nf-cavalier is a reproducible computational workflow implemented in Nextflow that automates end-to-end steps for genomic variant annotation, prioritization, and reporting in the context of suspected rare Mendelian diseases. The pipeline is intended to streamline the process from variant callsets through candidate review-ready artifacts, supporting clinical research teams by standardizing filtering and generating presentation-ready reports for rapid assessment.
The authors describe nf-cavalier as applicable to single individuals, family trios or larger rare disease cohorts. The design emphasizes configurable filters and outputs that facilitate human review rather than replacing expert interpretation.
The pipeline accepts variant callsets as its primary input together with a specification of the genomic focus: either a target gene panel or a phenotype of interest. This allows users to narrow analysis to genes or regions relevant to the clinical question. Use cases described include analysis of an individual proband, familial analyses where segregation is informative, and cohort-level workflows for rare disease research.
nf-cavalier applies a sequence of configurable filtering criteria to prioritize variants. The filters explicitly reported include predicted gene consequence, computational pathogenicity predictions, population allele frequency thresholds, and familial segregation patterns. Each of these criteria is customizable in the pipeline, enabling users to adjust sensitivity and specificity according to the research or clinical context.
The pipeline therefore operationalizes common prioritization strategies used in rare disease genomics: deprioritize common population variants, elevate variants predicted to have high impact on gene products, incorporate in silico pathogenicity scores, and use family structure to support or refute candidate variants.
For candidate variants that pass filtering, nf-cavalier renders sequencing data visualizations aimed at facilitating human review. These visualizations enable clinical researchers and variant curators to inspect read-level evidence supporting candidate calls before downstream clinical interpretation. The inclusion of visual review is emphasized as part of the pipeline’s support for rigorous variant assessment.
nf-cavalier returns candidate variant results in user-friendly output formats suitable for clinical research workflows. Report formats explicitly mentioned are interactive HTML reports and PowerPoint slide decks. Reports include embedded links to external resources to accelerate literature lookup and annotation, and are structured to support rapid review by clinical teams. The availability of both interactive web-style reports and slide decks facilitates communication between bioinformatics teams and clinical reviewers.
The pipeline is maintained on GitHub under the repository name bahlolab/nf-cavalier. It is released under the permissive MIT open-source licence, making the code and pipeline configuration available for community use and modification. The source location and licence were reported directly in the preprint.
The preprint lists Jacob E Munro, Joshua Reid, Melanie E Bahlo, and Mark F Bennett as authors, with affiliations that include the Walter and Eliza Hall Institute of Medical Research and the University of Melbourne. The manuscript was posted on bioRxiv on August 10, 2026. The report is a preprint and has not been certified by peer review; the authors note this in the article header.
The authors declared no competing interests in the preprint. Copyright for the preprint remains with the authors/funders, and the manuscript is made available under a CC-BY 4.0 International license as presented on bioRxiv. The pipeline code itself is distributed under the MIT licence as noted above.
Notes and limitations based on the source
All technical features, outputs, and configuration options summarized here are taken from the preprint abstract and article metadata. Detailed implementation specifics, performance metrics, benchmarking data, or case examples were not reported in the provided source excerpt; those details may be present in the full manuscript or repository but are not included in the source text used for this rewrite.