Prediction Of Retention Times Research Articles

Prosit-TMT: Deep Learning Boosts Identification of TMT-Labeled Peptides.

The prediction of fragment ion intensities and retention time of peptides has gained significant attention over the past few years. However, the progress shown in the accurate prediction of such properties focused primarily on unlabeled peptides. Tandem mass tags (TMT) are chemical peptide labels that are coupled to free amine groups usually after protein digestion to enable the multiplexed analysis of multiple samples in bottom-up mass spectrometry. It is a standard workflow in proteomics ranging from single-cell to high-throughput proteomics. Particularly for TMT, increasing the number of confidently identified spectra is highly desirable as it provides identification and quantification information with every spectrum. Here, we report on the generation of an extensive resource of synthetic TMT-labeled peptides as part of the ProteomeTools project and present the extension of the deep learning model Prosit to accurately predict the retention time and fragment ion intensities of TMT-labeled peptides with high accuracy. Prosit-TMT supports CID and HCD fragmentation and ion trap and Orbitrap mass analyzers in a single model. Reanalysis of published TMT data sets show that this single model extracts substantial additional information. Applying Prosit-TMT, we discovered that the expression of many proteins in human breast milk follows a distinct daily cycle which may prime the newborn for nutritional or environmental cues.

Identification of new Omega-3 very long chain poly-unsaturated fatty acids in meibomian gland secretions

Three new very long chain polyunsaturated fatty acids (VLC PUFA) belonging to the omega-3 family have been identified in meibum samples collected by Schirmer strips. These VLC PUFA, namely FA (32:3), FA (34:3) and FA (36:3), were detected in O-acyl-ω-hydroxy fatty acids using a molecular network approach, and as free fatty acids. Identification was supported by retention time prediction model, exact mass determination and isotopic patterns. Double bond location was determined using cross metathesis reaction associated to tandem mass spectrometry. In meibum, synthesis of these VLC PUFA is likely to be mediated by elongation of very long chain fatty acids 4 enzyme. The biological role of these newly VLC PUFA and their occurrence in other tissues and biological fluids remains to be elucidated.

Machine Learning-Based Retention Time Prediction of Trimethylsilyl Derivatives of Metabolites.

In gas chromatography–mass spectrometry-based untargeted metabolomics, metabolites are identified by comparing mass spectra and chromatographic retention time with reference databases or standard materials. In that sense, machine learning has been used to predict the retention time of metabolites lacking reference data. However, the retention time prediction of trimethylsilyl derivatives of metabolites, typically analyzed in untargeted metabolomics using gas chromatography, has been poorly explored. Here, we provide a rationalized framework for machine learning-based retention time prediction of trimethylsilyl derivatives of metabolites in gas chromatography. We compared different machine learning paradigms, in addition to exploring the influence of the computational molecular structure representation to train the prediction models: fingerprint class and fingerprint calculation software. Our study challenged predicted retention time when using chemical ionization and electron impact ionization sources in simulated and real cases, demonstrating a good correct identity ranking capability by machine learning, despite observing a limited false identity filtering power in cases where a spectrum or a monoisotopic mass match to multiple candidates. Specifically, machine learning prediction yielded median absolute and relative retention index (relative retention time) errors of 37.1 retention index units and 2%, respectively. In addition, fingerprint class and fingerprint calculation software, as well as the molecular structural similarity between the training and test or real case sets, showed to be critical modulators of the prediction performance. Finally, we leveraged the structural similarity between the training and test or real case set to determine the probability that the prediction error is below a specific threshold. Overall, our study demonstrates that predicted retention time can provide insights into the true structure of unknown metabolites by ranking from the most to the least plausible molecular identity, and sets the guidelines to assess the confidence in metabolite identification using predicted retention time data.

