The article presents Ansgar Korf’s views on analytical science, untargeted mass spectrometry, machine learning, and his career.
Approach:
Research focus: Korf’s team develops computational methods and open infrastructure through MZmine to process untargeted small-molecule mass spectrometry data and reduce software-related bottlenecks.
Data infrastructure: He describes FAIR-MS as an effort to make mass spectrometry data more findable and reusable for large-scale machine learning.
Career: After joining Bruker Daltonics, Korf co-founded mzio in early 2024 with Robin Schmid, Steffen Heuckeroth, and Tomáš Pluskal.
Key Findings:
Korf expects untargeted measurement to become more common as software improves, while noting that targeted methods remain important when validated answers are required.
He argues that data can be revisited to address questions not considered when they were collected.
He emphasizes that study design, sampling, preparation, acquisition, and processing jointly constrain analytical quality.
He cautions that AI cannot correct systematic errors such as matrix effects, ion suppression, batch drift, or sample degradation, and may learn them as signal.
He identifies the treatment of analytical science as support rather than a scientific contribution as a major field-wide challenge.
Interpretation:
Korf presents reusable, comparable data and end-to-end workflow quality as central concerns for applying machine learning to mass spectrometry.
Limitations:
This is an interview-style profile reporting Korf’s views and career account, not a research study presenting methods or quantitative results.
Conclusion:
Korf says he joined Bruker Daltonics before co-founding mzio GmbH in early 2024 with Robin Schmid, Steffen Heuckeroth, and Tomáš Pluskal. He describes the timing and founding team as central to making the venture possible.
