AlphaPept: a modern and open framework for MS-based proteomics.

In common with other omics technologies, mass spectrometry (MS)-based proteomics produces ever-increasing amounts of raw data, making efficient analysis a principal challenge. A plethora of different computational tools can process the MS data to derive peptide and protein identification and quantification. However, during the last years there has been dramatic progress in computer science, including collaboration tools that have transformed research and industry. To leverage these advances, we develop AlphaPept, a Python-based open-source framework for efficient processing of large high-resolution MS data sets. Numba for just-in-time compilation on CPU and GPU achieves hundred-fold speed improvements. AlphaPept uses the Python scientific stack of highly optimized packages, reducing the code base to domain-specific tasks while accessing the latest advances. We provide an easy on-ramp for community contributions through the concept of literate programming, implemented in Jupyter Notebooks. Large datasets can rapidly be processed as shown by the analysis of hundreds of proteomes in minutes per file, many-fold faster than acquisition. AlphaPept can be used to build automated processing pipelines with web-serving functionality and compatibility with downstream analysis tools. It provides easy access via one-click installation, a modular Python library for advanced users, and via an open GitHub repository for developers.

The social and structural architecture of the yeast protein interactome.

Cellular functions are mediated by protein-protein interactions, and mapping the interactome provides fundamental insights into biological systems. Affinity purification coupled to mass spectrometry is an ideal tool for such mapping, but it has been difficult to identify low copy number complexes, membrane complexes and complexes that are disrupted by protein tagging. As a result, our current knowledge of the interactome is far from complete, and assessing the reliability of reported interactions is challenging. Here we develop a sensitive high-throughput method using highly reproducible affinity enrichment coupled to mass spectrometry combined with a quantitative two-dimensional analysis strategy to comprehensively map the interactome of Saccharomyces cerevisiae. Thousand-fold reduced volumes in 96-well format enabled replicate analysis of the endogenous GFP-tagged library covering the entire expressed yeast proteome1. The 4,159 pull-downs generated a highly structured network of 3,927 proteins connected by 31,004 interactions, doubling the number of proteins and tripling the number of reliable interactions compared with existing interactome maps2. This includes very-low-abundance epigenetic complexes, organellar membrane complexes and non-taggable complexes inferred by abundance correlation. This nearly saturated interactome reveals that the vast majority of yeast proteins are highly connected, with an average of 16 interactors. Similar to social networks between humans, the average shortest distance between proteins is 4.2 interactions. AlphaFold-Multimer provided novel insights into the functional roles of previously uncharacterized proteins in complexes. Our web portal ( www.yeast-interactome.org ) enables extensive exploration of the interactome dataset.

Spatial single-cell mass spectrometry defines zonation of the hepatocyte proteome.

Single-cell proteomics by mass spectrometry is emerging as a powerful and unbiased method for the characterization of biological heterogeneity. So far, it has been limited to cultured cells, whereas an expansion of the method to complex tissues would greatly enhance biological insights. Here we describe single-cell Deep Visual Proteomics (scDVP), a technology that integrates high-content imaging, laser microdissection and multiplexed mass spectrometry. scDVP resolves the context-dependent, spatial proteome of murine hepatocytes at a current depth of 1,700 proteins from a cell slice. Half of the proteome was differentially regulated in a spatial manner, with protein levels changing dramatically in proximity to the central vein. We applied machine learning to proteome classes and images, which subsequently inferred the spatial proteome from imaging data alone. scDVP is applicable to healthy and diseased tissues and complements other spatial proteomics and spatial omics technologies.

AlphaPeptStats: an open-source Python package for automated and scalable statistical analysis of mass spectrometry-based proteomics.

