3.1 First-Generation Sequencing Technology

The foundation of modern sequencing technologies can be traced back to the mid-20th century. Early efforts at deciphering nucleic acid sequences relied on chemical degradation or enzymatic cleavage, fragmenting DNA and RNA molecules for subsequent analysis. A major milestone was achieved in 1964 when Robert Holley sequenced alanine tRNA using ribonuclease from S. cerevisiae.⁶⁰ Soon after, Walter Gilbert and Allan Maxam developed a chemical degradation method that enabled the sequencing of the complete genome of bacteriophage PhiX174.⁶¹ However, the most transformative advance in first-generation sequencing came from Frederick Sanger, who introduced the chain termination method in the late 1970s.⁴ This approach, which employed dideoxynucleotides to terminate DNA replication, allowed for the generation of sequence reads up to several hundred nucleotides in length.

Sanger sequencing, as it became known, revolutionized molecular biology by enabling rapid and precise DNA and RNA sequencing. The method's impact deepened with the introduction of the Applied Biosystems ABI 370, the first automated sequencing machine, in 1987.⁶² This machine used fluorescently labeled dideoxynucleotides and capillary electrophoresis, streamlining the Sanger method to produce faster, more accurate results. The ABI 370 quickly became the industry standard for DNA sequencing and laid the groundwork for sequencing the human genome in 2001,¹⁴ marking a critical turning point in cancer genomics. While first-generation sequencing has since been eclipsed by more advanced technologies, its foundational role in the study of genetics and the development of cancer diagnostics and therapies cannot be overstated.

3.2 Second-generation sequencing technologies

The advent of second-generation sequencing (also known as NGS) marked a quantum leap in the field by introducing high-throughput, parallel sequencing. In contrast to Sanger sequencing, which could only process a single DNA fragment at a time, second-generation sequencing platforms introduced the capability to sequence thousands to hundreds of millions of DNA fragments simultaneously. This advancement revolutionized the field by dramatically boosting both the speed and scale of data generation, allowing for a massive increase in sequencing throughput and enabling large-scale genomic studies to be completed far more efficiently.

One of the earliest and most prominent platforms in this generation was Roche's 454 sequencing, which employed pyrosequencing, a method that detects pyrophosphate release during nucleotide incorporation.⁶³ Another notable system is the Ion Torrent platform, which measures hydrogen ion release during DNA synthesis.⁶⁴ Illumina sequencing, arguably the most widely used second-generation platform, introduced a sequencing-by-synthesis approach involving reversible dye terminators.³⁹ Each of these technologies has enabled significant advances in cancer genomics, making it possible to analyze tumor heterogeneity, discover novel oncogenes, and perform comprehensive whole-genome and transcriptome analyses in cancer patients.

Another key player in this generation is the SOLiD (Sequencing by Oligonucleotide Ligation and Detection) platform,³⁹ which uses ligation-based chemistry to achieve high sequencing accuracy. These second-generation platforms have empowered researchers to explore genetic mutations, epigenetic modifications, and gene expression changes at an unprecedented scale. In cancer research, this has translated into the identification of critical biomarkers, enabling early diagnosis, targeted therapies, and personalized treatment strategies for malignancies like breast, lung, and colon cancers.

3.3 Third-generation sequencing technologies

Third-generation sequencing technologies represent the cutting edge of DNA sequencing, overcoming many of the limitations of earlier methods, particularly in terms of read length and real-time analysis. These platforms can sequence much larger DNA fragments, sometimes spanning tens of kilobases, with single-molecule precision.

PacBio sequencing, which utilizes an approach of single-molecule real-time (SMRT), is a leading third-generation technology.⁶⁵ By incorporating fluorescently labeled nucleotides and continuous real-time monitoring of DNA synthesis, PacBio can generate long-read sequences that provide insight into complex genomic regions previously inaccessible with second-generation methods. This is especially relevant in cancer research, where structural variations such as gene fusions, insertions, deletions, and CNVs play critical roles in oncogenesis and drug resistance, which have been explained later. Oxford Nanopore sequencing, another cutting-edge technology, relies on nanopore sensors to sequence single-stranded DNA molecules.⁶⁴ As DNA molecules pass through a nanopore, variations in the electrical current are monitored and analyzed. The portability, speed, and long-read capacity of Oxford Nanopore make it a powerful tool for real-time analysis of cancer genomes, potentially enabling clinicians to monitor tumor evolution and therapeutic response in a near-instantaneous manner.

