# Third-party notices

The project's original code is MIT-licensed. Third-party code and data retain
their own terms; the project license does not relicense those materials.

## GenomeSpy schema, browser runtime, and examples

The bundled `src/genome_spy/schema/genome-spy-schema.json`, generated schema
bindings and descriptions, browser runtime modules, and adapted official
visualization examples come from
[GenomeSpy](https://github.com/genome-spy/genome-spy), copyright 2018–2026 Kari
Lavikka, under MIT. The full license is in `LICENSES/GENOMESPY-MIT.txt`. The
browser files are generated from the pinned `@genome-spy/core` and
`@genome-spy/inspector` npm releases and include their bundled open-source
runtime dependencies.

The Core browser bundle embeds the Lato font bitmap and metrics, copyright
2010–2014 tyPoland Lukasz Dziedzic, with Reserved Font Name "Lato", under the
SIL Open Font License 1.1. Its full upstream notice is retained in
`LICENSES/LATO-OFL-1.1.txt` and in the Core module, including when that module
is embedded in exported HTML.

GenomeSpy's schema types include material adapted from Vega-Lite (axis, scale,
data, title, channel, and selection definitions) and Vega (input bindings).
Their notices also apply to the corresponding bundled schema and generated
descriptions: University of Washington Interactive Data Lab. All rights
reserved. The BSD-3-Clause terms and upstream copyright
notices are in `LICENSES/VEGA-LITE-BSD-3-Clause.txt` and
`LICENSES/VEGA-BSD-3-Clause.txt`. Expression API generation also reads
GenomeSpy and Vega expression documentation.

