Zubair Khalid

Virologist/Molecular Biologist | Veterinarian | Bioinformatician

Conventional & Molecular Virology • Vaccine Development • Computational Biology

Dr. Zubair Khalid is a veterinarian and virologist specializing in conventional and molecular virology, vaccine development, and computational biology. Dedicated to advancing animal health through innovative research and multi-omics approaches.

Dr. Zubair Khalid - Veterinarian, Virologist, and Vaccine Development Researcher specializing in Computational Biology, Multi-omics, Animal Health, and Infectious Disease Research

Category: Guides

Dna Sequencing

DNA sequencing is the process of determining the exact order of nucleotides in a DNA molecule. This guide is for biologists, clinicians, and bioinformaticians who need a practical, source bounded framework to understand and apply DNA sequencing. The core concept is straightforward: read the chemical letters of life. But the choice of method, the steps you take, and the quality controls you apply dramatically affect your results. Use this guide to navigate the decisions and avoid common pitfalls, drawing on authoritative resources like the NCBI Bookshelf [1] for foundational knowledge and the Galaxy Training Network [3] for practical workflows.

At its simplest, sequencing tells you the A, T, C, and G sequence of a sample. Whether you are identifying a mutation in a patient, profiling a microbial community, or mapping chromatin architecture, the same logical framework applies. The EMBL EBI Training portal [2] offers a structured introduction to the data types, while public repositories like the NCBI Sequence Read Archive [5] contain vast amounts of sequencing data for reference and reanalysis.

At a Glance

Aspect Details
What is DNA sequencing Determining the order of nucleotides (A, T, C, G) in a DNA molecule.
Main technologies Sanger (low throughput, long reads), Illumina (high throughput, short reads), PacBio / Oxford Nanopore (long reads, lower accuracy).
Key outputs Raw reads (FASTQ files), aligned reads (BAM files), consensus sequences, variant calls (VCF files).
Common applications Whole genome sequencing, exome sequencing, RNA seq, epigenetics, metagenomics, targeted panels.
Primary databases Sequence Read Archive (SRA), European Nucleotide Archive (ENA), GenBank.

Decision Criteria

Choosing a sequencing method depends on your biological question, budget, and required accuracy. The first decision is read length. Sanger sequencing offers up to 900 base pairs of high accuracy per read but low throughput. Illumina provides massive parallel output with 50 to 300 base pair reads and error rates below 1%. For resolving repetitive regions or structural variants, long read technologies like PacBio (accurate circular consensus reads) or Oxford Nanopore (real time sequencing) are essential.

Throughput matters. For a single gene, Sanger may suffice. For a human genome, you need a platform that generates hundreds of millions of reads. Consider the error profile. Illumina has very low per base error, but short reads struggle with homopolymers and GC rich regions. Long reads tolerate higher error but can be improved with consensus strategies. Cost per base is low for Illumina, but instrument cost is high. Bioconductor documentation [4] provides software tools for evaluating these trade offs computationally.

Also evaluate the need for paired ends. Paired end Illumina reads improve alignment across repeats. For structural variant detection, long reads or linked reads are superior. Each method has inherent biases. Review the EMBL EBI Training modules [2] on sequencing technology comparisons to align your goals with technique characteristics.

Practical Workflow

A standard DNA sequencing project follows these steps:

1. Sample preparation and library construction. Extract DNA, shear or fragment it, and ligate adapters. For short read Illumina, you need a library with size selection. For long read libraries, high molecular weight DNA is critical. Record all barcodes and indexes to avoid sample mix up. The Galaxy Training Network [3] offers step by step tutorials for library preparation quality control.

2. Sequencing run. Load your library onto the flow cell or nanopore array. Follow manufacturer protocols for cluster generation or pore loading. Monitor run metrics like cluster density or pore occupancy. Underclustering reduces yield, overclustering reduces quality.

3. Base calling and quality control. The instrument software converts raw signals to base calls and assigns quality scores (Phred scores). Examine per base quality, GC content, and adapter contamination. Tools like FastQC are standard. If you used a capture kit, check enrichment uniformity.

