How to Calculate Sequencing Depth and Coverage for Your NGS Run

Sequencing depth and coverage are the two most important quantitative measures for planning and evaluating any next-generation sequencing (NGS) experiment. Depth refers to the average number of times a nucleotide position is read during sequencing, while coverage describes the proportion of a target region that is sequenced at a defined depth. This article explains how to calculate both metrics, how to interpret them for different applications, and how to use online calculators to plan your runs. The guidance is written for laboratory students, technicians, researchers, and diagnostic professionals who need practical formulas and worked examples.

At a Glance

The table below summarizes the core calculations and their typical applications across common NGS workflows.

Understanding Sequencing Depth and Coverage

Sequencing depth, often expressed as "×" or fold coverage, is the average number of times each base in a target region is represented by sequencing reads. If a genome position is covered by 30 independent reads, that position has a depth of 30×. Coverage, in contrast, describes the fraction of the target region that achieves a minimum depth. A run that produces 30× mean depth across 95% of the exome has a breadth of coverage of 95% at the 30× threshold.

These two metrics answer different questions. Depth tells you how much redundancy exists at each position, which directly affects the confidence of base calls and variant detection. Coverage tells you how much of your target region is actually usable for analysis. A run can have high mean depth but poor coverage if reads are concentrated in a small portion of the target, such as GC-rich regions that amplify unevenly during library preparation.

The distinction matters for diagnostic applications. The World Health Organization Laboratory Quality Management System Handbook emphasizes that laboratory results must be reliable and that quality assurance practices should be documented and traceable. For NGS diagnostics, this means recording depth and coverage metrics as part of the validation and reporting process. Without these metrics, a laboratory cannot demonstrate that a negative result is trustworthy or that a variant call has sufficient support.

The Sequencing Depth Formula

The fundamental formula for calculating mean sequencing depth is:

Mean depth = (Total number of mapped bases) ÷ (Size of the target region)

Total mapped bases is calculated by multiplying the number of mapped reads by the read length. For example, if you have 10 million mapped reads of 150 base pairs each, the total mapped bases are 1.5 billion. If your target region is the human exome at approximately 30 million base pairs, the mean depth would be:

1,500,000,000 ÷ 30,000,000 = 50×

This calculation assumes all reads map to the target region. In practice, a portion of reads will map outside the target, fail quality filters, or represent duplicate molecules. The on-target rate, typically 60% to 80% for exome capture and higher for amplicon panels, must be factored into run planning.

For whole genome sequencing, the target region is the entire genome. The human genome is approximately 3.1 billion base pairs. To achieve 30× coverage for a human genome with 150 base pair reads, you would need:

(3,100,000,000 × 30) ÷ 150 = 620,000,000 reads

This calculation assumes perfect mapping and no duplicates. A realistic plan would add 20% to 30% excess reads to account for losses.

Calculating Coverage Breadth

Coverage breadth is calculated by dividing the number of bases that achieve a specified depth threshold by the total number of bases in the target region, then multiplying by 100.

Coverage breadth (%) = (Bases at or above threshold ÷ Total target bases) × 100

For example, if your exome panel targets 30 million base pairs and 28.5 million of those bases are covered at 20× or higher, the coverage breadth at 20× is:

(28,500,000 ÷ 30,000,000) × 100 = 95%

Coverage breadth is the more clinically relevant metric for diagnostic reporting. A mean depth of 50× is meaningless if a critical exon is covered at only 5×. Most diagnostic laboratories report both mean depth and the percentage of target bases covered at a defined threshold. The threshold varies by application and by the analytical requirements of the test.

For copy number variation detection from whole exome sequencing, the depth of coverage is the primary signal used to identify deletions and duplications. The CNVind pipeline demonstrates that depth of coverage calculations, quality control, normalization, and CNV calling are sequential steps in this analysis. The normalization process is performed separately for each sequencing region, which improves specificity for rare CNV detection. This means that uneven coverage across the exome directly affects the reliability of CNV calls, and laboratories must document coverage metrics to interpret CNV results correctly.

