# NGS Sequencing


## Key Takeaways

- Next-generation sequencing (NGS) fundamentally shifts from Sanger sequencing by performing massive parallel sequencing of millions of DNA or RNA fragments simultaneously, enabling high throughput and scalability. Key platforms like Illumina (short reads) and Oxford Nanopore (long reads) offer distinct advantages in read length and accuracy, influencing experimental design.
- NGS applications extend beyond whole genome sequencing to include targeted panels, RNA sequencing (RNA-seq), epigenomics, and metagenomics, providing versatile tools for diverse research and clinical questions. The choice of NGS approach hinges on biological questions, sample type, budget, and desired read length, with targeted panels offering cost-effectiveness for specific regions.
- Library preparation, including fragmentation, adapter ligation, and amplification, is a critical initial step in NGS, followed by sequencing and rigorous quality control. Phred quality scores (Q scores) are essential metrics for assessing base call accuracy, with Q30 indicating a 1 in 1000 error rate, and data often requires filtering for low-quality reads and adapter sequences.
- Downstream analysis involves aligning reads to a reference genome (e.g., using BWA MEM or STAR), variant calling (e.g., with GATK or FreeBayes), and annotation with databases like dbSNP and ClinVar. Depth of coverage is crucial, with typical requirements ranging from 30x for germline variants to over 100x for detecting low-frequency somatic mutations in circulating tumor DNA.
- Common pitfalls in NGS include insufficient coverage for the research question, overlooking batch effects, contamination, and misinterpreting low-frequency variants, which may necessitate orthogonal validation (e.g., ddPCR). Long-read sequencing, while powerful for resolving structural variants and repetitive regions, can still have lower raw accuracy than short-read technologies, often requiring polishing.
- Interpretation of NGS results requires careful consideration of inherent uncertainties, including base-calling errors in homopolymer regions and limitations in mapping short reads to repetitive sequences. Variants of uncertain significance (VUS) in noncoding regions present interpretation challenges, and clinical decision-making must integrate NGS findings with other diagnostic data.

---

Next generation sequencing (NGS) refers to high throughput DNA and RNA sequencing technologies that parallelize the reading of millions of nucleotide fragments simultaneously. This guide is written for researchers, clinicians, and bioinformatics beginners who need a practical, source bounded understanding of NGS: what it is, how to choose a method, how to execute an experiment, and how to interpret results with appropriate caution. Use this guide to build a working framework for planning your own NGS project or evaluating published studies.

First, grasp the core shift from older Sanger sequencing: NGS does not sequence a single DNA molecule, it sequences many short (or long) reads in parallel, then reassembles them computationally. [NCBI Bookshelf](https://www.ncbi.nlm.nih.gov/books/) offers authoritative technical references that detail the chemistry and instrumentation behind each platform. For example, Illumina sequencing uses reversible dye terminators and bridge amplification, while Oxford Nanopore reads ionic current changes across a membrane. These differences matter for your choice of platform.

Second, understand that the usefulness of NGS extends far beyond whole genome sequencing. Targeted panels, RNA seq, epigenomics, metagenomics, and even long read structural variant detection all fall under the NGS umbrella. [EMBL EBI Training](https://www.ebi.ac.uk/training/) provides structured courses that explain these applications and their analytical pipelines.

## At a Glance

| Feature | Description |
| --- | --- |
| Definition | Massive parallel sequencing of DNA or RNA fragments |
| Main platforms | Illumina (short reads), Oxford Nanopore (long reads), MGI, PacBio (long reads) |
| Typical read length | 50 600 bp (short) or 10 kb+ (long) |
| Output per run | 100 Mb to 6 Tb depending on platform |
| Common applications | Whole genome, exome, targeted panels, RNA seq, ChIP seq, metagenomics |
| Key advantages | Speed, scalability, cost per base |
| Key limitations | Error rates, read length tradeoffs, bioinformatics complexity |

## Core Concepts of NGS

NGS begins with library preparation: DNA or RNA is fragmented, adapter sequences are ligated, and the molecules are amplified. The library is then loaded onto a flow cell where clonal clusters (Illumina) or single molecules (Nanopore) are read. Sequencing by synthesis builds a complementary strand and records incorporated bases. [Galaxy Training Network](https://training.galaxyproject.org/) provides interactive tutorials that walk through this process from sample to data.

Quality control is vital. Each base receives a Phred quality score (Q score) that estimates the probability of an incorrect call. A Q30 score means 1 error per 1000 bases. Most NGS data require filtering to remove low quality reads and adapter sequences before analysis.

The data output is a FASTQ file containing raw sequences and quality scores. Downstream steps include alignment to a reference genome, variant calling, or de novo assembly. [Bioconductor](https://bioconductor.org/) hosts R packages for statistical analysis and visualization of NGS data.

