- Home
- Atlas
Oncotics Atlas
Oncotics Atlas is a map of oncology you can search. Look up a cancer, treatment, target or term for a plain-English summary with its primary sources and connections. Below it, find the right database for any question about a cancer variant.
Find the right variant database.
Interpreting a cancer variant means asking several questions, and each has its own source. Links were checked on 29 October 2026. Resources marked as searched run automatically in the Workspace.
Which resource answers which question?
Is it real, and is it germline?
Check the population frequency first. Common variants are rarely drivers.
Has anyone classified it?
For inherited risk, see submitted classifications and how they have changed.
Is it a known cancer hotspot?
See how often the change is found across tumours and cancer types.
Is it actionable?
Look for therapy, diagnostic and prognostic evidence with levels and citations. Compare at least two.
Which drugs touch this gene?
Drug-gene interactions and pharmacogenomic effects on response or toxicity.
Nobody has seen it. Could it matter?
Use predictors as supporting evidence only, never as proof.
I have a whole VCF or report.
Batch annotation and reporting instead of one-by-one lookups.
The directory
Clinical knowledge bases 11
OncoKB (opens in a new tab)
Memorial Sloan Kettering's precision oncology knowledge base. Assigns each somatic alteration a therapeutic level of evidence, from FDA-recognised biomarker to biological evidence.
CIViC (opens in a new tab)
Open, community-curated Clinical Interpretation of Variants in Cancer. Evidence items carry a type, direction, level and source citation, free to reuse (CC0).
Formerly civic.genome.wustl.edu.
Cancer Genome Interpreter (opens in a new tab)
Annotates a tumour's alterations, separates likely drivers from passengers and flags those with a known or predicted drug response.
JAX Clinical Knowledgebase (CKB) (opens in a new tab)
Curated gene and variant annotations with therapy, diagnostic and prognostic evidence and matching oncology trials.
Now hosted by Genomenon (formerly ckb.jax.org).
Molecular Oncology Almanac (opens in a new tab)
Dana-Farber's catalogue of alteration-to-action relationships from clinical, preclinical and inferential studies, with an open database on GitHub.
My Cancer Genome (opens in a new tab)
Clinician-oriented summaries of which alterations drive which cancers, their therapeutic implications and relevant trials. Some uses may require a licence.
Did not respond to our check.
Precision Medicine Knowledgebase (PMKB) (opens in a new tab)
Structured clinical interpretations of cancer variants, with community submissions reviewed by pathologists at Weill Cornell.
Did not respond to our check.
CanDL (opens in a new tab)
Expert-curated actionable driver mutations, built to help molecular pathologists annotate tumour sequencing results from the literature.
DoCM (Database of Curated Mutations) (opens in a new tab)
Highly curated list of known disease-causing mutations with direct links to source citations.
The website now redirects to an archived GitHub repository.
DGIdb (opens in a new tab)
Drug-gene interactions and the druggable genome, aggregated from dozens of sources. Useful for asking which drugs touch a mutated gene.
Formerly dgidb.genome.wustl.edu.
ClinPGx (formerly PharmGKB) (opens in a new tab)
How genetic variation changes drug response: curated pharmacogenomic annotations, dosing guidelines and drug labels.
pharmgkb.org now redirects here.
Somatic catalogues 2
COSMIC (opens in a new tab)
The Catalogue Of Somatic Mutations In Cancer from the Wellcome Sanger Institute: how often a mutation is seen across tumours, plus the Cancer Gene Census. Free for academic use; registration needed for downloads.
ISB Cancer Genomics Cloud (opens in a new tab)
TCGA and other NCI cohorts as query-ready cloud tables (Google BigQuery), exportable as CSV or JSON.
Germline and population 12
gnomAD (opens in a new tab)
The Genome Aggregation Database: allele frequencies from hundreds of thousands of exomes and genomes. The first stop for asking whether a variant is too common to be pathogenic.
dbSNP (opens in a new tab)
NCBI's reference catalogue of short human variants and the source of rs identifiers.
ClinVar (opens in a new tab)
NCBI's public archive of submitted variant classifications (benign to pathogenic) with the evidence and review status behind them.
