Diagnosing Viral Infections Through T-Cell Receptor Sequencing of Activated CD8+ T Cells.

T-cell-based diagnostic tools identify pathogen exposure but lack differentiation between recent and historical exposures in acute infectious diseases. Here, T-cell receptor (TCR) RNA sequencing was performed on HLA-DR+/CD38+CD8+ T-cell subsets of hospitalized coronavirus disease 2019 (COVID-19) patients (n = 30) and healthy controls (n = 30; 10 of whom had previously been exposed to severe acute respiratory syndrome coronavirus 2 [SARS-CoV-2]). CDR3α and CDR3ß TCR regions were clustered separately before epitope specificity annotation using a database of SARS-CoV-2-associated CDR3α and CDR3ß sequences corresponding to >1000 SARS-CoV-2 epitopes. The depth of the SARS-CoV-2-associated CDR3α/ß sequences differentiated COVID-19 patients from the healthy controls with a receiver operating characteristic area under the curve of 0.84 ± 0.10. Hence, annotating TCR sequences of activated CD8+ T cells can be used to diagnose an acute viral infection and discriminate it from historical exposure. In essence, this work presents a new paradigm for applying the T-cell repertoire to accomplish TCR-based diagnostics.

Unravelling the immune signature of herpes zoster: Insights into pathophysiology and the HLA risk profile.

The varicella-zoster virus (VZV) infects over 95% of the population. VZV reactivation causes herpes zoster (HZ), known as shingles, primarily affecting the elderly and immunocompromised individuals. However, HZ can also occur in otherwise healthy individuals. We analyzed the immune signature and risk profile in HZ patients using a genome-wide association study across different UK Biobank HZ cohorts. Additionally, we conducted one of the largest HZ HLA association studies to date, coupled with transcriptomic analysis of pathways underlying HZ susceptibility. Our findings highlight the significance of the MHC locus for HZ development, identifying five protective and four risk HLA alleles. This demonstrates that HZ susceptibility is largely governed by variations in the MHC. Furthermore, functional analyses revealed the upregulation of type I interferon and adaptive immune responses. These findings provide fresh molecular insights into the pathophysiology and the activation of innate and adaptive immune responses triggered by symptomatic VZV reactivation.

Machine learning-based peptide-spectrum match rescoring opens up the immunopeptidome.

Immunopeptidomics is a key technology in the discovery of targets for immunotherapy and vaccine development. However, identifying immunopeptides remains challenging due to their non-tryptic nature, which results in distinct spectral characteristics. Moreover, the absence of strict digestion rules leads to extensive search spaces, further amplified by the incorporation of somatic mutations, pathogen genomes, unannotated open reading frames, and post-translational modifications. This inflation in search space leads to an increase in random high-scoring matches, resulting in fewer identifications at a given false discovery rate. Peptide-spectrum match rescoring has emerged as a machine learning-based solution to address challenges in mass spectrometry-based immunopeptidomics data analysis. It involves post-processing unfiltered spectrum annotations to better distinguish between correct and incorrect peptide-spectrum matches. Recently, features based on predicted peptidoform properties, including fragment ion intensities, retention time, and collisional cross section, have been used to improve the accuracy and sensitivity of immunopeptide identification. In this review, we describe the diverse bioinformatics pipelines that are currently available for peptide-spectrum match rescoring and discuss how they can be used for the analysis of immunopeptidomics data. Finally, we provide insights into current and future machine learning solutions to boost immunopeptide identification.

Four layer multi-omics reveals molecular responses to aneuploidy in Leishmania.

Aneuploidy causes system-wide disruptions in the stochiometric balances of transcripts, proteins, and metabolites, often resulting in detrimental effects for the organism. The protozoan parasite Leishmania has an unusually high tolerance for aneuploidy, but the molecular and functional consequences for the pathogen remain poorly understood. Here, we addressed this question in vitro and present the first integrated analysis of the genome, transcriptome, proteome, and metabolome of highly aneuploid Leishmania donovani strains. Our analyses unambiguously establish that aneuploidy in Leishmania proportionally impacts the average transcript- and protein abundance levels of affected chromosomes, ultimately correlating with the degree of metabolic differences between closely related aneuploid strains. This proportionality was present in both proliferative and non-proliferative in vitro promastigotes. However, as in other Eukaryotes, we observed attenuation of dosage effects for protein complex subunits and in addition, non-cytoplasmic proteins. Differentially expressed transcripts and proteins between aneuploid Leishmania strains also originated from non-aneuploid chromosomes. At protein level, these were enriched for proteins involved in protein metabolism, such as chaperones and chaperonins, peptidases, and heat-shock proteins. In conclusion, our results further support the view that aneuploidy in Leishmania can be adaptive. Additionally, we believe that the high karyotype diversity in vitro and absence of classical transcriptional regulation make Leishmania an attractive model to study processes of protein homeostasis in the context of aneuploidy and beyond.

