If a gene panel is configured, PURPLE will compile a catalog of drivers including point mutations and copy number events. To include point mutations in the driver catalog, the somatic VCF must be annotated with PAVE.
Running LINX enriches the PURPLE driver catalog with the addition of disruption drivers. Fusions are also called but reported separately.
The gene panel is a configuration file of which genes to add to the driver catalog and which mutation types to report in the canonical transcript and any 'additionalReportedTranscripts' specified for each gene. The format is:
| Field | Values | Description |
|---|---|---|
| Gene | HGNC symbol of gene | |
| Report Missense | T/F | Report if any missense variant is found in the gene |
| Report Nonsense | T/F | Report if any nonsense or frameshift variant is found in the gene |
| Report Splice | T/F | Report if any canonical splice acceptor or donor variant is found in the gene [+1,+2,+5,-1,-2] Mutations affecting the last exonic base at a donor location as well as N>G variants only at the -3 acceptor base are also treated as SPLICE. |
| Report Amplification | T/F | Report amplification if min gene copy number > 3x sample ploidy and partial amplification if max gene copy number > 3x sample ploidy (for genes on chrX for males 1.5x sample ploidy is used). (Note if qcstatus = HIGH_CN_WARN_HIGH_COPY_NUMBER_NOISE, AMPS must be bounded on at least one side by an SV). |
| Report Deletion | T/F | Report if gene copy number < 0.5 (Note - If qcStatus in {WARN_DELETED_GENES,WARN_HIGH_COPY_NUMBER_NOISE} deletions must also be supported on both sides by SV OR (supported by SV + CENTROMERE/TELOMERE and be <10M bases) |
| Report Disruption | T/F | LINX will report ‘HOM DISRUPTION' where the exonic copy number of the gene is > 0.5 but where no intact copies of the gene are predicted to remain), or simply DISRUPTION for other structural variants which may disrupt the structure of the canonical transcript. |
| Report Hotspot | T/F | Report somatic hotspot mutation regardless of other rules |
| Likelihood Type | ONCO/TSG | Calculate driver likelihood as a tumor suppressor gene or onco gene |
| reportGermlineVariant | 'WILDTYPE_LOST','NONE', 'ANY','VARIANT_NOT_LOST' | Report any germline variants that meet pathogenic criteria based on specified tumor status |
| reportGermlineHotspot | 'WILDTYPE_LOST','NONE', 'ANY','VARIANT_NOT_LOST' | Report hotspot germline pathogenic variants based on specified tumor status |
| reportGermlineDisruption | 'WILDTYPE_LOST','NONE', 'ANY','VARIANT_NOT_LOST' | Report germline structural variant gene disruptions |
| reportGermlineDeletion | 'WILDTYPE_LOST','NONE', 'ANY','VARIANT_NOT_LOST' | Report germline gene deletions |
| additionalReportedTranscript | Ensembl transcripts to report on in addition to the ensembl canonical transcript (eg. CDKN2Ap14Arf). Any drivers on the additional transcript will be added to the driver catalog table as a separate record with canonicalTranscript=0. Copy number data for that transcript will also be added to the geneCopyNumber table and any somatic or germline point mutations will report the effect of the variant on the additional transcript in the otherTranscriptEffects column in somaticVariant and germlineVariant tables. Any breakend impacting the additional transcript will also be reported in the svBreakend table |
Up to 3 individual driver catalog records may be added per gene per sample if more than 1 type of event is present:
- Germline mutations (GERMLINE_MUTATION,GERMLINE_DELETION,GERMLINE_DISRUPTION)
- Copy number or disruption events (DEL,AMP,HOM_DISRUPTION)
- Somatic point mutations (MUTATION).
The Hartwig Medical Foundation curated gene panel is available from HMFTools-Resources > Gene Panel and is updated periodically. A detailed description of our gene discovery and initial construction of our gene panel is available in the supplementary information of our "Pan-cancer whole genome analyses of metastatic solid tumors" paper.
