DELPTM: A Statistical Algorithm to Identify Post-Translational Modifications from Tandem Mass Spectrometry (MS/MS) Data

• Activity Number: 641
• Type: Topic Contributed
• Date/Time: Thursday, August 2, 2012 : 10:30 AM to 12:20 PM
• Sponsor: Biometrics Section
• Author(s): Susmita Datta*+
• Companies: University of Louisville
• Keywords: PTM ; MS/MS ; Bayesian ; blind ; clustering ; phosphopeptides

Abstract

Post translational modification (PTM) of proteins play a significant role in complex diseases like cancer and diabetes etc. Hence, identification of PTMs on a genome-wide scale is important. Attempts have been made to detect PTM through analysis of tandem mass spectrometry data by several unrestrictive "blind" search algorithms. However, these methods suffer from mass measurement inaccuracy and uncertainty in predicting modification positions. In a recent published work, scientists have modified these "blind" search algorithms by post-processing the output of the "blind" search algorithm with Bayesian modeling of the probability distribution of PTM chemical process. This method improved the accuracy of estimation of PTM somewhat. In our work, we propose further modification to the methodology. We develop a self-validated clustering method for mixed data types through rank aggregation of the PTM data. We then use the clustering result as known groups of the PTM data and subsequently model the probability distribution of the PTM process. This reduces the complexity of the earlier model and also this simplification significantly improves the number of correct modification position pre