Page 93 - Read Online

Papadodima et al. J Transl Genet Genom 2019;3:7. I https://doi.org/10.20517/jtgg.2018.33 Page 3 of 12

technology have enabled the application of massively parallel sequencing, thus dramatically changing our
understanding of the somatic mutation landscape of melanoma. The first catalogue of somatic mutations of
[16]
a cancer genome, at the whole-genome level concerned a melanoma cell line , indicated the presence of a
great number of mutations per Mb and suggested a mutational signature related to UV exposure. Whole-
exome sequencing studies exploiting clinical samples demonstrated that NF1, ARID2, PPP6C, RAC1, SNX31,
TACC1, and STK19 are genes significantly mutated in melanoma [17,18] . The Cancer Genome Atlas Skin
Cutaneous Melanoma (SKCM-TCGA) project confirmed, through exome sequencing, previously reported
melanoma oncogenes and tumour suppressors (BRAF, NRAS, CDKN2A, TP53, and PTEN) and identified
[19]
several additional significantly mutated melanoma genes, namely, MAP2K1, IDH1, RB1, and DDX3X . The
study proposed the classification of CM into four major genomic subtypes, related to the presence of specific
mutations in established driver genes. In particular, the proposed genetic subtypes are the BRAF mutant,
RAS mutant, NF1 mutant, and the triple wild-type. Low-frequency mutations were identified in the triple
wild-type subtype in KIT, CTNNB1, GNA11, and GNAQ. More recently, the first large-scale study exploiting
whole-genome sequencing supported the involvement of the non-coding genome in melanoma pathogenesis
and revealed diverse carcinogenic processes across the different melanoma subtypes. Figure 1 summarizes
the research on melanoma during the last decades, pinpointing key milestones in understanding its
complexity.

MUTATION BURDEN AND SPECIFIC SIGNATURES IN MELANOMA
Sequencing of different cancers has revealed that the melanoma genome shows a substantial prevalence
of somatic mutations [16,20,21] . Particularly, in CM an increased abundance of cytidine to thymidine (C > T)
transitions is observed. This specific alteration is considered characteristic of a UV-light-induced mutational
signature. A recent study, exploiting whole-genome sequencing of cutaneous, acral and mucosal melanomas,
revealed distinct mutation profiles among these melanoma subtypes. The number of base substitutions and
short insertions and/or deletions in CM was generally much higher than of those observed in acral and
mucosal melanomas. In addition, the UV-related C > T transition was not observed in the latter melanoma
subtypes. In contrast, somatic structural rearrangements were more frequent in acral and mucosal
[22]
subtype . These data suggest that different etiologic pathways are involved in the manifestation of diverse
melanoma subtypes.

NEXT-GENERATION SEQUENCING APPROACHES AND BIOINFORMATICS
The significant progress towards the characterisation of the somatic mutational landscape of melanoma,
can be mainly attributed to the rapid evolution of sequencing technologies during the last fifteen years.
Nowadays, NGS has become the state-of-the-art tool in cancer research and is the most common and
advanced technology for de novo somatic mutation detection. NGS technologies are in continuous
development and improvement, both at the level of the applied protocols for library preparation and
sequencing chemistry, but also at the bioinformatics level. A large number of bioinformatics tools have
been developed for general pre-processing and basic analysis of NGS (WES/WGS) data with the aim of
revealing altered variants for the cases under investigation. In this part of our review, we will focus only
on tools developed for somatic mutation calling, bypassing those needed to reach this step of the analysis.
Furthermore, we will discuss most of the available tools for driver-mutation identification, including the
approaches that are used to achieve this step. Discriminating driver from passenger mutation remains a
challenge from the experimental as well as the bioinformatics points of view [17,23-26] . In the case of melanoma,
which is one of the cancers with the highest mutation burdens and heterogeneity, this problem is even more
difficult to address, due to the confounding impact of melanoma’s high mutation rate. The aim of this review
is not to perform a comparison of the tools since more detailed evaluations are available [26-28] .

The basic approach for somatic variance identification is to compare paired samples, i.e., analyse matched
tumour-normal samples collected from the same patient. Most callers are structured after this notion and