Aird

Aird provides a computation-oriented format and compression methods for mass spectrometry (MS) data to reduce storage size and accelerate data access for proteomics and metabolomics analyses.

Key Features:

• Controllable Precision: Allows adjustment of data precision for m/z values to match instrument specifications and processing algorithms.
• High Compression Rate: Employs the Zlib-Diff-PforDelta (ZDPD) compressor for m/z data, reducing file sizes by approximately 65% compared to Zlib-only methods.
• Computation-Friendly Design: Prioritizes decoding speed and efficient disk-read strategies with SIMD optimizations yielding decoding times around 33% of Zlib-only methods.
• Flexible Indexing Strategies: Incorporates multiple indexing techniques to accelerate random access of specific subsets of MS data.
• Metadata Storage with JSON: Stores metadata using JavaScript Object Notation (JSON).

Scientific Applications:

• Proteomics: Supports storage and rapid access of high-throughput MS datasets for protein identification and quantification.
• Metabolomics: Facilitates metabolite profiling and biomarker discovery by enabling efficient handling of large MS data volumes.
• Data-independent acquisition (DIA) workflows: Handles large DIA datasets and variable precision requirements from high-precision instruments.

Methodology:

Implemented in Java with an SDK; uses the Zlib-Diff-PforDelta (ZDPD) compressor and SIMD optimizations for encoding/decoding, stores metadata in JSON, employs multiple indexing strategies, and provides AirdPro to convert vendor-specific files (open-source under the MulanPSL2 license).