Using Online Sequencing Coverage Calculators

Several online calculators are available to simplify depth and coverage planning. These tools typically ask for the genome size, read length, desired coverage, and expected on-target rate. The calculator then returns the number of reads required and the estimated sequencing output.

When using an online calculator, verify the assumptions built into the tool. Key inputs to check include:

• Genome or panel size in base pairs
• Read length and whether paired-end reads are counted as one or two reads
• Expected on-target rate for your library preparation method
• Expected duplication rate, which varies with input amount and PCR cycles
• Quality filter pass rate for the sequencer and chemistry

A common error is entering the total number of reads when the calculator expects read pairs. For paired-end sequencing, each fragment produces two reads. If a calculator asks for read pairs and you enter the total read count, your planned depth will be double the actual value.

Another common error is ignoring the duplication rate. Duplicate reads arise when the same DNA fragment is amplified during library preparation and sequenced multiple times. These reads are computationally removed before variant calling because they do not represent independent observations. A high duplication rate inflates the apparent depth while reducing the effective depth after deduplication. For low input samples, such as cell-free DNA from liquid biopsy, duplication rates can be substantial and must be accounted for in run planning.

Depth Requirements by Application

Different applications require different depths because the analytical goals differ. The table below summarizes typical depth requirements and the rationale for each.

The depth requirement for low coverage genotyping deserves special attention. Genotype imputation from low coverage sequence data has been shown to be accurate at very low sequencing depth when the population structure is known. The imputation algorithm uses marker data and population design information efficiently, allowing accurate genotype calling and error correction even with substantial missing data. This approach is cost effective for breeding programs and population studies where individual genotype confidence is less critical than population level accuracy.

For deep mutational scanning experiments, the relationship between sequencing depth and measurement precision follows a different logic. Sampling more time points and extending the duration of the experiment improves precision more than increasing sequencing depth alone. This finding means that for experiments tracking mutant frequencies over time, the experimental design should prioritize time point sampling over raw depth. The sequencing depth still matters, but it is not the limiting factor in the same way it is for clinical variant detection.

Tumor Mutational Burden and Depth Considerations

Tumor mutational burden (TMB) measurement illustrates how depth requirements interact with the analytical method. TMB is detected primarily by whole exome sequencing or targeted panel sequencing on high throughput platforms. The lack of uniformity in detection methods, threshold settings, and reporting formats has hindered standardized application of this biomarker in clinical practice.

For TMB calculation, the sequencing depth must be sufficient to call somatic variants with confidence across the coding region of interest. Low depth increases the chance of missing true variants or calling false positives from sequencing errors. The consensus recommendations for TMB emphasize that laboratories must understand the testing standards and limitations of their chosen method. This includes documenting the depth and coverage achieved for each sample and establishing quality thresholds for reporting.

The depth requirement for TMB also depends on the tumor content of the sample. Samples with low tumor purity require higher depth to detect variants present at low allele fractions. A sample with 30% tumor content and a somatic variant present in all tumor cells will have a variant allele fraction of approximately 15%. Detecting this variant confidently requires depth sufficient to observe multiple copies of the alternate allele, typically 100× or higher at that position.

Short Tandem Repeat Genotyping and Coverage Thresholds

Short tandem repeat (STR) genotyping from NGS data presents a different coverage challenge. STR detection remains limited in routine short read NGS workflows because repeat regions are difficult to map and genotype. Computational tools such as ExpansionHunter can genotype disease associated STR loci from exome or genome data, but the coverage at these loci is often lower than the average coverage of the run.

A large Turkish neurodegeneration cohort study demonstrated that lowering the default locus coverage threshold from 10× to 3× markedly increased genotyping rates at low coverage loci in exome data. This adjustment allowed detection of pathogenic and intermediate expansions that would have been missed at the higher threshold. The study confirmed that computational STR genotyping can add diagnostic value to routine NGS workflows, but the coverage threshold must be optimized for the specific loci and data type.

This example shows that depth requirements are not uniform across all genomic regions. Repetitive and GC-rich regions often have lower coverage than the genome average. Laboratories performing tests that depend on these regions must verify coverage at the specific loci of interest instead of relying on the overall mean depth.