Building machine-learning-based models for retention time and resolution predictions in ion pair chromatography of oligonucleotides

Support vector regression models are created and used to predict the retention times of oligonucleotides separated using gradient ion-pair chromatography with high accuracy. The experimental dataset consisted of fully phosphorothioated oligonucleotides. Two models were trained and validated using two pseudo-orthogonal gradient modes and three gradient slopes. The results show that the spread in retention time differs between the two gradient modes, which indicated varying degree of sequence dependent separation. Peak widths from the experimental dataset were calculated and correlated with the guanine-cytosine content and retention time of the sequence for each gradient slope. This data was used to predict the resolution of the n – 1 impurity among 250 000 random 12- and 16-mer sequences; showing one of the investigated gradient modes has a much higher probability of exceeding a resolution of 1.5, particularly for the 16-mer sequences. Sequences having a high guanine-cytosine content and a terminal C are more likely to not reach critical resolution. The trained SVR models can both be used to identify characteristics of different separation methods and to assist in the choice of method conditions, i.e. to optimize resolution for arbitrary sequences. The methodology presented in this study can be expected to be applicable to predict retention times of other oligonucleotide synthesis and degradation impurities if provided enough training data.

Improving retention-time prediction in supercritical-fluid chromatography by multivariate modelling

The prediction of chromatographic retention under supercritical-fluid chromatography (SFC) conditions was studied, using established and novel theoretical models over ranges of modifier content, pressure and temperature. Whereas retention models used for liquid chromatography often only consider the modifier fraction, retention in SFC depends much more strongly on pressure and temperature. The viability of combining several retention models into surfaces that describe the effects of both modifier fraction and pressure was investigated. The ability of commonly used retention models to describe retention as a function of modifier fraction, expressed either as mass or volume fraction, pressure and density was assessed. Using the multivariate surfaces, retention-time prediction for isocratic separations at constant temperature improved significantly compared to univariate modelling when both pressure and modifier fractions were changed. The “mixed-mode” model with an additional exponential pressure or density parameter was able to predict retention times within 5%, with the majority of the predictions within 2%. The use of mass fraction and density further improves retention modelling compared to volume fraction and pressure. These variables however, do require extra computations.

MultiConditionRT: Predicting liquid chromatography retention time for emerging contaminants for a wide range of eluent compositions and stationary phases

Structural elucidation of compounds detected with liquid chromatography coupled to high resolution mass spectrometry is a challenging and time-consuming step in the workflow of non-targeted analysis and often requires manual validation of the results. Retention time, alongside exact mass, isotope pattern, fragmentation spectra, and collision cross-section, is valuable information for ruling out unlikely structures and increasing the confidence in others. Different approaches to predict retention times have been used previously for reversed phase chromatography and hydrophilic interaction liquid chromatography (HILIC), but application is limited to a small set of mobile phases and gradient profiles. Here, we expand the toolbox available for retention time predictions by developing a random forest regression model for predicting retention times for four column types and twenty mobile phase systems. MultiConditionRT was built using a dataset containing 78 compounds analyzed with C18 reversed phase, mixed mode, HILIC, and biphenyl columns. In addition, different eluent compositions were used: both methanol and acetonitrile were combined with different aqueous phases with pH from 2.1 to 10.0 (formic acid, acetic acid, trifluoroacetic acid, formate, acetate, bicarbonate, and ammonia). The root mean square error (RMSE) of the test set predictions was 1.55 min for C18 reversed phase, 1.79 min for mixed-mode, 1.93 min for HILIC, and 1.56 min for biphenyl column. Additionally, MultiConditionRT can be applied to different gradient profiles with a general additive model-based calibration approach. The approach of MultiConditionRT was validated externally and internally with 356 and 151 compounds respectively, yielding an RMSE of 2.68 and 2.32 min. 324 and 84 of these compounds were not in the dataset used in the model development.

MSLibrarian: Optimized Predicted Spectral Libraries for Data-Independent Acquisition Proteomics.

