2. Next-Generation Sequencing
2.1. First-Generation Sequencing
First-generation sequencing methods include the basic techniques that preceded the introduction of NGS. The most common methods are Sanger sequencing via synthesis (SBS) and Maxam–Gilbert chemical cleavage (MGC).
MGC is based on the specific chemical modification of DNA and the subsequent cleavage of the DNA backbone at the modified nucleotides. The obtained fragments are then separated in a gel according to their size to obtain a ladder of DNA fragments with a known nucleotide end 
In SBS, DNA or RNA sequencing is achieved by chain-terminating dideoxy nucleotides (ddNTP). The absence of 3′- OH interrupts the DNA chain during polymerization; therefore, phosphodiester bonds cannot be formed. Each ddNTP is labeled with a specific fluorophore to allow for the detection of DNA fragments of different sizes using capillary electrophoresis 
. Despite the efficiency of NGS techniques, SBS is still widely used for applications where a high throughput is not required. Its most common applications range from the validation of plasmid constructs or PCR products to testing enzyme action on DNA molecules 
2.2. Second-Generation Sequencing
The first properly named NGS techniques belong to this group. The development of these sequencing methods triggered the need for sequencing large genomes with a higher and cheaper throughput 
. Depending on the working principle, two categories of second-generation sequencing methods can be distinguished:
Next generation sequencing via hybridization (NGSH): this is based on a microarray of oligonucleotides with known sequences and positions and their hybridization with a fluorophore-labeled target DNA 
. Sequencing via hybridization is commonly proposed in routine clinical practice for the identification of disease-related SNPs (single-nucleotide polymorphisms) or chromosomal abnormalities (such as deletions and amplifications) 
Next-generation sequencing via synthesis (NGSS): this represents the evolution of Sanger sequencing via synthesis. These methods can only be applied to short DNA fragments (usually 300–500 nucleotides) and are characterized by a considerable error rate, but their high data throughput allows for comprehensive genome coverage and avoids the effects of the error rate 
NGSS techniques are based on different sequencing approaches, usually based on the individual isolation of DNA molecules in millions of chambers, wells, or specific spots on a chip. Among the many systems, the most common (schematically shown in Figure 1) are:
Figure 1. Schematic workflow of second-generation sequencing approaches.
454 Pyrosequencing: in this approach, each DNA fragment of 600–800 nucleotides is conjugated with an oligonucleotide adaptor complementary to a DNA sequence on a bead. Ideally, each bead can bind a single DNA fragment using specific adaptors. The DNA fragments on the beads are then amplified with an emulsion PCR (emPCR) and each bead is transferred to a picoliter-sized chamber. Each chamber is flooded alternately with one of the four nucleotides. When the correct nucleotide is incorporated, a pyrophosphate molecule is released and recorded using a light-generating reaction. Although this approach is rarely used, it represents an efficient choice for whole genome sequencing and metagenome analyses, due to the length of the analyzed DNA fragments 
Ion Torrent: this approach is quite similar to 454 pyrosequencing but differs in the method used to detect the correct insertion of nucleotides. When a nucleotide is inserted into a growing chain, it releases a hydrogen ion that changes the pH of the solution, which can be recorded as a voltage change by an ion sensor. This detection system is faster than that of 454 pyrosequencing because it does not require a camera or light source 
. For this reason, the ion torrent method is widely used for various applications, such as de novo sequencing or DNA alteration detection.