Clinical Decision Support for Improved Neonatal Care: The Development of a Machine Learning Model for the Prediction of Late-onset Sepsis and Necrotizing Enterocolitis.

OBJECTIVE: To develop an artificial intelligence-based software system for predicting late-onset sepsis (LOS) and necrotizing enterocolitis (NEC) in infants admitted to the neonatal intensive care unit (NICU). STUDY DESIGN: Single-center, retrospective cohort study, conducted in the NICU of the Antwerp University Hospital. Continuous monitoring data of 865 preterm infants born at <32 weeks gestational age, admitted to the NICU in the first week of life, were used to train an XGBoost machine learning (ML) algorithm for LOS and NEC prediction in a cross-validated setup. Afterward, the model's performance was assessed on an independent test set of 148 patients (internal validation). RESULTS: The ML model delivered hourly risk predictions with an overall sensitivity of 69% (142/206) for all LOS/NEC episodes and 81% (67/83) for severe LOS/NEC episodes. The model showed a median time gain of ≤10 hours (IQR, 3.1-21.0 hours), compared with historical clinical diagnosis. On the complete retrospective dataset, the ML model made 721 069 predictions, of which 9805 (1.3%) depicted a LOS/NEC probability of ≥0.15, resulting in a total alarm rate of <1 patient alarm-day per week. The model reached a similar performance on the internal validation set. CONCLUSIONS: Artificial intelligence technology can assist clinicians in the early detection of LOS and NEC in the NICU, which potentially can result in clinical and socioeconomic benefits. Additional studies are required to quantify further the effect of combining artificial and human intelligence on patient outcomes in the NICU.

Benchmarking of Small Molecule Feature Representations for hERG, Nav1.5, and Cav1.2 Cardiotoxicity Prediction.

In the field of drug discovery, there is a substantial challenge in seeking out chemical structures that possess desirable pharmacological, toxicological, and pharmacokinetic properties. Complications arise when drugs interfere with the functioning of cardiac ion channels, leading to serious cardiovascular consequences. The discontinuation and removal of numerous approved drugs from the market or at late development stages in the pipeline due to such inhibitory effects further highlight the urgency of addressing this issue. Consequently, the early prediction of potential blockers targeting cardiac ion channels during the drug discovery process is of paramount importance. This study introduces a deep learning framework that computationally determines the cardiotoxicity associated with the voltage-gated potassium channel (hERG), the voltage-gated calcium channel (Cav1.2), and the voltage-gated sodium channel (Nav1.5) for drug candidates. The predictive capabilities of three feature representations─molecular fingerprints, descriptors, and graph-based numerical representations─are rigorously benchmarked. Additionally, a novel training and evaluation data set framework is presented, enabling predictive model training of drug off-target cardiotoxicity using a comprehensive and large curated data set covering these three cardiac ion channels. To facilitate these predictions, a robust and comprehensive small molecule cardiotoxicity prediction tool named CToxPred has been developed. It is made available as open source under the permissive MIT license at https://github.com/issararab/CToxPred.

Open Modification Searching of SARS-CoV-2-Human Protein Interaction Data Reveals Novel Viral Modification Sites.

The outbreak of the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), the causative agent of the coronavirus 2019 disease, has led to an ongoing global pandemic since 2019. Mass spectrometry can be used to understand the molecular mechanisms of viral infection by SARS-CoV-2, for example, by determining virus-host protein-protein interactions through which SARS-CoV-2 hijacks its human hosts during infection, and to study the role of post-translational modifications. We have reanalyzed public affinity purification-mass spectrometry data using open modification searching to investigate the presence of post-translational modifications in the context of the SARS-CoV-2 virus-host protein-protein interaction network. Based on an over twofold increase in identified spectra, our detected protein interactions show a high overlap with independent mass spectrometry-based SARS-CoV-2 studies and virus-host interactions for alternative viruses, as well as previously unknown protein interactions. In addition, we identified several novel modification sites on SARS-CoV-2 proteins that we investigated in relation to their interactions with host proteins. A detailed analysis of relevant modifications, including phosphorylation, ubiquitination, and S-nitrosylation, provides important hypotheses about the functional role of these modifications during viral infection by SARS-CoV-2.

