Mutational signatures are characteristic combinations of mutation types arising from specific mutagenesis processes such as DNA replication infidelity, exogenous and endogenous genotoxin exposures, defective DNA repair pathways, and DNA enzymatic editing.

The term is used for two distinct concepts, often conflated: mutagen signatures and tumor signatures. Its original use, mutagen signature, referred to a pattern of mutations made in the laboratory by a known mutagen and not made by other mutagens – unique to the mutagen as a human signature is unique to the signer. Uniqueness allows the mutagen to be deduced from a cell's mutations Later, the phrase referred to a pattern of mutations characteristic of a tumor type, although usually not unique to the tumor type nor to a mutagen. If a tumor mutational signature matches a unique mutagen mutational signature, it is valid to deduce the carcinogen exposure or mutagenesis process that occurred in the patient's distant past. Increasingly refined tumor signatures are becoming assignable to mutagen signatures.

Deciphering mutational signatures in cancer provides insight into the biological mechanisms involved in carcinogenesis and normal somatic mutagenesis. Mutational signatures have shown their applicability in cancer treatment and cancer prevention. Advances in the fields of oncogenomics have enabled the development and use of molecularly targeted therapy, but such therapies historically focused on inhibition of oncogenic drivers (e.g. EGFR gain-of-function mutation and EGFR inhibitor treatment in colorectal cancer). More recently, mutational signatures profiling has proven successful in guiding oncological management and use of targeted therapies (e.g. immunotherapy in mismatch repair deficient of diverse cancer types, platinum and PARP inhibitor to exploit synthetic lethality in homologous recombination deficient breast cancer).

The biological mutagenesis mechanisms underlying mutational signatures (e.g. COSMIC Signatures 1 to 30) include, but are not limited to:

Cancer mutational signatures analyses require genomic data from cancer genome sequencing with paired-normal DNA sequencing in order to create the tumor mutation catalog (mutation types and counts) of a specific tumor. Different types of mutations (e.g. single nucleotide variants, indels, structural variants) can be used individually or in combination to model mutational signatures in cancer.

There are six classes of base substitution: C>A, C>G, C>T, T>A, T>C, T>G. The G>T substitution is considered equivalent to the C>A substitution because it is not possible to differentiate on which DNA strand (forward or reverse) the substitution initially occurred. Both the C>A and G>T substitutions are therefore counted as part of the "C>A" class. For the same reason the G>C, G>A, A>T, A>G and A>C mutations are counted as part of the "C>G", "C>T", "T>A", "T>C" and "T>G" classes respectively.

Taking the information from the 5' and 3' adjacent bases (also called flanking base pairs or trinucleotide context) lead to 96 possible mutation types (e.g. A[C>A]A, A[C>A]T, etc.). The mutation catalog of a tumor is created by categorizing each single nucleotide variant (SNV) (synonyms: base-pair substitution or substitution point mutation) in one of the 96 mutation types and counting the total number of substitutions for each of these 96 mutation types (see figure).

Once the mutation catalog (e.g. counts for each of the 96 mutation types) of a tumor is obtained, there are two approaches to decipher the contributions of different mutational signatures to tumor genomic landscape: