Abstract
The components of
complex protein mixtures fed into tandem Liquid Chromatography and Mass
Spectrometry are identified from the LC/MS generated experimental spectra using
database search. The peptide database search methods require a large search
time due to their computationally expensive spectral comparison. The current
search algorithms use database filtration methods and shared-peak counting to
reduce the search space to relevant components, but that requires high memory,
read/write, and caching overheads.FIU inventors
have developed an algorithm for compressing data to reduce index size (e.g. in
a fragment-ion index). Data in an index can be transformed for reducing entropy
and then encoded using a running counter technique to compress redundant information
in the index and thus reduce memory
footprint. The algorithm was able to achieve a 50% + reduction in memory
consumption compared to the related methods, along with improved peptide
inference speed.Benefit
· Reduced memory footprint during a search · Improved peptide inference speed from LC/MS spectraMarket Application
The method has applications in compressing data generated from mass spectrometry.
