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 Read source abstracts and check manuscript versions; preprints are not peer reviewed.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 145 records · Page 8Linked to original sources

bioRxiv2019

Thirty five novel nsSNPs may effect on ADAMTS13 protein leading to Thrombotic thrombocytopenic purpura (TTP) using bioinformatics approach

BackgroundGenetic polymorphisms in the ADAMTS13 gene are associated with thrombotic thrombocytopenic purpura or TTP, a life-threatening microangiopathic disorder. This study aims to predict the possible pathogenic SNPs of this gene and their impact on the protein structure and function using insilico methods.\n\nMethodsSNPs retrieved from the NCBI database were analyzed using several bioinformatics tools. The different algorithms applied collectively to predictthe effect of single nucleotide substitution on both structure and function of the ADAMTS13 protein.\n\nResultsFifty one mutations were found to be highly damaging to the structure and function of the protein. Of those, thirty five were novel nsSNPs not previously reported in the literature.\n\nConclusionAccording to our analysis we found thirty five nsSNPs effects on ADAMTS13 protein leading to thrombotic thrombocytopenic purpura using computational approach. Bioinformatics tools are vital in prediction analysis, making use of increasingly voluminous biomedical data thereby providing markers for screening or for genetic mapping studies.

bioinformatics↗

bioRxiv2026

Pipette: Encoding scientific literature into an executable Skill Graph for multi-agent bioinformatics

The cost of genomic sequencing has fallen by several orders of magnitude, yet data analysis remains a bottleneck concentrated among researchers with specialized computational expertise. While Large Language Models can generate bioinformatics code, they frequently produce incoherent multi-step workflows due to the absence of domain-specific analytical constraints. Here, we present Pipette, a multi-agent AI framework that orchestrates end-to-end bioinformatics workflows through natural language interaction, guided by a literature-derived Skill Graph. This directed, edge-weighted knowledge graph, extracted from over 20,000 peer-reviewed publications, constrains workflow generation to biologically valid analytical transitions, preventing incomplete or incoherent workflows. We benchmarked Pipette across four biological domains using published datasets: single-cell RNA-seq analysis of peripheral blood mononuclear cells and a human pancreas atlas, bulk RNA-seq differential expression in rice under environmental stress, and two structure-based computational drug design workflows. In ablation against two LLMs operating without Skill Graph constraints, Pipette matched or exceeded both baselines across all quantitative metrics while uniquely completing multi-step cross-domain transitions. We further evaluated Pipette on a clinical genomics task, where it executed an ACMG/AMP-compliant variant classification on a reference human genome. In all cases, Pipette recapitulated established biological and clinical findings while generating a fully reproducible, machine-readable provenance record. By reducing the computational expertise required to execute standard genomic analyses, Pipette lowers the barrier between sequencing data and biological insight for bench scientists. Pipette is available at https://pipette.bio.

bioinformatics↗

bioRxiv2026

BasalCell: A project scaffold generator for bioinformatics analysis

In the current bioinformatics landscape, where R-centric and Python-centric ecosystems coexist and overlap, there is an increasing demand for the organic integration of these disparate development environments. As bioinformatics practices become ubiquitous, it is crucial to lower the technical barriers for biologists to adopt software engineering standards--such as version control, environment reproducibility, code readability, continuous integration/continuous deployment (CI/CD), and comprehensive documentation--which often present significant implementation hurdles for biologists with limited programming experience. To address these challenges, we developed BasalCell (https://github.com/yo-aka-gene/BasalCell), a project scaffolding system designed to provide a standardized, easily reproducible template that integrates these essential features by default--enabling researchers to seamlessly manage multi-language environments and automate rigorous development workflows, ultimately fostering greater transparency and reliability in biological data science.

bioinformatics↗

bioRxiv2026

Reproducible and shareable bioinformatics pipelines from natural-language prompts

Large language models (LLMs) are increasingly used to generate bioinformatics pipelines and to carry out analyses from natural-language prompts. However, the resulting analyses are often difficult to reproduce across sessions, owing to the non-deterministic nature of LLM-driven conversations and heterogeneity of local execution environments, and cannot run on remote high-performance computing (HPC) servers or be shared and reused. We present Autopipe, a platform that guides any Model Context Protocol (MCP) - compatible LLM to produce, execute, and publish source-preserved, re-executable containerized pipelines. Autopipe enables users to execute bioinformatics pipelines on any on-premises remote servers - supported by comprehensive setup documentation aimed at researchers without prior server-administration experience - and to visualize results through an extensible web-based viewer. The Autopipe platform comprises four components: a desktop application with an embedded MCP server for pipeline management and remote execution, an online registry for pipeline and plugin discovery, a web-based result viewer, and a CLI tool for customizing viewer plugins. Autopipe turns conversational analysis into re-executable and shareable workflows. Autopipe is freely available at https://autopipe.org/.