Pushbuild ClinVar reclassification feed (opens in a new tab)
Month-to-month ClinVar classification changes since 2025, filterable by gene and direction (for example VUS to likely pathogenic). Commercial service with a free trial.
1000 Genomes / IGSR (opens in a new tab)
Population reference genomes from the 1000 Genomes Project, now maintained by the International Genome Sample Resource.
Formerly 1000genomes.org.
Kaviar (opens in a new tab)
A compilation of known human SNVs, indels and complex variants for testing whether an observed variant is novel and how frequent it is.
ExAC (opens in a new tab)
The Exome Aggregation Consortium browser, the predecessor of gnomAD.
Superseded by gnomAD; use gnomAD instead.
Exome Sequencing Project (ESP) (opens in a new tab)
NHLBI exome variant frequencies from about 6,500 individuals; historic, largely absorbed by gnomAD.
Did not respond to our check.
SweGen (opens in a new tab)
Variant frequencies from a cross-section of the Swedish population.
Genome of the Netherlands (opens in a new tab)
Whole-genome variation in Dutch families and population.
UK10K (opens in a new tab)
Rare-variant project combining 10,000 UK cohort and disease genomes and exomes; project archive.
GEUVADIS (opens in a new tab)
Genetic European Variation in Health and Disease: RNA sequencing of 1000 Genomes samples for linking variants to expression.
Aggregators 2
VarSome (opens in a new tab)
Search engine and aggregator for human variants that pulls many public databases into one page, with automated ACMG classification. Some features require an account.
MyVariant.info (opens in a new tab)
A fast, free API that returns one merged JSON document per variant from dbSNP, ClinVar, gnomAD, dbNSFP, CADD, COSMIC and more.
Annotation software 3
PCGR (Personal Cancer Genome Reporter) (opens in a new tab)
Open-source pipeline that annotates a tumour's SNVs, indels and copy-number changes (Ensembl VEP plus oncology sources) and writes an interactive clinical-style report.
UNMIRI NGS interpretation schema (opens in a new tab)
Open JSON schema (Apache-2.0) and sandbox API that normalise commercial NGS reports from different vendors into one FHIR-Genomics-aligned shape.
Kipoi (opens in a new tab)
Repository and API of ready-to-use trained genomics models, mostly for gene-regulation and splicing effects.
Effect predictors 20
AlphaMissense (opens in a new tab)
Google DeepMind's AlphaFold-based predictions of pathogenicity for every possible human missense change; also available through the AlphaMissenseR Bioconductor package.
CADD (opens in a new tab)
Combined Annotation Dependent Depletion: one deleteriousness score for any SNV or indel, genome-wide.
Also served at cadd.gs.washington.edu.
dbNSFP (opens in a new tab)
A database of precomputed functional predictions and conservation scores for every possible non-synonymous SNV, from dozens of tools in one table.
Formerly a Google Sites page.
MutationTaster (opens in a new tab)
Predicts whether a DNA change is disease-causing, now as MutationTaster2021.
PolyPhen-2 (opens in a new tab)
Predicts the effect of an amino-acid substitution on protein structure and function.
Did not respond to our check.
SIFT (opens in a new tab)
Predicts whether an amino-acid substitution is tolerated, based on sequence conservation.
Did not respond to our check (the original sift.jcvi.org is offline).
Mutation Assessor (opens in a new tab)
Functional impact of missense changes based on evolutionary conservation in protein families.
Did not respond to our check.
PROVEAN (opens in a new tab)
Predicts the effect of substitutions and indels on protein function.
Did not respond to our check.
CHASM / CRAVAT (opens in a new tab)
CHASM scores whether a missense mutation is a likely cancer driver; it now runs inside the CRAVAT / OpenCRAVAT annotation platform.
The CHASM wiki is retired.
CanDrA (opens in a new tab)
Cancer-specific driver missense prediction using 96 structural, evolutionary and gene features.
New MD Anderson address.
CAROL (opens in a new tab)
Combines SIFT and PolyPhen-2 into one weighted score for coding variants.
New Sanger address.
