Peak Intensity Prediction for PMF Mass Spectra Using Support Vector Regression

Abstract

With the increasing amount of data nowadays produced in the field of proteomics, automated approaches for reliable protein identification are highly desirable. One widely-used approach are protein mass fingerprints (PMFs) that allow database searching for the unknown protein, based on a MALDI-TOF mass spectrum of its tryptic digest. Current approaches and software packages for interpreting PMFs do rarely make use of peak intensities in the measured spectrum, mostly due to the difficulty of predicting peak intensities in the simulated mass spectra. In this work, we address the problem of predicting peak intensities in MALDI-TOF mass spectra, and we use regression support vector machines (ν-SVR) for this purpose. We compare the impact of different preprocessing and normalization modes such as binning and balancing data sets on prediction accuracy. Our preliminary results indicate that we can predict peak intensities using ν-SVR even from very small data sets. It is reasonable to assume that peak intensity prediction can greatly improve automated peptide identification.

Neural Network Approach for Mass Spectrometry Prediction by Peptide Prototyping

“…Our results show that the new LLM-approach combining data mining and supervised learning yields similar results in prediction accuracy to our first approach utilizing ν-SVR [TBTN06].…”

“…Other than for example support vector regression (SVR) it can be used for data mining once adapted in a straight forward manner, as demonstrated in this work. We propose a combination of unsupervised and supervised learning architecture with comparable results in predicting the peaks intensities to ν-Support Vector Regression (SVR) [TBTN06]. The mixture of linear experts derives implicit models for characterizing peptides and feature analysis as an unsupervised learning task.…”

Improved Mass Spectrometry Peak Intensity Prediction by Adaptive Feature Weighting

“…The learning architectures of Local Linear Map-type [16] and ν-Support Vector Regression (SVR) [8] have been proposed to model the non-linear relationship between peptide and peptide peak heights in MALDI-TOF mass spectra. Highdimensional numerical feature vectors are derived from the peptide sequence building the feature space as input for the learning architectures.…”

“…To overcome these obstacles to predict peak intensities in MALDI-TOF spectra based on a training set of peptide/peak intensity pairs, we considered an artificial neural net architecture, namely the Local Linear Map [7], since it combines unsupervised (a) and supervised (b) learning principles, with comparable results to those obtained by ν-Support Vector Regression (SVR) [8]. The LLM can learn global non-linear regression functions by fitting a set of local linear functions to the training data.…”