bioinformatics↗

bioRxiv2026

AutoZyme: An Autonomous Agentic Framework to Optimize Bioinformatics Software

Performance bottlenecks in widely used genomics and bioinformatics software present a substantial and growing burden as biological datasets continue to increase in size and number. Relieving these bottlenecks relies largely on expert manual optimization and therefore remains difficult to scale. Here we present AutoZyme, an agentic framework for scientific software optimization. Given a target function, AutoZyme builds benchmarks, identifies bottlenecks, and iteratively tests code changes, retaining only those that improve runtime while preserving output. We evaluated AutoZyme on 45 functions, improving runtime without substantial memory increases in over 95% of cases considered. Across 38 functions from Seurat, Scanpy and related packages in genomics and bioinformatics, AutoZyme reduced runtime by a median of 8.52-fold, with the largest reductions exceeding 676-fold. The optimized functions are distributed through AutoZyme-Library as drop-in replacements for existing analysis pipelines. We also release AutoZyme as a reusable framework for optimizing additional user-specified packages and functions.

bioinformatics↗

bioRxiv2026

FlowBench: separating planning, fault recovery and interpretation in agentic bioinformatics

AO_SCPLOWBSTRACTC_SCPLOWAgentic large language model (LLM) systems are being deployed in bioinformatics faster than they are understood, and single-metric evaluations conflate capabilities that fail independently. We introduce FlowBench, a benchmark that decomposes agentic bioinformatics performance into planning, fault recovery, biological interpretation, and end-to-end output-fidelity. Existing systems achieve high plan completeness, but their closed, single-provider designs prevent attribution of performance to scaffolding versus the underlying model. We therefore built FlowAgent, a modular, provider-agnostic framework whose components can be selectively disabled and whose backbone model can be swapped across providers on a shared harness, and used it to evaluate 23 models from three main providers. Three findings emerge. First, generating a valid workflow plan from a named toolchain is largely solved, whereas inferring an appropriate toolchain from biological intent alone is uniformly difficult regardless of model tier, compressing all models into a narrow 44-57% pass-rate band. Second, ablation shows that the dependency-structured plan and a completeness-reflection step drive performance, while adding a same-context validator-driven retry makes structural quality worse. Third, fault recovery and data-grounded interpretation remain unsolved. Models frequently propose fixes that force a clean exit while leaving the underlying data invalid, and data-grounded interpretation lags internal-knowledge recall by a consistent margin. Safety does not emerge from capability, and reasoning-tier models were among the least reliable at recognising unrecoverable faults. Once planning saturates, agent architecture and refusal calibration, not model scale, are the productive frontier. Availability and implementationFlowAgent and FlowBench are available under a GPLv3 licence at https://github.com/EnteloBio/flowagent Contactadam@entelo.bio

bioinformatics↗

bioRxiv2026

PathoBench: an open community-driven benchmark registry for pathogen bioinformatics tools

SummaryThe proliferation of pathogen bioinformatics pipelines has outpaced the communitys ability to compare them on common ground. Self-reported performance numbers, ad-hoc evaluation datasets, and inconsistent metrics make pipeline selection difficult for clinical and public-health researchers. We present PathoBench, an open web platform that addresses this gap through three coordinated mechanisms: (i) a curated registry of 26 standard benchmark datasets across 10 human pathogens, each with persistent identifiers and direct download links; (ii) pathogen-specific evaluation metrics that submissions must report, allowing direct head-to-head comparison only on the same dataset; and (iii) a credibility framework combining mandatory dataset attestation, ORCID-linked attribution, public peer comments, and administrator verification. As a case study, four published Mycobacterium tuberculosis drug-resistance pipelines were evaluated against the WHO TB mutation catalogue, demonstrating the frameworks discriminating power. PathoBench is open for community contributions across all ten supported pathogens. Availability and implementationPathoBench is freely available at https://pathobench.vercel.app. Source code is released under the MIT license at https://github.com/BPHL-Molecular/pathobench. The platform requires no installation for end users; programmatic access is available via a Supabase REST API. Contactyibo.dong@flhealth.gov Supplementary informationSupplementary data are available at Bioinformatics online.