Illumina technology: this is based on an innovative amplification method called “bridge PCR”. In this method, each DNA fragment of about 500 nucleotides is functionalized at both ends with two oligonucleotide adaptors complementary to several DNA sequences on a solid support. In this way, the DNA fragment is anchored to the solid support by the two adaptors like a bridge and can serve as a template for amplification in clusters of about 1000 replicates. The clusters are sequenced by chemically reversible chain-terminating nucleotides labeled with a fluorophore specific for each of the four nucleotides. The insertion of a modified nucleotide results in the temporary blockage of the growing chain, and the specific fluorescence is recorded. After the detection, the modified nucleotide is chemically unlocked and a new cycle of insertion/detection can begin 
. Illumina technology is by far the most widely used NGS technique and supports a variety of applications, including DNA or RNA sequencing, metagenomics, CHIP -seq, and methylome analyses 
2.3. Third-Generation Sequencing
In this generation, innovations in NGS approaches aim to increase the length of the DNA or RNA fragments analyzed. The leading method in this case is Pacbio sequencing, also called single-molecule real-time sequencing (SMRT), an innovative system that can sequence very long fragments (30–50 kilobases). Its workflow begins by functionalizing the DNA or RNA fragments to be sequenced with special adaptors that allow the molecule to circulate. The resulting circular DNA/RNA is then bound to a special polymerase and the complex is placed on the bottom of a zero-mode waveguide (ZMW) that is the size of a zeptoliter. The small size of the ZMW (1 zeptoliter = 1 × 10–21 L) directs the incident light to a small area so that only the bottom of the well can be illuminated and imaged. Each well is then flooded with a mixture of the four nucleotides labeled with a phosphor-bound fluorophore. In this way, the incorporation of a nucleotide into the growing DNA chain detaches the fluorophore, which then floats away from the illuminated region. Imaging synchronized with the nucleotide insertion rate thus allows for the detection of light from the fluorophore only at the time of insertion. In addition, this detection approach can provide information on the presence of various base modifications, such as adenine and cytosine methylation, as the latter alters the nucleotide incorporation rate 
. A scheme of this sequencing method is illustrated in Figure 2
Figure 2. Schematic workflow of Pacbio sequencing approaches. ZMW: zero-mode wave (1 zeptoliter-sized well).
It is essential to emphasize that SMRT sequencing is associated with a considerable error rate. However, its high number of parallel sequencing runs due to the small size of the ZMW can overcome this problem. The applications of SMRT technology are numerous, especially for the detection of base changes, and it is often used in combination with other NGS methods to increase the sequencing accuracy 
. The continuous optimization of the SMRT workflow makes this NGS technique a good choice for the analysis of large genomic samples.
2.4. Fourth-Generation Sequencing
Like the third generation, the sequencing methods of the fourth generation are focused on increasing the analyzed sample length. The fundamental principle of these sequencing approaches is the passage of the DNA or RNA molecule through a hole coupled with a detector system. Theoretically, around 1000 kb can pass through a single hole, which leads to a consistent sequencing throughput 
. According to the materials used, it is possible to distinguish two types of nanopore systems:
Solid-state sensor technology: this is based on metal or metal alloy chips with nanometer-sized pores, through which the DNA/RNA molecules pass 
Biological membrane systems: this approach uses an electrically resistant polymer membrane with nanopores generated by transmembrane proteins, usually the alpha-hemolysin or the Mycobacterium smegmatisporin A (MspA). The long double-strand DNA/cDNA molecule is initially coupled with an enzyme complex made of a highly processive DNA polymerase (phi29), a DNA helicase, and an exonuclease I, which is able to unwind and ratchet the DNA molecule through the pore at a constant rate. As a nucleotide passes through the pore, it generates an electronic signal characteristic of each of the four nucleotides. In this way, it is possible to sequence the analyzed DNA/cDNA sample knowing the oligonucleotide pass rate through the pore. The leader in biological membrane systems is the MinION platform commercialized by Oxford Nanopore Technologies (ONT), with a small handle size valuable device for its portability and low space requirement 
. A scheme of this sequencing method is illustrated in Figure 3
Figure 3. Scheme of working principle of nanopore sequencing approaches.
Like other NGS technologies, fourth-generation sequencing methods are characterized by a high error rate, which is circumvented by their high numbers of parallel reads. Nanopore systems are typically used for sequencing environmental and metagenomic samples; in space stations, for example, they are used to identify bacterial strains. Like SMRT sequencing, these methods can identify base changes by altering the standard electrical signals of the nucleotides passing through the nanopore 
In summary, NGS techniques are a useful tool for deepening the knowledge of genomics. The large pool of sequencing data generated by NGS has contributed to the discovery of new characteristic, diagnostic, and prognostic biomarkers for a variety of diseases and has improved the therapeutic options. Nevertheless, NGS techniques are characterized by a higher error rate and shorter analyzed DNA/RNA fragments than first-generation techniques, which retain their clinical utility, particularly in detecting known molecular aberrations. In addition, the results of NGS methods are more expensive than those of the first generation, which hinders their massive use. However, the advances in NGS techniques in terms of faster, more accurate, and less expensive sequencing, combined with their high throughput, are enabling their wider incorporation into routine clinical practice.