Calculating Depth for Copy Number Variation Detection

Copy number variation (CNV) detection from NGS data relies on read depth as the primary signal. The depth of coverage in sequencing regions is calculated, quality controlled, normalized, and then used for CNV calling. The accuracy of CNV detection depends on the uniformity of coverage across the genome or exome.

The shortest path based approach for CNV detection calculates the k nearest neighbors of each read depth and defines a relative score for each position. This method demonstrates that CNV detection accuracy is limited by bias, noise, and alignment errors in NGS data. Coverage normalization is essential to correct for these artifacts before CNV calling.

For clinical CNV detection, the depth must be sufficient to distinguish true copy number changes from technical noise. The required depth depends on the size of the CNV and the purity of the sample. Large CNVs spanning many exons can be detected at lower depth, while small CNVs affecting a single exon require higher depth and more uniform coverage.

The limitations of exome sequencing for CNV detection are well documented. Exome sequencing has restricted sensitivity for copy number variants, suboptimal coverage of highly homologous or GC-rich regions, and limited detection of deep intronic variants. These limitations mean that a negative CNV result from exome data must be interpreted with caution, particularly for genes in difficult to sequence regions.

Practical Steps for Calculating Depth and Coverage

Follow these steps to calculate depth and coverage for your NGS run.

Step 1: Determine the target region size. For whole genome sequencing, use the genome size. For exome or panel sequencing, use the total size of the capture or amplicon design. This information is available from the manufacturer or from the reference genome annotation.

Step 2: Calculate the total number of mapped bases. Multiply the number of mapped reads by the read length. For paired end reads, count each read separately. If you have 20 million read pairs at 150 base pairs each, the total is 40 million reads multiplied by 150 base pairs, which equals 6 billion bases.

Step 3: Divide total mapped bases by the target region size to obtain mean depth. Using the example above, if the target is the exome at 30 million base pairs, the mean depth is 6,000,000,000 divided by 30,000,000, which equals 200×.

Step 4: Calculate coverage breadth at your reporting threshold. Use a coverage analysis tool such as GATK DepthOfCoverage, mosdepth, or the coverage metrics from your sequencing platform. Count the bases at or above the threshold and divide by the total target bases.

Step 5: Record the metrics in your laboratory information system. Include the mean depth, the coverage breadth at the reporting threshold, the number of reads, the on-target rate, and the duplication rate. These records support quality assurance and troubleshooting.

Step 6: Compare the results to your validation data. If the coverage metrics fall below the validated thresholds, the run may need to be repeated or the sample reported with limitations.

Records and Measurements for Quality Assurance

Documentation of depth and coverage metrics is a core component of laboratory quality assurance. The World Health Organization Laboratory Quality Management System Handbook describes the requirements for a quality management system in medical laboratories, including documentation of test performance and quality control. For NGS, this documentation must include the sequencing metrics that support the analytical validity of the test.

The following records should be maintained for each sequencing run:

• Run identifier and date
• Instrument and reagent lot numbers
• Sample identifiers and library preparation method
• Total reads generated and reads passing quality filters
• Mapped reads and on-target rate
• Mean depth and coverage breadth at reporting thresholds
• Duplication rate
• Any run failures or anomalies

These records allow the laboratory to monitor performance over time and detect drift in the sequencing process. A gradual decline in on-target rate or coverage uniformity may indicate a problem with capture reagents, library preparation, or instrument performance. Early detection of these trends prevents reporting of unreliable results.

For diagnostic applications, the coverage metrics should be reviewed before results are released. A sample with coverage below the validated threshold should be flagged for review. The laboratory director or designated reviewer should determine whether the sample can be reported with limitations or must be repeated.

Common Failure Patterns in Depth and Coverage

Several recurring problems affect depth and coverage in NGS runs. Recognizing these patterns helps with troubleshooting and prevention.

Uneven coverage across GC-rich regions is a common failure pattern. Library preparation methods that use PCR amplification can bias against GC-rich and GC-poor regions. This bias reduces coverage in these regions and can create gaps in clinically important genes. Mitigation strategies include using PCR free library preparation, optimizing the capture design, and using enzymes that reduce GC bias.