SUMMARY: The widespread application of mass spectrometry (MS)-based proteomics in biomedical research increasingly requires robust, transparent, and streamlined solutions to extract statistically reliable insights. We have designed and implemented AlphaPeptStats, an inclusive Python package with currently with broad functionalities for normalization, imputation, visualization, and statistical analysis of label-free proteomics data. It modularly builds on the established stack of Python scientific libraries and is accompanied by a rigorous testing framework with 98% test coverage. It imports the output of a range of popular search engines. Data can be filtered and normalized according to user specifications. At its heart, AlphaPeptStats provides a wide range of robust statistical algorithms such as t-tests, analysis of variance, principal component analysis, hierarchical clustering, and multiple covariate analysis-all in an automatable manner. Data visualization capabilities include heat maps, volcano plots, and scatter plots in publication-ready format. AlphaPeptStats advances proteomic research through its robust tools that enable researchers to manually or automatically explore complex datasets to identify interesting patterns and outliers. AVAILABILITY AND IMPLEMENTATION: AlphaPeptStats is implemented in Python and part of the AlphaPept framework. It is released under a permissive Apache license. The source code and one-click installers are freely available and on GitHub at https://github.com/MannLabs/alphapeptstats.

Ångström-resolution fluorescence microscopy.

Fluorescence microscopy, with its molecular specificity, is one of the major characterization methods used in the life sciences to understand complex biological systems. Super-resolution approaches1-6 can achieve resolution in cells in the range of 15 to 20 nm, but interactions between individual biomolecules occur at length scales below 10 nm and characterization of intramolecular structure requires Ångström resolution. State-of-the-art super-resolution implementations7-14 have demonstrated spatial resolutions down to 5 nm and localization precisions of 1 nm under certain in vitro conditions. However, such resolutions do not directly translate to experiments in cells, and Ångström resolution has not been demonstrated to date. Here we introdue a DNA-barcoding method, resolution enhancement by sequential imaging (RESI), that improves the resolution of fluorescence microscopy down to the Ångström scale using off-the-shelf fluorescence microscopy hardware and reagents. By sequentially imaging sparse target subsets at moderate spatial resolutions of >15 nm, we demonstrate that single-protein resolution can be achieved for biomolecules in whole intact cells. Furthermore, we experimentally resolve the DNA backbone distance of single bases in DNA origami with Ångström resolution. We use our method in a proof-of-principle demonstration to map the molecular arrangement of the immunotherapy target CD20 in situ in untreated and drug-treated cells, which opens possibilities for assessing the molecular mechanisms of targeted immunotherapy. These observations demonstrate that, by enabling intramolecular imaging under ambient conditions in whole intact cells, RESI closes the gap between super-resolution microscopy and structural biology studies and thus delivers information key to understanding complex biological systems.

Transparent Exploration of Machine Learning for Biomarker Discovery from Proteomics and Omics Data.

Biomarkers are of central importance for assessing the health state and to guide medical interventions and their efficacy; still, they are lacking for most diseases. Mass spectrometry (MS)-based proteomics is a powerful technology for biomarker discovery but requires sophisticated bioinformatics to identify robust patterns. Machine learning (ML) has become a promising tool for this purpose. However, it is sometimes applied in an opaque manner and generally requires specialized knowledge. To enable easy access to ML for biomarker discovery without any programming or bioinformatics skills, we developed "OmicLearn" (http://OmicLearn.org), an open-source browser-based ML tool using the latest advances in the Python ML ecosystem. Data matrices from omics experiments are easily uploaded to an online or a locally installed web server. OmicLearn enables rapid exploration of the suitability of various ML algorithms for the experimental data sets. It fosters open science via transparent assessment of state-of-the-art algorithms in a standardized format for proteomics and other omics sciences.

The structural context of posttranslational modifications at a proteome-wide scale.

The recent revolution in computational protein structure prediction provides folding models for entire proteomes, which can now be integrated with large-scale experimental data. Mass spectrometry (MS)-based proteomics has identified and quantified tens of thousands of posttranslational modifications (PTMs), most of them of uncertain functional relevance. In this study, we determine the structural context of these PTMs and investigate how this information can be leveraged to pinpoint potential regulatory sites. Our analysis uncovers global patterns of PTM occurrence across folded and intrinsically disordered regions. We found that this information can help to distinguish regulatory PTMs from those marking improperly folded proteins. Interestingly, the human proteome contains thousands of proteins that have large folded domains linked by short, disordered regions that are strongly enriched in regulatory phosphosites. These include well-known kinase activation loops that induce protein conformational changes upon phosphorylation. This regulatory mechanism appears to be widespread in kinases but also occurs in other protein families such as solute carriers. It is not limited to phosphorylation but includes ubiquitination and acetylation sites as well. Furthermore, we performed three-dimensional proximity analysis, which revealed examples of spatial coregulation of different PTM types and potential PTM crosstalk. To enable the community to build upon these first analyses, we provide tools for 3D visualization of proteomics data and PTMs as well as python libraries for data accession and processing.