bioinformatics↗

bioRxiv2026

Survey and Evaluation of Applied Containerization Practices in Bioinformatics

Containerization enables portable, reproducible, and scalable scientific computing. However, container development, documentation, and deployment practices can vary widely, even within a single domain. Understanding patterns of how containers are implemented in real-world settings can inform community guidelines, quality scoring rubrics, and systems to automate software review. Focusing on bioinformatics as an exemplar domain, we surveyed published software tools to find applied containerization examples. After identifying >250 tools to review, we annotated metadata regarding version control, asset provenance, and general container image health and tested the ability to build and pull the container images in a purpose-built evaluation platform (https://github.com/vpnagraj/socr8s). The majority of images tested did not build successfully in our environment. Of the tool characteristics we tracked, the strongest predictor of build success was version control activity. Tools with commits in the preceding two years were roughly twice as likely to build. Deeper assessment of failures identified a variety of issues, including missing assets and broken dependency chains. While many tools had published images accessible in open registries, 24% could neither be built nor pulled. Images tended to be large, with median size of 1.57GB for those that were able to be built. Where image specification files were available, we found that more advanced container orchestration and build techniques were uncommon. Our results highlight areas of improvement in how containerized tools are built and maintained in bioinformatics and scientific computing software in general.

bioinformatics↗

bioRxiv2026

From Prompt to Pipeline: A Comparative Evaluation of Large Language Model Coding Agents for Reproducible Bioinformatics Pipeline Construction

Agentic coding systems are increasingly presented as a way to reduce the engineering burden of scientific software development. Bioinformatics is a strong test case for this claim because useful pipelines must combine domain-specific analysis choices, command-line software, sample metadata, workflow orchestration, container or HPC execution, and interpretable quality-control reporting. We evaluated three agentic systems - Biomni, Claude Code, and Codex - on the same task: constructing a Nextflow DSL2 pipeline for paired-end CUT&Tag data that included read QC, trimming, alignment, filtering, duplicate removal, signal track generation, per-sample and group-level peak calling, control-aware group merging, annotation, FRiP calculation, deepTools visualizations, and final MultiQC reporting. Each system received the same detailed CRAFT-style prompt and was assessed against a hand-coded reference pipeline developed by the authors. All three systems produced pipeline implementations that appeared plausible at the level of documentation and file structure, but none fully satisfied the requested analysis. The most consequential failure was shared: the agent-generated pipelines performed some form of group-level merging but did not produce the requested merged-group reporting outputs. Sample-level MultiQC reports also disagreed with the reference report. Codex was closest to the reference for primary mapped-read counts, although its total-read accounting and report structure still differed. Claude Code produced the broadest final report, but its mapping summary mixed stages and therefore could not be treated as numerically correct. Biomni produced the strongest subjective documentation, but its read-count agreement with the reference report was poor and several failures required substantial Nextflow expertise to diagnose. These results suggest that current coding agents can accelerate scaffolding, documentation, and routine implementation, but they do not eliminate the need for expert review in bioinformatics workflow construction. For complex sequencing workflows, prompts must specify not only the biological intent, but also the exact stage semantics, acceptance tests, metadata contracts, expected report sections, resource propagation rules, and failure criteria needed to distinguish a plausible pipeline from a correct one.

bioinformatics↗

bioRxiv2026

Characterization of heterogeneous bioinformatics training needs in the AI era

Bioinformatics training providers serve heterogeneous audiences, a challenge sharpened by fast-evolving technologies such as generative artificial intelligence (AI). Analyzing training-needs surveys in aggregate can mask this heterogeneity. We present a network-based clustering protocol that groups respondents by the similarity of their Likert-type topic ratings: respondents form a Spearman correlation network, and community detection groups them into learner profiles with shared training needs, independent of demographics. We demonstrate the protocol on the 2026 SIB Swiss Institute of Bioinformatics (SIB) training needs survey (n=260), covering topics, barriers, and formats. We identified four interpretable learner profiles: AI and data, infectious disease genomics, functional genomics, and biochemistry, imaging & AI. Each had distinct topic priorities, barriers, and format preferences that an aggregate analysis hides. For example, LLM-based knowledge extraction ranked 2nd overall but only 16th in the infectious disease genomics profile, whereas phylogenetics ranked 17th overall but first there. Across profiles, AI/LLM skills and reproducible data practices were in high demand, whereas AI-assisted learning was, unexpectedly, the least preferred format. Because the protocol requires only a shared Likert scale, it is reusable for other training-needs assessments; the implementation is openly available. Profile-based curricula can target each group's topics, formats, and barriers.

