Computational Quantification of Peptides from LC-MS data

Abstract

Liquid chromatography coupled to mass spectrometry (LC-MS) has become a major tool for the study of biological processes. High-throughput LC-MS experimentsare frequently conducted in modern laboratories, generating an enormous amountof data per day. A manual inspection is therefore no longer a feasible task. Consequently, there is a need for computational tools that can rapidly provide informationabout mass, elution time, and abundance of the compounds in a LC-MS sample. Wepresent an algorithm for the detection and quantification of peptides in LC-MS data. Our approach is flexible and independent of the MS technology in use. It is basedon a combination of the sweep line paradigm with a novel wavelet function tailoredto detect isotopic patterns of peptides. We propose a simple voting schema to usethe redundant information in consecutive scans for an accurate determination ofmonoisotopic masses and charge states. By explicitly modeling the instrument inaccuracy, we are also able to cope with data sets of different quality and resolution.We evaluate our technique on data from different instruments and show that we canrapidly estimate mass, centroid of retention time and abundance of peptides in a sound algorithmic framework. Finally, we compare the performance of our method to several other techniques on three data sets of varying complexity.