A driver likelihood estimate between 0 and 1 is calculated for each variant in the gene panel. High level amplifications, homozygous deletions, homozygous disruptions and TERT promoter mutations are all rare so have a likelihood of 1 when found affecting a driver gene. Note that non homozygous disruptions are also added to the driver catalog, but driver likelihood is set to 0.
For coding mutations we need to account for the large number of passenger point mutations that are present throughout the genome and thus also in driver genes. This is implemented as a model where the driver likelihood algorithm depends on the configured 'likelihood type' in the driver gene panel. The impacts of configuring ONCO vs TSG type is the following:
| Characteristic | ONCO |
TSG |
|---|---|---|
| High likelihood variants | Hotspot & Inframe (excluding repeat count > 7) assigned likelihood = 1 | Hotspot & Biallelic splice, indel and nonsense assigned likelihood = 1 |
| Biallelic mutations | Ignored in dnds calculations | For biallelic mutations, biallelic TMB only used in passenger likelihood. For non biallelic, the full TMB is used |
| Multi-hit | Maximum likelihood used. Highest ranked variant used only in dnds calculation (missense & inframe ranked first) | Multiple mutations are additive (product of probabilities used). For non biallelic variants highest 2 ranked variants used in dnds calculations (nonsense & splice ranked first) |
| Non-biallelic frameshift/Splice/Nonsense | No special treatment | driverLikelihood=max(selfLikelihood,missenseDriverLikelihood) |
For point mutations, we first determine a list of variants that are highly likely to be drivers and/or highly unlikely to have occurred as passengers as driver likelihood of 1 as per the table above. For the remaining point mutation variants these are only assigned a > 0 driver likelihood where there is a remaining excess of unallocated drivers based on the calculated dNdS rates in that gene across the cohort after applying the above rules. Any remaining point mutations are assigned a driver likelihood between 0 and 1 using a bayesian statistic to calculate a sample specific likelihood of each gene based on the type of variant observed (missense, nonsense, splice or INDEL) and taking into account the mutational load of the sample.
The principle behind the likelihood method is that the likelihood of a passenger variant occurring in a particular sample should be approximately proportional to the tumor mutational burden and hence variants in samples with lower mutational burden are more likely to be drivers.
The sample specific likelihood of a residual excess variant being a driver is estimated for each gene using the following formula:
P(Driver|Variant) = P(Driver) / (P(Driver) + P(Variant|Non-Driver) * (1-P(Driver)))
where P(Driver) in a given gene is assumed to be equal across all samples in the cohort, ie:
P(Driver) = (residual unallocated drivers in gene) / # of samples in cohort
And P(Variant|Non-Driver), the probability of observing n or more passenger variants of a particular variant type in a sample in a given gene, is assumed to vary according to tumor mutational burden, and is modelled as a poisson process:
P(Variant|Non-Driver) = 1 - poisson(λ = TMB(Sample) / TMB(Cohort) * (# of passenger variants in cohort),k=n-1)
Note that the TMB is calculated separately for INDELs and SNVs. For TSG dnds rates, we find that biallelic dnds rates are typically much higher than non-biallelic rates reflecting the biology that both alleles are normally required to be knocked out to cause a TSG driver. To accommodate for this, we use count biallelic TMB in the passenger likelihood for biallelic TSG VUS.
- Clustered variants - Pathogenic variants are likely to occur clustered in specific locations of genes that are functionally important. This should be reflected in driver likelihood
- Cancer type specificity - DNDS is calculated pan-cancer due to size of cohorts. Ideally this would be calculated cancer type specific, especially for large cohorts.
- INFRAME INDELS in repeats with 3 base microhomology - Generally inframe indels are extremely rare and are likely to be drivers and hence are given driver likelihood = 1. An exception should be made for 3 base indels found in microsatellites with a repeat length of 3 as this is a common passenger mutation, particularly in MSI samples.
- AMPLIFICATION PASSENGERS - All amplifications are given a likelihood of 1. Since nearby genes are amplified together, some of the amplification drivers may well be passengers, commonly amplified with nearby drivers.
- FUNCTIONAL ANNOTATION - may improve driver likelihood annotation