Statistical agglomeration: peak summarization for direct infusion lipidomics

• DOI: 10.1093/bioinformatics/btt376
• Publication Date: 2013-07-04T07:00:34Z
• Authors: Rob Smith; Tamil S. Anthonym...; Dan Ventura; John T. Prince
• Keywords: Identification

Abstract Motivation: Quantification of lipids is a primary goal in lipidomics. In direct infusion/injection (or shotgun) lipidomics, accurate downstream identification and quantitation requires summarization repetitive peak measurements. Imprecise multiplies error by propagating into species intensity estimation. To our knowledge, this the first analysis infusion literature. Results: We present two novel algorithms for samples compare them with an off-machine ad hoc algorithm as well propriety Xcalibur algorithm. Our statistical agglomeration reduces peakwise 38% mass/charge (m/z) 44% (intensity) compared method over three datasets. Pointwise reduced 23% (m/z). Compared Xcalibur, produces 68% less m/z 51% on average comparable Availability: The source code Statistical Agglomeration datasets used are freely available non-commercial purposes at https://github.com/optimusmoose/statistical_agglomeration. Modified Bin Aggolmeration MSpire, open mass spectrometry package https://github.com/princelab/mspire/. Contact: [e-mail] or [e-mail] Supplementary information: data Bioinformatics online.