Telescope: A Database Search Sandbox for Software Development

Database search remains a cornerstone of data-dependent proteomics. Its adaptation to real-time search (RTS) enables intelligent, on-the-fly acquisition decisions such as triggering validation or quantitation scans only when warranted. Because RTS must keep pace with ever-faster mass spectrometers using limited local compute, search algorithms must be fast and memory-efficient. Here, we present Telescope, a lightweight, open-source database search algorithm and sandbox designed for developing and testing RTS methods. Telescope uses a fragment ion index strategy that stores each indexed fragment in four bytes, supporting billions of peptidoforms while keeping memory tractable, and reproduces Comet's Xcorr scoring through a modular, heavily annotated codebase intended as a functional search tool and foundation for building more elaborate RTS strategies. Using a data-streaming application to replicate mass spectrometer acquisition at multiple scan rates, we show that Telescope replicates Comet's identifications while achieving per-spectrum search times in the microsecond range. Telescope scales to whole-proteome tryptic and semitryptic searches with combinations of methionine oxidation and phosphorylation. At per-scan search speeds ranging from 5,000 to 20,000 Hz, Telescope keeps pace beyond current instrument acquisition speeds, leaving computational headroom for added RTS capabilities such as retention-time alignment, quantification, and precursor prioritization.