MutPred2 (opens in a new tab)
Predicts pathogenicity and the molecular mechanism a missense change may disrupt.
SNPs&GO (opens in a new tab)
Predicts disease-related single-point protein variants using Gene Ontology context.
SNPs3D (opens in a new tab)
Structure- and sequence-based assessment of non-synonymous SNPs.
Panther cSNP scoring (opens in a new tab)
Evolutionary scoring of coding SNPs from PANTHER protein-family alignments.
FannsDB / Condel / TransFIC (opens in a new tab)
Precomputed scores from several predictors, including the Condel consensus and TransFIC.
Did not respond to our check.
PhyloP and PhastCons (opens in a new tab)
Conservation scores across species alignments, often used as a feature by other predictors.
Did not respond to our check.
GERP++ (opens in a new tab)
Constrained-element and rejected-substitution conservation scores.
Did not respond to our check.
SNAP2 (opens in a new tab)
Neural-network prediction of functional effects of sequence variants.
Old address returns not found.
nsSNPAnalyzer, AUTO-MUTE, stSNP, CoVEC (opens in a new tab)
Older academic predictors kept in the source list for completeness; their sites did not respond to our check.
See the source list for original links.
Where does the data come from?
From OnCo (onco.cc) (opens in a new tab), an open oncology knowledge graph licensed CC BY-NC 4.0. Oncotics hosts a compact search index; a record’s full detail loads from onco.cc’s public API when you open it.
How is it different from the Workspace?
The Atlas explains concepts and how they connect, as curated summaries. The Workspace queries live public evidence for a specific gene, variant, drug or trial. Orient yourself in the Atlas, then check current evidence in the Workspace.
Does the Atlas store what I search?
No. Searching runs in your browser. Opening a record sends only its identifier to onco.cc to fetch the detail. Oncotics saves nothing; your place is kept in the address bar so you can share a link.
Is it medical advice?
No. It is an educational reference built on a work-in-progress knowledge graph. Verify anything that matters at the primary source and with qualified clinicians.
Which database tells me if a cancer mutation is actionable?
Start with a clinical knowledge base: OncoKB assigns therapeutic levels of evidence, CIViC lists community-curated evidence items with citations, and the Cancer Genome Interpreter, JAX CKB and the Molecular Oncology Almanac add drug-response and trial context. Check more than one.
How do I know if a variant is a common germline polymorphism?
Look up its population frequency in gnomAD, and dbSNP for the rs identifier. A variant seen often in healthy people is unlikely to be a cancer driver or a high-penetrance pathogenic variant.
What should I use for a variant no database has seen?
Effect predictors such as AlphaMissense, CADD, PolyPhen-2, SIFT or the dbNSFP collection, and cancer-specific driver predictors such as CHASM or CanDrA. Treat predictions as supporting evidence only.
What do the "Moved" and "Check link" labels mean?
We checked every link on 29 October 2026. "Moved" means the resource has a new address, which is the one shown. "Check link" means the site did not respond to our check; it may be temporarily down, blocked from our network or retired.
Data from OnCo (onco.cc)
The knowledge graph is written and maintained by OnCo (opens in a new tab) and licensed CC BY-NC 4.0 (opens in a new tab): free for individual and educational use with attribution; commercial use needs a licence from OnCo. OnCo’s code is MIT-licensed (github.com/judegomila/OnCo (opens in a new tab)).
Changes by Oncotics: a new explorer interface, a trimmed search index (summaries shortened to about 260 characters) and links into the Workspace. The index file is shared under the same licence.
Built on a community list
The resource list is adapted from awesome-cancer-variant-resources (opens in a new tab) by Sean Davis and contributors (MIT licence), described in Frontiers in Oncology (2017) (opens in a new tab).
Oncotics rewrote the descriptions, grouped resources by question, updated moved addresses and marked Workspace coverage.
Not affiliated
Oncotics is not affiliated with or endorsed by OnCo or the databases listed. Each has its own terms, and some are free only for academic use. OnCo describes its content as a work in progress, checked against primary sources on the date shown on each record.
Educational use only. Not medical advice and not clinical guidance.