SEARCH · Search bioRxiv

Results for “Bioinformatics”

Search indexed bioRxiv preprints in genomics, neuroscience, cell biology and bioinformatics. Read source abstracts and check manuscript versions; preprints are not peer reviewed.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 829 records · Page 46Linked to original sources

bioRxiv2022

Hecatomb: An End-to-End Research Platform for Viral Metagenomics

BackgroundAnalysis of viral diversity using modern sequencing technologies offers extraordinary opportunities for discovery. However, these analyses present a number of bioinformatic challenges due to viral genetic diversity and virome complexity. Due to the lack of conserved marker sequences, metagenomic detection of viral sequences requires a non-targeted, random (shotgun) approach. Annotation and enumeration of viral sequences relies on rigorous quality control and effective search strategies against appropriate reference databases. Virome analysis also benefits from the analysis of both individual metagenomic sequences as well as assembled contigs. Combined, virome analysis results in large amounts of data requiring sophisticated visualization and statistical tools. ResultsHere we introduce Hecatomb, a bioinformatics platform enabling both read and contig based analysis. Hecatomb integrates query information from both amino acid and nucleotide reference sequence databases. Hecatomb integrates data collected throughout the workflow enabling analyst driven virome analysis and discovery. Hecatomb is available on GitHub at https://github.com/shandley/hecatomb. ConclusionsHecatomb provides a single, modular software solution to the complex tasks required of many virome analysis. We demonstrate the value of the approach by applying Hecatomb to both a host-associated (enteric) and an environmental (marine) virome data set. Hecatomb provided data to determine true- or false-positive viral sequences in both data sets and revealed complex virome structure at distinct marine reef sites.

bioinformatics↗

bioRxiv2022

On Weighted K-Mer Dictionaries

We consider the problem of representing a set of k-mers and their abundance counts, or weights, in compressed space so that assessing membership and retrieving the weight of a k-mer is efficient. The representation is called a weighted dictionary of k-mers and finds application in numerous tasks in Bioinformatics that usually count k-mers as a pre-processing step. In fact, k-mer counting tools produce very large outputs that may result in a severe bottleneck for subsequent processing. In this work we extend the recently introduced SSHash dictionary (Pibiri, Bioinformatics 2022) to also store compactly the weights of the k-mers. From a technical perspective, we exploit the order of the k-mers represented in SSHash to encode runs of weights, hence allowing (several times) better compression than the empirical entropy of the weights. We also study the problem of reducing the number of runs in the weights to improve compression even further and illustrate a lower bound for this problem. We propose an efficient, greedy, algorithm to reduce the number of runs and show empirically that it performs well, i.e., very similarly to the lower bound. Lastly, we corroborate our findings with experiments on real-world datasets and comparison with competitive alternatives. Up to date, SSHash is the only k-mer dictionary that is exact, weighted, associative, fast, and small.

bioinformatics↗

bioRxiv2022

In silico EsxG·EsxH Rational Epitope Selection: Candidate Epitopes for Vaccine Design against Pulmonary Tuberculosis

Rational design of new vaccines against pulmonary tuberculosis is imperative. EsxG{middle dot}EsxH complex is involved in metal uptake, drug resistance, and immune response evasion. These characteristics make it an ideal target for rational vaccine development. The aim of this study is to show the rational design of epitope-based peptide vaccines by using bioinformatics and structural vaccinology tools. A total of 4.15 {micro}s of Molecular Dynamics simulations were carried out to describe the behavior in solution of heterodimer, single epitopes, and epitopes loaded into MHC-II complexes. In order to predict T and B cell epitopes for antigenic activation, bioinformatic tools were used. Hence, we propose three epitopes with the potential to design pulmonary tuberculosis vaccines. The possible use of the proposed epitopes includes subunit vaccines, as a booster in BCG vaccination to improve its immune response, as well as the generation of antibodies that interfere with the mycobacterium tuberculosis homeostasis, affecting its survival.

bioinformatics↗

bioRxiv2022

CBP60-DB: An AlphaFold-predicted plant kingdom-wide database of the CALMODULIN-BINDING PROTEIN 60 (CBP60) protein family with a novel structural clustering algorithm

