Artificial Intelligence Understands Peptide Observability and Assists With Absolute Protein Quantification

Changing turn-over rates regulate abundance of tryptophan, GS biosynthesis, IAA transport and photosynthesis proteins in Arabidopsis growth defense transitions

BACKGROUND

Shifts in dynamic equilibria of the abundance of cellular molecules in plant-pathogen interactions need further exploration. We induced PTI in optimally growing Arabidopsis thaliana seedlings for 16 h, returning them to growth conditions for another 16 h.

METHODS

Turn-over and abundance of 99 flg22 responding proteins were measured chronologically using a stable heavy nitrogen isotope partial labeling strategy and targeted liquid chromatography coupled to mass spectrometry (PRM LC-MS). These experiments were complemented by measurements of mRNA and phytohormone levels.

RESULTS

Changes in synthesis and degradation rate constants (Ks and Kd) regulated tryptophane and glucosinolate, IAA transport, and photosynthesis-associated protein (PAP) homeostasis in growth/PTI transitions independently of mRNA levels. Ks values increased after elicitation while protein and mRNA levels became uncorrelated. mRNA returned to pre-elicitation levels, yet protein abundance remained at PTI levels even 16 h after media exchange, indicating protein levels were robust and unresponsive to transition back to growth. The abundance of 23 PAPs including FERREDOXIN-NADP( +)-OXIDOREDUCTASE (FNR1) decreased 16 h after PAMP exposure, their depletion was nearly abolished in the myc234 mutant. FNR1 Kd increased as mRNA levels decreased early in PTI, its Ks decreased in prolonged PTI. FNR1 Kd was lower in myc234, mRNA levels decreased as in wild type.

CONCLUSIONS

Protein Kd and Ks values change in response to flg22 exposure and constitute an additional layer of protein abundance regulation in growth defense transitions next to changes in mRNA levels. Our results suggest photosystem remodeling in PTI to direct electron flow away from the photosynthetic carbon reaction towards ROS production as an active defense mechanism controlled post-transcriptionally and by MYC2 and homologs. Target proteins accumulated later and PAP and auxin/IAA depletion was repressed in myc234 indicating a positive effect of the transcription factors in the establishment of PTI.

Proteomics and constraint-based modelling reveal enzyme kinetic properties of Chlamydomonas reinhardtii on a genome scale

Metabolic engineering of microalgae offers a promising solution for sustainable biofuel production, and rational design of engineering strategies can be improved by employing metabolic models that integrate enzyme turnover numbers. However, the coverage of turnover numbers for Chlamydomonas reinhardtii, a model eukaryotic microalga accessible to metabolic engineering, is 17-fold smaller compared to the heterotrophic cell factory Saccharomyces cerevisiae. Here we generate quantitative protein abundance data of Chlamydomonas covering 2337 to 3708 proteins in various growth conditions to estimate in vivo maximum apparent turnover numbers. Using constrained-based modeling we provide proxies for in vivo turnover numbers of 568 reactions, representing a 10-fold increase over the in vitro data for Chlamydomonas. Integration of the in vivo estimates instead of in vitro values in a metabolic model of Chlamydomonas improved the accuracy of enzyme usage predictions. Our results help in extending the knowledge on uncharacterized enzymes and improve biotechnological applications of Chlamydomonas.

DbyDeep: Exploration of MS-Detectable Peptides via Deep Learning

Predicting peptide detectability is useful in a variety of mass spectrometry (MS)-based proteomics applications, particularly targeted proteomics. However, most machine learning-based computational methods have relied solely on information from the peptide itself, such as its amino acid sequences or physicochemical properties, despite the fact that peptides detected by MS are dependent on many factors, including protein sample preparation, digestion, separation, ionization, and precursor selection during MS experiments. DbyDeep (Detectability by Deep learning) is an innovative end-to-end LSTM network model for peptide detectability prediction that incorporates sequence contexts of peptides and their cleavage sites (by protease). Utilizing the cleavage site contexts could improve the performance of prediction, and DbyDeep outperformed existing methods in predicting peptides recognizable from multiple MS/MS data sets with diverse species and MS instruments. We argue for the necessity of a learning model that encompasses several contexts associated with peptide detection, as opposed to depending just on peptide sequences. There is a Python implementation of DbyDeep at https://github.com/BISCodeRepo/DbyDeep.