3. NGS Technologies in TNBC Research
TNBC is a rare form of breast cancer with aggressive behavior and a high risk of recurrence, resulting in generally poor outcomes 
. In this perspective, the optimized sequencing technologies of the second generation can help to discover new diagnostic, predictive, and prognostic biomarkers for improving patient survival rates.
TNBC is characterized by a high molecular, histological, and clinical heterogeneity 
. Several efforts have focused on the identification of biomarkers using NGS techniques for TNBC diagnosis stratification.
3.1. NGS Analysis of Markers for TNBC
Twelve studies have analyzed TNBC markers using NGS, as shown in Table 1, most of which used formalin-fixed, paraffin-embedded samples. Four studies were performed in a liquid biopsy using peripheral blood.
Liang et al. investigated the mutation status of 91 breast cancer-related genes in 156 inflammatory breast cancer (IBC) patients (among which 51 were TNBC cases) using the Illumina platform. The sequencing results from the fresh or frozen biopsies were compared to a control group of 191 non-IBC primary breast cancer patients from the TGCA database. The Illumina analysis showed a higher somatic mutation frequency for TP53, NOTCH1/2, MYH9, BRCA1/2, ERBB4, POLE, FGFR3, ROS1, NOTCH4, LAMA2, EGFR, ESR1, THBS1
, and CASp8
, and a lower somatic mutation frequency for CDH1 in the IBC patients in comparison to the non-IBC ones. Upon grouping these genes in cellular pathways, it was possible to observe how the DNA repair, NOTCH, and RAS pathways were more altered in the IBC group than the non-IBC group, paving the way for specific IBC therapies. Moreover, the PIK3CA/AKT pathway was often altered in IBC (especially in TNBC subtypes) and was associated with a poor metastasis-free survival 
An exciting contribution to TNBC molecular profiling was provided by Srour et al., who compared the expression levels of 2567 cancer-related genes in 14 pairs of ER+ primary sites and paired axillary lymph node metastasis (ALN), and in 17 pairs of TNBC primary sites and paired ALN metastasis. In both subgroups, an Illumina analysis revealed multiple genes with different expression levels between the primary and metastatic sites. A comparison between the TNBC and ER+ results showed that 97 common upregulated genes and 115 common downregulated genes were found in the ALN metastasis in comparison to the primary sites 
. Another experiment by Srour et al. on the same patient pool showed an overexpression of anti-apoptosis genes (BIRC3, TCL1A, FLT3, and VCAM1
) and a downregulation of the genes regulating the microenvironment (MMP2, MMP 3, MMP 7, MMP 11, MMP14, COL1A1, COL1A2, COL3A1, COL5A1, COL5A2, COL6A6, COL11A1
, and COL17A1
) in ALN metastasis compared to the TNBC primary sites. Despite the small patient cohort, these results suggested a change in the transcriptome of TNBC-invasive cells that increases their metastatic potential 
Dillon et al. examined the mutation profiles of 20 TNBC patients using an NGS assay called JAX-CTP. This assay was based on a clinically validated panel of SNPs, copy number variations, insertions, and deletions commonly detected in 358 cancer-related genes. The Illumina sequencing of formalin-fixed, paraffin-embedded (FFPE) biopsies revealed MYC
amplification in 75% of the patients examined, while TP53, AURKA
, and KDR
mutations were present in 6 of 20 cases (30%) 
is a transcription factor that regulates approximately 15% of all human genes, including the genes related to cell proliferation and survival. MYC
dysregulation is indeed involved in tumorigenesis in various tissues, inducing epithelial–mesenchymal transition (EMT), angiogenesis, and tumor cell immortalization in the BLIS and BLIA subsets of TNBC 
. These results suggested that MYC
amplification is a molecular signature for basal-like TNBCs, as has been previously reported in the literature 
. The high mutational frequency of AURKA
in the studied cohort of patients was very interesting, as this plays a role in promoting MYC
expression in breast cancer stem cells 
. The AURKA inhibitors alisertib and TAS -119 have provided encouraging results in ongoing studies 
, suggesting AURKA
as an emerging target for the treatment of TNBC. However, the small patient population analyzed by Dillon et al. needs further investigation to confirm these results.
