Abstract
Mass spectrometry (MS) data consists mainly of either mass-to-charge (m/z) and intensity data pairs or retention time - intensity pairs enriched with additional metadata information. MS instruments store such data in vendor-specific formats which have to be converted to the open file formats mzML or mzXML for data transfer and import. In addition, MS data from spectral libraries or online annotation resources are usually stored and provided in a variety of different file formats ranging from the standard mgf file format to custom text-based file formats.
Learning from experiences gained in the MSnbase and xcms Bioconductor packages we implemeted the new Spectra package that provides a common, flexible and a high-performance infrastructure to represent and handle MS data in R. The introduction of dedicated backend classes makes the main Spectra object independent of the origin and storage of the data. Depending on the backend used, MS data can be kept in memory for fast processing, or retrieved from disk on demand for a low memory footprint hence enabling the analysis of large MS data sets. Backends interfacing directly SQL based spectral libraries can in future facilitate the comparison of spectral data from experiments with reference spectra for annotation purposes.