Insight on physicochemical properties governing peptide MS1 response in HPLC-ESI-MS/MS: A deep learning approach

Accurate and absolute quantification of peptides in complex mixtures using quantitative mass spectrometry (MS)-based methods requires foreground knowledge and isotopically labeled standards, thereby increasing analytical expenses, time consumption, and labor, thus limiting the number of peptides that can be accurately quantified. This originates from differential ionization efficiency between peptides and thus, understanding the physicochemical properties that influence the ionization and response in MS analysis is essential for developing less restrictive label-free quantitative methods. Here, we used equimolar peptide pool repository data to develop a deep learning model capable of identifying amino acids influencing the MS1 response. By using an encoder-decoder with an attention mechanism and correlating attention weights with amino acid physicochemical properties, we obtain insight on properties governing the peptide-level MS1 response within the datasets. While the problem cannot be described by one single set of amino acids and properties, distinct patterns were reproducibly obtained. Properties are grouped in three main categories related to peptide hydrophobicity, charge, and structural propensities. Moreover, our model can predict MS1 intensity output under defined conditions based solely on peptide sequence input. Using a refined training dataset, the model predicted log-transformed peptide MS1 intensities with an average error of 9.7 ± 0.5% based on 5-fold cross validation, and outperformed random forest and ridge regression models on both log-transformed and real scale data. This work demonstrates how deep learning can facilitate identification of physicochemical properties influencing peptide MS1 responses, but also illustrates how sequence-based response prediction and label-free peptide-level quantification may impact future workflows within quantitative proteomics.

One-helix protein 2 is not required for the synthesis of photosystem II subunit D1 in Chlamydomonas

In land plants and cyanobacteria, co-translational association of chlorophyll (Chl) to the nascent D1 polypeptide, a reaction center protein of photosystem II (PSII), requires a Chl binding complex consisting of a short-chain dehydrogenase (high chlorophyll fluorescence 244 [HCF244]/uncharacterized protein 39 [Ycf39]) and one-helix proteins (OHP1 and OHP2 in chloroplasts) of the light-harvesting antenna complex superfamily. Here, we show that an ohp2 mutant of the green alga Chlamydomonas (Chlamydomonas reinhardtii) fails to accumulate core PSII subunits, in particular D1 (encoded by the psbA mRNA). Extragenic suppressors arose at high frequency, suggesting the existence of another route for Chl association to PSII. The ohp2 mutant was complemented by the Arabidopsis (Arabidopsis thaliana) ortholog. In contrast to land plants, where psbA translation is prevented in the absence of OHP2, ribosome profiling experiments showed that the Chlamydomonas mutant translates the psbA transcript over its full length. Pulse labeling suggested that D1 is degraded during or immediately after translation. The translation of other PSII subunits was affected by assembly-controlled translational regulation. Proteomics showed that HCF244, a translation factor which associates with and is stabilized by OHP2 in land plants, still partly accumulates in the Chlamydomonas ohp2 mutant, explaining the persistence of psbA translation. Several Chl biosynthesis enzymes overaccumulate in the mutant membranes. Partial inactivation of a D1-degrading protease restored a low level of PSII activity in an ohp2 background, but not photoautotrophy. Taken together, our data suggest that OHP2 is not required for psbA translation in Chlamydomonas, but is necessary for D1 stabilization.

Toward an Integrated Machine Learning Model of a Proteomics Experiment

In recent years machine learning has made extensive progress in modeling many aspects of mass spectrometry data. We brought together proteomics data generators, repository managers, and machine learning experts in a workshop with the goals to evaluate and explore machine learning applications for realistic modeling of data from multidimensional mass spectrometry-based proteomics analysis of any sample or organism. Following this sample-to-data roadmap helped identify knowledge gaps and define needs. Being able to generate bespoke and realistic synthetic data has legitimate and important uses in system suitability, method development, and algorithm benchmarking, while also posing critical ethical questions. The interdisciplinary nature of the workshop informed discussions of what is currently possible and future opportunities and challenges. In the following perspective we summarize these discussions in the hope of conveying our excitement about the potential of machine learning in proteomics and to inspire future research.