bioinformatics↗

bioRxiv2020

Integrated bioinformatic analysis of miR-654-5p target genes and experimental validation of functional analyses

The pivotal roles of miRNAs in carcinogenesis, metastasis, and prognosis have been demonstrated recently in various cancers. This study intended to investigate the specific roles of hsa-miR-654-5p in lung cancer, which was poorly discussed. A series of closed-loop bioinformatic functional analyses were integrated with in vitro experimental validation to explore the overall biological functions and pan-cancer regulation pattern of miR-654-5p. We found that miR-654-5p abundance was significantly elevated in LUAD tissues and correlated with patients survival. 275 potential targets of miR-654-5p were then identified and the miR-654-5p-RNF8 regulation axis was validated in vitro as a proof of concept. Furthermore, we revealed the tumor-suppressing roles of miR-654-5p and demonstrated that miR-654-5p inhibited lung cancer cell epithelial-mesenchymal transition (EMT) process, cell proliferation, and migration using target-based, abundance-based and ssGSEA-based bioinformatic methods and in vitro validation. Following the construction of a protein-protein interaction network, 11 highly interconnected hub genes were identified and a five genes risk scoring model was developed to assess their potential prognostic ability. Our study will not only provide a basic miRNA-mRNA-phenotypes reference map for understanding the function of miR-654-5p in different cancers but also reveal the tumor-suppressing roles and prognostic values of miR-654-5p in lung cancer. Graphical abstract O_FIG O_LINKSMALLFIG WIDTH=199 HEIGHT=200 SRC="FIGDIR/small/423719v2_ufig1.gif" ALT="Figure 1"> View larger version (55K): org.highwire.dtl.DTLVardef@15da4caorg.highwire.dtl.DTLVardef@edfbb6org.highwire.dtl.DTLVardef@289028org.highwire.dtl.DTLVardef@1105a2e_HPS_FORMAT_FIGEXP M_FIG C_FIG

cancer biology↗

bioRxiv2020

Identification of potential and novel target genes in pituitary prolactinoma by bioinformatics analysis

Pituitary prolactinoma is one of the most complicated and fatally pathogenic pituitary adenomas. Therefore, there is an urgent need to improve our understanding of the underlying molecular mechanism that drives the initiation, progression, and metastasis of pituitary prolactinoma. The aim of the present study was to identify the key genes and signaling pathways associated with pituitary prolactinoma using bioinformatics analysis. Transcriptome microarray dataset GSE119063 was acquired from Gene Expression Omnibus datasets, which included 5 pituitary prolactinoma samples and 4 normal pituitaries samples. We screened differentially expressed genes (DEGs) with limma and investigated their biological function by pathway and Gene Ontology (GO) enrichment analysis. A protein-protein interaction (PPI) network of the up and down DEGs were constructed and analyzed by HIPPIE and Cytoscape software. Module analyses were performed. In addition, a target gene - miRNA network and target gene - TF network of the up and down DEGs were constructed by NetworkAnalyst and Cytoscape software. The set of DEGs exhibited an intersection consisting of 989 genes (461 up-regulated and 528 down-regulated), which may be associated with pituitary prolactinoma. Pathway enrichment analysis showed that the 989 DEGs were significantly enriched in the retinoate biosynthesis II, signaling pathways regulating pluripotency of stem cells, ALK2 signaling events, vitamin D3 biosynthesis, cell cycle and aurora B signaling. Gene Ontology (GO) enrichment analysis also showed that sensory organ morphogenesis, extracellular matrix, hormone activity, nuclear division, condensed chromosome and microtubule binding. In the PPI network and modules, SOX2, PRSS45, CLTC, PLK1, B4GALT6, RUNX1 and GTSE1 were considered as hub genes. In the target gene miRNA network and target gene - TF network, LINC00598, SOX4, IRX1 and UNC13A were considered as hub genes. Using integrated bioinformatics analysis, we identified candidate genes in pituitary prolactinoma, which may improve our understanding of the mechanisms of the pathogenesis and integration; genes may be therapeutic targets and prognostic markers for pituitary prolactinoma.

genomics↗

bioRxiv2021

Juvenile Huntington's Disease and Other PolyQ diseases, Update on Neurodevelopmental Character and Comparative Bioinformatic Review of Transcriptomic Data

