Explore Validation Strategies of DIA Hits with alphaDIA

[GeorgWa]

Title

Explore Validation Strategies of DIA Hits with alphaDIA

Abstract

Data-Independent Acquisition (DIA) proteomics has advanced significantly, yet its
black box algorithms and ML-driven scoring have hindered transparency and therefore adoption.
Concurrently, validation of individual DIA peptide identifications has remained largely
unexplored. Building on our open-source ecosystem for DIA with alphaDIA, alphaBase, and
alphaRaw, we aim to expand alphaDIA to investigate validation strategies for DIA hits, making
it more accessible to the growing DIA proteomics community.

Project Plan

In this hackathon, we aim to extend the alphaDIA ecosystem to improve DIA hit
validation. Specifically, we will:

1. Analyze Orbitrap Astral raw data using alphaDIA.
2. Develop interactive notebooks for spectrum-level identification exploration.
3. Integrate alphaDIA with alphaRaw and alphaBase for identification reliability assessment
   at the spectrum level.
4. Build protein-level metrics to explore evidence at the protein level.
5. Create user-friendly visualizations and metrics for DIA result interpretation.
6. Identify new features which can aid the automated scoring of alphaDIA

Post Hackathon:

• Incorporate new features and improvements into the alphaX packages.
• Create curated tutorials for exploring DIA validation.
• Incorporate additional features into the scoring algorithm of alphaDIA
  We anticipate contributions in code, documentation, and data, as well as valuable discussions
  between computational and experimental DIA proteomics researchers. All contributions will be
  formally acknowledged through package credits or authorship, depending on the output.

Technical Details

Programming language: Python and Jupyter for processing and interactive visualization.
The project will build on existing software: alphaDIA, alphaRaw, alphaPeptDeep,
directLFQ, and alphaBase.
We will use and extend publicly available DIA datasets.

Contact information

Georg Wallmann & Mann Labs
Max Planck Institute of Biochemistry
[email protected]