Molecular genetic analyses in the model species Arabidopsis thaliana have demonstrated the major roles of different CAM-BINDING PROTEIN 60 (CBP60) proteins in growth, stress signaling, and immune responses. Prominently, CBP60g and SARD1 are paralogous CBP60 transcription factors that regulate numerous components of the immune system, such as cell surface and intracellular immune receptors, MAP kinases, WRKY transcription factors, and biosynthetic enzymes for immunity-activating metabolites salicylic acid (SA) and N-hydroxypipecolic acid (NHP). However, their function, regulation and diversification in most species remain unclear. Here we have created CBP60-DB, a structural and bioinformatic database that comprehensively characterized 1052 CBP60 gene homologs (encoding 2376 unique transcripts and 1996 unique proteins) across 62 phylogenetically diverse genomes in the plant kingdom. We have employed deep learning-predicted structural analyses using AlphaFold2 and then generated dedicated web pages for all plant CBP60 proteins. Importantly, we have generated a novel clustering visualization algorithm to interrogate kingdom-wide structural similarities for more efficient inference of conserved functions across various plant taxa. Because well-characterized CBP60 proteins in Arabidopsis are known to be transcription factors with putative calmodulin-binding domains, we have integrated external bioinformatic resources to analyze protein domains and motifs. Collectively, we present a plant kingdom-wide identification of this important protein family in a user-friendly AlphaFold-anchored database, representing a novel and significant resource for the broader plant biology community.

bioinformatics↗

bioRxiv2022

Genozip Dual-Coordinate VCF format enables efficient genomic analyses and alleviates liftover limitations

We introduce Dual Coordinate VCF (DVCF), a file format that records genomic variants against two different reference genomes simultaneously and is fully compliant with the current VCF specification. As implemented in the Genozip platform, DVCF enables bioinformatics pipelines to seamlessly operate across two coordinate systems by leveraging the system most advantageous to each pipeline step, simplifying bioinformatics workflows and reducing file generation and associated data storage burden. Moreover, our benchmarking of Genozip DVCF shows that it produces more complete, less erroneous, and less biased translations across coordinate systems than two widely used alternative tools (i.e., LiftoverVcf and CrossMap). Availability and ImplementationAn open source (GPL) version of Genozip containing DVCF functionality but not compression functionality, and which includes scripts for reproducing the benchmarks presented here, is available at https://github.com/divonlan/dvcf. Documentation is available at https://genozip.com/dvcf.

bioinformatics↗

bioRxiv2022

GO Bench: Shared-hub for Universal Benchmarking of Machine Learning-Based Protein Functional Annotations

MotivationGene annotation is the problem of mapping proteins to their functions represented as Gene Ontology terms, typically inferred based on the primary sequences. Gene annotation is a multi-label multi-class classification problem, which has generated growing interest for its uses in the characterization of millions of proteins with unknown functions. However, there is no standard GO dataset used for benchmarking the newly developed new machine learning models within the bioinformatics community. Thus, the significance of improvements for these models remains unclear. SummaryThe Gene Benchmarking database is the first effort to provide an easy-to-use and configurable hub for the learning and evaluation of gene annotation models. It provides easy access to pre-specified datasets and takes the non-trivial steps of preprocessing and filtering all data according to custom presets using a web interface. The GO bench web application can also be used to evaluate and display any trained model on leaderboards for annotation tasks. Availability and ImplementationThe GO Benchmarking dataset is freely available at llp.berkeley.edu/GO_bench/dataset_form, and code is available at http://github.com/amdson/GO_pipeline. Contactamdickson@berkeley.edu Supplementary informationSupplementary data are available at Bioinformatics online.

bioinformatics↗

bioRxiv2022

Selection of an Ideal Machine Learning Framework for Predicting Perturbation Effects on Network Topology of Bacterial KEGGPathways

Biological networks for bacterial species are used to assign functional information to newly sequenced organisms but network quality can be largely affected by poor gene annotations. Current methods of gene annotation use homologous alignment to determine orthology, and have been shown to degrade network accuracy in non-model bacterial species. To address these issues in the KEGG pathway database, we investigated the ability for machine learning (ML) algorithms to re-annotate bacterial genes based on motif or homology information. The majority of the ensemble, clustering, and deep learning algorithms that we explored showed higher prediction accuracy than CD-hit in predicting EC ID, Map ID, and partial Map ID. Motif-based, machine-learning methods of annotation in new species were more accurate, faster, and had higher precisionrecall than methods of homologous alignment or orthologous gene clustering. Gradient boosted ensemble methods and neural networks also predicted higher connectivity of networks, finding twice as many new pathway interactions than blast alignment. The use of motif-based, machine-learning algorithms in annotation software will allow researchers to develop powerful network tools to interact with bacterial microbiomes in ways previously unachievable through homologous sequence alignment. CCS CONCEPTS* Applied computing [->] Computational biology; Life and medical sciences; Bioinformatics; * Computing methodologies [->] Machine learning algorithms; Machine learning approaches. ACM Reference FormatMichael Robben, Mohammad Sadegh Nasr, Avishek Das, Manfred Huber, Justyn Jaworski, Jon Weidanz, and Jacob Luber. 2022. Selection of an Ideal Machine Learning Framework for Predicting Perturbation Effects on Network Topology of Bacterial KEGG Pathways. In The 13th ACM Conference on Bioinformatics, Computational Biology, and Health Informatics, August 07-10, 2022, Chicago, IL. ACM, New York, NY, USA, 11 pages. https://doi.org/XXXXXXX.XXXXXXX