## Decision Criteria for Choosing an NGS Approach

Selecting the correct method depends on your biological question, sample type, and budget. Consider these factors:

**Read length** Short reads (150 bp) are cost effective for resequencing and RNA seq but struggle with repetitive regions and structural variants. Long reads (10 kb+) can span repeats and resolve complex rearrangements, as seen in a study of STRC variants in hereditary hearing impairment using long read sequencing with MLPA validation. [PubMed](https://pubmed.ncbi.nlm.nih.gov/42437880/) describes how long reads overcome limitations of short read mapping in highly homologous regions.

**Targeted versus whole genome** Targeted panels (amplicon or capture) focus on specific genes or regions, reducing cost and data volume. In infectious disease, 16S rRNA RC PCR NGS can identify pathogens from clinical samples with high sensitivity and specificity. A recent clinical impact study demonstrated that such targeted NGS improved management of infectious diseases. [PubMed](https://pubmed.ncbi.nlm.nih.gov/42439573/) provides evidence for this approach.

**Single versus paired end** Paired end sequencing reads both ends of a fragment, improving alignment accuracy and detection of insertions/deletions.

**Depth of coverage** More coverage increases confidence in variant calls. For germline variants, 30x is typical. For somatic or low frequency variants (e.g., in circulating tumor DNA), higher depth exceeding 100x is needed.

**Sample type and quality** Degraded DNA may require shorter targets or specialized library protocols. RNA seq requires poly A enrichment or ribodepletion.

## Practical Workflow for an NGS Experiment

Follow these steps for a typical targeted resequencing project:

1. **Sample collection and extraction** Isolate nucleic acid with appropriate quality controls. Use a fluorometric method (Qubit) and assess integrity (e.g., RIN for RNA, DNA integrity number for DNA).

2. **Library preparation** Fragment DNA (e.g., by sonication or enzymatic fragmentation). Repair ends, add A overhang, ligate adapters. For targeted approaches, perform hybrid capture or PCR amplicon amplification. Purify and quantify the final library.

3. **Pooling and sequencing** Pool barcoded libraries at equimolar concentrations. Load onto the sequencer according to the manufacturer's protocol. Monitor cluster density and quality metrics in real time.

4. **Primary data analysis** Convert raw data (BCL files for Illumina) to FASTQ. Demultiplex samples. Run FastQC for quality assessment. Trim adapters and low quality bases using tools like Trimmomatic or Cutadapt.

5. **Alignment** Map reads to a reference genome using a splice aware aligner (STAR for RNA seq) or a fast aligner (BWA MEM for DNA). Sort and index the BAM file.

6. **Variant calling** Use GATK, FreeBayes, or SAMtools for single nucleotide variants and small indels. For structural variants, consider long read based callers or specialized tools.

7. **Annotation and filtering** Annotate variants with databases (dbSNP, ClinVar, gnomAD). Apply quality filters (read depth, allele frequency, strand bias). [NCBI Sequence Read Archive](https://www.ncbi.nlm.nih.gov/sra) serves as a repository where you can deposit and access published sequencing data for reanalysis.

8. **Interpretation** Combine variant impact predictions (SIFT, PolyPhen) with clinical databases. For diagnostic applications, follow guidelines from professional societies.

## Quality Checks and Metrics

Monitor these metrics at each stage:

**Library level** Fragment size distribution (bioanalyzer trace), concentration, and adapter dimer contamination. Excess adapter dimer reduces effective sequencing yield.

**Sequencing level** Cluster density, percentage of bases above Q30, and read duplication rate. Low complexity libraries may show high duplication leading to wasted reads.

**Alignment level** Percentage of reads mapped, coverage uniformity, and insert size distribution. High mismatch rate may indicate contamination or sequencing error.

**Variant level** Read depth at variant sites, allele balance, and strand bias. False positives often arise from homopolymer artifacts or alignment errors.

Standardized reporting improves reproducibility. [EMBL EBI Training](https://www.ebi.ac.uk/training/) offers modules on quality control best practices.

## Common Mistakes and How to Avoid Them

**Mistake 1: Insufficient coverage for the question** A cancer panel run at 20x may miss subclonal mutations. Plan coverage based on variant frequency and required sensitivity.

**Mistake 2: Ignoring batch effects** Different library preparation runs can introduce bias. Use randomized sample layouts and include control samples.

**Mistake 3: Overlooking contamination** Bacterial reads in a human sample or human reads in a bacterial sample confuse analysis. Use FastQ Screen or similar tools to detect cross species contamination.

**Mistake 4: Using default parameters without validation** Trimming thresholds, mapping parameters, and variant filters should be tuned to your specific data type and platform. Pilot on a known sample.

**Mistake 5: Misinterpreting low frequency variants** A variant present in 1% of reads could be a true mutation, sequencing error, or alignment artifact. Require orthogonal validation (e.g., ddPCR) before clinical action.

