Aird 2.0: High-performance mass spectrometry data format optimized for computational analysis

Background: The escalating volume of data generated by mass spectrometry (MS) presents a significant bottleneck for large-scale computational analysis. While data compression is crucial, the primary challenge lies in accelerating data access for processing-intensive tasks. To address this, we introduce Aird 2.0, a high-performance data format engineered for rapid data retrieval and efficient storage. Results: Aird 2.0 features a novel, dynamic compression framework that intelligently selects from six specialized algorithms to optimize performance for any given MS file. Our comprehensive evaluation on 58 public datasets demonstrates that Aird 2.0 achieves a 2.8-fold increase in random-access speed and a 6-fold increase in full-file loading speed compared to the standard mzML format, while also reducing file sizes by an average of 12.3% over its predecessor, Aird 1.0. Additionally, it enhances metadata deserialization speed by 2.5-fold and halves metadata file size, further accelerating analysis workflows. Significance: Supported by a comprehensive conversion client (AirdPro) and SDKs for Java and C#. Aird 2.0 is a comprehensive, open-source solution for accelerating the analysis of large-scale Mass Spectrometry data. It provides a high-performance data format that can be seamlessly integrated into existing software ecosystems, offering a robust foundation for future research and development. • Aird 2.0: a high-performance MS data format for fast computational analysis. • Introduces adaptive combinatorial compression for speed and efficiency. • Achieves 6×faster loading and 2.7× faster access vs mzML. • Reduces metadata load time by 2.5× and improves compression ratio. • Open SDKs and tools enable broad integration in omics workflows.