Instead, Li et al. focused on the different histologic groups of TNBC, particularly invasive ductal carcinoma, without special type (NST) and special type (ST). The NST patients were distinguished from those with ST by the absence of special histological features found in a microscopic examination, which classify them into the typical ST subtypes such as medullary, metaplastic, and apocrine carcinomas 
. They performed an NGS survey using the Illumina platform on the plasma of 89 TNBC patients (72 NSTs and 17 STs) to investigate the mutation statuses of 520 cancer-related genes. The sequencing results showed a different mutation frequency between the two histological subtypes: in NST, the most frequently mutated genes were TP53
(26.8%), and MYC
(18.3%), whereas in ST, they were TP53
(18.8%), and KMT2C
(18.8%). Although TP53
are the most frequently mutated genes in both subgroups, significantly lower TP53
and higher PIK3CA
mutation rates are observed in ST in comparison to NST. This finding provided genetic evidence for the partially different molecular mechanisms underlying the tumor growth in these two TNBC subgroups. Moreover, the efficacy of drugs targeting the PIK3CA pathway may be higher for STs than for NSTs 
Beyond diagnostic stratification, the molecular profiling of TNBC with NGS techniques can provide novel biomarkers for investigating the clinical value of genomic alterations. An optimal example is provided by Pop et al., who analyzed the most frequent alterations of 46 genes that play well-defined roles in cancer and their clinical significance in a retrospective study of 96 TNBC patients. Pop et al. performed Ion Torrent sequencing on 30 FFPE tissues from the TNBC cohort and validated the results in all 96 cases. Consistent with other previously described studies, the analysis showed a high mutation frequency for TP53, KDR, PIK3CA, ATM, AKT1, KIT, ERBB4, FGFR3
, and MET
. In particular, the presence of AKT1
rs34231037 (c.1444T > C), KIT
rs3822214 (c.1621A > C), TP53
rs28934576 (c.818G > A), and BRCA1/2
class 5 mutations was associated with a worse survival in the patients studied 
. Of note, these results demonstrated the importance of discovering novel mutations in TNBC for the development of personalized therapeutic treatments.
Its high risk of recurrence is one of the main problems in the clinical management of TNBC. From this point of view, Balko et al. investigated the molecular footprint that determines the chemoresistant response of TNBC tumor cells. Accordingly, they performed a mutational analysis of 182 cancer-related genes in 85 TNBC patients after NAC and in 20 comparable biopsies before treatment with the Illumina platform. The study showed a high mutation frequency in several genes useful for targeted therapies in relapsed cases after NAC: PTEN
(PI3K and AKT inhibitors), JAK2
(ruxolitinib or tofacitinib), CDK6
(CDK4/6 inhibitors), and IGF1R
. This study underlined the importance of routine molecular analyses in follow-ups of TNBC patients in order to improve their outcomes with the appropriate adjuvant drugs. From this point of view, NGS can be a valuable choice, as it is compatible with clinical timing.
Similar to Balko et al., another group attempted to delve into the molecular signature of treatment-resistant TNBC. Lips et al. performed the molecular profiling of 56 TNBC biopsies before NAC using SOLiD 5500xl, an NGS platform based on library amplification using emulsion PCR and sequencing with fluorescent barcodes. The latter are libraries of specific oligonucleotides functionalized with different fluorophores. When an oligo anneals to its complementary sequence on the analyzed DNA/RNA, the fluorophore is enzymatically cleaved and a camera records the colored light emission. The image data are then converted into spatial data using software to sequence the analyzed molecule 
. In this study, the SOLiD system was used to analyze the pathogenic variants of 1977 tumorigenesis-related genes in the patient cohort and their correlation with the treatment responses. For each patient, normal DNA was extracted and sequenced from their plasma samples. Unfortunately, the analysis revealed no significant difference in the mutation rates between the patients who responded to treatment and those who were resistant. However, PIK3CA
mutations were observed exclusively in patients with the BRCA1
. This information may provide new predictive biomarkers for the efficacy of PARPi and drugs targeting the PIK3CA pathway.
