Skip to content

Citation and research-use limitations

Cite the exact version you used

OncoTracer does not yet have a formal article DOI. The repository's CITATION.cff is therefore the authoritative current citation metadata and explicitly marks itself as a placeholder until a formal citation is available.

A current software citation is:

Farkas, Carlos. (2026). OncoTracer: reproducible LP-WGS CNA analysis for ONT and Illumina data (version 2.1.0) [Computer software]. https://github.com/cfarkas/oncotracer

Record the embedded source identity used in your analysis:

oncotracer provenance --json

Suggested methods text:

Low-pass whole-genome sequencing copy-number analysis was performed with OncoTracer (version/commit: replace with exact value) using Illumina qDNAseq at replace-kb bins or ONT ichorCNA at replace-kb bins, followed by BAM-supported boundary refinement and CNA codification. The run used replace container digest/runtime, replace reference build, and the archived YAML/samplesheet.

Replace every bold placeholder. Do not cite only latest, because that tag can change. A dated image is also a tag: record its digest and embedded source commit. Current integration builds can retain software version 2.1.0 while containing newer code than the stable release; the version alone does not identify that code.

If variants were enabled, also state the caller/version, tumor-only versus matched-normal design, preservation, ONT model/preset where applicable, target intervals, native filters, extra assessment settings and annotation databases. Do not describe default Varlociraptor presence scoring as a somatic classifier.

GitHub can render CITATION.cff through its “Cite this repository” interface. If that file and this page differ, report the discrepancy and use the repository metadata from the exact commit analyzed.

Cite the methods you rely on

A reproducible report should also cite the relevant upstream methods/software, not only OncoTracer:

For small-variant analysis, cite only the components actually used:

Follow each project's citation instructions and record the actual version/model revision. For optional methylation, also cite the basecaller, modification tools and classifier used; see methylation configuration. For Sturgeon, acknowledge the UMC Utrecht team and cite Vermeulen, Pagès-Gallego, Kester et al., Nature (2023).

Use the citation/version recorded by the actual run manifest and stage-01 provenance, because tool versions may differ between releases. Cite Nextflow only when reproducing a historical v1.1 analysis; native v2 does not invoke it.

Cite public example data separately

Software citation does not replace dataset attribution. For the HCC1143 six-FASTQ example, cite:

For any other public data, record archive, project, sample/run accessions, retrieval date, checksums, and the associated study.

Minimum reproducibility record

Archive with the result:

oncotracer --version
oncotracer provenance --json
oncotracer doctor --backend conda

For Conda, preserve explicit package specifications for all five core prefixes and every selected optional environment, including separately installed ONT tools. For containers, preserve the immutable image digest and provenance emitted from inside that exact image, in addition to the host launcher's provenance.

For a stable release, retain its release-provenance.json, which binds its source, binary, image and complete public-data CNA parity audits. For a source checkout or dated integration image, retain that build's own provenance and validation evidence; do not attach an older stable release's parity bundle as validation of newer code.

Also preserve:

  • unedited run YAML and samplesheet/ONT mapping table;
  • input file checksums and source accessions;
  • reference assembly and identity/checksum (normally hg38; matching hg19 is supported for standalone BAM variants);
  • caller, analysis type, bin size, and refinement parameters;
  • workflow summary, native run manifest and .oncotracer-native/trace.tsv;
  • hardware/executor/runtime information;
  • any manual exclusions or reruns;
  • primary stage-02/03 tables and QC reports;
  • if variants ran, 08_variants/variant_status.json, variant_provenance.json, original/final caller VCFs, assessment tables and annotation database identities;
  • explicit matched-normal sample mapping for Strelka2 somatic, and reasons for skipped or unevaluated assessments.

Research-use scope

OncoTracer is a research workflow for CNA analysis with optional small-variant and methylation branches. It is not a standalone diagnostic system or a medical device. Its output must not be used by itself to diagnose disease, select treatment, establish prognosis, or report a clinical result.

Low-pass read-depth CNA analysis can support broad/focal gain and loss detection and CNA-burden/aneuploidy research. CNA output alone does not establish SNVs, indels, balanced rearrangements, most fusions, methylation class, expression/protein state, copy-neutral LOH, clonality, or biallelic status. Separate optional variant or methylation modules analyze their own evidence; their successful execution is not a sensitivity benchmark or clinical validation. Sensitivity depends on coverage, bin size, tumor fraction, ploidy, normal contamination, library quality, reference/mappability, and caller assumptions.

Optional classifier, literature, model, and pathology-concordance outputs are hypotheses or compatibility summaries. They do not validate pathology and must be reviewed against primary CNA tables, morphology, IHC, cytogenetics, clinical-grade sequencing, and other appropriate assays.

Data governance and privacy

  • Use de-identified research identifiers in samplesheets and pathology tables.
  • Include only pathology columns needed for the planned analysis.
  • Do not send identifiable clinical text to network services or public issue trackers.
  • Confirm institutional approval, consent/data-use conditions, and computing policy before analysis.
  • Treat public cell-line/tutorial results as software demonstrations, not clinical validation.

Licensing and reuse

The repository currently does not include a standalone LICENSE file. Do not assume that public visibility grants unrestricted redistribution or commercial reuse. Contact the repository owner for licensing clarification, and follow the licenses/citation requirements of every bundled or downloaded dependency and dataset.

FFPERASE has additional conditions: its pinned license requires MSK's express written permission to publish research results. OncoTracer's download acknowledgment does not grant that permission. ANNOVAR software/databases remain separately obtained under their own terms.