UCSC Genome Browser · Tutorial 2

Cancer Data

The cancer & clinical databases, worked on BRAF V600E

A hands-on tour of the oncology tracks · genome.ucsc.edu

What we'll cover

How to analyse variants using the cancer database tracks.
The main tracks: CIViC, COSMIC, ClinVar, TCGA (somatic) and GenCC, OMIM, gnomAD (germline).
A running example throughout: BRAF V600E, the well-known melanoma driver.
Where each database fits when you ask “is this variant real, and does it matter?”

Two kinds of question We split the databases into somatic (acquired, in cancer) and germline (inherited) diagnosis.

Where the cancer data lives

The key datasets and what each is good for. Click a name for its track description page.

Clinical & cancer databases

ClinVar: archive of lab and expert-panel variant classifications. Is it already called pathogenic, benign, or VUS (and now its somatic oncogenicity)?
COSMIC: catalogue of somatic mutations in cancer. Shows how recurrently a site is mutated across tumours, flagging likely drivers.
CIViC: crowd-sourced, expert-curated clinical interpretations. Is the variant oncogenic and linked to a therapy?
GenCC: curated gene-disease validity with evidence strength. Is the gene genuinely tied to a heritable disease?
OMIM: genes and Mendelian phenotypes. The reference for known gene-disease relationships and disease alleles.

Population frequency & evidence

TCGA Pan-Cancer: somatic mutations across ~10,000 tumours of many types, drawn as a lollipop. Shows recurrence and hotspots by cancer type.
gnomAD: population allele frequencies from large healthy cohorts. Common implies benign, and it filters germline variants out of somatic calls.
UniProt: protein domains and key residues. Does the variant fall in a functional region (ACMG PM1)?
REVEL: ensemble missense pathogenicity score (0 to 1). High values support a damaging computational call (PP3).
Conservation (100-way): cross-species constraint. Conserved positions are more likely functionally important.

Somatic diagnosis

Somatic variants

acquired mutations: SNVs, de novo changes, and cancer drivers

CIViC: Clinical Interpretation of Variants in Cancer

Collaborative and expert-reviewed: submissions are vetted by editors and the community.
Links each variant to diseases, drugs and therapies.
Every claim carries a cited publication as evidence.
Best for: is this variant oncogenic and linked to a therapy?

Try it, ▶ open the BRAF CIViC session Hover a variant to preview its therapies and diseases.

CIViC track with two variants pointing to their mouse-over tooltips — Hover a variant in the CIViC track to see its associated therapies and diseases.

CIViC: click a feature for the details

CIViC variant details page with the link to CIViC boxed in red — Click a variant to open its details page; the boxed link goes straight to the CIViC page for that item.

CIViC: the page for BRAF V600E

CIViC web page for BRAF V600E, annotated — The CIViC page for BRAF V600E: click a variant for more, and read the diseases and therapies associated with each.

CIViC: the evidence behind each call

Clicking an Evidence ID opens its detailed evidence record — Click an Evidence ID (EID) in the list to open its full evidence record.

Each disease and therapy links to Evidence Items (EIDs); each EID cites the supporting publication and rates its strength: the “show me the evidence” layer behind actionability.

COSMIC: Catalogue of Somatic Mutations in Cancer

The reference catalogue of somatic mutations seen across tumours.
Shows how recurrently a position is mutated, and in which cancer types.
Best for: spotting recurrent hotspots and likely drivers.

Try it, ▶ open COSMIC at the BRAF hotspot The V600 hotspot stands out by recurrence.

COSMIC track in the Genome Browser — COSMIC in the Browser: recurrent somatic mutations across cancer types.

ClinVar: variant-disease classifications

Public archive of submitted clinical classifications, pathogenic to benign, with review status.
Increasingly carries somatic oncogenicity and clinical-impact calls.
Click a variant for its interpretation and submitters.

Try it, ▶ open the BRAF ClinVar interpretation track Variants coloured by clinical significance.

The ClinVar track across a gene, coloured by clinical significance.

ClinVar at BRAF V600E — At BRAF V600E: pathogenic submissions (T>A).

ClinVar database link

Clicking a ClinVar interp item in the Browser opens its full ClinVar record — Click a variant in the **ClinVar interp** track for its details page, then follow the link out to the variant's full record on ClinVar.

TCGA Pan-Cancer

TCGA lollipop track, with a click to a variant's mouse-over and details page — Lollipop track of somatic mutations; click a lollipop for its mouse-over and the item details (here C>T, seen in 38 samples).

Somatic variants across thousands of tumours and many cancer types; useful for variant frequency by cancer type.
Drawn as a lollipop plot; mouse-over identifies the variant and its sample count.

COSMIC vs TCGA Both show recurrence. TCGA is one uniform cohort with comparable, by-cancer-type frequencies; COSMIC is the broader aggregate catalogue (bigger counts plus the driver Census), but heterogeneous.

Try it, ▶ open TCGA Pan-Cancer at BRAF

Germline diagnosis

Germline variants

inherited disease: gene-level validity and population frequency

GenCC: the Gene Curation Coalition

GenCC track in the Genome Browser — The GenCC track: gene-disease links coloured by evidence strength.

Aggregates gene-disease validity from many expert groups.
Germline / Mendelian focus: genes to inherited diseases, not specific variants or somatic mutations.
Colour = strength of the link (definitive, strong, moderate, limited); spans the whole gene, mouse-over for the curated detail.

Try it, ▶ open the GenCC BRAF session The gene-disease links for BRAF, coloured by evidence.

GenCC: click through to the source

Clicking a GenCC feature opens its details, with a link to the GenCC page — Click a feature for its details, then follow the boxed link straight to the GenCC page for the gene.

OMIM: Online Mendelian Inheritance in Man

The classic catalogue of human genes and their phenotypes.
Mostly inherited disease; well-known genes (e.g. BRAF) can list somatic phenotypes too.
Covers whole-gene phenotype links and specific alleles.

OMIM vs GenCC OMIM gives per-gene depth (narrative biology and specific disease alleles); GenCC gives the standardized gene-disease validity tier and is fed partly by OMIM.

Try it, ▶ open OMIM at BRAF

OMIM information in the Genome Browser — OMIM in the Browser: gene-phenotype relationships and known disease alleles.

gnomAD: Genome Aggregation Database

Allele frequencies from large, broadly healthy populations (the “normal” database).
A variant common here is most likely benign.
In cancer, used to filter germline and common variants out of somatic calls.

gnomAD track in the Genome Browser — gnomAD variants across a gene.

gnomAD mouse-over detail — Mouse-over: allele frequency, filter status and ancestry.

Where to next

These databases come together in Tutorial 3: Clinical Examples & Variant Interpretation, where we work full examples end to end.
Help → Interactive Tutorials, or the docs at genome.ucsc.edu/docs.
Email the team: genome@soe.ucsc.edu.

Thank you!

Questions? · genome@soe.ucsc.edu

UCSC Genome Browser · genome.ucsc.edu