Acquired non-thermal plasma resistance mediates a shift towards aerobic glycolysis and ferroptotic cell death in melanoma.

AIMS: To gain insights into the underlying mechanisms of NTP therapy sensitivity and resistance, using the first-ever NTP-resistant cell line derived from sensitive melanoma cells (A375). METHODS: Melanoma cells were exposed to NTP and re-cultured for 12 consecutive weeks before evaluation against the parental control cells. Whole transcriptome sequencing analysis was performed to identify differentially expressed genes and enriched molecular pathways. Glucose uptake, extracellular lactate, media acidification, and mitochondrial respiration was analyzed to determine metabolic changes. Cell death inhibitors were used to assess the NTP-induced cell death mechanisms, and apoptosis and ferroptosis was further validated via Annexin V, Caspase 3/7, and lipid peroxidation analysis. RESULTS: Cells continuously exposed to NTP became 10 times more resistant to NTP compared to the parental cell line of the same passage, based on their half-maximal inhibitory concentration (IC50). Sequencing and metabolic analysis indicated that NTP-resistant cells had a preference towards aerobic glycolysis, while cell death analysis revealed that NTP-resistant cells exhibited less apoptosis but were more vulnerable to lipid peroxidation and ferroptosis. CONCLUSIONS: A preference towards aerobic glycolysis and ferroptotic cell death are key physiological changes in NTP-resistance cells, which opens new avenues for further, in-depth research into other cancer types.

Semisupervised Machine Learning for Sensitive Open Modification Spectral Library Searching.

A key analysis task in mass spectrometry proteomics is matching the acquired tandem mass spectra to their originating peptides by sequence database searching or spectral library searching. Machine learning is an increasingly popular postprocessing approach to maximize the number of confident spectrum identifications that can be obtained at a given false discovery rate threshold. Here, we have integrated semisupervised machine learning in the ANN-SoLo tool, an efficient spectral library search engine that is optimized for open modification searching to identify peptides with any type of post-translational modification. We show that machine learning rescoring boosts the number of spectra that can be identified for both standard searching and open searching, and we provide insights into relevant spectrum characteristics harnessed by the machine learning model. The semisupervised machine learning functionality has now been fully integrated into ANN-SoLo, which is available as open source under the permissive Apache 2.0 license on GitHub at https://github.com/bittremieux/ANN-SoLo.

A new Plasmodium vivax reference genome for South American isolates.

BACKGROUND: Plasmodium vivax is the second most important cause of human malaria worldwide, and accounts for the majority of malaria cases in South America. A high-quality reference genome exists for Papua Indonesia (PvP01) and Thailand (PvW1), but is lacking for South America. A reference genome specifically for South America would be beneficial though, as P. vivax is a genetically diverse parasite with geographical clustering. RESULTS: This study presents a new high-quality assembly of a South American P. vivax isolate, referred to as PvPAM (P. vivax Peruvian AMazon). The genome was obtained from a low input patient sample from the Peruvian Amazon and sequenced using PacBio technology, resulting in a highly complete assembly with 6497 functional genes. Telomeric ends were present in 17 out of 28 chromosomal ends, and additional (sub)telomeric regions are present in 12 unassigned contigs. A comparison of multigene families between PvPAM and the PvP01 genome revealed remarkable variation in vir genes, and the presence of merozoite surface proteins (MSP) 3.6 and 3.7. Three dhfr and dhps drug resistance associated mutations are present in PvPAM, similar to those found in other Peruvian isolates. Mapping of publicly available South American whole genome sequencing (WGS) data to PvPAM resulted in significantly fewer variants and truncated reads compared to the use of PvP01 or PvW1 as reference genomes. To minimize the number of core genome variants in non-South American samples, PvW1 is most suited for Southeast Asian isolates, both PvPAM and PvW1 are suited for South Asian isolates, and PvPAM is recommended for African isolates. Interestingly, non-South American samples still contained the least subtelomeric variants when mapped to PvPAM, indicating high quality of the PvPAM subtelomeric regions. CONCLUSIONS: Our findings show that the PvPAM reference genome more accurately represents South American P. vivax isolates in comparison to PvP01 and PvW1. In addition, PvPAM has a high level of completeness, and contains a similar number of annotated genes as PvP01 or PvW1. The PvPAM genome therefore will be a valuable resource to improve future genomic analyses on P. vivax isolates from the South American continent.