The application of third-generation sequencing technologies in cancer genomics has opened new avenues for understanding the full spectrum of genetic alterations driving cancer progression. Long-read sequencing is particularly beneficial for studying complex genomic regions such as telomeres, centromeres, and repetitive sequences, which are often altered in cancer cells. Additionally, third-generation methods allow for the direct sequencing of RNA molecules without the need for reverse transcription, providing a more comprehensive view of the cancer transcriptome, including the role of Noncoding RNAs, alternative splicing events, and fusion transcripts in tumorigenesis.

4.1 Role of NGS in detecting MRD

MRD is the term used to describe cancer cells that remain in a patient's body after therapy, even if typical clinical evaluations are unable to identify them.^{31, 91} Residual cancers following treatment present a considerable challenge for oncologists. These cancers tend to be not only more resistant to therapeutic interventions but also often exhibit greater aggressiveness than the original tumor. Evaluating MRD after chemotherapy is crucial for planning subsequent treatment strategies. NGS plays a pivotal role in the detection of MRD, serving as an essential component in both cancer research and clinical management. NGS has transformed the field by enabling comprehensive genomic analyses, making it an invaluable tool for elucidating the genetic underpinnings of cancer, identifying mutations, and characterizing tumor heterogeneity. The detection of MRD is critical in oncology care, as it informs treatment decisions and provides insights into the likelihood of cancer recurrence, ultimately helping to tailor patient management strategies more effectively.^{31, 91} NGS is necessary for identifying MRD because of its excellent sensitivity and ability to identify even minute amounts of altered DNA or RNA.^{91, 92} Additionally, the sensitivity of MRD detection can be increased by using focused NGS panels that concentrate on known mutations or gene fusions linked with malignancy. Treatment response and the establishment of resistance mutations can be understood by tracking changes in the tumor's genetic profile with serial NGS assessments conducted both during and after treatment. NGS-MRD detection can provide helpful prognostic data, assisting physicians in assessing the possibility of novel drug delivery.

The use of NGS-MRD assessment as an endpoint in clinical studies to evaluate treatment effectiveness and guide medication development is growing in popularity. Physicians can assess the chance of a disease recurrence by using prognostic information obtained from NGS MRD detection. Several cancer types have been identified with NGS-based MRD detection, including solid tumors, lymphomas, leukemia, and multiple myeloma.^{29, 92-94}

Several methods and markers may be used depending on the type of cancer and the genetic alterations taking place. Because NGS makes comprehensive genome analysis possible, it has changed clinical treatment and cancer research. When used for MRD detection, NGS exhibits exceptional sensitivity and specificity, facilitating the monitoring of cancer patients' prognosis and response to treatment.^{31, 94} Combining NGS with MRD evaluation is viable for integrating into one health approach.

4.2 Application of NGS to detect CNVs in cancer

The genomes of many species, including humans, have CNVs, which can be found and characterized using NGS, a potent and adaptable technique.⁷⁹ DNA segments that range in size from kilobases to megabases are either deleted or duplicated in CNVs, a kind of structural variation, as compared to the reference genome.⁹⁵ These changes to the genome may have significant ramifications for oncology and other genetics-related professions. NGS provides many methods, each with pros and cons, for identifying CNVs. Whole-genome sequencing (WGS) is a widely used technique in which all genomes is sequenced and read-depth variations are examined. A change in read depth within a specific area denotes the existence of duplications or deletions, respectively. WGS is valuable for identifying CNVs across the entire genome but can be computationally intensive and expensive.^{32, 96} Protein-coding portions of the genome are the focus of targeted sequencing techniques like exome sequencing.⁸⁰ This method is frequently applied in clinical settings and is reasonably priced. It might, however, overlook CNVs outside of the designated areas. Targeted sequencing and genome-wide CNV analysis are used in hybrid capture-based approaches to balance specificity and genome coverage.

4.3 Application of NGS in detecting gene fusion

NGS has revolutionized detecting and characterizing fusion genes, which are chimeric genes formed by the abnormal fusion of two genes.^{70, 100} These fusion events occur in various cancers and can serve as critical oncogenic drivers or therapeutic targets. In this academic discussion, we will explore how NGS is utilized in detecting fusion genes and its significance in cancer research and clinical practice.