Data-independent acquisition-mass spectrometry (DIA-MS) is the method of choice for deep, consistent, and accurate single-shot profiling in bottom-up proteomics. While classic workflows for targeted quantification from DIA-MS data require auxiliary data-dependent acquisition (DDA) MS analysis of subject samples to derive prior-knowledge spectral libraries, library-free approaches based on in silico prediction promise deep DIA-MS profiling with reduced experimental effort and cost. Coverage and sensitivity in such analyses are however limited, in part, by the large library size and persistent deviations from the experimental data. We present MSLibrarian, a new workflow and tool to obtain optimized predicted spectral libraries by the integrated usage of spectrum-centric DIA data interpretation via the DIA-Umpire approach to inform and calibrate the in silico predicted library and analysis approach. Predicted-vs-observed comparisons enabled optimization of intensity prediction parameters, calibration of retention time prediction for deviating chromatographic setups, and optimization of the library scope and sample representativeness. Benchmarking via a dedicated ground-truth-embedded experiment of species-mixed proteins and quantitative ratio-validation confirmed gains of up to 13% on peptide and 8% on protein level at equivalent FDR control and validation criteria. MSLibrarian is made available as an open-source R software package, including step-by-step user instructions, at https://github.com/MarcIsak/MSLibrarian.

Automated retention time prediction of new psychoactive substances in gas chromatography

New psychoactive substances (NPSs) are drugs that exhibit chemical structures and pharmacological actions similar to illegal drugs. Because the health hazards caused by these drugs of abuse have become a social problem, the quick identification of NPSs is necessary for rapid regulation. We developed an automated machine learning system to predict the retention time of the gas chromatography-mass spectrometry for NPSs. The use of an automated machine learning system and Mordred that calculates molecular descriptors enables accurate retention time prediction. Because the retention time prediction is crucial for identifying NPSs, our research can contribute to the prevention of health hazards.

3D chemical structures allow robust deep learning models for retention time prediction

We present a robust deep learning method CPORT to predict retention time from 3D molecular structures. It generates 4D tensor representations of 3D conformers, that are processed by a neural network with 3D convolutional and fully-connected layers.

Deep learning for retention time prediction in reversed-phase liquid chromatography

Retention time prediction in high-performance liquid chromatography (HPLC) is the subject of many studies since it can improve the identification of unknown molecules in untargeted profiling using HPLC coupled with high-resolution mass spectrometry. Lots of approaches were developed for retention time prediction in liquid chromatography for a different number of molecules considering various molecular properties and machine learning algorithms. The recently built large retention time data set of standard compounds from the Metabolite and Chemical Entity Database (METLIN) allows researchers to create a model that can be used for retention time prediction of small molecules with wide varieties of structures and physicochemical properties. The ability to predict retention times using the largest data set was studied for different architectures of deep learning models that were trained on molecular fingerprints, and SMILES (string representation of a molecule) represented as one-hot matrices. The best result was achieved with a one-dimensional convolutional neural network (1D CNN) that uses SMILES as an input. The proposed model reached the mean absolute error and the median absolute error equal to 34.7 and 18.7 s, respectively, which outperformed the results previously obtained for this data set. The pre-trained 1D CNN on the METLIN SMRT data set was transferred on five other data sets to evaluate the generalization ability.

Development and Application of Multidimensional Lipid Libraries to Investigate Lipidomic Dysregulation Related to Smoke Inhalation Injury Severity.

The implication of lipid dysregulation in diseases, toxic exposure outcomes, and inflammation has brought great interest to lipidomic studies. However, lipids have proven to be analytically challenging due to their highly isomeric nature and vast concentration ranges in biological matrices. Therefore, multidimensional techniques such as those integrating liquid chromatography, ion mobility spectrometry, collision-induced dissociation, and mass spectrometry (LC-IMS-CID-MS) have been implemented to separate lipid isomers as well as provide structural information and increased identification confidence. These data sets are however extremely large and complex, resulting in challenges for data processing and annotation. Here, we have overcome these challenges by developing sample-specific multidimensional lipid libraries using the freely available software Skyline. Specifically, the human plasma library developed for this work contains over 500 unique lipids and is combined with adapted Skyline functions such as indexed retention time (iRT) for retention time prediction and IMS drift time filtering for enhanced selectivity. For comparison with other studies, this database was used to annotate LC-IMS-CID-MS data from a NIST SRM 1950 extract. The same workflow was then utilized to assess plasma and bronchoalveolar lavage fluid (BALF) samples from patients with varying degrees of smoke inhalation injury to identify lipid-based patient prognostic and diagnostic markers.

Deep Neural Network Pretrained by Weighted Autoencoders and Transfer Learning for Retention Time Prediction of Small Molecules.