MeRgeION: a Multifunctional R Pipeline for Small Molecule LC-MS/MS Data Processing, Searching, and Organizing.

Small molecule structure elucidation using tandem mass spectrometry (MS/MS) plays a crucial role in life science, bioanalytical, and pharmaceutical research. There is a pressing need for increased throughput of compound identification and transformation of historical data into information-rich spectral databases. Meanwhile, molecular networking, a recent bioinformatic framework, provides global displays and system-level understanding of complex LC-MS/MS data sets. Herein we present meRgeION, a multifunctional, modular, and flexible R-based toolbox to streamline spectral database building, automated structural elucidation, and molecular networking. The toolbox offers diverse tuning parameters and the possibility to combine various algorithms in the same pipeline. As an open-source R package, meRgeION is ideally suited for building spectral databases and molecular networks from privacy-sensitive and preliminary data. Using meRgeION, we have created an integrated spectral database covering diverse pharmaceutical compounds that was successfully applied to annotate drug-related metabolites from a published nontargeted metabolomics data set as well as reveal the chemical space behind this complex data set through molecular networking. Moreover, the meRgeION-based processing workflow has demonstrated the usefulness of a spectral library search and molecular networking for pharmaceutical forced degradation studies. meRgeION is freely available at: https://github.com/daniellyz/meRgeION2.

Current challenges for unseen-epitope TCR interaction prediction and a new perspective derived from image classification.

The prediction of epitope recognition by T-cell receptors (TCRs) has seen many advancements in recent years, with several methods now available that can predict recognition for a specific set of epitopes. However, the generic case of evaluating all possible TCR-epitope pairs remains challenging, mainly due to the high diversity of the interacting sequences and the limited amount of currently available training data. In this work, we provide an overview of the current state of this unsolved problem. First, we examine appropriate validation strategies to accurately assess the generalization performance of generic TCR-epitope recognition models when applied to both seen and unseen epitopes. In addition, we present a novel feature representation approach, which we call ImRex (interaction map recognition). This approach is based on the pairwise combination of physicochemical properties of the individual amino acids in the CDR3 and epitope sequences, which provides a convolutional neural network with the combined representation of both sequences. Lastly, we highlight various challenges that are specific to TCR-epitope data and that can adversely affect model performance. These include the issue of selecting negative data, the imbalanced epitope distribution of curated TCR-epitope datasets and the potential exchangeability of TCR alpha and beta chains. Our results indicate that while extrapolation to unseen epitopes remains a difficult challenge, ImRex makes this feasible for a subset of epitopes that are not too dissimilar from the training data. We show that appropriate feature engineering methods and rigorous benchmark standards are required to create and validate TCR-epitope predictive models.

Variants in Bedaquiline-Candidate-Resistance Genes: Prevalence in Bedaquiline-Naive Patients, Effect on MIC, and Association with Mycobacterium tuberculosis Lineage.

Studies have shown that variants in bedaquiline-resistance genes can occur in isolates from bedaquiline-naive patients. We assessed the prevalence of variants in all bedaquiline-candidate-resistance genes in bedaquiline-naive patients, investigated the association between these variants and lineage, and the effect on phenotype. We used whole-genome sequencing to identify variants in bedaquiline-resistance genes in isolates from 509 bedaquiline treatment naive South African tuberculosis patients. A phylogenetic tree was constructed to investigate the association with the isolate lineage background. Bedaquiline MIC was determined using the UKMYC6 microtiter assay. Variants were identified in 502 of 509 isolates (98.6%), with the highest (85%) prevalence of variants in the Rv0676c (mmpL5) gene. We identified 36 unique variants, including 19 variants not reported previously. Only four isolates had a bedaquiline MIC equal to or above the epidemiological cut-off value of 0.25 µg/mL. Phylogenetic analysis showed that 14 of the 15 variants observed more than once occurred monophyletically in one Mycobacterium tuberculosis (sub)lineage. The bedaquiline MIC differed between isolates belonging to lineage 2 and 4 (Fisher's exact test, P = 0.0004). The prevalence of variants in bedaquiline-resistance genes in isolates from bedaquiline-naive patients is high, but very few (<2%) isolates were phenotypically resistant. We found an association between variants in bedaquiline resistance genes and Mycobacterium tuberculosis (sub)lineage, resulting in a lineage-dependent difference in bedaquiline phenotype. Future studies should investigate the impact of the presence of variants on bedaquiline-resistance acquisition and treatment outcome.