bioinformatics↗

bioRxiv2022

Metagenomic profiling pipelines improve taxonomic classification for 16S amplicon sequencing data

Most experiments studying bacterial microbiomes rely on the PCR amplification of all or part of the gene for the 16S rRNA subunit, which serves as a biomarker for identifying and quantifying the various taxa present in a microbiome sample. Several computational methods exist for analyzing 16S amplicon-based metagenomics. However, the most-used bioinformatics tools cannot produce quality genus-level or species-level taxonomic calls and may underestimate the degree to which these calls are possible. We used 16S sequencing data from mock bacterial communities to evaluate the sensitivity and specificity of several bioinformatics pipelines and genomic reference libraries used for microbiome analyses, concentrating on measuring the accuracy of species-level taxonomic assignments of 16S amplicon reads. We evaluated the tools Qiime 2, Mothur, PathoScope 2, and Kraken 2 in conjunction with reference libraries from Greengenes, Silva, Kraken, and RefSeq. Profiling tools were compared using publicly available mock community data from several sources, comprising 136 samples with varied species richness and evenness, several different amplified regions within the 16S gene, and both DNA spike-ins and cDNA from collections of plated cells. PathoScope 2 and Kraken 2, both tools designed for whole-genome metagenomics, outperformed Qiime 2 using the DADA2 plugin and Mothur, both of which are theoretically specialized for 16S analyses.

bioinformatics↗

bioRxiv2022

WITCH-NG: Efficient and Accurate Alignment of Datasets with Sequence Length Heterogeneity

Multiple sequence alignment (MSA) is a basic part of many bioinformatics pipelines, including in phylogeny estimation, prediction of structure for both RNAs and proteins, and metagenomic sequence analysis. Yet many sequence datasets exhibit substantial sequence length heterogeneity, both because of large insertions and deletions (indels) in the evolutionary history of the sequences and the inclusion of sequencing reads or incompletely assembled sequences in the input. A few methods have been developed that can be highly accurate in aligning datasets with sequence length heterogeneity, with UPP (Nguyen et al., 2015) one of the first methods to achieve good accuracy, and WITCH (Shen et al., Bioinformatics 2021) an improvement on UPP for accuracy, In this paper, we show how we can speed up WITCH. Our improvement includes replacing a critical step in WITCH (currently performed using a heuristic search) by a polynomial time exact algorithm using Smith-Waterman. Our new method, WITCH-NG (i.e., "next generation WITCH", pronounced "witching") achieves the same accuracy but is substantially faster. WITCH-NG is available in open source form at https://github.com/RuneBlaze/WITCH-NG.

bioinformatics↗

bioRxiv2022

Leveraging Scaffold Information to Predict Protein-ligand Binding Affinity with an Empirical Graph Neural Network

MotivationProtein-ligand binding affinity prediction is an important task in structural bioinformatics for drug discovery and design. Although various scoring functions have been proposed, it remains challenging to accurately evaluate the binding affinity of a protein-ligand complex with known bound structure due to the potential preference of scoring system. In recent years, deep learning techniques have been applied to scoring functions without sophisticated feature engineering. Nevertheless, existing methods cannot model the differential contribution of atoms in various regions of proteins, and the relationship between atom properties and intermolecular distance is also not fully explored. ResultsWe propose a novel empirical graph neural network for accurate protein-ligand binding affinity prediction (EGNA). Graphs of protein, ligand and their interactions are constructed based on different regions of each bound complex. Proteins and ligands are effectively represented by graph convolutional layers, enabling the EGNA to capture interaction patterns precisely by simulating empirical scoring functions. The contributions of different factors on binding affinity can thus be transparently investigated. EGNA is compared with the state-of-the-art machine learning-based scoring functions on two widely used benchmark datasets. The results demonstrate the superiority of EGNA and its good generalization capability. Availability and implementationThe web server and source code of EGNA is available at www.csbio.sjtu.edu.cn/bioinf/EGNA and https://github.com/chunqiux/EGNA. Contacthbshen@sjtu.edu.cn or 2008xypan@sjtu.edu.cn Supplementary informationSupplementary data are available at Bioinformatics online.