Xiangmei et al. focused on the molecular characterization of TNBC patients who had not achieved a pathological complete response (non-pCR) after NAC. Here, the 14 post-NAC TNBC patients analyzed were divided into 7 cases with a short disease-free survival (DFS, within 12 months) and 7 others with a DFS of more than one year. Illumina sequencing was used to examine 422 cancer-related genes in the DNA of FFPE, while matched plasma samples were sequenced to detect germline mutations. The analysis revealed a higher mean number of mutations in the short DFS group than in the long DFS group (6 vs. 4.3), indicating tumor mutation burden (TMB) as a possible marker of recurrence. Moreover, mutations of PTPN13
were found only in patients from the short DFS group. A parallel JARID2
knockout experiment at MDA-MBA -231 (a TNBC cell line) showed a decrease in E-cadherin and an increase in vimentin, MMP7
, and MMP9
expression, suggesting a possible role of JARID2
in epithelial–mesenchymal transition (EMT) 
. These results highlight the prognostic value of TMB and JARID2
and represent a new potential therapeutic target for TNBC. The plasma analysis showed no returned germline mutations among the 72 pathogenic mutations found in both groups. Like Lips et al. 
, in this study, the DNA from the plasma was sequenced using NGS techniques and used as a control group. Liquid biopsies represent a useful clinical tool due to their non-invasive collection, contrary to standard solid biopsies. The faster sampling of peripheral blood and the high throughput of NGS techniques allow for a solid control data group, resulting in a more accurate analysis in a non-invasive way. However, plasma can also be used as a principal tumor DNA source, as Li et al. showed 
. These three studies highlight the potentiality of liquid biopsies for tumor research with NGS techniques.
Heeke et al. contributed to TNBC molecular profiling for predicting responses to treatment. Their study was performed on 4647 breast cancer cases (including 1568 TNBCs) using the NGS-592 Sure-Selected XT, an Illumina-based FFPE-specific assay targeting a pan-carcinoma of 592 genes. Specifically, the mutation statuses of DNA-repair-associated genes (ARID1A, ATM, ATRX, BAP1, BARD1, BLM, BRCA1/2, BRIP1, CHEK1/2, FANCA/C/D2/E/F/G/L, KMT2D, MRE11, NBN, PALB2, RAD50/51/51B
, and WRN
) and chemoresistant-related genes (AKT1, AR, ARID2, ATR, AURKA/B, BCL7A, BCL11A/B, BRAF, CDK4/6, CDKN2A, EGFR, ERBB2, ERBB3, ERBB4, ESR1, IDH1, IDH2, JAK1, JAK2, KIT, MET, MTOR, NTRK1/2/3, PBRM1, PD -L1, PIK3CA, POLE, RB1, RET, SMARCB1, SMARCE1, SMARCA4, SMO,
) were performed. The analysis revealed that 17.9% of all cases had at least one mutation in the genes involved in the recombination mechanisms of DNA repair homologs. These mutations were found in 18.2% of the TNBC cases 
. Moreover, breast cancers with homologs recombination deficiencies (HRD) are characterized by a higher tumor mutational burden, PD-L1 expression, and PIK3CA pathway alteration rate in comparison to ones without HRD. The frequent co-occurrence of DNA repair dysfunctions with other response markers to immunotherapy suggests a new possible combinatorial therapeutic approach for TNBCs with HRD.
Immune checkpoint inhibitors (ICPIs) represent a group of drugs that are increasingly used in the treatment of malignancies 
. In TNBC, several clinical trials of PD-L1-targeted drugs have produced positive results, with atezolizumab being approved by the FDA in 2019 
. The study by Hoda et al. aimed to investigate the frequency of PD-L1 alterations in TNBC and their associated genomic landscape. To this end, 164 FFPE samples from 158 primary and metastatic TNBC patients were analyzed using MSK-IMPACT, an FDA-approved assay from Illumina that detects 468 major cancer-related genes. Immunohistochemistry (IHC) was used to evaluate PD-L1 positivity. The analysis showed that 47.4% of cases had PD-L1 positivity and that the latter was more common in primary TNBC than metastatic. Interestingly, Hoda et al. reported several benefits of anti-PD-L1 therapy in metastatic patients with PD-L1 negativity. According to the authors, IHC results may be influenced by the different methods of collecting primary and metastatic samples (resection and core needle biopsies, respectively). The NGS analysis revealed no significant difference in the genomic assets between the PD-L1-positive and PD-L1-negative TNBCs, except for CBFB
. In fact, mutations in this gene are more frequent in patients without PD-L1 expression than in PD-L1-positive ones 
. It has been reported that CBFB
mutations usually occur together with mutations in the genes associated with the luminal androgen receptor subset (LAR) of TNBC, such as AKT1
and CDH1 
. Finally, the study indicated that CBFB
mutations and PD-L1 negativity may represent novel molecular signatures for LAR TNBCs.
