Evaluation of Time-of-Flight Secondary Ion Mass Spectrometry Spectra of Peptides by Random Forest with Amino Acid Labels: Results from a Versailles Project on Advanced Materials and Standards Interlaboratory Study

Aoyagi, Satoka; Fujiwara, Yukio; Takano, Ayumi; Vorng, Jean‐Luc; Gilmore, Ian S.; Wang, Yung-Chen; Tallarek, Elke; Hagenhoff, Birgit; Iida, Shin‐ichi; Luch, Andreas; Jungnickel, Harald; Lang, Yusheng; Shon, Hyun Kyong; Lee, Tae Geol; Li, Zhanping; Matsuda, Kazuhiro; Mihara, Ichiro; Miisho, Ako; Murayama, Yohei; Nagatomi, T.; Ikeda, Reiko; Okamoto, Masayuki; Saiga, Kunio; Tsuchiya, Toshihiko; Uemura, Shigeaki

doi:10.1021/acs.analchem.0c04577

Cited by 14 publications

(18 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…14 Machine and deep learning methods have been employed for interpreting complex mass spectral and imaging data. [15][16][17][18][19][20][21][22][23][24][25] For example, an autoencoder is an ANN-based unsupervised method. It has been applied to interpret complex ToF-SIMS data 20,21,24 and to extract important sample features.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Quantitative analysis of ToF‐SIMS data of a two organic compound mixture using an autoencoder and simple artificial neural networks

Aoyagi

Matsuda

2022

Rapid Comm Mass Spectrometry

Self Cite

View full text Add to dashboard Cite

Rationale Matrix effects cause a nonlinear relationship between ion intensities and concentrations in mass spectrometry, including time‐of‐flight secondary ion mass spectrometry (ToF‐SIMS). Here, two artificial neural network (ANN)‐based methods, autoencoder‐based and simple ANN methods, were employed for the quantitative and qualitative analyses of a two organic compound mixture via ToF‐SIMS. Methods The multilayer model sample contained a mixture of Irganox 1010 and Fmoc‐pentafluoro‐L‐phenylalanine (Fmoc‐PFLPA). The sample's positive and negative ion depth profiles were collected through ToF‐SIMS. ToF‐SIMS‐derived cross‐sectional image datasets were analyzed using three unsupervised methods, namely principal component analysis (PCA), multivariate curve resolution (MCR), and use of a sparse autoencoder (SAE). The supervised simple ANN method was optimized based on the spectra and validated by predicting the test dataset ratios of Irganox 1010. Results The results obtained using the SAE demonstrated linear calibration curves and appropriate material distribution images. The Irganox 1010 and Fmoc‐PFLPA positive and negative ion datasets exhibited >0.97 correlation coefficients. The PCA and MCR results demonstrated lower linearity than that of SAE. Moreover, SAE weights indicated the ions important for each organic compound. The simple ANN method accurately predicted the ratios in the test dataset and indicated the important ions. Conclusions Both the supervised and unsupervised methods based on ANN, which were employed in regulating nonlinear relationships, were effective in the quantitative and qualitative analyses of the ToF‐SIMS data of the two organic compound mixtures. Regarding qualitative analysis, both ANN‐based methods indicated specific ions from the molecules in the sample.

show abstract

Section: Introductionmentioning

confidence: 99%

“…Artificial neural network (ANN)‐based flexible methods allow the interpretation of complex ToF‐SIMS data despite the matrix effects 14 . Machine and deep learning methods have been employed for interpreting complex mass spectral and imaging data 15–25 . For example, an autoencoder is an ANN‐based unsupervised method.…”

Section: Introductionmentioning

confidence: 99%

Quantitative analysis of ToF‐SIMS data of a two organic compound mixture using an autoencoder and simple artificial neural networks

Aoyagi

Matsuda

2022

Rapid Comm Mass Spectrometry

Self Cite

View full text Add to dashboard Cite

show abstract

“…Due to the complexity of the abovementioned systems, the classical mass-interpretation methods were replaced with advanced data analysis techniques. For instance, Aoyagi et al [47] conducted an interlaboratory study on the identification of the peptide sample ToF-SIMS data by machine learning where the spectra of the test peptide sample were predicted by random forest. Also, in medical studies, Zhang et al [48] used machine learning algorithms on serum blueprints extracted from LDI-MS for the diagnosis of coronary heart disease.…”