bioinformatics↗

bioRxiv2022

Identification of potential biomarker candidates of drug-induced vascular injury (DIVI) in rats using gene expression and histopathology data

Drug-induced vascular injury (DIVI) observed in non-clinical species often leads to significant delays or termination of compounds in drug development due to the lack of translatable biomarkers and unknown relevance to humans. This study focused on the identification of potential biomarker candidates of drug-induced vascular injury, or more specifically mesenteric medial arterial necrosis (MAN), in rats. To do so, an adapted bioinformatic filtering pipeline was applied to previously generated gene expression data obtained from laser capture microdissected endothelial and vascular smooth muscle cells and corresponding histopathological annotations from mesenteric arteries following treatment of rats with known vascular toxicants, non-vasotoxic vasoactive comparator compounds, and corresponding vehicle. A novel gene panel including 33 genes with consistent, specific, and dose-responsive dysregulation across multiple treatments inducing MAN was identified. The degree to which these reflect injury progression was characterized and changes were identified in samples from animals where injury was anticipated but not yet histologically observed. The most predictive candidates (AUC > 0.9) with the strongest changes (logFC > 1.7) in MAN encode secreted proteins (Tnc, Vcan, Timp1 and Fn1) which are strongly interlinked with each other and other potential candidate biomarkers encoding cell surface proteins. Although further validation is required for biomarker qualification, the adapted bioinformatic approach utilized in this study provides informed data-driven starting points for further DIVI biomarker discovery and development in rats, as well as potential mechanistic insight into MAN pathogenesis.

bioinformatics↗

bioRxiv2022

Cloud-native distributed genomic pileup operations

MotivationPileup analysis is a building block of many bioinformatics pipelines, including variant calling and genotyping. This step tends to become a bottleneck of the entire assay since the straightforward pileup implementations involve processing of all base calls from all alignments sequentially. On the other hand, a distributed version of the algorithm faces the intrinsic challenge of splitting reads-oriented file formats into self-contained partitions to avoid costly data exchange between computation nodes. ResultsHere, we present a scalable, distributed, and efficient implementation of a pileup algorithm that is suitable for deploying in cloud computing environments. In particular, we implemented: (i) our custom data-partitioning algorithm optimized to work with the alignment reads, (ii) a novel and unique approach to process alignment events from sequencing reads using the MD tags, (iii) the source code micro-optimizations for recurrent operations, and (iv) a modular structure of the algorithm. We have proven that our novel approach consistently and significantly outperforms other state-of-the-art distributed tools in terms of execution time (up to 6.5x faster) and memory usage (up to 2x less), resulting in a substantial cloud cost reduction. SeQuiLa is a cloud-native solution that can be easily deployed using any managed Kubernetes and Hadoop services available in public clouds, like Microsoft Azure Cloud, Google Cloud Platform, or Amazon Web Services. Together with the already implemented distributed range joins and coverage calculations, our package provides end-users with an unified SQL interface for convenient analyzing of population-scale genomic data in an interactive way. Availabilityhttps://biodatageeks.github.io/sequila/ Contacttomasz.gambin@pw.edu.pl Supplementary informationSupplementary data are available at Bioinformatics online.