PeptideRanger: An R Package to Optimize Synthetic Peptide Selection for Mass Spectrometry Applications

Targeted and semitargeted mass spectrometry-based approaches are reliable methods to consistently detect and quantify low abundance proteins including proteins of clinical significance. Despite their potential, the development of targeted and semitargeted assays is time-consuming and often requires the purchase of costly libraries of synthetic peptides. To improve the efficiency of this rate-limiting step, we developed PeptideRanger, a tool to identify peptides from protein of interest with physiochemical properties that make them more likely to be suitable for mass spectrometry analysis. PeptideRanger is a flexible, extensively annotated, and intuitive R package that uses a random forest model trained on a diverse data set of thousands of MS experiments spanning a variety of sample types profiled with different chromatography setups and instruments. To support a variety of applications and to leverage rapidly growing public MS databases, PeptideRanger can readily be retrained with experiment-specific data sets and customized to prioritize and filter peptides based on selected properties.

PD-BertEDL: An Ensemble Deep Learning Method Using BERT and Multivariate Representation to Predict Peptide Detectability

Peptide detectability is defined as the probability of identifying a peptide from a mixture of standard samples, which is a key step in protein identification and analysis. Exploring effective methods for predicting peptide detectability is helpful for disease treatment and clinical research. However, most existing computational methods for predicting peptide detectability rely on a single information. With the increasing complexity of feature representation, it is necessary to explore the influence of multivariate information on peptide detectability. Thus, we propose an ensemble deep learning method, PD-BertEDL. Bidirectional encoder representations from transformers (BERT) is introduced to capture the context information of peptides. Context information, sequence information, and physicochemical information of peptides were combined to construct the multivariate feature space of peptides. We use different deep learning methods to capture the high-quality features of different categories of peptides information and use the average fusion strategy to integrate three model prediction results to solve the heterogeneity problem and to enhance the robustness and adaptability of the model. The experimental results show that PD-BertEDL is superior to the existing prediction methods, which can effectively predict peptide detectability and provide strong support for protein identification and quantitative analysis, as well as disease treatment.

Surrogate peptide selection and internal standardization for accurate quantification of endogenous proteins

Relative quantification techniques have dominated the field of proteomics. However, biomarker discovery, mathematical model development and studies on transporter-mediated drug disposition still need absolute quantification of proteins. The quality of data of trace-level protein quantification is solely dependent on the specific selection of surrogate peptides. Selection of surrogate peptides has a major impact on the accuracy of the method. In this article, the advanced approaches for selection of surrogate peptides, which can provide absolute quantification of the proteins are discussed. In addition, internal standardization, which accounts for variations in the quantitation process to achieve absolute protein quantification is discussed.

Generalized Property-Based Encoders and Digital Signal Processing Facilitate Predictive Tasks in Protein Engineering

Computational methods in protein engineering often require encoding amino acid sequences, i.e., converting them into numeric arrays. Physicochemical properties are a typical choice to define encoders, where we replace each amino acid by its value for a given property. However, what property (or group thereof) is best for a given predictive task remains an open problem. In this work, we generalize property-based encoding strategies to maximize the performance of predictive models in protein engineering. First, combining text mining and unsupervised learning, we partitioned the AAIndex database into eight semantically-consistent groups of properties. We then applied a non-linear PCA within each group to define a single encoder to represent it. Then, in several case studies, we assess the performance of predictive models for protein and peptide function, folding, and biological activity, trained using the proposed encoders and classical methods (One Hot Encoder and TAPE embeddings). Models trained on datasets encoded with our encoders and converted to signals through the Fast Fourier Transform (FFT) increased their precision and reduced their overfitting substantially, outperforming classical approaches in most cases. Finally, we propose a preliminary methodology to create de novo sequences with desired properties. All these results offer simple ways to increase the performance of general and complex predictive tasks in protein engineering without increasing their complexity.

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.