6.1 Single-cell RNA sequencing (scRNA-seq)

While scRNA-seq is not strictly a spatial technique, it can provide insights into the gene expression profiles of individual cells, which can be integrated with spatial data to infer cellular interactions and relationships (Figure 4).¹²³

Integrating NGS and spatial analysis techniques allows researchers to Map Gene Expression spatially. Researchers can create spatial gene expression maps by combining spatial transcriptomics with NGS,^{25, 112} revealing how genes are expressed within specific tissue regions. Spatial analysis can also uncover the heterogeneity of tissues, showing variations in gene expression or genetic alterations across different areas. Further researchers gain insights into the spatial distribution of disease-associated markers or pathways within tissues, aiding in disease diagnosis and understanding. Integrating scRNA-seq with spatial data can elucidate cellular interactions and communication networks within tissues.¹²⁴ Spatial genomics can guide the development of targeted therapies by identifying regions of interest for drug targeting.^{125, 126}

In recent years, significant advancements have been made in the development of scRNA-seq techniques across both laboratory and computational domains. A key aspect of this methodology involves the meticulous isolation of individual cells from tissue samples. Following isolation, the cells are lysed using specialized reagents to extract the target RNA molecules, which are subsequently purified. To eliminate ribosomal RNA, the RNA molecules are polyadenylated using poly(T) primers, as analyzing non-polyadenylated mRNA poses challenges and necessitates specific protocols for effective analysis.¹²⁷ The extracted total RNA is used to synthesize cDNA via reverse transcription, with specific adapter sequences (unique molecular identifiers) added for RNA detection by NGS platforms. The cDNA is then amplified using PCR or in vitro transcription, followed by another round of reverse transcription or nucleotide barcode-tagging amplification. Finally, the tagged cDNA from each cell is pooled and sequenced using NGS, employing library preparation techniques, sequencing platforms, and genomic-alignment tools similar to those used for bulk samples.¹²⁸

Several methods are employed for single-cell transcriptome sequencing, with prominent techniques including SCRB-seq, CEL-seq. 2, MARS-seq, Drop-seq, Smart-seq. 1, Smart-seq. 2, and 10× Genomics. Among these, Smart-seq. 2 is notable for detecting the maximum number of transcripts in individual cells, while CEL-seq. 2, MARS-seq, Drop-seq, and SCRB-seq effectively quantify mRNA with minimal noise using unique molecular identifiers (UMIs). Drop-seq is the most cost-effective for analyzing a large number of cells, whereas SCRB-seq, MARS-seq, and Smart-seq. 2 are more effective for profiling the transcriptome of fewer cells.

6.2 Spatial Transcriptomics

Spatial transcriptome analysis is a significant breakthrough in medical biotechnology, enabling the mapping of analytes, such as RNA, within their physical locations in tissue sections. Various techniques, including in situ sequencing, fluorescent in situ hybridization (FISH), in situ capture, and in silico methods, have since been used for spatial RNA mapping.¹⁴⁰

6.3 Integration of spatial transcriptomics and scRNA-seq

ScRNA-seq has revolutionized the analysis of gene expression by preserving the histological context of tissues, allowing for the identification of thousands of genes in sampled tissues. This technology offers critical clinical tools for early cancer detection and the development of targeted treatment strategies.¹⁵² Moreover, scRNA-seq can pinpoint the locations of mutated or aggressive clones within tumors, differentiate between the tumor's core and its invasive edges, evaluate molecular changes in the stroma both inside and outside the tumor, and detect epithelial-mesenchymal transitions.¹⁵³

Recent advancements demonstrate that scRNA-seq can analyze RNA from both fresh-frozen and fixed tumor cells without losing their spatial context.¹⁵³ Additionally, it can integrate spatial features directly into specific genetic elements in organoids or native tissues using image analysis. This integration characterizes and maps sequences from tumor tissues, revealing the complex arrangement of various cell types regulated by single-cell interactions. The primary goal of spatial mapping techniques is to acquire comprehensive genomic or transcriptomic data from entire tissue slides, enhancing it with sequencing methods. Combining scRNA-seq with other omics technologies provides a more thorough analysis of RNA analytes within their spatial context.¹⁵⁴ In situ visualization of the transcriptome through scRNA-seq offers multiplexed information on gene expression, cell types, and disease progression. For instance, ISS utilizes padlock probes and rolling circle amplification to spatially display RNA expressions across tissue sections.