Nanopore Sequencing for Mycobacterium tuberculosis: a Critical Review of the Literature, New Developments, and Future Opportunities.

The next-generation, short-read sequencing technologies that generate comprehensive, whole-genome data with single nucleotide resolution have already advanced tuberculosis diagnosis, treatment, surveillance, and source investigation. Their high costs, tedious and lengthy processes, and large equipment remain major hurdles for research use in high tuberculosis burden countries and implementation into routine care. The portable next-generation sequencing devices developed by Oxford Nanopore Technologies (ONT) are attractive alternatives due to their long-read sequence capability, compact low-cost hardware, and continued improvements in accuracy and throughput. A systematic review of the published literature demonstrated limited uptake of ONT sequencing in tuberculosis research and clinical care. Of the 12 eligible articles presenting ONT sequencing data on at least one Mycobacterium tuberculosis sample, four addressed software development for long-read ONT sequencing data with potential applications for M. tuberculosis. Only eight studies presented results of ONT sequencing of M. tuberculosis, of which five performed whole-genome and three did targeted sequencing. Based on these findings, we summarize the standard processes, reflect on the current limitations of ONT sequencing technology, and the research needed to overcome the main hurdles. The low capital cost, portable nature and continued improvement in the performance of ONT sequencing make it an attractive option for sequencing for research and clinical care, but limited data are available on its application in the tuberculosis field. Important research investment is needed to unleash the full potential of ONT sequencing for tuberculosis research and care.

ClusTCR: a python interface for rapid clustering of large sets of CDR3 sequences with unknown antigen specificity.

MOTIVATION: The T-cell receptor (TCR) determines the specificity of a T-cell towards an epitope. As of yet, the rules for antigen recognition remain largely undetermined. Current methods for grouping TCRs according to their epitope specificity remain limited in performance and scalability. Multiple methodologies have been developed, but all of them fail to efficiently cluster large datasets exceeding 1 million sequences. To account for this limitation, we developed ClusTCR, a rapid TCR clustering alternative that efficiently scales up to millions of CDR3 amino acid sequences, without knowledge about their antigen specificity. RESULTS: Benchmarking comparisons revealed similar accuracy of ClusTCR as compared to other TCR clustering methods, as measured by cluster retention, purity and consistency. ClusTCR offers a drastic improvement in clustering speed, which allows the clustering of millions of TCR sequences in just a few minutes through ultraefficient similarity searching and sequence hashing. AVAILABILITY AND IMPLEMENTATION: ClusTCR was written in Python 3. It is available as an anaconda package (https://anaconda.org/svalkiers/clustcr) and on github (https://github.com/svalkiers/clusTCR). SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

A treatment recommender clinical decision support system for personalized medicine: method development and proof-of-concept for drug resistant tuberculosis.

BACKGROUND: Personalized medicine tailors care based on the patient's or pathogen's genotypic and phenotypic characteristics. An automated Clinical Decision Support System (CDSS) could help translate the genotypic and phenotypic characteristics into optimal treatment and thus facilitate implementation of individualized treatment by less experienced physicians. METHODS: We developed a hybrid knowledge- and data-driven treatment recommender CDSS. Stakeholders and experts first define the knowledge base by identifying and quantifying drug and regimen features for the prototype model input. In an iterative manner, feedback from experts is harvested to generate model training datasets, machine learning methods are applied to identify complex relations and patterns in the data, and model performance is assessed by estimating the precision at one, mean reciprocal rank and mean average precision. Once the model performance no longer iteratively increases, a validation dataset is used to assess model overfitting. RESULTS: We applied the novel methodology to develop a treatment recommender CDSS for individualized treatment of drug resistant tuberculosis as a proof of concept. Using input from stakeholders and three rounds of expert feedback on a dataset of 355 patients with 129 unique drug resistance profiles, the model had a 95% precision at 1 indicating that the highest ranked treatment regimen was considered appropriate by the experts in 95% of cases. Use of a validation data set however suggested substantial model overfitting, with a reduction in precision at 1 to 78%. CONCLUSION: Our novel and flexible hybrid knowledge- and data-driven treatment recommender CDSS is a first step towards the automation of individualized treatment for personalized medicine. Further research should assess its value in fields other than drug resistant tuberculosis, develop solid statistical approaches to assess model performance, and evaluate their accuracy in real-life clinical settings.

