Post-Translational Modifications (PTMs) are changes that occur in proteins after synthesis, influencing their structure, function, and cellular behavior. PTMs are essential in cell biology; they regulate protein function and stability, are involved in various cellular processes, and are linked to numerous diseases. A particularly interesting class of PTMs are chemical modifications such as phosphorylation introduced on amino acid side chains because they can drastically alter the physicochemical properties of the peptides once they are present. One or more PTMs can be attached to each amino acid of the peptide sequence. The most commonly applied technique to detect PTMs on proteins is bottom-up Mass Spectrometry-based proteomics (MS), where proteins are digested into peptides and subsequently analyzed using Tandem Mass Spectrometry (MS/MS). While an increasing number of machine learning models are published focusing on MS/MS-related property prediction of unmodified peptides, high-quality reference data for modified peptides is missing, impeding model development for this important class of peptides. To enable researchers to train machine learning models that can accurately predict the properties of modified peptides, we introduce four high-quality labeled datasets for applying machine and deep learning to tasks in MS-based proteomics. The four datasets comprise several subgroups of peptides with 1.2 million unique modified peptide sequences and 30 unique pairs of (amino-acid, PTM), covering both experimentally introduced and naturally occurring modifications on various amino acids. We evaluate the utility and importance of the dataset by providing benchmarking results on models trained with and without modifications and highlighting the impact of including modified sequences on downstream tasks. We demonstrate that predicting the properties of modified peptides is more challenging but has a broad impact since they are often the core of protein functionality and its regulation, and they have a potential role as biomarkers in clinical applications. Our datasets contribute to applied machine learning in proteomics by enabling the research community to experiment with methods to encode PTMs as model inputs and to benchmark against reference data for model comparison. With a proper data split for three common tasks in proteomics, we provide a robust way to evaluate model performance and assess generalization on unseen modified sequences.

Proteomics is the interdisciplinary field focusing on the large-scale study of proteins. Proteins essentially organize and execute all functions within organisms. Today, the bottom-up analysis approach is the most commonly used workflow, where proteins are digested into peptides and subsequently analyzed using Tandem Mass Spectrometry (MS/MS). MS-based proteomics has transformed various fields in life sciences, such as drug discovery and biomarker identification. Today, proteomics is entering a phase where it is helpful for clinical decision-making. Computational methods are vital in turning large amounts of acquired raw MS data into information and, ultimately, knowledge. Deep learning has proved its success in multiple domains as a robust framework for supervised and unsupervised machine learning problems. In proteomics, scientists are increasingly leveraging the potential of deep learning to predict the properties of peptides based on their sequence to improve their confident identification. However, a reference dataset is missing, covering several proteomics tasks, enabling performance comparison, and evaluating reproducibility and generalization. Here, we present a large labeled proteomics dataset spanning several tasks in the domain to address this challenge. We focus on two common applications: peptide retention time and MS/MS spectrum prediction. We review existing methods and task formulations from a machine learning perspective and recommend suitable evaluation metrics and visualizations. With an accessible dataset, we aim to lower the entry barrier and enable faster development in machine learning for proteomics.

MOTIVATION: Comprehensive 2D gas chromatography-mass spectrometry is an established method for the analysis of complex mixtures in analytical chemistry and metabolomics. It produces large amounts of data that require semiautomatic, but preferably automatic handling. This involves the location of significant signals (peaks) and their matching and alignment across different measurements. To date, there exist only a few openly available algorithms for the retention time alignment of peaks originating from such experiments that scale well with increasing sample and peak numbers, while providing reliable alignment results. RESULTS: We describe BiPACE 2D, an automated algorithm for retention time alignment of peaks from 2D gas chromatography-mass spectrometry experiments and evaluate it on three previously published datasets against the mSPA, SWPA and Guineu algorithms. We also provide a fourth dataset from an experiment studying the H2 production of two different strains of Chlamydomonas reinhardtii that is available from the MetaboLights database together with the experimental protocol, peak-detection results and manually curated multiple peak alignment for future comparability with newly developed algorithms. AVAILABILITY AND IMPLEMENTATION: BiPACE 2D is contained in the freely available Maltcms framework, version 1.3, hosted at http://maltcms.sf.net, under the terms of the L-GPL v3 or Eclipse Open Source licenses. The software used for the evaluation along with the underlying datasets is available at the same location. The C.reinhardtii dataset is freely available at http://www.ebi.ac.uk/metabolights/MTBLS37.

BACKGROUND: Modern analytical methods in biology and chemistry use separation techniques coupled to sensitive detectors, such as gas chromatography-mass spectrometry (GC-MS) and liquid chromatography-mass spectrometry (LC-MS). These hyphenated methods provide high-dimensional data. Comparing such data manually to find corresponding signals is a laborious task, as each experiment usually consists of thousands of individual scans, each containing hundreds or even thousands of distinct signals. In order to allow for successful identification of metabolites or proteins within such data, especially in the context of metabolomics and proteomics, an accurate alignment and matching of corresponding features between two or more experiments is required. Such a matching algorithm should capture fluctuations in the chromatographic system which lead to non-linear distortions on the time axis, as well as systematic changes in recorded intensities. Many different algorithms for the retention time alignment of GC-MS and LC-MS data have been proposed and published, but all of them focus either on aligning previously extracted peak features or on aligning and comparing the complete raw data containing all available features. RESULTS: In this paper we introduce two algorithms for retention time alignment of multiple GC-MS datasets: multiple alignment by bidirectional best hits peak assignment and cluster extension (BIPACE) and center-star multiple alignment by pairwise partitioned dynamic time warping (CeMAPP-DTW). We show how the similarity-based peak group matching method BIPACE may be used for multiple alignment calculation individually and how it can be used as a preprocessing step for the pairwise alignments performed by CeMAPP-DTW. We evaluate the algorithms individually and in combination on a previously published small GC-MS dataset studying the Leishmania parasite and on a larger GC-MS dataset studying grains of wheat (Triticum aestivum). CONCLUSIONS: We have shown that BIPACE achieves very high precision and recall and a very low number of false positive peak assignments on both evaluation datasets. CeMAPP-DTW finds a high number of true positives when executed on its own, but achieves even better results when BIPACE is used to constrain its search space. The source code of both algorithms is included in the OpenSource software framework Maltcms, which is available from http://maltcms.sf.net. The evaluation scripts of the present study are available from the same source.