AlphaMap: an open-source Python package for the visual annotation of proteomics data with sequence-specific knowledge.

SUMMARY: Integrating experimental information across proteomic datasets with the wealth of publicly available sequence annotations is a crucial part in many proteomic studies that currently lacks an automated analysis platform. Here, we present AlphaMap, a Python package that facilitates the visual exploration of peptide-level proteomics data. Identified peptides and post-translational modifications in proteomic datasets are mapped to their corresponding protein sequence and visualized together with prior knowledge from UniProt and with expected proteolytic cleavage sites. The functionality of AlphaMap can be accessed via an intuitive graphical user interface or-more flexibly-as a Python package that allows its integration into common analysis workflows for data visualization. AlphaMap produces publication-quality illustrations and can easily be customized to address a given research question. AVAILABILITY AND IMPLEMENTATION: AlphaMap is implemented in Python and released under an Apache license. The source code and one-click installers are freely available at https://github.com/MannLabs/alphamap. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

AlphaTims: Indexing Trapped Ion Mobility Spectrometry-TOF Data for Fast and Easy Accession and Visualization.

High-resolution MS-based proteomics generates large amounts of data, even in the standard LC-tandem MS configuration. Adding an ion mobility dimension vastly increases the acquired data volume, challenging both analytical processing pipelines and especially data exploration by scientists. This has necessitated data aggregation, effectively discarding much of the information present in these rich datasets. Taking trapped ion mobility spectrometry (TIMS) on a quadrupole TOF (Q-TOF) platform as an example, we developed an efficient indexing scheme that represents all data points as detector arrival times on scales of minutes (LC), milliseconds (TIMS), and microseconds (TOF). In our open-source AlphaTims package, data are indexed, accessed, and visualized by a combination of tools of the scientific Python ecosystem. We interpret unprocessed data as a sparse four-dimensional matrix and use just-in-time compilation to machine code with Numba, accelerating our computational procedures by several orders of magnitude while keeping to familiar indexing and slicing notations. For samples with more than six billion detector events, a modern laptop can load and index raw data in about a minute. Loading is even faster when AlphaTims has already saved indexed data in an HDF5 file, a portable scientific standard used in extremely large-scale data acquisition. Subsequently, data accession along any dimension and interactive visualization happens in milliseconds. We have found AlphaTims to be a key enabling tool to explore high-dimensional LC-TIMS-Q-TOF data and have made it freely available as an open-source Python package with a stand-alone graphical user interface at https://github.com/MannLabs/alphatims or as part of the AlphaPept 'ecosystem'.

Artificial intelligence for proteomics and biomarker discovery.

There is an avalanche of biomedical data generation and a parallel expansion in computational capabilities to analyze and make sense of these data. Starting with genome sequencing and widely employed deep sequencing technologies, these trends have now taken hold in all omics disciplines and increasingly call for multi-omics integration as well as data interpretation by artificial intelligence technologies. Here, we focus on mass spectrometry (MS)-based proteomics and describe how machine learning and, in particular, deep learning now predicts experimental peptide measurements from amino acid sequences alone. This will dramatically improve the quality and reliability of analytical workflows because experimental results should agree with predictions in a multi-dimensional data landscape. Machine learning has also become central to biomarker discovery from proteomics data, which now starts to outperform existing best-in-class assays. Finally, we discuss model transparency and explainability and data privacy that are required to deploy MS-based biomarkers in clinical settings.

High-resolution serum proteome trajectories in COVID-19 reveal patient-specific seroconversion.

