Uncertainty estimation of predictions of peptides' chromatographic retention times in shotgun proteomics

A prediction model for blood-brain barrier penetrating peptides based on masked peptide transformers with dynamic routing

Blood-brain barrier penetrating peptides (BBBPs) are short peptide sequences that possess the ability to traverse the selective blood-brain interface, making them valuable drug candidates or carriers for various payloads. However, the in vivo or in vitro validation of BBBPs is resource-intensive and time-consuming, driving the need for accurate in silico prediction methods. Unfortunately, the scarcity of experimentally validated BBBPs hinders the efficacy of current machine-learning approaches in generating reliable predictions. In this paper, we present DeepB3P3, a novel framework for BBBPs prediction. Our contribution encompasses four key aspects. Firstly, we propose a novel deep learning model consisting of a transformer encoder layer, a convolutional network backbone, and a capsule network classification head. This integrated architecture effectively learns representative features from peptide sequences. Secondly, we introduce masked peptides as a powerful data augmentation technique to compensate for small training set sizes in BBBP prediction. Thirdly, we develop a novel threshold-tuning method to handle imbalanced data by approximating the optimal decision threshold using the training set. Lastly, DeepB3P3 provides an accurate estimation of the uncertainty level associated with each prediction. Through extensive experiments, we demonstrate that DeepB3P3 achieves state-of-the-art accuracy of up to 98.31% on a benchmarking dataset, solidifying its potential as a promising computational tool for the prediction and discovery of BBBPs.

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.

Evaluation of Machine Learning Models for Proteoform Retention and Migration Time Prediction in Top-Down Mass Spectrometry

Reversed-phase liquid chromatography (RPLC) and capillary zone electrophoresis (CZE) are two primary proteoform separation methods in mass spectrometry (MS)-based top-down proteomics. Proteoform retention time (RT) prediction in RPLC and migration time (MT) prediction in CZE provide additional information for accurate proteoform identification and quantification. While existing methods are mainly focused on peptide RT and MT prediction in bottom-up MS, there is still a lack of methods for proteoform RT and MT prediction in top-down MS. We systematically evaluated eight machine learning models and a transfer learning method for proteoform RT prediction and five models and the transfer learning method for proteoform MT prediction. Experimental results showed that a gated recurrent unit (GRU)-based model with transfer learning achieved a high accuracy (R = 0.978) for proteoform RT prediction and that the GRU-based model and a fully connected neural network model obtained a high accuracy of R = 0.982 and 0.981 for proteoform MT prediction, respectively.

Challenges and Opportunities for Bayesian Statistics in Proteomics

Proteomics is a data-rich science with complex experimental designs and an intricate measurement process. To obtain insights from the large data sets produced, statistical methods, including machine learning, are routinely applied. For a quantity of interest, many of these approaches only produce a point estimate, such as a mean, leaving little room for more nuanced interpretations. By contrast, Bayesian statistics allows quantification of uncertainty through the use of probability distributions. These probability distributions enable scientists to ask complex questions of their proteomics data. Bayesian statistics also offers a modular framework for data analysis by making dependencies between data and parameters explicit. Hence, specifying complex hierarchies of parameter dependencies is straightforward in the Bayesian framework. This allows us to use a statistical methodology which equals, rather than neglects, the sophistication of experimental design and instrumentation present in proteomics. Here, we review Bayesian methods applied to proteomics, demonstrating their potential power, alongside the challenges posed by adopting this new statistical framework. To illustrate our review, we give a walk-through of the development of a Bayesian model for dynamic organic orthogonal phase-separation (OOPS) data.

Evaluation of an integrative Bayesian peptide detection approach on a combinatorial peptide library

Detection of peptides lies at the core of bottom-up proteomics analyses. We examined a Bayesian approach to peptide detection, integrating match-based models (fragments, retention time, isotopic distribution, and precursor mass) and peptide prior probability models under a unified probabilistic framework. To assess the relevance of these models and their various combinations, we employed a complete- and a tail-complete search of a low-precursor-mass synthetic peptide library based on oncogenic KRAS peptides. The fragment match was by far the most informative match-based model, while the retention time match was the only remaining such model with an appreciable impact--increasing correct detections by around 8 %. A peptide prior probability model built from a reference proteome greatly improved the detection over a uniform prior, essentially transforming de novo sequencing into a reference-guided search. The knowledge of a correct sequence tag in advance to peptide-spectrum matching had only a moderate impact on peptide detection unless the tag was long and of high certainty. The approach also derived more precise error rates on the analyzed combinatorial peptide library than those estimated using PeptideProphet and Percolator, showing its potential applicability for the detection of homologous peptides. Although the approach requires further computational developments for routine data analysis, it illustrates the value of peptide prior probabilities and presents a Bayesian approach for their incorporation into peptide detection.

Deep learning neural network tools for proteomics

Mass-spectrometry-based proteomics enables quantitative analysis of thousands of human proteins. However, experimental and computational challenges restrict progress in the field. This review summarizes the recent flurry of machine-learning strategies using artificial deep neural networks (or "deep learning") that have started to break barriers and accelerate progress in the field of shotgun proteomics. Deep learning now accurately predicts physicochemical properties of peptides from their sequence, including tandem mass spectra and retention time. Furthermore, deep learning methods exist for nearly every aspect of the modern proteomics workflow, enabling improved feature selection, peptide identification, and protein inference.

