Beyond Trial-and-Error: Data-Driven Optimization for Lytic Virus Production

Discover how we accelerated viral vector process development, shortening lab work by months and process time by 24 hours per run.

At bespark*bio, we believe that modern bioprocess development must go beyond trial-and-error experimentation. In a case study with our strategic partner Novasign, we demonstrate how combining deep process expertise with advanced hybrid modeling can fundamentally transform viral vector process development – making it faster, more cost-efficient, and significantly lowering risk.

The Challenge

Rethinking Viral Vector Production

Traditional bioprocess development often relies on resource-intensive trial-and-error.

In this project, bespark*bio led the development of a Vero cell–based lytic virus production process.

 

The goal:

Move beyond conventional Design of Experiments (DoE)

Reduce experimental burden

Gain deeper mechanistic insight without sacrificing accuracy

The Solution

Smart Design Meets Hybrid Modeling

We integrated our hands-on bioprocess expertise with Novasign Studio, an advanced hybrid modeling platform.

Instead of a standard 22-run DoE, we executed 8 strategically selected experiments to characterize the process.

We focused on three critical process parameters.

Multiplicity of Infection

Temperature

pH

The result:
Model predictions based on 8 experiments were identical to those using all 22.

Time-Resolved Insights

Multi-Objective Optimization

Unlike endpoint measurements, our hybrid approach enables continuous evaluation of process dynamics. We answer critical questions in real-time:

Optimal timing: When exactly should the process be stopped to maximize yield?

Balancing act: How do we optimize virus titer while minimizing impurities like dsDNA and HCP?

We found that increasing MOI did not improve titer but significantly increased dsDNA levels – a crucial insight for downstream efficiency.

The Impact

Better Decisions, Better Economics

By combining data-driven modeling with expert process understanding, we achieved measurable real-world gains:

Experimental effort

64% fewer runs (8 vs. 22)

Development timeline

12 weeks saved in the lab

Operational costs

€ 40.000 savings in materials

Process efficiency

24 hours shorter process time per run

The Future of Bioprocessing

This case study proves that modern bioprocessing doesn’t require more resources – it requires more insight.

By reducing the experimental burden while increasing data depth, bespark*bio enables clients to make confident decisions early, when they matter most.