Polyglutamine (PolyQ) diseases are neurodegenerative disorders caused by the CAG repeat expansion mutation in affected genes resulting in toxic proteins containing a long chain of glutamines. There are nine PolyQ diseases: Huntingtons disease (HD), spinocerebellar ataxias (types 1, 2, 3, 6, 7, and 17), dentatorubral-pallidoluysian atrophy (DRPLA), and spinal bulbar muscular atrophy (SBMA). In general, longer CAG expansions and longer glutamine tracts lead to earlier disease presentations in PolyQ patients. Rarely, cases of extremely long expansions are identified for PolyQ diseases, and they consistently lead to juvenile or sometimes very severe infantile-onset polyQ syndromes. In apparent contrast to the very long CAG tracts, shorter CAGs and PolyQs in proteins seems to be the evolutionary factor enhancing human cognition. Therefore, polyQ tracts in proteins can be modifiers of brain development and disease drivers, which contribute neurodevelopmental phenotypes in juvenile- and adult-onset PolyQ diseases. Therefore we performed a bioinformatics review of published RNAseq polyQ expression data resulting from the presence of polyQ genes in search of neurodevelopmental expression patterns and comparison between diseases. The expression data were collected from cell types reflecting stages of development such as iPSC, neuronal stem cell, neurons, but also the adult patients and models for PolyQ disease. Our comparative bioinformatic review highlighted several (neuro)developmental pathways and genes identified within PolyQ diseases and mouse models responsible for neural growth, synaptogenesis, and synaptic plasticity.

neuroscience↗

bioRxiv2021

Combining IP3 affinity chromatography and bioinformatics reveals a novel protein-IP3 binding site on Plasmodium falciparum MDR1 transporter

Intracellular Ca2+ mobilization induced by second messenger IP3 controls many cellular events in most of the eukaryotic groups. Despite the increasing evidence of IP3-induced Ca2+ in apicomplexan parasites like Plasmodium, responsible for malaria infection, no protein with potential function as an IP3-receptor has been identified. The use of bioinformatic analyses based on previously known sequences of IP3-receptor failed to identify potential IP3-receptor candidates in any Apicomplexa. In this work, we combine the biochemical approach of an IP3 affinity chromatography column with bioinformatic meta-analyses to identified potential vital membrane proteins that present binding with IP3 in Plasmodium falciparum. Our analyses reveal that PF3D7_0523000, a gene that codes a transport protein associated with multidrug resistance, as a potential target for IP3. This work provides a new insight for probing potential candidates for IP3-receptor in Apicomplexa.

biochemistry↗

bioRxiv2021

Proteome, bioinformatic and functional analyses reveal a distinct and conserved metabolic pathway for bile salt degradation in the Sphingomonadaceae

Bile salts are amphiphilic steroids with a C5 carboxylic side chain with digestive functions in vertebrates. Upon excretion, they are degraded by environmental bacteria. Degradation of the bile-salt steroid skeleton resembles the well-studied pathway for other steroids like testosterone, while specific differences occur during side-chain degradation and the initiating transformations of the steroid skeleton. Of the latter, two variants via either {Delta}1,4- or {Delta}4,6-3-ketostructures of the steroid skeleton exist for 7-hydroxy bile salts. While the {Delta}1,4- variant is well-known from many model organisms, the {Delta}4,6-variant involving a 7-hydroxysteroid dehydratase as key enzyme has not been systematically studied. Here, combined proteomic, bioinformatic and functional analyses of the {Delta}4,6-variant in Sphingobium sp. strain Chol11 were performed. They revealed a degradation of the steroid rings similar to the {Delta}1,4-variant except for the elimination of the 7-OH as key difference. In contrast, differential production of the respective proteins revealed a putative gene cluster for side-chain degradation encoding a CoA-ligase, an acyl-CoA dehydrogenase, a Rieske monooxygenase, and an amidase, but lacking most canonical genes known from other steroid-degrading bacteria. Bioinformatic analyses predicted the {Delta}4,6-variant to be widespread among the Sphingomonadaceae, which was verified for three type strains which also have the predicted side-chain degradation cluster. A second amidase in the side-chain degradation gene cluster of strain Chol11 was shown to cleave conjugated bile salts while having low similarity to known bile-salt hydrolases. This study signifies members of the Sphingomonadaceae remarkably well-adapted to the utilization of bile salts via a partially distinct metabolic pathway. ImportanceThis study highlights the biochemical diversity of bacterial degradation of steroid compounds, in particular bile salts. Furthermore, it substantiates and advances knowledge of a variant pathway for degradation of steroids by sphingomonads, a group of environmental bacteria that are well-known for their broad metabolic capabilities. Biodegradation of bile salts is a critical process due to the high input of these compounds from manure into agricultural soils and wastewater treatment plants. In addition, these results may also be relevant for the biotechnological production of bile salts or other steroid compounds with pharmaceutical functions.