Retention time (RT) prediction contributes to identification of small molecules measured by high-performance liquid chromatography coupled with high-resolution mass spectrometry. Deep learning algorithms based on big data can enhance the accuracy of RT prediction. But at different chromatographic conditions, RTs of compounds are different, and the number of compounds with known RTs is small in most cases. Therefore, the transfer of big data is necessary. In this work, a strategy using a deep neural network (DNN) pretrained by weighed autoencoders and transfer learning (DNNpwa-TL) was proposed to efficiently predict RTs of compounds. The loss function in the autoencoders was calculated with features weighted by mutual information. Then, a DNN pretrained by weighted autoencoders (DNNpwa) was produced. For other specific chromatographic methods, the transfer learning model DNNpwa-TLs were built through fine-tuning the DNNpwa with the help of some compounds with known RTs to conduct the RT prediction. With the above strategy, a DNNpwa was first built with the METLIN small molecule retention time data set containing 80 038 small molecule compounds. A median relative error of 3.1% and a mean relative error of 4.9% were achieved. Then, 17 data sets from different chromatographic methods were studied, and the results showed that the performance of DNNpwa-TL was better than those of other deep learning models. Besides, DNNpwa-TL outperformed random forest, gradient boost, least absolute shrinkage and selection operator regression, and DNN for most of the 17 data sets. Therefore, DNNpwa-TL can provide an efficient method to perform RT prediction of small molecule compounds for different chromatographic methods and conditions.

Graph Convolutional Networks for Improved Prediction and Interpretability of Chromatographic Retention Data.

Machine learning is a popular technique to predict the retention times of molecules based on descriptors. Descriptors and associated labels (e.g., retention times) of a set of molecules can be used to train a machine learning algorithm. However, descriptors are fixed molecular features which are not necessarily optimized for the given machine learning problem (e.g., to predict retention times). Recent advances in molecular machine learning make use of so-called graph convolutional networks (GCNs) to learn molecular representations from atoms and their bonds to adjacent atoms to optimize the molecular representation for the given problem. In this study, two GCNs were implemented to predict the retention times of molecules for three different chromatographic data sets and compared to seven benchmarks (including two state-of-the art machine learning models). Additionally, saliency maps were computed from trained GCNs to better interpret the importance of certain molecular sub-structures in the data sets. Based on the overall observations of this study, the GCNs performed better than all benchmarks, either significantly outperforming them (5-25% lower mean absolute error) or performing similar to them (<5% difference). Saliency maps revealed a significant difference in molecular sub-structures that are important for predictions of different chromatographic data sets (reversed-phase liquid chromatography vs hydrophilic interaction liquid chromatography).

DeepLC can predict retention times for peptides that carry as-yet unseen modifications.

The inclusion of peptide retention time prediction promises to remove peptide identification ambiguity in complex liquid chromatography-mass spectrometry identification workflows. However, due to the way peptides are encoded in current prediction models, accurate retention times cannot be predicted for modified peptides. This is especially problematic for fledgling open searches, which will benefit from accurate retention time prediction for modified peptides to reduce identification ambiguity. We present DeepLC, a deep learning peptide retention time predictor using peptide encoding based on atomic composition that allows the retention time of (previously unseen) modified peptides to be predicted accurately. We show that DeepLC performs similarly to current state-of-the-art approaches for unmodified peptides and, more importantly, accurately predicts retention times for modifications not seen during training. Moreover, we show that DeepLC's ability to predict retention times for any modification enables potentially incorrect identifications to be flagged in an open search of a wide variety of proteome data.

Bioassay-Guided Interpretation of Antimicrobial Compounds in Kumu, a TCM Preparation From Picrasma quassioides' Stem via UHPLC-Orbitrap-Ion Trap Mass Spectrometry Combined With Fragmentation and Retention Time Calculation.

