Targeted sequencing is a DNA sequencing strategy that focuses on specific, pre-selected regions of the genome rather than reading the entire thing. Instead of sequencing all 3 billion base pairs of human DNA, it zeros in on dozens to thousands of genes already known to be relevant to a particular disease or trait. This makes it faster, cheaper, and capable of reading those selected regions at much greater depth, which is critical for catching rare or low-frequency mutations.
How Targeted Sequencing Works
The core idea is enrichment: before sequencing begins, the DNA regions of interest are physically separated or amplified from the rest of the genome. A sample of DNA is collected, broken into fragments, and then only the fragments matching the target regions are kept for sequencing. Everything else is discarded. The result is a concentrated pool of DNA from just the genes or regions that matter for the question at hand.
This enrichment step is what distinguishes targeted sequencing from whole genome sequencing (WGS), which reads everything, and whole exome sequencing (WES), which reads all protein-coding regions. Targeted panels are narrower still, covering anywhere from a handful of genes to a few hundred, depending on the application.
Two Main Enrichment Methods
There are two primary ways to isolate the DNA regions you want: hybridization capture and amplicon-based sequencing. Each has trade-offs that matter in different clinical and research settings.
Hybridization Capture
In this approach, genomic DNA is first broken into small fragments using high-frequency sound waves (sonication). Synthetic DNA probes that are complementary to the target regions are then mixed with those fragments. The probes bind to matching sequences, and magnetic beads pull the probe-target complexes out of solution, leaving everything else behind. The captured fragments are then sequenced.
Hybridization capture produces more uniform coverage across target regions, meaning each region gets read a similar number of times. This matters because uneven coverage can cause some areas to be read too shallowly to detect mutations. The downside is that it works best on fragments around 500 base pairs long, which can make it harder to capture closely spaced regions like some human exons.
Amplicon-Based Sequencing
This method uses PCR (a standard DNA copying technique) with primers designed to amplify only the target regions. It’s simpler to set up, requires less starting DNA, and tends to have a higher on-target rate, meaning a greater percentage of the sequencing reads actually land on the intended regions rather than being wasted on off-target DNA.
The trade-off is consistency. Amplicon methods are more prone to missing variants that other approaches catch, particularly near the edges of reads or in regions with limited coverage. Some of these missed calls appear to stem from the amplification process itself introducing artifacts. For detecting mutations present at low frequencies, amplicon-based panels can reliably find variants down to about 5% variant allele frequency. Pushing below that threshold typically requires adding unique molecular identifiers, short tags that help distinguish true mutations from errors introduced during amplification.
Why Depth of Coverage Matters
One of targeted sequencing’s biggest advantages is depth: because you’re reading a small slice of the genome, you can afford to read it hundreds or thousands of times over. This repetition, called coverage depth, is what makes it possible to detect mutations that exist in only a small fraction of cells in a sample.
For inherited (germline) mutations, which are present in every cell, moderate coverage is sufficient. Detecting mutations in tumors is harder because cancer samples are a mixture of normal and tumor cells, and the mutation might only appear in a subset of the tumor cells. Clinical guidelines call for at least 80x unique coverage across 95% of targeted regions to confidently detect a mutation present at 10% frequency. The matched normal sample (used to filter out inherited variants) needs at least 60x coverage. Multi-center validation studies have shown that well-designed panels achieve over 96% concordance for variants above 0.5% frequency.
Clinical Uses in Cancer and Beyond
Targeted sequencing panels have become a standard tool in oncology. Over the past decade, drugs that target specific mutations have improved survival for patients whose tumors carry those mutations. Panels test for these “actionable” mutations across many genes simultaneously, replacing the older approach of testing one gene at a time.
In a study of 394 patients with advanced solid tumors at a tertiary care center, sequencing panel results directly influenced treatment decisions for about 16% of patients. Of those whose treatment changed, 60% enrolled in a clinical trial matched to their mutation, 21% received an approved drug, 6% got off-label therapy, and 13% avoided a treatment that wouldn’t have worked. Among patients who received mutation-matched treatment, 71% received targeted therapy.
Beyond oncology, targeted panels are used for hereditary disease screening (testing known genes for conditions like inherited heart disease or familial cancer syndromes), prenatal genetics, pharmacogenomics (predicting how someone will respond to a medication), and infectious disease identification.
Practical Advantages Over Whole Genome Sequencing
Cost is the most obvious advantage. Sequencing an entire human genome generates roughly 90 gigabytes of raw data. A targeted panel covering a few hundred genes generates a fraction of that, which translates directly into lower sequencing costs, less data storage, and faster analysis. For large studies or clinical labs processing hundreds of samples, this difference is substantial.
Targeted sequencing also allows heavy multiplexing. Because each sample produces relatively little data, many samples can be pooled onto a single sequencing run. Researchers have demonstrated pooling at least 95 individually barcoded samples for simultaneous target capture without meaningful loss in efficiency. By combining internal barcodes with a second barcode added during amplification, the number of samples per run becomes nearly unlimited. This is what makes large-scale screening programs economically feasible.
Library preparation, the process of getting DNA ready for sequencing, has also been streamlined for targeted approaches. While older protocols were labor-intensive and required large amounts of starting DNA, newer methods can be completed in a single day without gel purification steps.
What Targeted Sequencing Can Miss
The focused nature of targeted panels is both their strength and their limitation. You can only find what you look for. If a patient’s disease is caused by a mutation in a gene not included on the panel, it will be invisible. Novel genes, unexpected structural rearrangements, and mutations in non-coding regions (the vast stretches of DNA between genes) are all outside the scope of most targeted panels.
Even within targeted regions, certain types of variation are difficult to detect. Short-read sequencing technology, which most panels rely on, struggles with large structural variants, repeat expansions, DNA methylation changes, and regions with high sequence similarity or unusual GC content. These technical blind spots contribute to the fact that more than half of patients with rare genetic diseases remain undiagnosed even after standard sequencing.
Long-Read Targeted Sequencing
Newer long-read sequencing platforms from Oxford Nanopore and PacBio are being adapted for targeted work, aiming to combine the cost efficiency of targeted approaches with the ability to read through complex genomic regions. Several enrichment strategies are in use: long-range PCR with target-specific primers, hybridization capture adapted for longer fragments, and CRISPR-Cas9 systems that cut DNA at precise locations flanking a region of interest.
One particularly novel approach is adaptive sampling, which is computational rather than biochemical. As DNA passes through the sequencing instrument in real time, software evaluates whether each fragment matches a target region. Non-target reads are ejected from the pore, and only target reads are fully sequenced. This eliminates the need for any physical enrichment step before sequencing, though it currently requires significant computational resources running alongside the sequencer.
Choosing Between Targeted and Broader Sequencing
The choice depends on what you need to find. If the question is well-defined, such as “does this tumor carry a mutation that matches an available drug,” a targeted panel is the most efficient tool. It delivers deep, reliable coverage of known genes at lower cost and faster turnaround. If the question is open-ended, such as “what is causing this patient’s undiagnosed condition,” whole genome or whole exome sequencing casts a wider net, trading depth for breadth. In practice, many clinical workflows start with a targeted panel and escalate to broader sequencing only if the panel comes back negative.