Low on-target rate reduces effective depth. If a large fraction of reads map outside the target region, the mean depth will be lower than planned. Causes include poor capture efficiency, contamination with non target DNA, and incorrect hybridization conditions. The on-target rate should be monitored for each run and compared to the expected range for the library preparation method.

High duplication rate reduces effective depth after deduplication. Duplicates arise from PCR amplification of the same fragment. High duplication rates are more common with low input DNA, high PCR cycle numbers, and degraded samples. The duplication rate should be calculated and reported, particularly for samples with limited DNA.

Coverage dropouts in specific regions can occur even when the overall metrics are acceptable. These dropouts may affect a single exon or a repetitive region. For diagnostic tests, the laboratory must verify coverage at the clinically relevant positions, beyond the overall metrics. This verification is particularly important for genes with pseudogenes or highly homologous regions, where reads may map incorrectly.

Welfare and Safety Context for Laboratory Work

Laboratory safety is a separate but essential consideration for NGS workflows. The World Health Organization Laboratory Biosafety Manual provides guidance on biosafety practices for laboratories handling biological materials. NGS workflows involve multiple steps where exposure to biological samples, chemicals, and equipment can occur.

Sample preparation involves handling blood, tissue, or other clinical specimens that may contain infectious agents. Standard precautions should be followed, including the use of gloves, laboratory coats, and eye protection. Sample processing should occur in a designated area with appropriate containment for the risk level of the specimens.

Library preparation involves the use of enzymes, buffers, and reagents that may be hazardous. Material safety data sheets should be reviewed for each reagent, and appropriate personal protective equipment should be worn. PCR amplification steps may involve handling of amplified DNA, which is a contamination risk for other samples.

Sequencing instruments contain lasers, high voltage components, and moving parts. Instrument operation should follow the manufacturer's safety instructions. Maintenance and troubleshooting should be performed by trained personnel with the instrument powered down and locked out as appropriate.

Chemical waste from NGS workflows, including used reagent cartridges and flow cells, should be disposed of according to institutional and local regulations. Biological waste should be decontaminated before disposal.

Professional Escalation Criteria

Laboratory personnel should know when to escalate problems with depth and coverage to a supervisor or laboratory director. The following situations warrant escalation:

• Mean depth falls below the validated threshold for the test
• Coverage breadth at the reporting threshold falls below the validated threshold
• A clinically relevant region has zero coverage or coverage below the threshold for confident variant calling
• The duplication rate exceeds the expected range for the library preparation method
• The on-target rate falls below the expected range
• Coverage metrics show a trend of decline over multiple runs
• A sample fails coverage metrics and cannot be repeated due to insufficient DNA

When escalating, provide the run records, the coverage metrics, and the relevant quality thresholds. This information allows the supervisor to determine whether the run can be reported, must be repeated, or requires investigation of the underlying cause.

For diagnostic samples, the decision to report a result with suboptimal coverage should be made by the laboratory director or a designated qualified professional. The limitations of the result should be documented and communicated to the requesting clinician when appropriate.

Limitations of Depth and Coverage Calculations

Depth and coverage calculations have inherent limitations that affect their interpretation. Mean depth is sensitive to outliers. A few regions with very high coverage can inflate the mean, masking poor coverage elsewhere. The median depth and the distribution of depth across the target should be reviewed alongside the mean.

Coverage breadth depends on the chosen threshold. A sample can have 99% coverage at 10× but only 90% coverage at 30×. The reporting threshold should be selected based on the analytical requirements of the test and validated during test development.

Depth calculations assume that reads are independent observations. Duplicate reads violate this assumption because they represent the same DNA fragment. After deduplication, the effective depth is lower than the raw depth. Laboratories should report both raw and deduplicated depth when relevant.

Coverage calculations depend on the reference genome and the alignment method. Reads that map to multiple locations, such as reads from repetitive regions, may be excluded or assigned ambiguously. This affects the coverage at these regions and can create false gaps or false coverage.