**Mistake 6: Believing that long reads are error free** Oxford Nanopore raw accuracy is improving but still lower than Illumina. Polishing with short reads or using consensus strategies is often necessary.

## Limits of Interpretation and Uncertainty

Every NGS result has inherent uncertainty. Base calling errors, especially in homopolymer runs, can produce false variants. Short reads often fail to map uniquely to repetitive regions, leading to gaps in coverage. Structural variant detection from short reads is limited to deletions and duplications that are sufficiently short.

Interpretation of variants in noncoding regions is challenging because functional impact is less understood. Even with good coverage, a variant may be classified as a variant of uncertain significance (VUS). Clinical decision making must integrate NGS findings with other laboratory and clinical data.

A narrative review on integrating molecular and conventional diagnostics in native vertebral osteomyelitis noted that while NGS can detect pathogens, its false positive rate and inability to distinguish active from past infection must be considered. [PubMed](https://pubmed.ncbi.nlm.nih.gov/42441123/) highlights this limitation.

In forensic and anti doping contexts, detection of recombinant AAV based gene doping using NGS requires careful control for contamination and interpretation of vector sequences. [PubMed](https://pubmed.ncbi.nlm.nih.gov/42438284/) discusses how sequencing depth and bioinformatic filters affect detection limits.

Species identification from metagenomic NGS depends on database completeness. Unknown organisms may not be detected. For pediatric rheumatology, skeletal dysplasias can mimic juvenile idiopathic arthritis, and NGS panels may miss novel or deep intronic variants. [PubMed](https://pubmed.ncbi.nlm.nih.gov/42439985/) advises caution in purely NGS based diagnoses.

## Frequently Asked Questions

**What is the difference between NGS and Sanger sequencing?** Sanger sequencing reads one fragment at a time and is suitable for small targeted regions. NGS sequences millions of fragments in parallel, making it much faster and cheaper per base for large projects.

**How do I choose between Illumina and Oxford Nanopore?** If you need high accuracy for short variants and high throughput, choose Illumina. If you need long reads to resolve structural variants, repetitive regions, or full length transcripts, choose Oxford Nanopore. Consider also cost, instrument availability, and bioinformatics support.

**What does read depth mean and why is it important?** Read depth (or coverage) is the number of times a base is sequenced. Higher depth increases confidence in variant calls and allows detection of low frequency alleles. Minimum depth depends on your application: 10 30x for germline, 100x or more for somatic.

**Can NGS detect RNA viruses directly?** Yes, RNA seq can detect viral transcripts, but the library preparation must include random priming or rRNA depletion. Sensitivity is limited by viral load and host background. Additional validation by qPCR or other methods is recommended.

## Related Clinical & Scientific Guides

* [Observational vs. Experimental Studies: How to Tell Them Apart](/blog/guides/observational-vs-experimental-studies-how-to-tell-them-apart)
* [Astrocyte Single Cell Rna Seq](/blog/guides/astrocyte-single-cell-rna-seq)
* [Structural Genes](/blog/guides/structural-genes)


## References and Further Reading

- [NCBI Bookshelf: Molecular Biology of the Cell](https://www.ncbi.nlm.nih.gov/books/) , foundational reference for NGS chemistry and genomics.
- [EMBL EBI Training: NGS Analysis](https://www.ebi.ac.uk/training/) , structured courses covering every step of NGS analysis.
- [Galaxy Training Network: NGS Data Analysis](https://training.galaxyproject.org/) , hands on tutorials with real datasets.
- [Bioconductor: NGS Resources](https://bioconductor.org/) , R software packages for sequence analysis.
- [NCBI Sequence Read Archive](https://www.ncbi.nlm.nih.gov/sra) , public repository for raw NGS data.
- [Clinical Impact of 16S rRNA RC PCR NGS](https://pubmed.ncbi.nlm.nih.gov/42439573/) , demonstrates targeted NGS in infectious disease.
- [Long Read Sequencing for STRC Variants](https://pubmed.ncbi.nlm.nih.gov/42437880/) , shows value of long reads in hearing loss genes.
- [NGS in Vertebral Osteomyelitis](https://pubmed.ncbi.nlm.nih.gov/42441123/) , discusses diagnostic integration and limitations.
- [NGS for Gene Doping Detection](https://pubmed.ncbi.nlm.nih.gov/42438284/) , application in anti doping with cautionary notes.

## Related Articles

- [Protein Synthesis](/blog/guides/protein-synthesis)
- [Incomplete Dominance Gene](/blog/guides/incomplete-dominance-gene)
- [Cell Membrane Function Biology](/blog/guides/cell-membrane-function-biology)
- [Dna Structure](/blog/guides/dna-structure)
- [Protein Structure](/blog/guides/protein-structure)