Storing Mass-Spectrometry Data in Simple Databases Enables Flexible and Intuitive Exploration without Time or Space Penalties

Mass spectrometry (MS) generates large data sets that are stored in increasingly optimized and complex file types, demanding technical expertise to extract information rapidly and easily. We wondered whether a simple structured query language (SQL) database could hold raw MS data and allow for easily readable queries without incurring major penalties in the read time or disk space relative to other popular MS formats. Here, we describe a basic MS schema with intuitive database tables and fields that can outperform other formats for exploratory and interactive analysis according to six data subsets commonly extracted: single scans (both MS¹ and MS²), ion chromatograms, retention time ranges, and fragmentation searches (both precursor and fragment search). Additionally, we compare SQLite, DuckDB, and Parquet implementations and find that they can perform these tasks in under a second, even when the files occupy over a gigabyte of data on the disk. We believe that this tidy data schema expands nicely to most forms of MS data and offers a way to transparently query data sets while preserving computational performance.