Overview of the alphaDIA framework. From Wallmann et al., 2025. “AlphaDIA enables DIA transfer learning for feature-free proteomics“, Nat Biotechnol (2025). https://doi.org/10.1038/s41587-025-02791-w; Licensed under the terms of the Creative Commons CC-BY 4.0 license.
AlphaDIA is a new one-stop processing platform for raw DIA proteomics data. Introduced by Wallmann et al, from the Mann lab, AlphaDIA addresses challenges in processing complex spectral data from advanced time of flight instrumentation, offering a feature-free approach that aggregates signals of up to four dimensions, without loss of retention or ion mobility signal, before making discrete identifications. This method enhances sensitivity and accuracy, particularly with complex time-of-flight (TOF) detector data.
The framework demonstrates competitive or superior performance compared to existing tools across various platforms and experimental designs. It identified over 120,000 precursors and 9,500 protein groups in a 60 samples per day (SPD) format on the Orbitrap Astral.
A key innovation is alphaDIA’s end-to-end transfer learning capability, which adapts to experiment-specific conditions, significantly improving peptide identifications. This was showcased on dimethylated HeLa peptides, resulting in a 48% increase in unique precursor identifications and a 25% increase in protein groups.
This study is important as it advances DIA proteomics analysis, potentially enabling more comprehensive and accurate proteomic characterization in clinical and translational research.
The IonOpticks Aurora Rapid 8×150 XT C18 UHPLC column was used to separate peptides by the Whisper 60SPD method gradient, while samples were separated by the Whisper 40SPD method using the IonOpticks Aurora Elite 15×75 C18 UHPLC column.
Publication
Nature Biotechnology
Authors
Georg Wallmann, Patricia Skowronek, Vincenth Brennsteiner, Mikhail Lebedev, Marvin Thielert, Sophia Steigerwald, Mohamed Kotb, Oscar Despard, Tim Heymann, Xie-Xuan Zhou, Maximilian T. Strauss, Constantin Ammar, Sander Willems, Magnus Schwörer, Wen-Feng Zeng & Matthias Mann;
Title
AlphaDIA enables DIA transfer learning for feature-free proteomics