DeepRescore: Leveraging Deep Learning to Improve Peptide Identification in Immunopeptidomics

The identification of major histocompatibility complex (MHC)-binding peptides in mass spectrometry (MS)-based immunopeptideomics relies largely on database search engines developed for proteomics data analysis. However, because immunopeptidomics experiments do not involve enzymatic digestion at specific residues, an inflated search space leads to a high false positive rate and low sensitivity in peptide identification. In order to improve the sensitivity and reliability of peptide identification, a post-processing tool named DeepRescore is developed. DeepRescore combines peptide features derived from deep learning predictions, namely accurate retention timeand MS/MS spectra predictions, with previously used features to rescore peptide-spectrum matches. Using two public immunopeptidomics datasets, it is shown that rescoring by DeepRescore increases both the sensitivity and reliability of MHC-binding peptide and neoantigen identifications compared to existing methods. It is also shown that the performance improvement is, to a large extent, driven by the deep learning-derived features. DeepRescore is developed using NextFlow and Docker and is available at https://github.com/bzhanglab/DeepRescore.

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.

Machine learning to predict retention time of small molecules in nano-HPLC

Retention time is an important parameter for identification in untargeted LC-MS screening. Precise retention time prediction facilitates the annotation process and is well known for proteomics. However, the lack of available experimental information for a long time has limited the prediction accuracy for small molecules. Recently introduced large databases for small-molecule retention times make possible reliable machine learning-based predictions for the whole diversity of compounds. Applying simple projections may expand these predictions on various LC systems and conditions. In our work, we describe a complex approach to predict retention times for nano-HPLC that includes the consequent deployment of binary and regression gradient boosting models trained on the METLIN small-molecule dataset and simple projection of the results with a small number of easily available compounds onto nano-HPLC separations. The proposed model outperforms previous attempts to use machine learning for predictions with a 46-s mean absolute error. The overall performance after transfer to nano-LC conditions is less than 155 s (10.8%) in terms of the median absolute (relative) error. To illustrate the applicability of the described approach, we successfully managed to eliminate averagely 25 to 42% of false-positives with a filter threshold derived from ROC curves. Thus, the proposed approach should be used in addition to other well-established in silico methods and their integration may broaden the range of correctly identified molecules.

Cancer neoantigen prioritization through sensitive and reliable proteogenomics analysis

Genomics-based neoantigen discovery can be enhanced by proteomic evidence, but there remains a lack of consensus on the performance of different quality control methods for variant peptide identification in proteogenomics. We propose to use the difference between accurately predicted and observed retention times for each peptide as a metric to evaluate different quality control methods. To this end, we develop AutoRT, a deep learning algorithm with high accuracy in retention time prediction. Analysis of three cancer data sets with a total of 287 tumor samples using different quality control strategies results in substantially different numbers of identified variant peptides and putative neoantigens. Our systematic evaluation, using the proposed retention time metric, provides insights and practical guidance on the selection of quality control strategies. We implement the recommended strategy in a computational workflow named NeoFlow to support proteogenomics-based neoantigen prioritization, enabling more sensitive discovery of putative neoantigens.

Identifying mutation-derived neoantigens by proteogenomics requires robust strategies for quality control. Here, the authors propose peptide retention time as an evaluation metric for proteogenomics quality control methods, and develop a deep learning algorithm for accurate retention time prediction.

Improved Peptide Retention Time Prediction in Liquid Chromatography through Deep Learning

The accuracy of peptide retention time (RT) prediction model in liquid chromatography (LC) is still not sufficient for wider implementation in proteomics practice. Herein, we propose deep learning as an ideal tool to considerably improve this prediction. A new peptide RT prediction tool, DeepRT, was designed using a capsule network model, and the public data sets containing peptides separated by reverse-phase liquid chromatography were used to evaluate the DeepRT performance. Compared with other prevailing RT predictors, DeepRT attained overall improvement in the prediction of peptide RTs with an R² of ∼0.994. Moreover, DeepRT was able to accommodate to the peptides that were separated by different types of LC, such as strong cation exchange (SCX) and hydrophilic interaction liquid chromatography (HILIC) and to reach the RT prediction with R² values of ∼0.996 for SCX and ∼0.993 for HILIC, respectively. If a large peptide data set is available for one type of LC, DeepRT can be promoted to DeepRT(+) using transfer learning. Based on a large peptide data set gained from SWATH, DeepRT(+) further elevated the accuracy of RT prediction for peptides in a small data set and enabled a satisfactory prediction upon limited peptides approximating hundreds. Further, DeepRT automatically learns retention-related properties of amino acids under different separation mechanisms, which are well consistent with retention coefficients (Rc) of the amino acids. DeepRT was thus proven to be an improved RT predictor with high flexibility and efficiency. DeepRT is available at https://github.com/horsepurve/DeepRTplus .