For low depth sequencing, the uncertainty in genotype calls is higher. Methods that estimate relatedness from low depth whole genome sequencing data use genotype likelihoods instead of called genotypes to account for this uncertainty. This approach provides accurate relatedness estimates while being robust and computationally efficient. Laboratories performing low depth studies should use analysis methods that account for genotype uncertainty instead of relying on hard genotype calls.

Interpreting Depth for Species Identification and Metagenomics

Depth and coverage calculations also apply to species identification and metagenomic analysis. A comparative genomic approach for distinguishing closely related species uses species specific regions that exhibit distinctive coverage patterns. These regions show average coverage when samples are aligned to their matching reference genome but absent or low coverage when aligned to non matching references.

This approach demonstrates that coverage patterns can be used as a diagnostic signal beyond simple depth calculations. For species identification, the presence or absence of coverage at curated species specific regions provides a binary classification that is more reliable than overall alignment rates. The method achieved high accuracy in validation datasets and can be extended to other closely related species.

For metagenomic samples, depth calculations are complicated by the mixed composition of the sample. The depth at each taxonomic group depends on the abundance of that organism in the sample and the efficiency of sequencing. Coverage calculations for metagenomes require assembly or reference based mapping to determine which reads belong to which organism.

Depth Considerations for Digital PCR and Non Invasive Prenatal Testing

Digital PCR and NGS have different depth requirements for the same clinical question. A fetal fraction optimized digital PCR assay for non invasive prenatal testing detected trisomies at fetal fractions of 3% or higher using 5 nanograms of DNA. The assay measured fetal fraction irrespective of fetal sex and achieved high sensitivity and specificity in a large sample set.

For NGS based non invasive prenatal testing, the depth requirement is driven by the need to detect small differences in chromosome representation. Trisomy 21 samples have a 50% increase in chromosome 21 reads relative to disomic samples, but the fetal fraction dilutes this signal. A sample with 10% fetal fraction has only a 5% increase in chromosome 21 reads. Detecting this difference requires high depth and careful statistical analysis.

The comparison between digital PCR and NGS for this application shows that the optimal technology depends on the specific clinical question and the available resources. Digital PCR can achieve high sensitivity with lower sequencing depth because the assay targets specific loci. NGS provides broader coverage but requires higher depth to achieve the same analytical sensitivity for rare variants.

Frequently Asked Questions

What is the difference between sequencing depth and coverage?

Sequencing depth is the average number of times each base in the target region is read. Coverage, also called breadth of coverage, is the percentage of the target region that achieves a specified depth threshold. A run can have high depth but low coverage if reads are concentrated in a portion of the target.

How do I calculate sequencing depth from my run data?

Divide the total number of mapped bases by the size of the target region. Total mapped bases equals the number of mapped reads multiplied by the read length. For example, 10 million mapped reads at 150 base pairs gives 1.5 billion bases. Dividing by a 30 million base pair exome gives 50× mean depth.

What depth do I need for whole genome sequencing?

For human whole genome sequencing, 30× is a common target for germline variant detection. This depth balances cost with the ability to call heterozygous variants confidently. Lower depth can be used for population studies with imputation, while higher depth is needed for somatic variant detection in tumor samples.

How does the on-target rate affect my depth calculation?

The on-target rate is the fraction of reads that map to the target region. If your on-target rate is 70%, you need approximately 43% more total reads to achieve the same mean depth as a run with 100% on-target rate. Plan your read count using the expected on-target rate for your library preparation method.

What is a good coverage breadth for a diagnostic exome test?

Most diagnostic exome tests aim for at least 95% of the target region covered at 20× or 30×. The specific threshold depends on the test requirements and the validation data. Coverage below the validated threshold at clinically relevant positions should be flagged for review.

How do duplicate reads affect my depth calculation?

Duplicate reads are copies of the same DNA fragment and do not represent independent observations. They inflate the raw depth but are removed before variant calling. The effective depth after deduplication is lower than the raw depth. High duplication rates reduce the value of additional sequencing.

Can I use an online calculator to plan my sequencing run?

Online calculators can estimate the number of reads needed for a target depth. Verify the assumptions built into the calculator, including genome size, read length, on-target rate, and duplication rate. Enter read pairs instead of total reads when the calculator asks for read pairs.

