Somatic variant calling
Tumor-normal and tumor-only somatic SNV and indel calling with panel-of-normals filtering, tuned for low allele fractions.
Research Area
Oncology & Oncogenomics
Cancer is a genomic disease, and its data is uniquely demanding—somatic signal buried in noise, tumor heterogeneity, and subclonal evolution. We bring the computational depth to make sense of it: somatic variants, copy-number, signatures, ctDNA, and single-cell heterogeneity, for research and clinical research.
WGS · WES · panels
Bulk & single-cell RNA
ctDNA / liquid biopsy
Research-use · not a diagnosis
Illustrative sample output
missense
truncating
copy-number
Overview
Deep computational oncology, from tumor genome to biology
Cancer genomes break the assumptions ordinary pipelines make. Somatic mutations sit at low allele fractions against a normal background; tumors are mixtures of evolving subclones; copy-number and structural rearrangements reshape whole chromosomes; and purity and ploidy confound every measurement. Getting real biology out of this needs methods built for cancer—and the judgement to know when a signal is a driver, an artifact, or a passenger.
We bring that depth to your cancer data. From tumor-normal somatic calling and copy-number profiling to mutational signatures, ctDNA, and single-cell tumor heterogeneity, we apply established, peer-reviewed methods against the field's reference cohorts and knowledge bases—and return results documented for reproducibility. This is research and clinical-research work, not a diagnostic service: clinical decisions stay with treating oncologists and accredited laboratories.
Applications
What we analyse in cancer
The computational analyses that turn tumor sequencing into biology—from somatic calling to single-cell heterogeneity and immunogenomics.
Copy-number & structural
Allele-specific copy-number, purity and ploidy estimation, and structural-variant and fusion detection across the tumor genome.
TMB, MSI & signatures
Tumor mutational burden, microsatellite-instability status, and SBS/ID/DBS mutational-signature deconvolution against COSMIC.
Heterogeneity & evolution
Subclonal reconstruction and clonal-evolution analysis to map tumor heterogeneity and how it changes over time or treatment.
ctDNA & liquid biopsy
Error-corrected, UMI-aware detection of low-frequency variants for tumor profiling and minimal-residual-disease research.
Single-cell & microenvironment
Single-cell RNA analysis for tumor and immune-cell states, plus deconvolution of the tumor microenvironment from bulk data.
Immunogenomics & neoantigens
HLA typing, neoantigen prediction, and TCR/BCR repertoire analysis to study tumor–immune interactions and immunotherapy response.
Driver discovery & pathways
Significantly mutated genes, driver detection, and pathway-level analysis to move from a mutation list to cancer biology.
The Pipeline
A representative oncogenomics workflow
A transparent, cancer-aware sequence from tumor reads to interpreted biology—each step tuned for somatic signal and documented for reproducibility.
Adapted to your study: tumor-normal vs. tumor-only, WGS vs. panel vs. ctDNA, DNA-only vs. paired RNA and single-cell. We confirm the plan with you before any compute begins.
Intake & QC
Tumor and matched-normal raw reads are quality-checked—coverage, contamination, and tumor-in-normal—before analysis.
Tools: FastQC · MultiQC · VerifyBamID
Alignment & preprocessing
Reads are aligned and prepared—duplicate marking, base recalibration, and, for ctDNA, UMI consensus.
Tools: BWA-MEM2 · GATK · fgbio
Somatic calling
SNVs and indels are called against the normal (or a panel-of-normals), then filtered for artifacts and germline leakage.
Tools: Mutect2 · Strelka2 · PoN
Copy-number, SV & purity
Allele-specific copy-number, purity and ploidy, and structural variants and fusions across the tumor genome.
Tools: FACETS · Manta · GRIDSS
Annotation & drivers
Variants are annotated and linked to cancer knowledge bases; significantly mutated genes and drivers are identified.
Tools: VEP · OncoKB · dNdScv
Signatures, TMB & MSI
Mutational signatures are deconvolved, and tumor mutational burden and microsatellite-instability status are computed.
Tools: SigProfiler · MSIsensor
Integration & reporting
Results become an oncoprint, signature and copy-number figures, tables, and reproducible methods with every tool version.