In a study conducted by Gyllborg et al., hybridization-based in situ sequencing (HybISS) was utilized to investigate the spatial localization of RNA within human and mouse brain tissues. This enhanced version of the HybISS platform has significantly improved combinatorial barcoding and detection limits, enabling precise mapping of RNA transcripts in their respective spatial contexts. By leveraging this advanced methodology, researchers were able to obtain detailed insights into the distribution and organization of RNA within the complex architecture of brain tissues.¹⁵⁵ Asp et al. integrated scRNA sequencing (scRNA-seq) with spatial transcriptomics and in situ sequencing (ISS) to provide a comprehensive overview of the cellular architecture of the human heart across different developmental stages. This innovative approach allowed for a detailed examination of the cellular components, while ISS offered insights into the subcellular features and their specific physical locations within the heart tissue. Through this combination of techniques, the study illuminated the intricate organization of cardiac cells and their developmental dynamics.¹⁵⁶

Moncada et al. integrated scRNA-seq with microarray-based spatial transcriptomics to uncover spatial gene expression variations in pancreatic ductal adenocarcinomas, identifying various cell types and subsets, including cancer cells and immune cells, within the tumors.¹⁴³ Another approach, geographical positional sequencing (Geo-seq), combined LCM techniques with scRNA-seq to analyze cellular heterogeneity in pancreatic ductal adenocarcinomas, differentiating various subpopulations of cells spatially.¹⁵⁷ Further, spatial transcriptomics involves mounting tissue sections on glass slides fixed with arrays of transcription primers, followed by permeabilization and reverse transcription to synthesize cDNA. This process allows for the spatial mapping of RNA transcripts on a two-dimensional plane. Advances in spatial profiling of transcriptomes, particularly in cancer research, have been meticulously reviewed, detailing methods and applications, including liver histology analysis.¹⁴¹ In a significant investigation conducted by Ji et al., a combination of single-cell RNA sequencing (scRNA-seq), spatial transcriptomics, and multiplexed ion beam imaging techniques was employed to study the spatial localization of tumor-specific keratinocytes (TSK) in breast carcinoma. This integrated approach allowed for the detailed visualization of TSK populations, along with immune infiltrates, revealing the cellular heterogeneity present at the tumor margins. By utilizing these advanced techniques, the study provided valuable insights into the complex interactions between tumor cells and the immune environment in breast cancer.¹⁴² ScRNA-seq has identified distinct cancer cell populations within pancreatic ductal adenocarcinoma (PDAC-A), revealing significant histological differences. For example, PDAC-A sub-region 1 showed an abundance of endothelial cells, monocytes, and fibroblasts, but a notable scarcity of cancer cells.¹⁵⁸

7.1 Pros of NGS

NGS technologies can sequence millions to billions of DNA fragments simultaneously, allowing for the analysis of entire genomes, transcriptomes, or metagenomes in a single run are much faster than traditional Sanger sequencing and produce results in hours to days instead of weeks to months.⁶⁴ Although this may seem expensive, over time, the cost of sequencing has dropped tremendously, making NGS accessible to researchers and clinicians.⁶⁴ NGS provides high coverage for genomic regions. Thus increasing detection rates for rare variant calls and expanding knowledge about genetic diversity across populations. Some long-read-length sequencing platforms like PacBio and Nanopore sequencing can resolve complex genomic regions, detect structural variations, and investigate epigenetic modifications.^{20, 159, 160} This means that NGS is a widely applicable research tool in different scientific spheres, such as genomics, transcriptomics, epigenomics, and metagenomics, among others, which makes it a versatile tool for studying scientists and clinicians use it (specificities given in Figure 5). It has changed the way that we find disease-associated genetic variation; it has become an essential part of clinical diagnostics for personalized medicine and early disease detection.¹⁶¹ It is also used to evaluate complicated microbial communities in environmental samples, with the human microbiome being one of those ecosystems.