Tan et al. 
investigated the genomic landscape underlying the response to ICPI treatment. To this end, they performed an Illumina assay of 457 cancer-related genes on plasma samples from 11 ICPI-treated TNBC patients. Specifically, they collected and analyzed circulating cell-free tumor DNA (ctDNA) from each patient before the ICPI treatment. Although the number of patients studied was small, this work is interesting because the analysis revealed a shorter progression-free survival after ICPI treatment in patients with deletions of CYP2D6
and gains in CNV of NAS, BCL2L1, H3F3C, LAG3, FGF23, CCND2, SESN1, SNHG16, MYC, HLA-E,
. The study suggested these 12 genes as novel predictive biomarkers of ICPI treatment efficacy in TNBC. Notably, among these genes, GNAS, BCL2L1, LAG3, CCND2, SNHG16, MYC, HLA-E
, and MCL-1
were involved in cell progression, metastasis, apoptosis inhibition, and immune evasion in breast cancer 
Another contribution to the identification of the genes associated with responses to immunotherapy in TNBC was made by Sivapiragasam et al. They performed the comprehensive genomic profiling of 3831 metastatic breast cancers (1237 ER+, 1953 HER2+, and 641 TNBC) using FoundationOne CDx, an Illumina-based assay, to detect genetic alterations in 324 genes, as well as genomic signatures such as microsatellite instability (MSI) and tumor mutation burden (TMB). The analysis showed that 42.6% of the ER + cases, 12.1% of the HER2+ cases, and 56.4% of the TNBC cases had genomic alterations associated with a response to ICPIs. The coupled IHC assay indicated PD-L1 positivity in 13%, 33%, and 47% of the ER+, HER2+, and TNBC cases, respectively, confirming the literature data on the high frequency of PD-L1 expression in the triple-negative subtype 
. An NGS validation of the PD-L1 status revealed a low CNA rate (1–2% of all cases), ruling out the amplification of this gene as the main cause of its high expression in metastatic breast cancer 
. An overall analysis of the breast cancer tissues showed genomic alterations in STK11
in 2% of the TNBC cases and in MDM2
in 3% of the TNBC cases. The latter has already been associated with resistance to ICPIs, as it attenuates the immune response 
, but few studies have considered it in breast cancer. In conclusion, this trial was useful for demonstrating how comprehensive genomic profiling using NGS can aid with ICPI treatment by identifying biomarkers of resistance, such as STK11
Tyrosine receptor kinases (TRKs) are an important class of transmembrane receptors involved in cell proliferation and survival pathways. Among them, neurotrophic TRK (NTRK
) is of interest in tumors because the aberrant fusion of its kinase domains with other genes can lead to a ligand-independent activation of the receptor 
. In light of this, RTK inhibitors (RTKi) may offer clinical benefits as effective adjuvant drugs in patients with NRTK fusion. Wu et al. conducted a retrospective study to evaluate the NTRK statuses in 305 TNBC patients and to investigate the potential clinical applications of RTKi in TNBC treatment. These NTRK
statuses were assessed using IHC, FISH, and Illumina NGS sequencing. The IHC and FISH analyses showed that 11.15% of cases had NRTK fusion, but an NGS validation showed no positivity 
. Although the results precluded the use of RTKi in TNBC, this trial represents an optimal example of the use of NGS for the validation of routine clinical procedures.
Table 1. Summary of the experimental methods and main findings from reported next-generation sequencing (NGS) studies in triple-negative breast cancer (TNBC) research. Abbreviations: CNV (copy number variation); DFS (disease-free survival); ER+ (estrogen receptor positive breast cancer); FFPE (formalin-fixed paraffin-embedded); HER2+ (HER2 positive breast cancer); IBC (inflammatory breast cancer); ICPI (immune checkpoint inhibitor); IHC (immunohistochemistry); LAR (luminal androgen receptor); MSI (microsatellite instability); NAC (neo-adjuvant chemotherapy); NR (not reported); NST (no special-type tumor); PFS (progression-free survival); PR+ (progesterone receptor positive breast cancer); ST (special-type tumor); TMB (tumor mutational burden); TNBC (triple-negative breast cancer); and yrs (years).
