Depth and Coverage – What Is the Difference?

During the sequencing process and the subsequent bioinformatic analysis, different parameters can be captured. Two of these parameters are the depth and the coverage. These two terms are often used interchangeably in the context of sequencing, although they are generally not the same. It becomes even more complicated as – dependent on the context – they can describe sometimes similar and sometimes rather different things.

The coverage, also known as genomic coverage or coverage of the sequencing, describes the percentage of the genome or intended target region that is sequenced to a certain threshold. For some applications, the genomic coverage is calculated based on the target region being covered at least once. For others, the target region needs to be covered at least 30 times. In the example in figure 1, the coverage of the target region is 95%; thus, 5% of the reference sequence is not covered by at least one read. The genomic coverage is an important parameter to ensure that the entirety, or at least as much as possible, of the target region is sequenced at least once. It is calculated as the fraction of the number of nucleotides covered by reads and the number of nucleotides in the reference area.

The genomic coverage can be influenced by

• the quality of the DNA sample,
• the library preparation,
• sequencing bias,
• a high GC content,
• repetitive elements
• or other genomic complexities.

In contrast, the sequencing depth, also known as read depth, sequencing coverage, or depth of coverage, describes how often a specific base in the reference is read during the sequencing process. Based on these numbers per base, an average sequencing coverage – or sequencing depth – is calculated. Thus, it needs to be distinguished whether the sequencing coverage is given for a certain position in the reference or as an average over the whole reference. In general, the sequencing coverage is given as an average, for example, 50x. This means that on average, each base of the reference is read 50 times. The average depth of the sequencing coverage can theoretically be calculated by C=LN/G, where L is the read length, N is the number of reads, and G is the haploid genome length. In the example in figure 2, the average sequencing coverage is 2.36x.

The average sequencing coverage and the genomic coverage are connected. However, a high average sequencing coverage does not guarantee a high genomic coverage. Challenging regions in the genome, such as repetitive elements or regions with a high GC content, remain challenging, even with a higher sequencing depth.

The depth of the coverage can be affected

• by the accuracy of the genome alignment algorithms,
• by the uniqueness of the sequencing reads,
• and by how well they can be mapped to the target genome.

Errors in sequencing are common. Increasing the sequencing depth helps to distinguish actual variants from sequencing errors. Sequencing errors themselves are not reduced by higher sequencing depths, but they can be distinguished more easily from actual variants. Thus, by increasing the sequencing depth, the confidence in calling a variant at a specific location increases. Especially for rare variants with low allele frequencies, rare alleles, or for very heterogenous samples, such as tumor tissue, a high sequencing depth and increased confidence is essential to distinguish the variants from sequencing errors. Thus, the sequencing coverage level often determines whether variants can be discovered with a certain degree of confidence at a specific base position.

Table 1 | Comparison of coverage and depth. For both terms, synonyms, a definition, an example, the purpose, and a graphical example are provided.