CONSTANd: An Efficient Normalization Method for Relative Quantification in Small- and Large-Scale Omics Experiments in R BioConductor and Python.

For differential expression studies in all omics disciplines, data normalization is a crucial step that is often subject to a balance between speed and effectiveness. To keep up with the data produced by high-throughput instruments, researchers require fast and easy-to-use yet effective methods that fit into automated analysis pipelines. The CONSTANd normalization method meets these criteria, so we have made its source code available for R/BioConductor and Python. We briefly review the method and demonstrate how it can be used in different omics contexts for experiments of any scale. Widespread adoption across omics disciplines would ease data integration in multiomics experiments.

Open Science Resources for the Mass Spectrometry-Based Analysis of SARS-CoV-2.

The SARS-CoV-2 virus is the causative agent of the 2020 pandemic leading to the COVID-19 respiratory disease. With many scientific and humanitarian efforts ongoing to develop diagnostic tests, vaccines, and treatments for COVID-19, and to prevent the spread of SARS-CoV-2, mass spectrometry research, including proteomics, is playing a role in determining the biology of this viral infection. Proteomics studies are starting to lead to an understanding of the roles of viral and host proteins during SARS-CoV-2 infection, their protein-protein interactions, and post-translational modifications. This is beginning to provide insights into potential therapeutic targets or diagnostic strategies that can be used to reduce the long-term burden of the pandemic. However, the extraordinary situation caused by the global pandemic is also highlighting the need to improve mass spectrometry data and workflow sharing. We therefore describe freely available data and computational resources that can facilitate and assist the mass spectrometry-based analysis of SARS-CoV-2. We exemplify this by reanalyzing a virus-host interactome data set to detect protein-protein interactions and identify host proteins that could potentially be used as targets for drug repurposing.

Sorted B cell transcriptomes point towards actively regulated B cell responses during ongoing chronic hepatitis B infections.

The natural course of chronic hepatitis B virus (HBV) infections follows distinct clinical disease phases, characterized by fluctuating levels of serum HBV DNA and ALT. The immune cells and their features that govern these clinical disease transitions remain unknown. In the current study, we performed RNA sequencing on purified B cells from blood (n = 42) and liver (n = 10) of healthy controls and chronic HBV patients. We found distinct gene expression profiles between healthy controls and chronic HBV patients, as evidenced by 190 differentially expressed genes (DEG), but also between the clinical phenotypes of a chronic HBV infection (17-110 DEG between each phase). Numerous immune pathways, including the B cell receptor pathway were upregulated in liver B cells when compared to peripheral B cells. Further investigation of the detected DEG suggested an activation of B cells during HBeAg seroconversion and an active regulation of B cell signalling in the liver.

Large-scale tandem mass spectrum clustering using fast nearest neighbor searching.

RATIONALE: Advanced algorithmic solutions are necessary to process the ever-increasing amounts of mass spectrometry data that are being generated. In this study, we describe the falcon spectrum clustering tool for efficient clustering of millions of MS/MS spectra. METHODS: falcon succeeds in efficiently clustering large amounts of mass spectral data using advanced techniques for fast spectrum similarity searching. First, high-resolution spectra are binned and converted to low-dimensional vectors using feature hashing. Next, the spectrum vectors are used to construct nearest neighbor indexes for fast similarity searching. The nearest neighbor indexes are used to efficiently compute a sparse pairwise distance matrix without having to exhaustively perform all pairwise spectrum comparisons within the relevant precursor mass tolerance. Finally, density-based clustering is performed to group similar spectra into clusters. RESULTS: Several state-of-the-art spectrum clustering tools were evaluated using a large draft human proteome data set consisting of 25 million spectra, indicating that alternative tools produce clustering results with different characteristics. Notably, falcon generates larger highly pure clusters than alternative tools, leading to a larger reduction in data volume without the loss of relevant information for more efficient downstream processing. CONCLUSIONS: falcon is a highly efficient spectrum clustering tool, which is publicly available as an open source under the permissive BSD license at https://github.com/bittremieux/falcon.