## Bundled reference and example data

- **`p53_sequence_comparison_aligned.fasta.gz`:** 34 UniProt
  Consortium p53 sequences, obtained through
  [Plotly's FASTA](https://github.com/plotly/datasets/blob/0c447c47b757ad74edecab31f0d72f849d2e67c2/Dash_Bio/Genetic/alignment_viewer_p53.fasta).
  The accession and sequence version are retained in each header, including
  P02340 sequence version 3. GenomeSpy Python aligned the sequences with MAFFT
  7.526 L-INS-i and gzip-compressed the result. UniProt's copyrightable data
  are [CC BY 4.0](https://www.uniprot.org/help/license/); retain attribution,
  the license link, and this modification notice. UniProt provides no warranty
  of correctness and does not grant patent or other third-party rights.
- **`pik3ca_mutations.json`:** Plotly's NeedlePlot PIK3CA fixture,
  copied without modification. Its 173 feature rows match UniProt P42336
  entry version 193 (also 197, 201 and 203), grouped by feature type and
  coordinates; values count annotations, not tumor samples. The separate
  domain track comes from Pfam through the
  [demo's parser](https://github.com/plotly/dash-bio/blob/master/dash_bio/utils/mutation_data_parser.py).
  Credit the UniProt Consortium (CC BY 4.0), Pfam/InterPro at EMBL-EBI
  ([CC0 1.0](https://www.ebi.ac.uk/interpro/about/license/)), and Plotly for
  selecting and reformatting this historical teaching fixture.
- **`hapmap_gwas.csv`:** byte-identical to Sahir Bhatnagar's
  [manhattanly table](https://github.com/sahirbhatnagar/manhattanly/blob/f959be07f7c1df747ed5fe0169d72a97a49b343a/data-raw/HapMap.csv),
  via Plotly. Coordinates and rs identifiers are from the HapMap 3 draft
  release 2 PLINK map (2009-01, build 36); the
  [HapMap project released its data into the public domain](https://www.genome.gov/11511175/about-the-international-hapmap-project-fact-sheet).
  Gene names/distances are UCSC hg18 annotations, not HapMap observations.
  P-values and effect sizes are simulated; z-scores are calculated from those
  p-values. Retain the manhattanly/Plotly MIT contribution notices and credit
  HapMap and UCSC. No individual genotypes are in this file.
- **`refseq_gene_bodies.csv.gz`:** independently prepared from
  UCSC hg19/hg38 `refGene` downloads dated 2026-08-26; overlapping transcripts
  are collapsed and counted. Credit UCSC and NCBI RefSeq. The
  [UCSC data terms](https://genome.ucsc.edu/license/) permit public and
  commercial use of these reference tables.
- **`tal1_alphagenome_reference.json.gz`:** UCSC hg38/GRCh38
  chr1:47173759–47304831 (zero-based, half-open), retrieved 2026-08-10 and
  packaged with project-authored interval metadata. Credit UCSC and the Genome
  Reference Consortium under the same UCSC data terms. This is reference
  sequence, not AlphaGenome code, weights, or predictions.
- **`mutation_impact_reference.json`:** project-authored synthetic
  teaching data, under the project's MIT license.
- **`pik3ca_tcga_brca_lollipop.json`:** named datasets copied from
  [GenomeSpy's documented example](https://github.com/genome-spy/genome-spy/blob/3b55caf5111ab769da222a93990ca38480495f9e/docs/examples/genomic-data/pik3ca-tcga-brca-lollipop.md).
  Open-access GDC TCGA-BRCA masked somatic calls were aggregated into distinct
  tumor counts for recurrent canonical-transcript mutations, omitting
  singletons. Credit the TCGA Research Network and retain the no-reidentification
  requirement. Protein length and selected domains are from UniProt P42336,
  CC BY 4.0; credit the UniProt Consortium and GenomeSpy's selection/reformatting.
- **`tcga_ov_gistic_scores.tsv.gz`, `tcga_ov_gistic_lesions.tsv.gz`:**
  [GenomeSpy's documented open-access Broad/TCGA outputs](https://genomespy.app/docs/examples/genomic-data/tcga-ov-gistic/),
  byte-identical after decompression to Firehose OV-TP CopyNumber GISTIC2
  Level 4, run 2016-01-28, GISTIC2.0.22, hg19. Only compression changed.
  Credit the TCGA Research Network and Broad Institute TCGA Genome Data
  Analysis Center (2016), retain the source archive citation and respect
  TCGA responsible-use and no-reidentification requirements.
- **`rnf7_direct_rna.json.gz`:** a reduced RNF7 transcript-window derivative
  of the [xPore demo archive](https://doi.org/10.5281/zenodo.4587661) and
  Supplementary Table 6 from Hendra et al.,
  [*Nature Methods* 2022](https://doi.org/10.1038/s41592-022-01666-1).
  Both sources are CC BY 4.0. GenomeSpy selected one transcript and replicate
  per condition, retained mapped primary alignment geometry and alignment
  events, pseudonymized read labels, calculated coverage, and
  selected/reformatted published m6Anet site probabilities. The displayed
  GRCh38 RNF7-201 reference sequence was reconstructed from three exon regions
  returned by Ensembl, whose data are available without restriction. Credit
  Pratanwanich et al. for xPore, Hendra et al. for m6Anet, and Ensembl for the
  sequence service; link CC BY 4.0 and retain this modification notice.

The full CC BY 4.0 and CC0 texts are in `LICENSES/CC-BY-4.0.txt` and
`LICENSES/CC0-1.0.txt`. The source-repository MIT notices in
`LICENSES/DATA-SOURCE-REPOSITORIES-MIT.txt` accompany the corresponding
selection, formatting and example contributions.

## Bundled TCGA example tables

| Files | Source and processing |
| --- | --- |
| `tcga.tsv` | Pyoncoprint revision `31e6d6de78b64070a9c6c582ce3ef571b14e4c71`, a cBioPortal LUAD PanCancer Atlas 2018 tabular export. See the provider notices below. |
| `oncoprint_dataset3.json` | Plotly Dash Bio alteration fixture at revision `20088fe2168789d0d5f5ba9820df9bd46f73231f`, copied without modification. |
| `tcga_laml.maf.gz`, `tcga_laml_annot.tsv`, `brca.maf.gz` | Maftools files at `015a4cf8c69ba89a55a3fdcea911421509e9a198`, copied without modification. LAML includes mutation/VAF and clinical annotations; BRCA contains one sample, TCGA-A8-A08B. |
| `tcga_laml_combined_oncoplot.json.gz` | Python-prepared tables derived from six maftools inputs at that revision. Custom pathway groups come from its vignette; GISTIC and MutSig inputs match Firehose LAML-TB 2016-01-28. Processing joins clinical data and derives alteration, recurrence, VAF and MutSig summaries. |

Credit TCGA and the contributing studies, Plotly, pyoncoprint and maftools
(Anand Mayakonda), as applicable.

