Georg Wallman

Summary: A preview of 2026 separation science trends, from PFAS and SFC to metabolomics, next-generation HPLC columns, oligonucleotides, and ADC analysis.

Wallman argues proteomics is moving from niche tool to high-throughput platform core, with the greatest untapped potential in covalent drug programs and targeted protein degradation.

Wallman explains that PeptDeepKontext is designed to generalize without lab-specific calibration, though highly unusual setups may still benefit from additional fine-tuning in the future.

Wallman unpacks how graph-level fragmentation in fragDETR captures internal fragments and neutral losses, with peak and intensity improvements concentrated in challenging non-standard peptides.

Wallman explains how spectral library accuracy, retention time prediction, and instrument-specific variation make deep learning essential yet difficult in data-independent acquisition proteomics.

Wallman details PeptDeepKontext, a model built to predict peptide properties across diverse instruments and PTMs by embracing rather than eliminating inter-laboratory variability.