The stem of Picrasma quassioides (PQ) was recorded as a prominent traditional Chinese medicine, Kumu, which was effective for microbial infection, inflammation, fever, and dysentery, etc. At present, Kumu is widely used in China to develop different medicines, even as injection (Kumu zhusheye), for combating infections. However, the chemical basis of its antimicrobial activity has still not been elucidated. To examine the active chemicals, its stem was extracted to perform bioassay-guided purification against Staphylococcus aureus and Escherichia coli. In this study, two types of columns (normal and reverse-phase) were used for speedy bioassay-guided isolation from Kumu, and the active peaks were collected and identified via an UHPLC-Orbitrap-Ion Trap Mass Spectrometer, combined with MS Fragmenter and ChromGenius. For identification, the COCONUT Database (largest database of natural products) and a manually built PQ database were used, in combination with prediction and calculation of mass fragmentation and retention time to better infer their structures, especially for isomers. Moreover, three standards were analyzed under different conditions for developing and validating the MS method. A total of 25 active compounds were identified, including 24 alkaloids and 1 triterpenoid against S. aureus, whereas only β-carboline-1-carboxylic acid and picrasidine S were active against E. coli. Here, the good antimicrobial activity of 18 chemicals was reported for the first time. Furthermore, the spectrum of three abundant β-carbolines was assessed via their IC50 and MBC against various human pathogens. All of them exhibited strong antimicrobial activities with good potential to be developed as antibiotics. This study clearly showed the antimicrobial chemical basis of Kumu, and the results demonstrated that HRMS coupled with MS Fragmenter and ChromGenius was a powerful tool for compound analysis, which can be used for other complex samples. Beta-carbolines reported here are important lead compounds in antibiotic discovery.

Ion Mobility Coupled to a Time-of-Flight Mass Analyzer Combined With Fragment Intensity Predictions Improves Identification of Classical Bioactive Peptides and Small Open Reading Frame-Encoded Peptides.

Bioactive peptides exhibit key roles in a wide variety of complex processes, such as regulation of body weight, learning, aging, and innate immune response. Next to the classical bioactive peptides, emerging from larger precursor proteins by specific proteolytic processing, a new class of peptides originating from small open reading frames (sORFs) have been recognized as important biological regulators. But their intrinsic properties, specific expression pattern and location on presumed non-coding regions have hindered the full characterization of the repertoire of bioactive peptides, despite their predominant role in various pathways. Although the development of peptidomics has offered the opportunity to study these peptides in vivo, it remains challenging to identify the full peptidome as the lack of cleavage enzyme specification and large search space complicates conventional database search approaches. In this study, we introduce a proteogenomics methodology using a new type of mass spectrometry instrument and the implementation of machine learning tools toward improved identification of potential bioactive peptides in the mouse brain. The application of trapped ion mobility spectrometry (tims) coupled to a time-of-flight mass analyzer (TOF) offers improved sensitivity, an enhanced peptide coverage, reduction in chemical noise and the reduced occurrence of chimeric spectra. Subsequent machine learning tools MS2PIP, predicting fragment ion intensities and DeepLC, predicting retention times, improve the database searching based on a large and comprehensive custom database containing both sORFs and alternative ORFs. Finally, the identification of peptides is further enhanced by applying the post-processing semi-supervised learning tool Percolator. Applying this workflow, the first peptidomics workflow combined with spectral intensity and retention time predictions, we identified a total of 167 predicted sORF-encoded peptides, of which 48 originating from presumed non-coding locations, next to 401 peptides from known neuropeptide precursors, linked to 66 annotated bioactive neuropeptides from within 22 different families. Additional PEAKS analysis expanded the pool of SEPs on presumed non-coding locations to 84, while an additional 204 peptides completed the list of peptides from neuropeptide precursors. Altogether, this study provides insights into a new robust pipeline that fuses technological advancements from different fields ensuring an improved coverage of the neuropeptidome in the mouse brain.