## Airway data

**`airway_scaledcounts.csv`, `airway_metadata.csv`:** exact Bioconnector
workshop copies. Counts were prepared with kallisto and rounded tximport
`lengthScaledTPM` values from the Himes et al. experiment.

Airway source attribution: Stephen Turner and the UVA Bioconnector workshop
contributors, [scaled counts](https://github.com/bioconnector/workshops/blob/d0c4d0cca5f21dc2c6e12912dc09d5e279e706ce/data/airway_scaledcounts.csv)
and [metadata](https://github.com/bioconnector/workshops/blob/b563237b8feb18857d048b74975de6e753667fad/data/airway_metadata.csv),
from Himes et al., *PLoS One* 2014,
[doi:10.1371/journal.pone.0099625](https://doi.org/10.1371/journal.pone.0099625),
GEO GSE52778. The workshop's [CC BY-NC-SA 4.0 license](https://github.com/bioconnector/workshops/blob/b563237b8feb18857d048b74975de6e753667fad/LICENSE)
is retained in `LICENSES/CC-BY-NC-SA-4.0.txt` for its contributions. It requires
attribution, noncommercial use, modification notices, and share-alike for
adaptations.

## Provider terms and scientific context

The LUAD export includes Dicipivirus microbiome values from Poore et al. (2020),
[retracted in 2024](https://doi.org/10.1038/s41586-024-07656-x).
cBioPortal [removed the profile](https://github.com/cBioPortal/datahub/commit/18edd66a52b4d1b10a697e8d2fee79e1b1452b73)
on 2024-07-08. These values must not be presented as validated biological
findings. Credit TCGA, cBioPortal, the contributing studies and pyoncoprint.

cBioPortal's [data policy](https://docs.cbioportal.org/user-guide/faq/) specifies
ODbL unless otherwise noted; the historical LUAD
[study notice](https://github.com/cBioPortal/datahub/blob/2ba97208ff41206df19e860cee5e44236c3cc93b/public/luad_tcga_pan_can_atlas_2018/LICENSE)
refers to Broad GDAC terms. Preserve the applicable provider notices. ODbL's
requirements include attribution, database notices, and share-alike and
machine-readable availability for qualifying derivative databases.

Results using TCGA-derived data are in whole or part based upon data generated
by the [TCGA Research Network](https://www.cancer.gov/ccg/research/genome-sequencing/tcga/using-tcga-data/citing).
GDC's
[analysis policy](https://gdc.cancer.gov/analyze-data/data-analysis-policies)
prohibits attempts to reidentify participants, including for open data.

## Code adapted from Vega-Altair

Parts of this project are adapted from
[Vega-Altair](https://github.com/vega/altair), copyright 2015–2025 the
Vega-Altair Developers, under the BSD-3-Clause license. The complete license
is in [`LICENSES/ALTAIR-BSD-3-Clause.txt`](LICENSES/ALTAIR-BSD-3-Clause.txt).

The adapted areas are:

- `src/genome_spy/schemapi.py`, a reduced schema-wrapper runtime based on
  `altair/utils/schemapi.py`;
- `src/genome_spy/_expressions.py`, expression authoring primitives adapted
  from `altair/expr/core.py`;
- `tools/schemapi/codegen.py`, whose schema-wrapper generation architecture is
  adapted from Altair's `tools/schemapi/` package and
  `tools/generate_schema_wrapper.py`;
- the multifeature penguins and cars strip-plot cases in `tests/test_chart.py`,
  adapted from Altair's example suite and mark documentation.