4. Alignment or assembly. Map reads to a reference genome (e.g., with BWA or minimap2) or perform de novo assembly. For bacterial genomes, short read assemblies produce many contigs. Long reads can produce a single contig. For metagenomics, careful taxonomic binning is required. Use Galaxy workflows [3] for reproducible pipelines.

5. Variant calling and annotation. Identify single nucleotide variants, insertions, deletions, or structural variants. Filter low quality calls. Annotate variants with effect prediction (e.g., SnpEff or VEP). For clinical applications, follow guidelines from the Sequence Read Archive [5] for minimum depth and allele frequency thresholds.

6. Interpretation. Compare results with known databases. For rare diseases, check allele frequencies in population datasets. For functional studies, confirm variants affect protein structure or regulation. A recent study on PRKAB1 regulation of phosphatidylcholine metabolism in IBD [8] demonstrates how multi omics integration with sequencing can uncover mechanistic links.

Common Mistakes

Insufficient coverage depth. A human genome requires 30x coverage for confident homozygous calls and more for low frequency variants. Always calculate needed coverage based on expected variant frequency and platform error rate.

Ignoring base quality scores. Phred scores are calibrated. Do not blindly accept all calls. Filter low quality bases and reads, especially for downstream biological interpretation.

Index and barcode errors. Multiplexing libraries requires careful index assignment. Index hopping (cross talk between samples) can occur on patterned flow cells. Use unique dual indexes to mitigate.

Adapter contamination. If adapters are not fully trimmed, they can cause false alignments or assembly breaks. Always perform adapter removal with tools like Cutadapt.

Assuming a perfect reference genome. Human reference lacks many population specific variants. In non model organisms, reference quality may be poor. This can lead to variant calling bias. Consider using pangenome references or long reads to resolve difficult regions.

Overlooking batch effects. If you sequence multiple batches, include technical replicates and use proper normalization. Batch effects can mimic biological differences.

Limits of Interpretation

Sequencing data is always an approximation. Sequencing by synthesis (Illumina) has trouble with highly repetitive sequences and extreme GC content. Long read technologies have higher per base error, especially in homopolymer runs. No method can sequence every base in the genome with absolute certainty. Structural variants in complex regions often require orthogonal validation.

Reference bias is another limitation. When aligning reads to a standard reference, you may miss novel sequences or large structural variants that differ from the reference. This is especially important in studies of disease associated structural variation. For example, mitochondrial DNA mutations in hepatocellular carcinoma [10] can be difficult to detect because of heteroplasmy and varying mitochondrial copy number.

Interpretation also depends on database completeness. Many variants are of unknown significance. A negative result does not prove absence of a variant. For metagenomics, short read approaches cannot easily resolve species level diversity in complex communities, and functional interpretation relies on annotation databases that may be incomplete. The Molecular Epidemiology study on thalassemia [6] shows how population specific variants require targeted detection methods, because standard short read panels may miss large deletions.

Finally, sequencing reveals genotype but not always phenotype. Variants may be in non coding regions with unknown effects. Protein structure prediction (see related guide [Protein Structure]) can help, but functional validation remains essential. Use multiple lines of evidence and public databases to bound your confidence.

Frequently Asked Questions

What is the difference between Sanger sequencing and next generation sequencing?
Sanger reads are longer (up to 900 bases) and highly accurate, but yield only one sequence per reaction. NGS (Illumina, Ion Torrent) produces millions of short reads in parallel, enabling whole genome assays. Use Sanger for targeted low throughput projects, NGS for large scale discovery.

How much sequencing coverage do I need?
Coverage depth depends on your goal. For human genome variant calling, 30x is standard. For detecting somatic mutations in cancer, 100x or more may be needed to see low allele fractions. For bacteria, 50x often suffices. Always sequence a positive control to confirm sensitivity.

Can I sequence RNA directly?
RNA sequencing typically involves reverse transcription to cDNA before sequencing. Direct RNA sequencing is available with Oxford Nanopore technology, but it remains less common. Most RNA seq data is derived from DNA sequencing of cDNA libraries.

How do I choose between short and long reads?
Short reads are best for base level accuracy, expression quantification, and small variant detection. Long reads excel at resolving repeats, phasing haplotypes, and detecting structural variants. If your genome has high repeat content or you need a complete assembly, include long reads.

References and Further Reading

Related Articles