ID2 · GRCh37 · COSMIC v93

Mutational profile

The height of each mutational profile bar represents the proportion of one ID mutation type among all ID mutation types in the signature. Although there is no single intuitive and naturally constrained set of ID mutation types (as there arguably are for SBSs and DBSs), an 83 subclass categorisation of ID mutations was designed.

The 83 ID classification incorporates the prior knowledge that IDs commonly have sizes of 1-10 bps, that both insertions and deletions exist, that IDs of C and T occur at different rates, that IDs preferentially occur at repetitive elements, that the length of the repeat unit may influence the likelihood of an ID occurring, that the number of repeat units in a repeat stretch may influence the likelihood of an ID occurring, that IDs are also fostered in some instances by overlapping sequence microhomologies at the ID boundaries and that different mutational processes may, in principle, be differently influenced by these features. We therefore designed an 83 subclass categorisation of IDs that allows some exploration of all the above possibilities, while constraining the number of categories in order to accommodate the relatively small numbers of IDs (compared to substitutions) found in most genomes. This classification categorises IDs of lengths from 1bp to >5bp, for 1bp IDs classifies them as T or C and the number of single base repeats they occur in from 0 to >5, categorises lengths of non-single base repeat units from 2bp to >5bp and the number of repeats from 1 to >5 and size of microhomology from 0bp to >5bp. We recognise that different classifications of IDs may be preferred by others. The ID mutation types are enumerated in the following Excel document.


Genome: GRCh37

Proposed aetiology

Slippage during DNA replication of the template DNA strand. This signature is found in almost all samples, however, substantial number of mutations of this signature are found in cancers with DNA mismatch repair deficiency.


ID2 is found in most cancer samples. The ID2 mutation burden is, in some series, correlated with the age of cancer diagnosis in non-hypermutated samples and this clock-like behaviour suggests that ID2 mutations may accumulate in normal cells. This mutational signature also tends to be highly elevated in cancer samples with defective DNA mismatch repair and microsatellite instability (MSI). These MSI samples generally exhibit one or more of the single base substitution mutational signatures: SBS6, SBS14, SBS15, SBS20, SBS21, SBS26, and/or SBS44.

Acceptance criteria

Summary of the technical and experimental evidence available in the scientific literature regarding the validation of the mutational signature.

Supporting evidence for mutational signature validity

Validated evidence for real signature
Unclear evidence for real signature
Evidence for artefact signature
Background Identification study First included in COSMIC
Alexandrov et al. 2020 Nature v3
Identification NGS technique Different variant callers Multiple sequencing centres
WES & WGS Yes Yes
Technical validation Validated in orthogonal techniques Replicated in additional studies Extended context enrichment
Yes Yes -
Proposed aetiology Mutational process Support
Slippage of template strand during DNA replication Statistical association
Experimental validation Experimental study Species
- -

Tissue distribution

Numbers of mutations per megabase attributed to the mutational signature across the cancer types in which the signature was found. Each dot represents an individual sample and only samples where the signature is found are shown. The number of mutations per megabase was calculated by assuming that an average whole-exome has 30 Mb with sufficient coverage, whereas an average whole-genome has 2,800 Mb with sufficient coverage.

The numbers below the dots for each cancer type indicate the number of high confidence tumours in which the signature was attributed (above the blue horizontal line) and the total number of high confidence tumours analysed (below the blue horizontal line). Only high confidence data are displayed: samples with reconstruction accuracy >0.90.


Associated signatures

Associated with SBS1 in non-hypermutated samples.