7.2 Cons of NGS

The massive amount of data generated by NGS requires substantial computational resources and expertize for data storage, processing, and analysis.¹⁴ Sample preparation for NGS can be time-consuming and sensitive to DNA quality and quantity. Inadequate sample preparation can lead to data artifacts.¹⁰⁴ This can be a significant challenge for smaller research groups (Figure 5). Despite advances, NGS technologies can still produce sequencing errors, especially in regions with repetitive sequences or GC-rich regions. Careful data analysis and quality control are essential to mitigate these errors. NGS platforms are also technically demanding, requiring specialized equipment and highly trained personnel for operation and maintenance. Besides the technical challenges, the bioinformatics challenges are also significant, and manpower needs to be trained to harness the complete data information. NGS data requires bioinformatics expertize and specialized software tools, which may pose a barrier for researchers without a computational background.¹⁴ Data output often runs in several gigabytes and terabytes.¹⁶²

While some NGS technologies offer long read lengths, many platforms produce short reads, making it challenging to assemble complex genomes or accurately characterize repetitive regions. While NGS has become more cost-effective in recent years, achieving deep genome coverage, and other specialized assay like single-cell genomics can still be quite expensive. The ability to generate extensive genomic information also raises ethical and privacy concerns related to the confidentiality and potential misuse of personal genetic data. The probability of widespread adoption in the future is heightened by the long-turn-around times and lower capital investment required for short-read sequencing technologies. However, the considerable costs and extended timelines of metagenomic research, coupled with its ambiguous correlation to conventional oncology methods, present major challenges to its routine or frequent application. Most commercial platforms currently rely on short-read sequencing, and since these tests are typically outsourced, the process of shipping specimens introduces further logistical delays, in addition to the runtime of the sequencing process itself and the subsequent bioinformatics analysis.¹⁶³

These factors also complicate the timely reporting of data, making it difficult for clinicians—and, in some cases, even patients—to properly interpret the results. This issue could be mitigated by developing localized sequencing infrastructure, streamlining bioinformatics pipelines, and enhancing clinician education regarding metagenomic analysis, thereby improving both turnaround times and the clarity of results interpretation.

7.3 Future directions

Future research directions for NGS in oncology are expanding its precision and application to deepen our understanding of cancer's molecular underpinnings. Key areas of development include:

Single-cell NGS for Clonal Dynamics and Tumor Heterogeneity: Single-cell sequencing offers high-resolution insights into the clonal architecture of tumors, enabling the identification of subclonal populations and their roles in tumor evolution, immune escape, and therapeutic resistance. This will provide critical data for developing targeted therapies that account for clonal diversity within tumors.

Liquid Biopsy and MRD Detection: NGS applied to liquid biopsies—particularly in analyzing ctDNA and circulating tumor cells—is critical for early detection of MRD. Enhanced sensitivity will allow for the identification of ultra-low-frequency variants, crucial for predicting recurrence and guiding adjuvant therapies, even when tumors are not detectable via traditional imaging.

Immunogenomics and Neoantigen Discovery: NGS will continue to enhance immunogenomic profiling, allowing for the identification of tumor-specific neoantigens that inform personalized immunotherapy approaches. This includes leveraging tumor mutational burden and neoantigen prediction for optimizing immunotherapies, including checkpoint inhibitors and personalized cancer vaccines.

CNVs and single nucleotide polymorphisms (SNPs) Detection: Research will further enhance NGS for the detection of CNVs and SNPs, both critical in cancer development and progression. Accurate identification of CNVs will improve our understanding of gene dosage effects, such as oncogene amplification and tumor suppressor gene loss. Similarly, SNP profiling will reveal key germline and somatic mutations that influence cancer susceptibility, treatment response, and resistance mechanisms, with implications for risk stratification and therapeutic targeting.

Spatial Genomics and Transcriptomics for Tumor Microenvironment: Spatial NGS techniques are being developed to map gene expression and genomic alterations in situ within the tumor microenvironment. Understanding the spatial distribution of tumor cells, stromal elements, and immune infiltrates will provide crucial insights into localized interactions that drive metastasis, therapy resistance, and tumor immune evasion.

Noncoding RNA and Epigenomic Sequencing: Expanding NGS to noncoding RNA (ncRNA) profiling, including microRNAs and long noncoding RNAs (lncRNAs), will unveil regulatory networks that contribute to cancer progression. Epigenomic sequencing will provide additional insights into DNA methylation, histone modifications, and chromatin architecture, offering new therapeutic targets within the cancer epigenome.

AI-Enhanced NGS Data Analysis: Machine learning algorithms are poised to revolutionize NGS data interpretation, improving the accuracy of variant calling for SNPs, CNVs, and structural variants. AI-driven models will enable the integration of multi-dimensional genomic data, predicting therapeutic responses and identifying novel biomarkers, streamlining clinical decision-making in precision oncology.

These advances will further refine NGS's role in cancer biology, driving innovations in early detection, diagnostics, and tailored treatment strategies, cementing its importance in precision oncology.