A deeper understanding of COVID-19 on human molecular pathophysiology is urgently needed as a foundation for the discovery of new biomarkers and therapeutic targets. Here we applied mass spectrometry (MS)-based proteomics to measure serum proteomes of COVID-19 patients and symptomatic, but PCR-negative controls, in a time-resolved manner. In 262 controls and 458 longitudinal samples of 31 patients, hospitalized for COVID-19, a remarkable 26% of proteins changed significantly. Bioinformatics analyses revealed co-regulated groups and shared biological functions. Proteins of the innate immune system such as CRP, SAA1, CD14, LBP, and LGALS3BP decreased early in the time course. Regulators of coagulation (APOH, FN1, HRG, KNG1, PLG) and lipid homeostasis (APOA1, APOC1, APOC2, APOC3, PON1) increased over the course of the disease. A global correlation map provides a system-wide functional association between proteins, biological processes, and clinical chemistry parameters. Importantly, five SARS-CoV-2 immunoassays against antibodies revealed excellent correlations with an extensive range of immunoglobulin regions, which were quantified by MS-based proteomics. The high-resolution profile of all immunoglobulin regions showed individual-specific differences and commonalities of potential pathophysiological relevance.

Deep learning the collisional cross sections of the peptide universe from a million experimental values.

The size and shape of peptide ions in the gas phase are an under-explored dimension for mass spectrometry-based proteomics. To investigate the nature and utility of the peptide collisional cross section (CCS) space, we measure more than a million data points from whole-proteome digests of five organisms with trapped ion mobility spectrometry (TIMS) and parallel accumulation-serial fragmentation (PASEF). The scale and precision (CV < 1%) of our data is sufficient to train a deep recurrent neural network that accurately predicts CCS values solely based on the peptide sequence. Cross section predictions for the synthetic ProteomeTools peptides validate the model within a 1.4% median relative error (R > 0.99). Hydrophobicity, proportion of prolines and position of histidines are main determinants of the cross sections in addition to sequence-specific interactions. CCS values can now be predicted for any peptide and organism, forming a basis for advanced proteomics workflows that make full use of the additional information.

Quantitative single-protein imaging reveals molecular complex formation of integrin, talin, and kindlin during cell adhesion.

Single-molecule localization microscopy (SMLM) enabling the investigation of individual proteins on molecular scales has revolutionized how biological processes are analysed in cells. However, a major limitation of imaging techniques reaching single-protein resolution is the incomplete and often unknown labeling and detection efficiency of the utilized molecular probes. As a result, fundamental processes such as complex formation of distinct molecular species cannot be reliably quantified. Here, we establish a super-resolution microscopy framework, called quantitative single-molecule colocalization analysis (qSMCL), which permits the identification of absolute molecular quantities and thus the investigation of molecular-scale processes inside cells. The method combines multiplexed single-protein resolution imaging, automated cluster detection, in silico data simulation procedures, and widely applicable experimental controls to determine absolute fractions and spatial coordinates of interacting species on a true molecular level, even in highly crowded subcellular structures. The first application of this framework allowed the identification of a long-sought ternary adhesion complex-consisting of talin, kindlin and active ß1-integrin-that specifically forms in cell-matrix adhesion sites. Together, the experiments demonstrate that qSMCL allows an absolute quantification of multiplexed SMLM data and thus should be useful for investigating molecular mechanisms underlying numerous processes in cells.

Urinary proteome profiling for stratifying patients with familial Parkinson's disease.

The prevalence of Parkinson's disease (PD) is increasing but the development of novel treatment strategies and therapeutics altering the course of the disease would benefit from specific, sensitive, and non-invasive biomarkers to detect PD early. Here, we describe a scalable and sensitive mass spectrometry (MS)-based proteomic workflow for urinary proteome profiling. Our workflow enabled the reproducible quantification of more than 2,000 proteins in more than 200 urine samples using minimal volumes from two independent patient cohorts. The urinary proteome was significantly different between PD patients and healthy controls, as well as between LRRK2 G2019S carriers and non-carriers in both cohorts. Interestingly, our data revealed lysosomal dysregulation in individuals with the LRRK2 G2019S mutation. When combined with machine learning, the urinary proteome data alone were sufficient to classify mutation status and disease manifestation in mutation carriers remarkably well, identifying VGF, ENPEP, and other PD-associated proteins as the most discriminating features. Taken together, our results validate urinary proteomics as a valuable strategy for biomarker discovery and patient stratification in PD.

Complex multicomponent patterns rendered on a 3D DNA-barrel pegboard.