Prediction of Peptide Detectability Based on CapsNet and Convolutional Block Attention Module

According to proteomics technology, as impacted by the complexity of sampling in the experimental process, several problems remain with the reproducibility of mass spectrometry experiments, and the peptide identification and quantitative results continue to be random. Predicting the detectability exhibited by peptides can optimize the mentioned results to be more accurate, so such a prediction is of high research significance. This study builds a novel method to predict the detectability of peptides by complying with the capsule network (CapsNet) and the convolutional block attention module (CBAM). First, the residue conical coordinate (RCC), the amino acid composition (AAC), the dipeptide composition (DPC), and the sequence embedding code (SEC) are extracted as the peptide chain features. Subsequently, these features are divided into the biological feature and sequence feature, and separately inputted into the neural network of CapsNet. Moreover, the attention module CBAM is added to the network to assign weights to channels and spaces, as an attempt to enhance the feature learning and improve the network training effect. To verify the effectiveness of the proposed method, it is compared with some other popular methods. As revealed from the experimentally achieved results, the proposed method outperforms those methods in most performance assessments.

Data Management and Modeling in Plant Biology

The study of plant-environment interactions is a multidisciplinary research field. With the emergence of quantitative large-scale and high-throughput techniques, amount and dimensionality of experimental data have strongly increased. Appropriate strategies for data storage, management, and evaluation are needed to make efficient use of experimental findings. Computational approaches of data mining are essential for deriving statistical trends and signatures contained in data matrices. Although, current biology is challenged by high data dimensionality in general, this is particularly true for plant biology. Plants as sessile organisms have to cope with environmental fluctuations. This typically results in strong dynamics of metabolite and protein concentrations which are often challenging to quantify. Summarizing experimental output results in complex data arrays, which need computational statistics and numerical methods for building quantitative models. Experimental findings need to be combined by computational models to gain a mechanistic understanding of plant metabolism. For this, bioinformatics and mathematics need to be combined with experimental setups in physiology, biochemistry, and molecular biology. This review presents and discusses concepts at the interface of experiment and computation, which are likely to shape current and future plant biology. Finally, this interface is discussed with regard to its capabilities and limitations to develop a quantitative model of plant-environment interactions.

Multiomic Big Data Analysis Challenges: Increasing Confidence in the Interpretation of Artificial Intelligence Assessments

The need for holistic molecular measurements to better understand disease initiation, development, diagnosis, and therapy has led to an increasing number of multiomic analyses. The wealth of information available from multiomic assessments, however, requires both the evaluation and interpretation of extremely large data sets, limiting analysis throughput and ease of adoption. Computational methods utilizing artificial intelligence (AI) provide the most promising way to address these challenges, yet despite the conceptual benefits of AI and its successful application in singular omic studies, the widespread use of AI in multiomic studies remains limited. Here, we discuss present and future capabilities of AI techniques in multiomic studies while introducing analytical checks and balances to validate the computational conclusions.

mRNAs, proteins and the emerging principles of gene expression control

Gene expression involves transcription, translation and the turnover of mRNAs and proteins. The degree to which protein abundances scale with mRNA levels and the implications in cases where this dependency breaks down remain an intensely debated topic. Here we review recent mRNA-protein correlation studies in the light of the quantitative parameters of the gene expression pathway, contextual confounders and buffering mechanisms. Although protein and mRNA levels typically show reasonable correlation, we describe how transcriptomics and proteomics provide useful non-redundant readouts. Integrating both types of data can reveal exciting biology and is an essential step in refining our understanding of the principles of gene expression control.

Overexpression of Sedoheptulose-1,7-Bisphosphatase Enhances Photosynthesis in Chlamydomonas reinhardtii and Has No Effect on the Abundance of Other Calvin-Benson Cycle Enzymes

The productivity of plants and microalgae needs to be increased to feed the growing world population and to promote the development of a low-carbon economy. This goal can be achieved by improving photosynthesis via genetic engineering. In this study, we have employed the Modular Cloning strategy to overexpress the Calvin-Benson cycle (CBC) enzyme sedoheptulose-1,7-bisphosphatase (SBP1) up to threefold in the unicellular green alga Chlamydomonas reinhardtii. The protein derived from the nuclear transgene represented ∼0.3% of total cell protein. Photosynthetic rate and growth were significantly increased in SBP1-overexpressing lines under high-light and elevated CO2 conditions. Absolute quantification of the abundance of all other CBC enzymes by the QconCAT approach revealed no consistent differences between SBP1-overexpressing lines and the recipient strain. This analysis also revealed that the 11 CBC enzymes represent 11.9% of total cell protein in Chlamydomonas. Here, the range of concentrations of CBC enzymes turned out to be much larger than estimated earlier, with a 128-fold difference between the most abundant CBC protein (rbcL) and the least abundant (triose phosphate isomerase). Accordingly, the concentrations of the CBC intermediates are often but not always higher than the binding site concentrations of the enzymes for which they act as substrates. The enzymes with highest substrate to binding site ratios might represent good candidates for overexpression in subsequent engineering steps.