bioinformatics↗

bioRxiv2022

AcaFinder: genome mining for anti-CRISPR associated genes

Anti-CRISPR (Acr) proteins are encoded by (pro)viruses to inhibit their hosts CRISPR-Cas systems. Genes encoding Acr and Aca (Acr associated) proteins often co-localize to form acr-aca operons. Here, we present AcaFinder as the first Aca genome mining tool. AcaFinder can: (i) predict Acas and their associated acr-aca operons using guilt-by-association (GBA); (ii) identify homologs of known Acas using an HMM (Hidden Markov model) database; (iii) take input genomes for potential prophages, CRISPR-Cas systems, and self-targeting spacers (STSs); and (iv) provide a standalone program (https://github.com/boweny920/AcaFinder) and a web server (http://aca.unl.edu/Aca). AcaFinder was applied to mining over 16,000 prokaryotic and 142,000 gut phage genomes. After a multi-step filtering, 36 high-confident new Aca families were identified, which is three times of the 12 known Aca families. Seven new Aca families were from major human gut bacteria (Bacteroidota, Actinobacteria, Fusobacteria) and their phages, while most known Aca families were from Proteobacteria and Firmicutes. A complex association network between Acrs and Acas was revealed by analyzing their operonic co-localizations. It appears very common in evolution that the same aca genes can recombine with different acr genes and vice versa to form diverse acr-aca operon combinations. ImportanceAt least four bioinformatics programs have been published for genome mining of Acrs since 2020. In contrast, no bioinformatics tools are available for automated Aca discovery. As the self-transcriptional repressor of acr-aca operons, Aca can be viewed as anti-anti-CRISPRs, with a great potential in the improvement of CRISPR-Cas technology. Although all the 12 known Aca proteins contain a conserved Helix-Turn-Helix (HTH) domain, not all HTH-containing proteins are Acas. However, HTH-containing proteins with an adjacent Acr homologs encoded in the same genetic operon are likely Aca proteins. AcaFinder implements this guilt-by-association (GBA) idea and the idea of using HMMs of known Acas for homologs into one software package. Applying AcaFinder in screening prokaryotic and gut phage genomes reveals a complex acr-aca operonic co-localization network between different families of Acrs and Acas.

bioinformatics↗

bioRxiv2022

TransporterPAL: An integrative database Transporter Prediction ALgorithm

MotivationNatural products are used as drugs, cosmetic ingredients, pigments, flavors, and agricultural products. The compounds are retrievable as extracts from natural sources, but the yields are often low, and the final product may contain various impurities. These challenges can be solved by expressing the biosynthetic pathway in microbial cell factories and ensuring product secretion from the cell by using an appropriate transporter. However, insufficient knowledge of transporters for specific compounds often obstructs efficient secretion of the natural product. Therefore, our goal was to develop an algorithm that predicts transporters for a given compound using available public data. ResultsThe web application TransporterPAL predicts suitable transporters for compounds by interconnecting data for biosynthetic genes and their interactions with transporters. The web application queries the STITCH, STRING, and UniProtKB databases via their respective APIs and returns a set of potential transporters based on a compound and, optionally, the organism as input. For a test set of 61 transporter systems, each containing one or more transporters, a total of 90 unique transporters with a known substrate, we could retrieve 45% of the transporters. To our knowledge, this is the first bioinformatics tool for predicting transporter candidates for a given molecule. Availabilityhttps://transporterpal.com Contactirbo@biosustain.dtu.dk Supplementary informationSupplementary data are available at Bioinformatics online.

bioinformatics↗

bioRxiv2022

AYUKA: A toolkit for fast viral genotyping using whole genome sequencing.

Technological advances enabled the frequent use of whole genome sequencing in the clinical microbiology laboratory. While generating data is now easier than ever, the computational resources and expertise required for analysis are still a challenge for clinical applications. Since it is not always possible to collect clinical specimens at the peak viral load, sequencing results are also not always amenable for analysis with bioinformatics pipelines that always require high quality data. Here we present a fast and reliable method, we named AYUKA, for analysis of viral sequencing data that does not require data pre-processing and provides quality control metrics including estimates for sequencing depth and genome coverage, as well as identifying the viral genotypes in a sample and distinguishing mixed infection from recombinants. This method can be applied to any virus where a classification by genotype is employed and determining it is relevant. We generated a validation dataset composed of cultured and sequenced reference adenoviruses from distinct species, that we compared with the gold standard clinical processing pipeline currently implemented to demonstrate reliability. The validation shows better sensitivity than mapping and perfect specificity in detecting the correct genotypes and in a wide range of adenovirus species. Run time was consistently under one minute per sample on a standard laptop, allowing the analysis of more than 100 samples per hour. This open-source method is available at https://github.com/afonsoguerra/AYUKA and precomputed databases are available at https://zenodo.org/record/6521576 allowing analysis of raw data straight from the sequencer within minutes on a standard computer, with minimum setup or expertise required to perform the analysis. The information contained within the AYUKA report can be of use for both the clinical team that collected the sample, but also for guiding the bioinformatics analysis team in the in-depth downstream analyses and genetic epidemiology investigations.

bioinformatics↗

bioRxiv2022

DEGoldS: a workflow to assess the accuracy of differential expression analysis pipelines through gold-standard construction