microbiology↗

bioRxiv2021

Methylothon: a versatile course-based high school research experience in microbiology and bioinformatics-- with pink bacteria

Methylothon is an inquiry-based high school learning module in microbial ecology, molecular biology, and bioinformatics that centers around pink-pigmented plant-associated methylotrophic bacteria. Here we present an overview of the modules learning goals, describe course resources (available for public use on http://methylothon.com), and relate lessons learned from adapting Methylothon for remote learning during the pandemic in spring of 2021. The original in-person version of the module allows students to isolate their own strains of methylotrophic bacteria from plants they sample from the environment, to identify these using PCR, sequencing, and phylogenetic analysis, and to contribute their strains to original research in a university lab. The adapted version strengthens the focus on bioinformatics and increases its flexibility and accessibility by making the lab portion optional and adopting free web-based tools. Student feedback and graded assignments from Spring 2021 revealed that the lesson was especially effective at introducing the concepts of BLAST and phylogenetic trees, and that students valued and felt inspired by the opportunity to conduct hands-on work and to participate in community science.

scientific communication and education↗

bioRxiv2021

Construction of a solid Cox model for AML patients based on multiomics bioinformatic analysis

Acute myeloid leukemia (AML) is a highly heterogeneous hematological malignancy. The bone marrow (BM) microenvironment in AML plays an important role in leukemogenesis, drug resistance and leukemia relapse. In this study, we aimed to identify reliable immune-related biomarkers for AML prognosis by multiomics analysis. We obtained expression profiles from The Cancer Genome Atlas (TCGA) database and constructed a LASSO-Cox regression model to predict the prognosis of AML using multiomics bioinformatic analysis data. This was followed by independent validation of the model in the GSE106291 (n=251), GSE12417 (n=163) and GSE37642 (n=137) datasets and mutated genes in clinical samples for predicting overall survival (OS). Molecular docking was performed to predict the most optimal ligands to these hub genes. The single-cell RNA sequence dataset GSE116256 was used to clarify the expression of the hub genes in different immune cell types. According to their significant differences in immune gene signatures and survival trends, we concluded that the immune infiltration-lacking subtype (IL type) is associated with better prognosis than the immune infiltration-rich subtype (IR type). Using the LASSO model, we built a classifier based on 5 hub genes to predict the prognosis of AML (risk score = -0.086xADAMTS3 + 0.180xCD52 + 0.472xCLCN5 - 0.356xHAL + 0.368xICAM3). In summary, we constructed a prognostic model of AML using integrated multiomics bioinformatic analysis that could serve as a therapeutic classifier.

cancer biology↗

bioRxiv2022

Bridging the gap between bioinformatics and the clinical and public health microbiology laboratory: An ISO-accredited genomics workflow for antimicrobial resistance

Realising the promise of genomics to revolutionise routine AMR diagnosis and surveillance has been a long-standing challenge in clinical and public health microbiology. We have directly addressed this issue by creating and validating abritAMR, an ISO-accredited bioinformatics platform for genomics-based bacterial AMR gene detection. abritAMR utilises the NCBI AMRFinderPlus for detection of AMR genes and mutations, with additional features to classify AMR determinants into an antibiotic class. We validated abritAMR by comparing with multiplex PCR or gold-standard reference genomes, together representing 1500 different bacteria across 29 genera and covering 415 antibiotic resistance alleles. We also assessed inference of phenotypic resistance by comparing genomic predictions with agar dilution results for 864 Salmonella spp. Performance of abritAMR was excellent, detecting AMR genes with 99.9% accuracy (95% CI 99.9-99.9%), 97.9% sensitivity (95% CI 97.5-98.4%) and 100% specificity (100-100%). Phenotypic inference of resistance for Salmonella spp. was equally impressive, with 98.9% accuracy (98.7-99.1%). Validation data were submitted to the governing authority and ISO15189 accreditation was achieved. Implementation of abritAMR resulted in streamlined bioinformatics and reporting pathways, and it was readily updated and re-verified with database revisions or changes in reporting requirements. abritAMR is publicly and freely available to assist clinical and public health microbiology laboratories everywhere harness the power of AMR genomics in their professional practice.

microbiology↗