Tools: maftools · ComplexHeatmap · versioned manifest
Tools & Technologies
Established, peer-reviewed tools—matched to your data
We select from the field's standard toolkit rather than forcing every dataset through one pipeline. A representative set of what we work with:
Somatic calling
Mutect2
Strelka2
VarScan2
Lancet
SAGE
Copy-number & structural
FACETS
CNVkit
GATK CNV
Manta
GRIDSS
Signatures, TMB & MSI
SigProfiler
deconstructSigs
MSIsensor
TMB tools
Single-cell & deconvolution
Seurat
Scanpy
CIBERSORTx
inferCNV
Immunogenomics
OptiType
pVACtools
NetMHCpan
MiXCR
Heterogeneity & evolution
PyClone-VI
SciClone
PhyloWGS
fishplot
Cancer knowledge bases
OncoKB
CIViC
COSMIC
cBioPortal
CGI
Visualization & reporting
maftools
ComplexHeatmap
GenVisR
IGV
Key Databases
The cancer-genomics resources we draw on
Cancer analysis is powered by the field's reference cohorts and knowledge bases. We build on the community's authoritative, versioned resources.
- TCGA
- The Cancer Genome Atlas—multi-omic profiles across cancer types.
- ICGC / PCAWG
- International and pan-cancer whole-genome reference cohorts.
- COSMIC
- Catalogue of somatic mutations and mutational-signature references.
- OncoKB
- Precision-oncology knowledge base linking variants to therapies.
- CIViC
- Open, expert-curated clinical interpretations of cancer variants.
- cBioPortal
- Interactive exploration and visualization of cancer-genomics data.
- GDC
- NCI Genomic Data Commons—harmonised cancer sequencing data.
- DepMap
- Cancer dependency maps linking genotype to vulnerabilities.
- gnomAD
- Population allele frequencies for germline filtering in tumor-only work.
Choosing an Approach
Tumor-normal vs. tumor-only vs. ctDNA
Different sample setups give different somatic confidence. A quick orientation; we will help you match it to your study.
| Dimension | Tumor-normal | Tumor-only | ctDNA / liquid biopsy |
|---|---|---|---|
| Input | Tumor + matched normal | Tumor tissue only | Cell-free DNA from blood |
| Somatic confidence | Highest—germline subtracted | Good, with PoN & filters | High with UMI error-correction |
| Allele fractions | Clonal & subclonal | Clonal & subclonal | Very low (down to ~0.1%) |
| Watch for | Needs a normal sample | Germline leakage | Low input; sampling |
| Best suited to | Definitive somatic profiling | Archival / no-normal cohorts | Monitoring & MRD research |
What You Receive
A complete, documented deliverable
Not just a VCF dropped in a folder—a coherent picture of the tumor, documented so it reproduces and stands up to review.
- Annotated somatic variant tables (MAF) with driver flags
- Copy-number, purity, ploidy & structural-variant calls
- TMB, MSI status & mutational-signature contributions
- Cohort oncoprint & publication-quality figures
- Single-cell / microenvironment results, where applicable
- Immunogenomics outputs (HLA, neoantigens), where applicable
- Reproducible methods with every tool & reference version
Built for reproducibility, not just a result
Cancer analysis follows documented best practices—matched-normal subtraction or panel-of-normals filtering, artifact removal, purity and ploidy correction, and reference cohorts and knowledge-base versions all pinned—so a called driver is real signal, not a mapping artifact or a germline variant in disguise. This is research and clinical-research work that your team interprets; it is not a clinical diagnostic service and not a medical diagnosis, and treatment decisions and clinical reporting remain with treating oncologists and accredited laboratories.
The practical payoff: your methods section writes itself, a reviewer can re-run the analysis, and a result from today can be reproduced a year from now. We will also tell you honestly when a design or sample size won't support the conclusion you're after.
FAQ
Oncology & oncogenomics questions
What cancer researchers and oncology teams most often ask before starting.
Tumor and matched-normal whole-genome, whole-exome, and targeted panels; bulk and single-cell RNA-seq; and ctDNA/liquid biopsy — starting from raw reads or aligned data from your core or a public cohort like TCGA.
Matched tumor-normal gives the cleanest somatic calls, but we also run tumor-only analysis with a panel-of-normals and population filtering when a normal isn't available — and we're explicit about the trade-offs.
Yes. We calculate tumor mutational burden and microsatellite-instability status, and deconvolve SBS, ID, and DBS mutational signatures against COSMIC references to characterise the mutational processes at work.
Yes. We run error-corrected, UMI-aware ctDNA workflows for low-frequency variants, and single-cell analyses for tumor heterogeneity, clonal structure, and the tumor microenvironment.
No. We provide research and clinical-research bioinformatics — analysis your team interprets. We are not a clinical diagnostic laboratory and results are not a medical diagnosis; clinical decisions and reporting stay with treating oncologists and accredited labs.
Oncology is a research area we support through our core services — chiefly Genomics & Variant Analysis, Biomarker & Variant Interpretation, and Multi-Omics Integration — applied to cancer data. This page shows how they come together for oncology.
Related
Services that power oncology work
Have cancer data to make sense of?
Tell us your organism, data type, and question—we'll scope it honestly, including if a different design would serve you better.