DNA origami, in which a long scaffold strand is assembled with a many short staple strands into parallel arrays of double helices, has proven a powerful method for custom nanofabrication. However, currently the design and optimization of custom 3D DNA-origami shapes is a barrier to rapid application to new areas. Here we introduce a modular barrel architecture, and demonstrate hierarchical assembly of a 100 megadalton DNA-origami barrel of ~90 nm diameter and ~250 nm height, that provides a rhombic-lattice canvas of a thousand pixels each, with pitch of ~8 nm, on its inner and outer surfaces. Complex patterns rendered on these surfaces were resolved using up to twelve rounds of Exchange-PAINT super-resolution microscopy. We envision these structures as versatile nanoscale pegboards for applications requiring complex 3D arrangements of matter, which will serve to promote rapid uptake of this technology in diverse fields beyond specialist groups working in DNA nanotechnology.

Proteome profiling in cerebrospinal fluid reveals novel biomarkers of Alzheimer's disease.

Neurodegenerative diseases are a growing burden, and there is an urgent need for better biomarkers for diagnosis, prognosis, and treatment efficacy. Structural and functional brain alterations are reflected in the protein composition of cerebrospinal fluid (CSF). Alzheimer's disease (AD) patients have higher CSF levels of tau, but we lack knowledge of systems-wide changes of CSF protein levels that accompany AD. Here, we present a highly reproducible mass spectrometry (MS)-based proteomics workflow for the in-depth analysis of CSF from minimal sample amounts. From three independent studies (197 individuals), we characterize differences in proteins by AD status (> 1,000 proteins, CV < 20%). Proteins with previous links to neurodegeneration such as tau, SOD1, and PARK7 differed most strongly by AD status, providing strong positive controls for our approach. CSF proteome changes in Alzheimer's disease prove to be widespread and often correlated with tau concentrations. Our unbiased screen also reveals a consistent glycolytic signature across our cohorts and a recent study. Machine learning suggests clinical utility of this proteomic signature.

The proteome landscape of the kingdoms of life.

Proteins carry out the vast majority of functions in all biological domains, but for technological reasons their large-scale investigation has lagged behind the study of genomes. Since the first essentially complete eukaryotic proteome was reported1, advances in mass-spectrometry-based proteomics2 have enabled increasingly comprehensive identification and quantification of the human proteome3-6. However, there have been few comparisons across species7,8, in stark contrast with genomics initiatives9. Here we use an advanced proteomics workflow-in which the peptide separation step is performed by a microstructured and extremely reproducible chromatographic system-for the in-depth study of 100 taxonomically diverse organisms. With two million peptide and 340,000 stringent protein identifications obtained in a standardized manner, we double the number of proteins with solid experimental evidence known to the scientific community. The data also provide a large-scale case study for sequence-based machine learning, as we demonstrate by experimentally confirming the predicted properties of peptides from Bacteroides uniformis. Our results offer a comparative view of the functional organization of organisms across the entire evolutionary range. A remarkably high fraction of the total proteome mass in all kingdoms is dedicated to protein homeostasis and folding, highlighting the biological challenge of maintaining protein structure in all branches of life. Likewise, a universally high fraction is involved in supplying energy resources, although these pathways range from photosynthesis through iron sulfur metabolism to carbohydrate metabolism. Generally, however, proteins and proteomes are remarkably diverse between organisms, and they can readily be explored and functionally compared at www.proteomesoflife.org.

nanoTRON: a Picasso module for MLP-based classification of super-resolution data.

MOTIVATION: Classification of images is an essential task in higher-level analysis of biological data. By bypassing the diffraction limit of light, super-resolution microscopy opened up a new way to look at molecular details using light microscopy, producing large amounts of data with exquisite spatial detail. Statistical exploration of data usually needs initial classification, which is up to now often performed manually. RESULTS: We introduce nanoTRON, an interactive open-source tool, which allows super-resolution data classification based on image recognition. It extends the software package Picasso with the first deep learning tool with a graphic user interface. AVAILABILITY AND IMPLEMENTATION: nanoTRON is written in Python and freely available under the MIT license as a part of the software collection Picasso on GitHub (http://www.github.com/jungmannlab/picasso). All raw data can be obtained from the authors upon reasonable request. CONTACT: jungmann@biochem.mpg.de. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.