3.2. NGS in Ethnic-Specific Molecular Profiling of TNBC
As reported by the American Cancer Society in 2019, the incidence of TNBC is twice as high in women of African American and Hispanic descent than in white women 
. These incidence rates highlight the genetic population landscape as a potential TNBC risk factor. NGS technologies can to help investigate the role of ethnicity-specific pathogenic alterations in the incidence and progression of TNBC. The related studies are shown in Table 2
From this perspective, Anwar et al. characterized the copy number variations of 409 cancer genes in 11 Ghanaian metastatic breast cancers (among which 9 were TNBC) using Ion Torrent. Despite the small cohort analyzed, the analysis showed 17 genes with frequent copy number alterations (SDHC, RECQL4, TFE3, BCL11A, BCL2L1, PDGFRA, DEK, SMUG1, AKT3, SMARCA4, VHL, KLF6, CCNE1, G6PD, FGF3, ABL1
, and CCND1
), among which, the most common were RECQL4
(50%) and SDHC
(60%). A network analysis indicated the involvement of these genes in cell proliferation, apoptosis, and the PIK3CA pathways 
. Moreover, 13 out of the 17 genes interact with EZH2
, an epigenetic regulator whose overexpression is associated with metastasis and a negative status of ER and PR in breast tissues 
. Indeed, another report showed a strong relationship between EZH2
overexpression and high-grade basal-like breast cancers in a cohort of 169 Ghanaian women 
. These data provide a basis for further explorations of the pathobiology of breast cancer and TNBC in African and American African women.
The study by Ben Ayed-Guerfali et al. focused on the mutational characterization of BRCA1/2
in 110 Tunisian women with breast cancer (26 TNBC) and 24 Tunisian women with ovarian cancer. A sequencing analysis was performed using the Illumina platform on plasma samples from the patient cohort. Overall, BRCA 1/2
mutations occurred in 14.17% of cases and were mainly frameshift (76.9%). The most common genetic alterations were c.1310_1313 delAAGA in the BRCA2
gene and c.5030_5033 delCTAA in the BRCA1
gene, which were found in 4% of the breast cancer patients and 20% of the ovarian cancer patients. Interestingly, there was a strong correlation between BRCA
mutation carriers and TNBC. In fact, 5 of the 26 TNBC patients (19.23%) carried BRCA
alterations, of which, 4 carried BRCA1
mutations. This result suggested a higher frequency of BRCA1 carriers than BRCA2
carriers in the TNBC cases, at least in the Tunisian population 
. Despite the small number of TNBC cases analyzed, these data are consistent with those reported in other studies 
, highlighting the potential of NGS to deepen population-specific molecular TNBC profiles.
Laraqui et al. instead performed a molecular characterization of 30 Moroccan women with early-stage TNBC. Specifically, Illumina-based sequencing was used to investigate the pathogenic variants of 63 cancer-related genes: AIP, APC, ATM, BAP1, BARD1, BMPR1A, BRCA1, BRCA2, BRIP1, CASR, CDC73, CDH1, CDK4, CDKN2A, CHEK2, EPCAM, FANCM, FH, FLCN, MAX, MCIR, MEN1, MET, MITF, MLH1, MSH2, MSH6, MUTYH, NBN, NF1, NF2, PALB2, PMS2, POLD1, POLE, PTEN, RAD50, RAD51C, RAD51D, RET, SDHA, SDHAF2, SDHB, SDHC, SDHD, SMAD4, SMARCA4, STK11, TMEM127, TP53, VHL, RK1, FAM175, GREM1, MLH3, MRE11, MMSH2, NTHL1, PMS1, RAD51, RAD51B, RINTI, RNF3, RNF43,
. The analysis revealed that 6 of the 30 patients (20%) had known pathogenic variants, the most common of which were associated with BRCA
(3 patients with BRCA1
mutations and 2 with BRCA2
. As in the previously described Ben Ayed-Guerfali effort, the pathogenic variant c.1310_1313 delAAGA was detected in the BRCA2
gene in the patients studied, suggesting that this alteration represents a possible founding mutation for TNBC in the North African population. In addition, the analysis revealed 42 variants of unknown/uncertain significance (VUS) in 70% of the patients (21/30), with ATM
being the gene with the highest frequency of VUS (4/30). This result is to be expected with multigene panel testing and large genes such as ATM 
. Apart from technical limitations, the high number of VUS may indicate a possible new molecular signature of TNBC.
