A Machine Learning Approach to Predicting Peptide Fragmentation Spectra

Arnold, Randy J.; Jayasankar, Narmada; Aggarwal, Divya; Tang, Haixu; Radivojac, Predrag

doi:10.1142/9789812701626_0021

Cited by 54 publications

(73 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To find the optimal values for the SVR regularization parameter C, the Gaussian kernel's bandwidth γ, and ν, grid searches were performed in the three-dimensional parameter space log 2 (C) ∈ [−5, 15], log 2 (γ) ∈ [− 15,7], and ν ∈ [0.1, 0.9] for each training set separately.…”

Section: Prediction Of Peak Intensities By Machine Learningmentioning

confidence: 99%

See 1 more Smart Citation

Peak Intensity Prediction for PMF Mass Spectra Using Support Vector Regression

Timm¹,

Böcker²,

Twellmann³

et al. 2006

Applied Artificial Intelligence

View full text Add to dashboard Cite

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.

show abstract

Section: Prediction Of Peak Intensities By Machine Learningmentioning

confidence: 99%

“…As noted above, software for PMF interpretation usually ignores intensity information. Modeling of intensities in tandem mass spectra has been dealt with by multiple authors 7,8,9,10 . Such models mainly consider the fragmentation probabilities of molecules.…”

Section: Introductionmentioning

confidence: 99%

Peak Intensity Prediction for PMF Mass Spectra Using Support Vector Regression

Timm¹,

Böcker²,

Twellmann³

et al. 2006

Applied Artificial Intelligence

View full text Add to dashboard Cite

show abstract

“…At present, the greatest focus in improving protein identification software is on the following: (1) developing better scoring metrics or including additional information [41][42]; (2) improving fragmentation models. The inclusion of new metrics [43] and use of new techniques [44] applied to fragmentation modeling allows for better prediction of theoretical spectra. This, in turn, leads to more discriminating scoring systems; (3) data representations for clustering or filtering to improve speed and efficiency [45,46].…”

Section: Discussionmentioning

confidence: 99%

Computational Diagnostics Based on Proteomic Data- Review on Approaches and Algorithms

AN¹,

Badrunnisa²,

M³

et al. 2010

Int J of Bioi Res

View full text Add to dashboard Cite

-Protein identification using mass spectrometry is an indispensable tool for proteomics which in recent days has evolved to give better understanding of the biology of cell and its functioning. Proteomics has wide application in diagnosing diseases such as cancer, Alzheimer's disease etc. The data obtained from the diagnostic tools like LC-MS is to be interpreted accurately so as to obtain the correct qualitative and quantitative information about the peptides present in the biological sample. Such interpretation requires and exhaustive knowledge and review about different tools that can be employed and their comparison. This article focuses on comparison of different proteomic tools available for the MS data processing and interpretation. The accuracy demanded during protein identification can be fulfilled by tag based approaches, than PMF or PFF systems. Although, there is a need of standardized matrices for the comparison of the protein identification tools, identifying the single best package for each application from the available literature is at present extremely difficult as each package has its own advantage over other. The datasets and thresholds used in these kinds of comparisons have a critical importance on the outcome of such experiments, and that the high variability in machine and experimental setups complicates analysis. The state of data standards and lack of benchmarks therefore makes it difficult to make an effective comparison. While the increasing availability of data in public repositories and tightening standards will no doubt ameliorate the problem, until this basic benchmarking problem is overcome, no single package or approach can conclusively be declared to outperform all others, expect, perhaps, in the specific circumstances used in particular studies.

show abstract

“…Whereas early peptide fragmentation prediction tools such as MassAnalyzer [99] implement a deductive physicochemical model of peptide fragmentation based on this knowledge, current state-of-the-art prediction tools such as PeptideART [100] and MS2PIP [101] employ a fully data-driven machine learning approach to compute accurate peptide fragmentation models from the amino acid properties in a peptide. This information can be beneficial both to increase the protein identification (coverage) in discovery experiments as well as the proteotypicity of the fragment ions for SRM targeted assays.…”

Section: Fragmentation Modelingmentioning

confidence: 99%

Designing biomedical proteomics experiments: state-of-the-art and future perspectives

Maes

Kelchtermans

Bittremieux

et al. 2016

Expert Review of Proteomics

View full text Add to dashboard Cite

With the current expanded technical capabilities to perform mass spectrometry-based biomedical proteomics experiments, an improved focus on the design of experiments is crucial. As it is clear that ignoring the importance of a good design leads to an unprecedented rate of false discoveries which would poison our results, more and more tools are developed to help researchers designing proteomic experiments. In this review, we apply statistical thinking to go through the entire proteomics workflow for biomarker discovery and validation and relate the considerations that should be made at the level of hypothesis building, technology selection, experimental design and the optimization of the experimental parameters.

show abstract

A Machine Learning Approach to Predicting Peptide Fragmentation Spectra

Cited by 54 publications

References 0 publications

Peak Intensity Prediction for PMF Mass Spectra Using Support Vector Regression

Peak Intensity Prediction for PMF Mass Spectra Using Support Vector Regression

Computational Diagnostics Based on Proteomic Data- Review on Approaches and Algorithms

Designing biomedical proteomics experiments: state-of-the-art and future perspectives

Contact Info

Product

Resources

About