Prediction of Peptide Fragment Ion Mass Spectra by Data Mining Techniques

Prediction of peptide mass spectral libraries with machine learning

The recent development of machine learning methods to identify peptides in complex mass spectrometric data constitutes a major breakthrough in proteomics. Longstanding methods for peptide identification, such as search engines and experimental spectral libraries, are being superseded by deep learning models that allow the fragmentation spectra of peptides to be predicted from their amino acid sequence. These new approaches, including recurrent neural networks and convolutional neural networks, use predicted in silico spectral libraries rather than experimental libraries to achieve higher sensitivity and/or specificity in the analysis of proteomics data. Machine learning is galvanizing applications that involve large search spaces, such as immunopeptidomics and proteogenomics. Current challenges in the field include the prediction of spectra for peptides with post-translational modifications and for cross-linked pairs of peptides. Permeation of machine-learning-based spectral prediction into search engines and spectrum-centric data-independent acquisition workflows for diverse peptide classes and measurement conditions will continue to push sensitivity and dynamic range in proteomics applications in the coming years.

Multiaspect Examinations of Possible Alternative Mappings of Identified Variant Peptides: A Case Study on the HEK293 Cell Line

Adopting proteogenomics approach to validate single nucleotide variation events by identifying corresponding single amino acid variant peptides from mass spectrometry (MS)-based proteomics data facilitates translational and clinical research. Although variant peptides are usually identified from MS data with a stringent false discovery rate (FDR), FDR control could fail to eliminate dubious results caused by several issues; thus, postexamination to eliminate dubious results is required. However, comprehensive postexaminations of identification results are still lacking. Therefore, we propose a framework of three bottom-up levels, peptide-spectrum match, peptide, and variant event levels, that consists of rigorous 11-aspect examinations from the MS perspective to further confirm the reliability of variant events. As a proof of concept and showing feasibility, we demonstrate 11 examinations on the identified variant peptides from an HEK293 cell line data set, where various database search strategies were applied to maximize the number of identified variant PSMs with an FDR <1% for postexaminations. The results showed that only FDR criterion is insufficient to validate identified variant peptides and the 11 postexaminations can reveal low-confidence variant events detected by shotgun proteomics experiments. Therefore, we suggest that postexaminations of identified variant events based on the proposed framework are necessary for proteogenomics studies.

Deep learning approaches for data-independent acquisition proteomics

INTRODUCTION

Data-independent acquisition (DIA) is an emerging technology for large-scale proteomic studies. DIA data analysis methods are evolving rapidly, and deep learning has cut a conspicuous figure in this field.

AREAS COVERED

This review discusses and provides an overview of the deep learning methods that are used for DIA data analysis, including spectral library prediction, feature scoring, and statistical control in peptide-centric analysis, as well as de novo peptide sequencing. Literature searches were performed for articles, including preprints, up to December 2021 from PubMed, Scopus, and Web of Science databases.

EXPERT OPINION

While spectral library prediction has broken through the limitation on proteome coverage of experimental libraries, the statistical burden due to the large query space is the remaining challenge of utilizing proteome-wide predicted libraries. Analysis of post-translational modifications is another promising direction of deep learning-based DIA methods.

Deep Learning in Proteomics

Proteomics, the study of all the proteins in biological systems, is becoming a data-rich science. Protein sequences and structures are comprehensively catalogued in online databases. With recent advancements in tandem mass spectrometry (MS) technology, protein expression and post-translational modifications (PTMs) can be studied in a variety of biological systems at the global scale. Sophisticated computational algorithms are needed to translate the vast amount of data into novel biological insights. Deep learning automatically extracts data representations at high levels of abstraction from data, and it thrives in data-rich scientific research domains. Here, a comprehensive overview of deep learning applications in proteomics, including retention time prediction, MS/MS spectrum prediction, de novo peptide sequencing, PTM prediction, major histocompatibility complex-peptide binding prediction, and protein structure prediction, is provided. Limitations and the future directions of deep learning in proteomics are also discussed. This review will provide readers an overview of deep learning and how it can be used to analyze proteomics data.

A Frequency-Based Approach to Predict the Low-Energy Collision-Induced Dissociation Fragmentation Spectra

Peptide identification algorithms rely on the comparison between the experimental tandem mass spectrometry spectrum and the theoretical spectrum to identify a peptide from the tandem mass spectra. Hence, it is important to understand the fragmentation process and predict the tandem mass spectra for high-throughput proteomics research. In this study, a novel method was developed to predict the theoretical ion trap collision-induced dissociation (CID) tandem mass spectra of the singly, doubly, and triply charged tryptic peptides. The fragmentation statistics of the ion trap CID spectra were used to predict the theoretical tandem mass spectra of the peptide sequence. The study estimated the relative cleavage frequency for each pair of adjacent amino acids along the peptide length. The study showed that the cleavage frequency can be directly used to predict the tandem mass spectra. The predicted spectra show a high correlation with the experimental spectra used in this study; 99.73% of the high-quality reference spectra have correlation scores greater than 0.8. The new method predicts the theoretical spectrum and correlates significantly better with the experimental spectrum as compared to the existing spectrum prediction tools OpenMS_Simulator, MS2PIP, and MS2PBPI, where only 80, 85.76, and 85.80% of the spectral count, respectively, has a correlation score greater than 0.8.

High-quality MS/MS spectrum prediction for data-dependent and data-independent acquisition data analysis

Peptide fragmentation spectra are routinely predicted in the interpretation of mass-spectrometry-based proteomics data. However, the generation of fragment ions has not been understood well enough for scientists to estimate fragment ion intensities accurately. Here, we demonstrate that machine learning can predict peptide fragmentation patterns in mass spectrometers with accuracy within the uncertainty of measurement. Moreover, analysis of our models reveals that peptide fragmentation depends on long-range interactions within a peptide sequence. We illustrate the utility of our models by applying them to the analysis of both data-dependent and data-independent acquisition datasets. In the former case, we observe a q-value-dependent increase in the total number of peptide identifications. In the latter case, we confirm that the use of predicted tandem mass spectrometry spectra is nearly equivalent to the use of spectra from experimental libraries.

Improving Peptide-Spectrum Matching by Fragmentation Prediction Using Hidden Markov Models

Tandem mass spectrometry has become the method of choice for high-throughput, quantitative analysis in proteomics. Peptide spectrum matching algorithms score the concordance between the experimental and the theoretical spectra of candidate peptides by evaluating the number (or proportion) of theoretically possible fragment ions observed in the experimental spectra without any discrimination. However, the assumption that each theoretical fragment is just as likely to be observed is inaccurate. On the contrary, MS2 spectra often have few dominant fragments. Using millions of MS/MS spectra we show that there is high reproducibility across different fragmentation spectra given the precursor peptide and charge state, implying that there is a pattern to fragmentation. To capture this pattern we propose a novel prediction algorithm based on hidden Markov models with an efficient training process. We investigated the performance of our interpolated-HMM model, trained on millions of MS2 spectra, and found that our model picks up meaningful patterns in peptide fragmentation. Second, looking at the variability of the prediction performance by varying the train/test data split, we observed that our model performs well independent of the specific peptides that are present in the training data. Furthermore, we propose that the real value of this model is as a preprocessing step in the peptide identification process. The model can discern fragment ions that are unlikely to be intense for a given candidate peptide rather than using the actual predicted intensities. As such, probabilistic measures of concordance between experimental and theoretical spectra will leverage better statistics.