Shotgun lipidomics relies on the direct infusion of total lipid extracts into a high resolution tandem mass spectrometer. A single shotgun analysis produces several hundred of densely populated FT MS and FT MS/MS spectra, each of which might comprise thousands of peaks although a very small percentage of those belong to lipids. Eliminating noise by adjusting a minimal peak intensity threshold is biased and inefficient since lipid species and classes vary in their natural abundance and ionization capacity. We developed a method of peak intensity-independent noise filtering in shotgun FT MS and FT MS/MS spectra that capitalizes on a stable composition of the infused analyte leading to consistent time-independent detection of its bona fide components. Repetition rate filtering relies on a single quantitative measure of peaks detection reproducibility irrespectively of their absolute intensities, masses, or assumed elemental compositions. In comparative experiments, it removed more than 95% of signals detectable in shotgun spectra without compromising the accuracy and scope of lipid identification and quantification. It also accelerated spectra processing by 15-fold and increased the number of simultaneously processed spectra by approximately 500-fold hence eliminating the major bottleneck in high-throughput bottom-up shotgun lipidomics.