The studies described so far are characterized by a relatively small number of analyzed patients, highlighting the challenges encountered in optimal sample collection from several African environments. Despite this, NGS technologies have achieved promising results in the molecular characterization of TNBC in African women.
Table 2. Summary of the experimental methods and main findings of next-generation sequencing (NGS) studies aimed at exploring ethnic-specific molecular signatures in triple-negative breast cancer (TNBC). Abbreviations: BC (breast cancer); CNA (copy number alteration); FFPE (formalin-fixed paraffin-embedded); TNBC (triple-negative breast cancer); and yrs (years).
3.3. NGS of Third and Fourth Generation in TNBC Research
As mentioned before, only the Illumina and Ion Torrent platforms have been used for sequencing in the previous chapters. Only few studies have used third- and fourth-generation NGS technologies for a molecular characterization of TNBC, highlighting the need for their optimization for clinical applications. The few TNBC studies in which Pacbio and nanopore sequencing have been used are briefly reported below.
Aganezov et al. performed the whole-genome sequencing of SKBR3, a TNBC cell line 
, using the Illumina platform, Pacbio sequencing, and Oxford Nanopore Technology (ONT) to compare the sequencing efficiency of each method. In addition, the same comparison was performed on organoids from two breast cancer patients obtained from cancer and normal tissues. For all the samples, the analysis showed an excellent accordance of the genetic variants (GVs) detected using the Pacbio and ONT methods, with more than 90% of the GVs being detected by both sequencing techniques. In addition, the long-read methods of Pacbio and ONT detected a large number of GVs missed by the Illumina short-read method. This finding suggests introducing long-read sequencing techniques into routine clinical practice to improve cancer risk assessment, analysis, and treatment. However, the cost of full genomic characterizations using Pacbio and ONT remains significantly higher than that of Illumina sequencing. From the perspective of cost reduction, the study by Aganezov et al. reported the optimal GV detection of long-read methods even at a low average read depth coverage 
. Unfortunately, this approach may underestimate the heterozygous germline SNPs, which are common in samples from cancer patients 
. This problem can be circumvented by combining short-read and long-read techniques and continuously reducing the cost of Pacbio nanopore sequencing.
Weirather et al. investigated the possibility of integrating the results of long-read and short-read sequencing methods to allow for a higher precision analysis of fusion genes. This combinatorial method, called IDP-fusion (Isoform Detection and Precision), aims to detect fusion genes, determine fusion sites, and identify fusion isoforms. The flowchart of the IDP-fusion approach begins with the identification of fusion genes using Pacbio wide sequencing. The resulting long reads are aligned to the reference genome to discover the pair fragments that can be mapped to two gene loci. Unfortunately, the boundaries of the selected long fragments cannot be considered as true fusion sites due to the high error rate of Pacbio sequencing. To circumvent this problem, the long fragments are extended by about 2000 base pairs beyond the alignment ends and concatenated to obtain an “artificial reference sequence” (ARS). The ARS thus obtained is then used as a reference to map the high-quality short reads from second-generation sequencing methods, such as Illumina, to identify the true fusion sites. Ultimately, fusion isoform calling involves three steps: the identification of splice sites from long reads, building a candidate isoform library, and estimating the isoform abundance. Weirather et al. tested IDP fusion using a genomic study of MCF-7, a TNBC cell line 
. Using a panel of 71 common fusion genes in TNBC, the analysis showed a higher fusion gene detection accuracy for the hybrid method (68%) in comparison to the short-read and long-read methods alone (14% and 18%, respectively). In addition, the IDP fusion analysis identified novel fusion genes with tumor interest, such as AIB1-chr1:107073407 (tamoxifen resistance) or TPD52L2-chr17:60952559 (breast cancer proliferation). In summary, long-read and short-read sequencing methods have several inherent limitations, such as a high error rate and miscalling in repetitive genomic regions, respectively 
. IDP fusion can overcome these problems by combining both techniques, providing a new NGS approach for the molecular characterization of TNBC.
Despite the paucity of efforts to characterize TNBC using third- and fourth-generation sequencing methods, the reported studies have highlighted the possibility of innovative approaches that hold promise for the future use of these technologies in the clinical treatment of TNBC.