Section: Introductionmentioning

confidence: 99%

Sorption of Fulvic Acids onto Titanium Dioxide Nanoparticles Extracted from Commercial Sunscreens: ToF-SIMS and High-Dimensional Data Analysis

et al. 2022

View full text Add to dashboard Cite

Titanium dioxide nanoparticles (n-TiO2) are common ingredients of sunscreens and are often released into surface waters during usage. Once released, the surface chemistry of n-TiO2 changes by interacting with dissolved organic matter (DOM). In previous studies, these interactions were investigated using model n-TiO2 and; therefore, do not account for the complex composition of the coating of n-TiO2 aged in sunscreens. Taking advantage of a mild extraction method to provide more realistic nanoparticles, we investigated the potentials of time of flight-secondary ion mass spectrometry (ToF-SIMS) combined with high-dimensional data analysis to characterize the sorption of fulvic acids, as a model for DOM, on titanium dioxide nanoparticles extracted from ten different commercial sunscreens (n-TiO2 ⸦ sunscreen). Clustering analysis confirmed the ability of ToF-SIMS to detect the sorption of fulvic acids. Moreover, a unique sorption pattern was recognized for each n-TiO2 ⸦ sunscreen, which implied different fractionation of fulvic acids based on the initial specifications of nanoparticles, e.g., size, coating, etc. Furthermore, random forest was used to extract the most important fragments for predicting the presence of fulvic acids on the surface of n-TiO2 ⸦ sunscreen. Finally, we evaluate the potential of ToF-SIMS for characterizing the sorption layer.

show abstract

“…Low surface amounts (submonolayer coverage) of single amino acids (e.g., Phe, Gly and Leu 22 ) and peptides such as RGD, 23,24 RGDS 25 and phosphorylated RGDS 25 were detectable by ToF‐SIMS and with the recent development of MS–MS capabilities for SIMS, proteins can also be identified 26 . Increasingly sophisticated tools such as machine learning algorithms 27 emerge to aid qualitative SIMS data interpretation and identification of amino acids and peptides. Due to the matrix effect, that is, the dependence of secondary ion intensities on the presence of other materials in the sample ion intensities derived from ToF‐SIMS are generally not considered quantitative 28 .…”

Section: Introductionmentioning

confidence: 99%

Quantifiable correlation of ToF‐SIMS and XPS data from polymer surfaces with controlled amino acid and peptide content

Taylor

Simões

Smith

et al. 2022

Surface & Interface Analysis

View full text Add to dashboard Cite

Peptide‐coated surfaces are widely employed in biomaterial design, but quantifiable correlation between surface composition and biological response is challenging due to, for example, instrumental limitations, a lack of suitable model surfaces or limitations in quantitatively correlating data from different surface analytical techniques. Here, we first establish a reference material that allows control over amino acid content. Reversible addition‐fragmentation chain‐transfer (RAFT) polymerisation is used to prepare a copolymer containing alkyne and furan units with well‐defined chain length and composition. Huisgen Cu(I)‐catalysed azide‐alkyne cycloaddition reaction is used to attach the model azido‐polyethyleneglycol‐amide‐modified pentafluoro‐l‐phenylalanine to the polymer. Different compositional ratios of the polymer provide a surface with varying amino acid content that is analysed by X‐ray photoelectron spectroscopy (XPS) and time‐of‐flight secondary ion mass spectrometry (ToF‐SIMS). Nitrogen‐related signals are compared with fluorine signals from both techniques. Fluorine and nitrogen signals from both techniques are found to be related to the copolymer compositions, but the homopolymer data deviate from this trend. The approach is then translated to a heparin‐binding peptide that supports cell adhesion. Human embryonic stem cells cultured on copolymer surfaces presenting different amounts of heparin‐binding peptide show strong cell growth while maintaining pluripotency after 72 h of culture. The early cell adhesion at 24 h can be correlated to the logarithm of the normalised CH4N+ ion intensity from ToF‐SIMS data, which is established as a suitable and generalisable marker ion for amino acids and peptides. This work contributes to the ability to use ToF‐SIMS in a more quantitative manner for the analysis of amino acid and peptide surfaces.

show abstract

Evaluation of Time-of-Flight Secondary Ion Mass Spectrometry Spectra of Peptides by Random Forest with Amino Acid Labels: Results from a Versailles Project on Advanced Materials and Standards Interlaboratory Study

Cited by 14 publications

References 31 publications

Quantitative analysis of ToF‐SIMS data of a two organic compound mixture using an autoencoder and simple artificial neural networks

Quantitative analysis of ToF‐SIMS data of a two organic compound mixture using an autoencoder and simple artificial neural networks

Sorption of Fulvic Acids onto Titanium Dioxide Nanoparticles Extracted from Commercial Sunscreens: ToF-SIMS and High-Dimensional Data Analysis

Quantifiable correlation of ToF‐SIMS and XPS data from polymer surfaces with controlled amino acid and peptide content

Contact Info

Product

Resources

About