brühlmann, host of the Smart Biotech Scientist Podcast, reveals how traditional "one factor at a time" screening in biosimilar development can take over 12 months, while the parallel group design massively accelerates discovery by grouping up to five factors per experiment and applying a multivariate analysis pipeline.Topics discussed:The limitations of traditional and large DoE designs and the advantages of parallel group design (00:08)Best practices for grouping compounds by biological mechanism with four essential rules (00:53)The importance of anchor compounds, separating strong modulators, and initial univariate screens for unknown compounds (01:43)Guidance on managing practical issues, including evaporation, liquid handling, osmolality, and replicating production processes (06:42)The use of multivariate analysis tools: Principal Component Analysis, Mahalanobis distance, and decision trees for candidate selection (10:14)Key results and outcomes from applying the parallel group method, including faster and more cost-effective quality modulator identification (12:46)Three improvements David would recommend today: prequalifying compounds, broader quality analytics, and hybrid modeling integration (13:49)The shift in mindset from “time problem” to “information problem” in process development (16:50)Extending the parallel group and multivariate approach to other areas like clone selection and scale-up decisions (17:52)Smart insight:Process development is fundamentally about generating actionable information, not just running more experiments. The parallel group, multivariate pipeline lets teams ask better questions, in parallel, with dramatically improved data yield. This mindset and methodology extend well beyond biosimilar media development into clone selection, feed design, and process characterization, wherever complexity would paralyze traditional approaches.If you want more detail, you can read the full article “Parallel experimental design and multivariate analysis provides efficient screening of cell culture media supplements to improve biosimilar product quality” published in Biotechnology and Bioengineering, which outlines the methods and findings behind this approach.If you’re interested in hybrid modeling, here’s what previous podcast guests have shared on the topic, offering perspectives from fundamentals to real-world applications.Episodes 05 - 06: Hybrid Modeling: The Key to Smarter Bioprocessing with Michael SokolovEpisodes 99 - 100: From Raw Data to Actionable Insights: Unlocking the Power of Process Models with Fabian FeidlEpisodes 137 - 138: Skip 90% of Bioreactor Runs: The In Silico Revolution in Bioprocess Development with Yossi QuintEpisodes 173 - 174: Mastering Hybrid Model Digital Twins: From Lab Scale to Commercial Bioprocessing with Krist GernaeyNext step: If this was useful, leave a review on Apple Podcasts or Spotify. It helps other scientists find this content, and it genuinely matters.Support the show