Chromatograms: A Modular Infrastructure for Scalable Chromatographic Data Handling in R

Chromatographic traces are central to LC-MS analysis, yet most software treats them as transient intermediates inside a fixed pipeline rather than as data objects users can inspect and manipulate directly. In R, xcms provides the established infrastructure for large-scale LC-MS analysis, but the way it represents chromatograms limits what can be done with them once extracted. We present chromatograms, an open-source R/Bioconductor package and a core component of the R for mass spectrometry ecosystem, built on two design principles: interchangeable storage backends, which keep the analysis code independent of where the data reside (in memory, on disk, etc.), and delayed evaluation, which applies queued operations only to the chromatograms actually requested and processed at a time. Chromatograms was fully integrated into xcms hence enabling the efficient extraction of EICs for identified features, exemplified on public data, for which extraction of 4980 features across 4000 samples took only a few hours, against a full day with the legacy infrastructure. For targeted data analysis or inspection, a peak table from any external software, or published with a study, is enough to recover the corresponding chromatograms from the raw files, with annotation columns preserved; four published biomarkers were recovered from all 4063 samples of a cohort even without parallel processing in just under an hour. Chromatograms does not perform peak detection: it provides the infrastructure for working with chromatographic data once peaks are defined, giving a scalable foundation for targeted and untargeted LC-MS analysis in R.