Retention time prediction in hydrophilic interaction liquid chromatography with graph neural network and transfer learning

The combination of retention time (RT), accurate mass and tandem mass spectra can improve the structural annotation in untargeted metabolomics. However, the incorporation of RT for metabolite identification has received less attention because of the limitation of available RT data, especially for hydrophilic interaction liquid chromatography (HILIC). Here, the Graph Neural Network-based Transfer Learning (GNN-TL) is proposed to train a model for HILIC RTs prediction. The graph neural network was pre-trained using an in silico HILIC RT dataset (pseudo-labeling dataset) with ∼306 K molecules. Then, the weights of dense layers in the pre-trained GNN (pre-GNN) model were fine-tuned by transfer learning using a small number of experimental HILIC RTs from the target chromatographic system. The GNN-TL outperformed the methods in Retip, including the Random Forest (RF), Bayesian-regularized neural network (BRNN), XGBoost, light gradient-boosting machine (LightGBM), and Keras. It achieved the lowest mean absolute error (MAE) of 38.6 s on the test set and 33.4 s on an additional test set. It has the best ability to generalize with a small performance difference between training, test, and additional test sets. Furthermore, the predicted RTs can filter out nearly 60% false positive candidates on average, which is valuable for the identification of compounds complementary to mass spectrometry.

Development of a single retention time prediction model integrating multiple liquid chromatography systems: Application to new psychoactive substances

Database-driven suspect screening has proven to be a useful tool to detect new psychoactive substances (NPS) outside the scope of targeted screening; however, the lack of retention times specific to a liquid chromatography (LC) system can result in a large number of false positives. A singular stream-lined, quantitative structure-retention relationship (QSRR)-based retention time prediction model integrating multiple LC systems with different elution conditions is presented using retention time data (n = 1281) from the online crowd-sourced database, HighResNPS. Modelling was performed using an artificial neural network (ANN), specifically a multi-layer perceptron (MLP), using four molecular descriptors and one-hot encoding of categorical labels. Evaluation of test set predictions (n = 193) yielded coefficient of determination (R2) and mean absolute error (MAE) values of 0.942 and 0.583 min, respectively. The model successfully differentiated between LC systems, predicting 54%, 81% and 97% of the test set within ±0.5, ±1 and ±2 min, respectively. Additionally, retention times for an analyte not previously observed by the model were predicted within ±1 min for each LC system. The developed model can be used to predict retention times for all analytes on HighResNPS for each participating laboratory's LC system to further support suspect screening.

Deep Learning Based Prediction of Gas Chromatographic Retention Indices for a Wide Variety of Polar and Mid-Polar Liquid Stationary Phases.

Prediction of gas chromatographic retention indices based on compound structure is an important task for analytical chemistry. The predicted retention indices can be used as a reference in a mass spectrometry library search despite the fact that their accuracy is worse in comparison with the experimental reference ones. In the last few years, deep learning was applied for this task. The use of deep learning drastically improved the accuracy of retention index prediction for non-polar stationary phases. In this work, we demonstrate for the first time the use of deep learning for retention index prediction on polar (e.g., polyethylene glycol, DB-WAX) and mid-polar (e.g., DB-624, DB-210, DB-1701, OV-17) stationary phases. The achieved accuracy lies in the range of 16–50 in terms of the mean absolute error for several stationary phases and test data sets. We also demonstrate that our approach can be directly applied to the prediction of the second dimension retention times (GC × GC) if a large enough data set is available. The achieved accuracy is considerably better compared with the previous results obtained using linear quantitative structure-retention relationships and ACD ChromGenius software. The source code and pre-trained models are available online.

Accurate Retention Time Prediction Based on Monolinked Peptide Information to Confidently Identify Cross-Linked Peptides.

Cross-linking mass spectrometry methods have not been successfully applied to protein-protein interaction discovery at a proteome-wide level mainly due to the computation complexity (O (n2)) issue. In a previous report, we proposed a decision tree searching strategy (DTSS), which can reduce complexity by orders of magnitude. In this study, we further found that the monolinked peptides carry out the information on the retention time of the corresponding cross-linked pairs; therefore, the retention time of cross-linked peptide pairs can be predicted accurately. By utilizing the retention time as an extra filter, the false positive rate can be reduced by around 86% with a sensitivity loss of 10%. The method combined with DTSS (T-DTSS) not only benefits improving identification confidence but also leads to lower cutoff scores and facilitates substantially increasing inter-cross-link identification. T-DTSS was successfully applied to the identification of inter-cross-links obtained from Escherichia coli cell lysate cross-linked by a newly synthesized enrichable cross-linker, pDSBE. The approach can be applicable to both cleavable and noncleavable methods.