Why is my coverage lower in GC-rich regions?

Library preparation methods that use PCR can bias against GC-rich and GC-poor regions. This bias reduces coverage in these regions and can create gaps in clinically important genes. PCR free library preparation, optimized capture designs, and enzymes that reduce GC bias can improve uniformity.

Related Diagnostic Guides

• Calibration Report: How to Document and Interpret Calibration Results
• NanoDrop DNA Concentration: Formula and Calculation
• How to Calculate the Specific Activity of an Enzyme: Formula and Examples
• How to Calculate and Interpret qPCR Efficiency from a Standard Curve
• How to Interpret Blue-White Screening Results in Cloning

References and Further Reading

• Laboratory Quality Management System Handbook. World Health Organization.
• Laboratory Biosafety Manual. World Health Organization.
• Assay Guidance Manual. National Center for Advancing Translational Sciences.
• Bioanalytical Method Validation Guidance. U.S. Food and Drug Administration.
• NCBI Literature Resources. National Center for Biotechnology Information.
• Moment estimators of relatedness from low-depth whole-genome sequencing data.. BMC bioinformatics, 2022.
• Super-Resolution Imaging Through the Human Skull.. IEEE transactions on ultrasonics, ferroelectrics, and frequency control, 2020.
• CNVind: an open source cloud-based pipeline for rare CNVs detection in whole exome sequencing data based on the depth of coverage.. BMC bioinformatics, 2022.
• Refractive changes after vitrectomy and phacovitrectomy for macular hole and epiretinal membrane.. Journal of cataract and refractive surgery, 2013.
• A shortest path-based approach for copy number variation detection from next-generation sequencing data.. Frontiers in genetics, 2022.
• Decoding Complexity in Synthetic Oligonucleotides: Unraveling Coeluting Isobaric Impurity Ions by High Resolution Mass Spectrometry.. Analytical chemistry, 2024.
• A Statistical Guide to the Design of Deep Mutational Scanning Experiments.. Genetics, 2016.
• Accurate Genotype Imputation in Multiparental Populations from Low-Coverage Sequence.. Genetics, 2018.
• Clinical Utility of NGS-Based Diagnosis in Primary Ciliary Dyskinesia: Experience from a Brazilian Pediatric Cohort at a Reference Center for Rare Diseases.. 2026.
• Real-World Evaluation of Uromonitor&lt,sup&gt,®&lt,/sup&gt, for Bladder Cancer Detection and Surveillance.. 2026.
• Computational Short Tandem Repeat Genotyping Reveals Clinically Relevant Expansions in a Large Turkish Neurodegeneration Disease Cohort.. 2026.
• A genomic approach for accurate identification of closely related species with next-generation sequencing samples.. 2026.
• A Fetal Fraction Optimized 106-Plex Digital PCR Assay for Non-Invasive Prenatal Testing of Fetal Trisomy.. 2026.
• Expert consensus on the detection and clinical application of tumor mutational burden.. 2026.
• MPCI: A novel metric for quantifying DNA methylation patterns in NGS data.. 2026.
• Identification and Characterization of a Novel Bovine Adenovirus Which Represents a Distinct Evolutionary Branch.. 2026.
• Intraocular lens calculation formula selection for short eyes: based on axial length and anterior chamber depth. BMC Ophthalmology, 2025.
• A simple calculation formula for the insertion depth of catheter of the central venous access port in Chinese patients. Heliyon, 2024.
• Review on Applicability of Local Scour Depth Calculation Formula in River. 2019.
• Effect of anterior chamber depth on the choice of intraocular lens calculation formula. PLoS ONE, 2017.
• Effect of capsular tension ring implantation on intraocular lens calculation formula selection for long axial myopia. BMC Ophthalmology, 2024.
• LOcating Non-Unique matched Tags (LONUT) to Improve the Detection of the Enriched Regions for ChIP-seq Data. Plos One, 2013.

This article is educational and does not replace validated laboratory procedures, institutional biosafety review, manufacturer instructions, or professional interpretation.