Intelligent Inclusion as a Means of Improving Targeted Proteomics in Presence of a Dominant Background
Currently, mass-spectrometry (MS) based clinical proteomics is impeded by a critical limitation in that there is a disproportionate presence of highly abundant proteins in complex samples such as human serum. Thus, targeted analysis of peptides of interest (e.g. from a pathogen infection or from a diagnostic cancer marker) is often impossible without extensive experimental protocols, as the dominant human proteome background usually obscures any informative signal. We propose using novel optimization-based algorithms for overcoming this drawback.
Using constraints based on comparison of target peptide features (such as mass, elution time, and detectability) against background peptide features, we identify an optimal set of target peptides to include for selective analysis by the MS instrument - particularly in the presence of a complex dominant protein background. These peptides are compiled into an inclusion list for targeted proteomics, which we term as “intelligent inclusion”. Simulation studies show an increase in the number of target molecules being analyzed using our algorithms. From characteristic patterns, we are also able to predict peptide “inclusivity”, which is the probability that a target peptide can be added to an inclusion list given a particular background.
Anoop Mayampurath received his PhD. in Bioinformatics in 2013 from Indiana University. His research interests are in algorithm and statistical applications to mass-spectrometry based proteomics. His doctoral work involved developing a computational framework for discovering glycoprotein biomarkers in esophageal cancer using mass spectrometry data.
Prior to graduate school, Dr. Mayampurath spent two years at Pacific Northwest National Laboratories developing open source software applications for analyzing proteomics data. Within the Computation Institute, Dr. Mayampurath, under guidance of Dr. Sam Volchenboum, focuses chiefly on improving targeted proteomics using computational algorithms and in characterizing phosphorylation and glycosylation changes in neuroblastoma.