Identification of Chloroplast Envelope Proteins with Critical Importance for Cold Acclimation

The ability of plants to withstand cold temperatures relies on their photosynthetic activity. Thus, the chloroplast is of utmost importance for cold acclimation and acquisition of freezing tolerance. During cold acclimation, the properties of the chloroplast change markedly. To provide the most comprehensive view of the protein repertoire of the chloroplast envelope, we analyzed this membrane system in Arabidopsis (Arabidopsis thaliana) using mass spectrometry-based proteomics. Profiling chloroplast envelope membranes was achieved by a cross comparison of protein intensities across the plastid and the enriched membrane fraction under both normal and cold conditions. We used multivariable logistic regression to model the probabilities for the classification of an envelope localization. In total, we identified 38 envelope membrane intrinsic or associated proteins exhibiting altered abundance after cold acclimation. These proteins comprise several solute carriers, such as the ATP/ADP antiporter nucleotide transporter2 (NTT2; substantially increased abundance) or the maltose exporter MEX1 (substantially decreased abundance). Remarkably, analysis of the frost recovery of ntt loss-of-function and mex1 overexpressor mutants confirmed that the comparative proteome is well suited to identify key factors involved in cold acclimation and acquisition of freezing tolerance. Moreover, for proteins with known physiological function, we propose scenarios explaining their possible roles in cold acclimation. Furthermore, spatial proteomics introduces an additional layer of complexity and enables the identification of proteins differentially localized at the envelope membrane under the changing environmental regime.

Quality control of imbalanced mass spectra from isotopic labeling experiments

BACKGROUND

Mass spectra are usually acquired from the Liquid Chromatography-Mass Spectrometry (LC-MS) analysis for isotope labeled proteomics experiments. In such experiments, the mass profiles of labeled (heavy) and unlabeled (light) peptide pairs are represented by isotope clusters (2D or 3D) that provide valuable information about the studied biological samples in different conditions. The core task of quality control in quantitative LC-MS experiment is to filter out low-quality peptides with questionable profiles. The commonly used methods for this problem are the classification approaches. However, the data imbalance problems in previous control methods are often ignored or mishandled. In this study, we introduced a quality control framework based on the extreme gradient boosting machine (XGBoost), and carefully addressed the imbalanced data problem in this framework.

RESULTS

In the XGBoost based framework, we suggest the application of the Synthetic minority over-sampling technique (SMOTE) to re-balance data and use the balanced data to train the boosted trees as the classifier. Then the classifier is applied to other data for the peptide quality assessment. Experimental results show that our proposed framework increases the reliability of peptide heavy-light ratio estimation significantly.

CONCLUSIONS

Our results indicate that this framework is a powerful method for the peptide quality assessment. For the feature extraction part, the extracted ion chromatogram (XIC) based features contribute to the peptide quality assessment. To solve the imbalanced data problem, SMOTE brings a much better classification performance. Finally, the XGBoost is capable for the peptide quality control. Overall, our proposed framework provides reliable results for the further proteomics studies.

DeepMSPeptide: peptide detectability prediction using deep learning

Summary

The protein detection and quantification using high-throughput proteomic technologies is still challenging due to the stochastic nature of the peptide selection in the mass spectrometer, the difficulties in the statistical analysis of the results and the presence of degenerated peptides. However, considering in the analysis only those peptides that could be detected by mass spectrometry, also called proteotypic peptides, increases the accuracy of the results. Several approaches have been applied to predict peptide detectability based on the physicochemical properties of the peptides. In this manuscript, we present DeepMSPeptide, a bioinformatic tool that uses a deep learning method to predict proteotypic peptides exclusively based on the peptide amino acid sequences.

Availability and implementation

DeepMSPeptide is available at https://github.com/vsegurar/DeepMSPeptide.

Supplementary information

Supplementary data are available at Bioinformatics online.