RNA sequencing (RNA-seq) is a high throughput sequencing method that has become one the most employed tools in transcriptomics. The implementation of optimal bioinformatic analyses required in RNA-seq experiments may be complicated due to the large amounts of data generated by the sequencing platforms, along with the intrinsic nature of these data types. In the last years many programs and pipelines have been developed for differential expression (DE) analyses, but their effectiveness can be reduced when working with non-model species lacking public genomic resources. Moreover, there is not a universal recipe for all the experiments and datasets and the modification of standard RNA-seq bioinformatic pipelines through parameter tuning and the use of alternative software may have a strong impact in the outcome of DE analysis. Therefore, although the selection of the most accurate DE pipeline configuration and the evaluation of how these changes could affect the final DE results in RNA-seq experiments is mandatory to reduce bias, the lack of gold-standard datasets with known expression patterns hampers its implementation. In the present manuscript we present DEGoldS, a workflow consisting on sequential Bash and R scripts to construct gold-standards for simulation-based benchmarking of user selected pipelines for DE analysis and the computation of the accuracy of the pipelines. We validated the workflow with a case study consisting on real RNA-seq libraries of radiata pine, an important forest tree species with no publicly available reference genome. The results showed that slight pipeline modifications produced remarkable differences in the outcome of DE analysis.

bioinformatics↗

bioRxiv2022

PathExpSurv: Pathway Expansion for Explainable Survival Analysis and Disease Gene Discovery

MotivationIn the field of biology and medicine, the interpretability and accuracy are both important when designing predictive models. The interpretability of many machine learning models such as neural networks is still a challenge. Recently, many researchers utilized prior information such as biological pathways to develop bioinformatics methods based on neural networks, so that the prior information can provide some insights and interpretability for the models. However, the prior biological knowledge may be incomplete and there still exists some unknown information to be explored. ResultsWe proposed a novel method, named PathExpSurv, to gain an insight into the black-box model of neural network for cancer survival analysis. We demonstrated that PathExpSurv could not only incorporate the known prior information into the model, but also explore the unknown possible expansion to the existing pathways. We performed downstream analyses based on the expanded pathways and successfully identified some key genes associated with the diseases and original pathways. AvailabilityPython source code of PathExpSurv is freely available at https://github.com/Wu-Lab/PathExpSurv. Contactlywu@amss.ac.cn Supplementary informationSupplementary data are available at Bioinformatics online.

bioinformatics↗

bioRxiv2022

cazy_webscraper: local compilation and interrogation of comprehensive CAZyme datasets

Carbohydrate Active enZymes (CAZymes) are pivotal in biological processes including energy metabolism, cell structure maintenance, signalling and pathogen recognition. Bioinformatic prediction and mining of CAZymes improves our understanding of these activities, and enables discovery of candidates of interest for industrial biotechnology, particularly the processing of organic waste for biofuel production. CAZy (www.cazy.org) is a high-quality, manually-curated and authoritative database of CAZymes that is often the starting point for these analyses. Automated querying, and integration of CAZy data with other public datasets would constitute a powerful resource for mining and exploring CAZyme diversity. However, CAZy does not itself provide methods to automate queries, or integrate annotation data from other sources (except by following hyperlinks) to support further analysis. To overcome these limitations we developed cazy_webscraper, a command-line tool that retrieves data from CAZy and other online resources to build a local, shareable, and reproducible database that augments and extends the authoritative CAZy database. cazy_webscrapers integration of curated CAZyme annotations with their corresponding protein sequences, up to date taxonomy assignments, and protein structure data facilitates automated large-scale and targeted bioinformatic CAZyme family analysis and candidate screening. This tool has found widespread uptake in the community, with over 20,000 downloads. We demonstrate the use and application of cazy_webscraper to: (i) augment, update and correct CAZy database accessions; (ii) explore taxonomic distribution of CAZymes recorded in CAZy, identifying underrepresented taxa and unusual CAZy class distributions; and (iii) investigate three CAZymes having potential biotechnological application for degradation of biomass, but lacking a representative structure in the PDB database. We describe in general how cazy_webscraper facilitates functional, structural and evolutionary studies to aid identification of candidate enzymes for further characterisation, and specifically note that CAZy provides supporting evidence for recent expansion of the Auxiliary Activities (AA) CAZy family in eukaryotes, consistent with functions potentially specific to eukaryotic lifestyles. Supplementary informationcazy_webscraper source code is available at https://github.com/HobnobMancer/cazy_webscraper, and online documentation is provided at https://cazywebscraper.readthedocs.io.

bioinformatics↗