Decoding Intelligence: AI, Language, Computational Biology

The production of biologics such as therapeutic antibodies using Chinese Hamster Ovary (CHO) cells is a cornerstone of modern biopharmaceutical manufacturing. One of the most critical quality attributes of these antibodies is correct N-glycosylation, which affects both their safety and efficacy. However, glycosylation is influenced by a complex network of cellular processes that vary over time and under different culture conditions.
To better understand these dynamics, we are analyzing multi-omics time series data from the NISTCHO cell line, which produces the cNISTmAb—an industry-relevant reference antibody with high production yields. Rather than relying on black-box predictive models, this study focuses on statistically grounded hypothesis testing using spline-based modelling in the limma framework, enabling us to detect significant temporal changes in gene and protein expression.
We then apply enrichment analysis to identify key pathways and processes involved in glycosylation and antibody production. This approach aims to generate interpretable, mechanistically meaningful insights into how high-producing CHO cells behave over time, helping to bridge the gap between systems biology and bioprocess optimization in industrial settings.

Have you ever thought about how much language can influence the way we experience the world, what we pay attention to and which information we express in what way?
If German is your native (or dominant) language, you will probably expect words such as gerade, wieder, weiter etc. if you want to say that an action is going on right now, is repeated or continued. In languages such as French, Italian or Spanish, however, many of those things are expressed in a different way. How can we deal with this if we have the ‘German pattern’ in mind? And what happens when other languages come in?

Artificial intelligence is powerful in learning relationships. However, it has - so far - met its match in understanding “molecular intelligence”. Molecular intelligence is the ability of cells to information to grow and fight off threats - without a brain or computer chip. The molecular processes behind the intelligent behavior of cells have been optimized through evolution. I will present selected areas of research, where biologists try to understand molecular intelligence by collecting big molecular datasets and using artificial intelligence algorithms to better understand how cells function in healthy conditions and how they dysfunction in disease.