Peak intensity prediction in MALDI-TOF mass spectrometry: A machine learning study to support quantitative proteomics

Timm, Wiebke; Scherbart, Alexandra; Böcker, Sebastian; Kohlbacher, Oliver; Nattkemper, Tim Wilhelm

doi:10.1186/1471-2105-9-443

Cited by 33 publications

(24 citation statements)

References 38 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…[39][40][41][42][43] The resulting quantitative count data was fit using a quasipoisson model and a log 2 fold change was calculated. 44,45 A run effect was also applied as one sample each of human and macaque milk was ran several months prior to the remaining samples.…”

Section: Homolog Identification and Quantitative Analysismentioning

confidence: 99%

Comparative Proteomics of Human and Macaque Milk Reveals Species-Specific Nutrition during Postnatal Development

Beck¹,

Weber²,

Phinney³

et al. 2015

J. Proteome Res.

View full text Add to dashboard Cite

Milk has been well established as the optimal nutrition source for infants, yet there is still much to be understood about its molecular composition. Therefore, our objective was to develop and compare comprehensive milk proteomes for human and rhesus macaques to highlight differences in neonatal nutrition. We developed a milk proteomics technique that overcomes previous technical barriers including pervasive post-translational modifications and limited sample volume. We identified 1606 and 518 proteins in human and macaque milk, respectively. During analysis of detected protein orthologs, we identified 88 differentially abundant proteins. Of these, 93% exhibited increased abundance in human milk relative to macaque and include lactoferrin, polymeric immunoglobulin receptor, alpha-1 antichymotrypsin, vitamin D-binding protein, and haptocorrin. Furthermore, proteins more abundant in human milk compared with macaque are associated with development of the gastrointestinal tract, the immune system, and the brain. Overall, our novel proteomics method reveals the first comprehensive macaque milk proteome and 524 newly identified human milk proteins. The differentially abundant proteins observed are consistent with the perspective that human infants, compared with nonhuman primates, are born at a slightly earlier stage of somatic development and require additional support through higher quantities of specific proteins to nurture human infant maturation.

show abstract

Section: Homolog Identification and Quantitative Analysismentioning

confidence: 99%

Comparative Proteomics of Human and Macaque Milk Reveals Species-Specific Nutrition during Postnatal Development

Beck¹,

Weber²,

Phinney³

et al. 2015

J. Proteome Res.

View full text Add to dashboard Cite

show abstract

“…The confidence interval (95% of confidence level) is better for the SVM prediction compares to the values from the Cargile and Bjellqvist methods. The use of SVM algorithms and machine learning techniques in general, give the possibility to find a new model to predict the isoelectric point given some background knowledge (reliable identifications) from all fractions [20][21][22][23]. Another published dataset was also used to demonstrate that the model can predict accurately across diverse dataset and experimental settings.…”

mentioning

confidence: 99%

Isoelectric point optimization using peptide descriptors and support vector machines

Perez‐Riverol

Audain

Millan

et al. 2012

Journal of Proteomics

View full text Add to dashboard Cite

“…It is known that peak intensity is additive. The -trimmed mean has been used to deal with one-to-many (15). For the many-to-one case, two or more theoretical peptides that are matched to the same experimental peak should share the intensity of the latter.…”

Section: Newly Applied Parameters Of Spectrum Alignmentmentioning

confidence: 99%

“…Additional parameters such as pI, molecular mass, post-modification, and chemical composition (7), methods such as a proteotypic peptide library (8) and integrated information (9) have been taken into account. The introduction of new parameters, such as spectrum similarity (10), mass spectra alignment (11), peak bagging (12), negative ionization (13), probability-based scoring function (14), peak intensity prediction (15,16), mass accuracy (17,18), mass tolerance (19), and a validation system (20) has improved the accuracy of PMF identification.…”

mentioning

confidence: 99%

Feature-matching Pattern-based Support Vector Machines for Robust Peptide Mass Fingerprinting

Hao

Zhang

et al. 2011

Molecular & Cellular Proteomics

View full text Add to dashboard Cite

Peptide mass fingerprinting, regardless of becoming complementary to tandem mass spectrometry for protein identification, is still the subject of in-depth study because of its higher sample throughput, higher level of specificity for single peptides and lower level of sensitivity to unexpected post-translational modifications compared with tandem mass spectrometry. In this study, we propose, implement and evaluate a uniform approach using support vector machines to incorporate individual concepts and conclusions for accurate PMF.We focus on the inherent attributes and critical issues of the theoretical spectrum (peptides), the experimental spectrum (peaks) and spectrum (masses) alignment. Eighty-one feature-matching patterns derived from cleavage type, uniqueness and variable masses of theoretical peptides together with the intensity rank of experimental peaks were proposed to characterize the matching profile of the peptide mass fingerprinting procedure. We developed a new strategy including the participation of matched peak intensity redistribution to handle shared peak intensities and 440 parameters were generated to digitalize each feature-matching pattern. A high performance for an evaluation data set of 137 items was finally achieved by the optimal multi-criteria support vector machines approach, with 491 final features out of a feature vector of 35,640 normalized features through cross training and validating a publicly available "gold standard" peptide mass fingerprinting data set of 1733 items. Compared with the Mascot, MS-Fit, ProFound and Aldente algorithms commonly used for MS-based protein identification, the feature-matching patterns algorithm has a greater ability to clearly separate correct identifications and random matches with the highest values for sensitivity (82%), precision (97%) and F1-measure (89%) of protein identification.Several conclusions reached via this research make general contributions to MS-based protein identification. Firstly, inherent attributes showed comparable or even greater robustness than other explicit. As an inherent attribute of an experimental spectrum, peak intensity should receive considerable attention during protein identification. Secondly, alignment between intense experimental peaks and properly digested, unique or non-modified theoretical peptides is very likely to occur in positive peptide mass fingerprinting. Finally, normalization by several types of harmonic factors, including missed cleavages and mass modification, can make important contributions to the performance of the procedure. Molecular

show abstract

Peak intensity prediction in MALDI-TOF mass spectrometry: A machine learning study to support quantitative proteomics

Cited by 33 publications

References 38 publications

Comparative Proteomics of Human and Macaque Milk Reveals Species-Specific Nutrition during Postnatal Development

Comparative Proteomics of Human and Macaque Milk Reveals Species-Specific Nutrition during Postnatal Development

Isoelectric point optimization using peptide descriptors and support vector machines

Feature-matching Pattern-based Support Vector Machines for Robust Peptide Mass Fingerprinting

Contact Info

Product

Resources

About