The integration of next-generation sequencing (NGS) in the field of haematology has revolutionised diagnostic and monitoring capabilities, particularly for haematological malignancies such as leukaemia and lymphoma.
One of the critical aspects of utilising NGS effectively is understanding the relationship between sequencing depth and Variant Allele Frequency (VAF) sensitivity, which plays a significant role in accurately detecting genetic variants, especially in Measurable Residual Disease (MRD) detection.
Sequencing depth, also known as coverage, refers to the number of times a particular region of the genome is read during sequencing. In targeted NGS panels, such as those offered by OGT, sequencing depth is particularly important as it determines the ability of the panel to detect low-frequency variants.
It is worth noting that different analysis methods calculate the sequencing depth slightly differently, for OGT’s NGS analysis solution Interpret, we use quality filtered, mapped, de-duplicated unique reads over the region of interest. This approach is associated with higher quality reads less prone to error to provide the most accurate value.
Shallow sequencing reads each base a few times and is used in technologies such as Whole Genome Sequencing (WGS) which is useful for broader studies like population genetics. However, deep sequencing, an approach used in targeted panels, reads each base multiple times to detect rare variants, this is particularly important for haematological malignancies where clonal mutations may be present at low frequencies.
Variant Allele Frequency (VAF) is the proportion of sequence reads that contain a specific variant relative to the total number of reads at that position.
Basic calculation:
If a targeted NGS panel yields 1,000 reads for a given position and 50 of those reads show a variant, the VAF would be 5%.
Because of several factors, including the sequencing process being affected by sampling effect where, for instance, if your coverage is not very high, you may over or underestimate (or miss) certain variants, with a lower number of reads, we may not be sure that the reads that we observe are representative of all DNA fragments in the sample. Consider an example where you have a 1% VAF with only 100x coverage (i.e. 100 total reads at that position), in practice this means you have detected 1 read but it is entirely possibly that you could miss a single read call if your sampling of 100 reads over any specific base position is inaccurate. Instead, if we had higher number of reads, such as 100 reads with the variant and 10,000 reads in total (10000x coverage), we would be more confident that the 1% VAF calculation is accurate.
Higher sequencing depth improves VAF sensitivity, making it possible to detect low-frequency variants. In haematological malignancies, where detection of clonal mutations is crucial, we can increase the sequencing depth to allow for a more accurate detection of rare variants, including those present in a small percentage of cells.
However, there is a trade-off, while deeper sequencing improves VAF sensitivity, it also increases costs. Researchers and diagnostic labs must balance the need for sensitivity with their available budget.
While we are discussing specificity, we should also consider the specificity of next-generation sequencing. In this context, specificity is the ability to detect the exact targeted sequences in our assay. If your sequencing has a low specificity then we would expect a high rate of false positive calls for targeted regions (even with high sensitivity), that if undetected could lead to the misclassification of samples based on non-existent variants detected.
Many factors influence specificity and so some general considerations include:
Targeted NGS panels, such as those used for the detection and research of haematological malignancies, focus on specific genes known to be associated with the disease.
In the detection of MRD (the small number of cancer cells that may remain in a patient after they have received treatment) VAF sensitivity is crucial. Detecting MRD requires high sequencing depth to capture ultra low-frequency variants that indicate residual disease, which can then be used to help guide further decisions.
Check out OGT’s SureSeq™ Myeloid MRD Panel to see how we have developed an NGS panel capable of capturing ultra low-frequency variants
Technical aspects that influence VAF sensitivity and sequencing depth include:
Clinical context matters! Sequencing depth should be tailored to the specific clinical question. For example, an examination of germline variants where we may be examining a ~50% VAF or ~100% VAF would only require low coverage over the region of interest to sufficiently quantify VAF with confidence. Conversely for MRD we would aim for higher coverage to obtain the same level of confidence.
NGS users can also consider their panel design, including planning on how to successfully customise panels without sacrificing sensitivity. In these cases users could choose to add additional relevant targets or expand the coverage of existing regions while balancing this against the sequencing cost for the necessary depth to account for expected variants in these biomarkers. But the benefits of generating a customised targeted panel can include to reduce the sequencing and data analysis implications for unneeded biomarkers, so you can focus only on your specific regions of interest.
Understanding the relationship between sequencing depth and VAF sensitivity is essential for advancing diagnostics and research in haematological malignancies. By optimising these factors, clinicians and researchers can improve the detection of low-frequency variants, particularly in the context of MRD, leading to better patient outcomes and more targeted therapeutic strategies.
As one of the first steps in many NGS data analysis pipelines, accurate variant calling is often critical to downstream analysis and interpretation. Here, we take a look at variant calling best practice through a modern lens.
Read
A high-quality sequencing library is the linchpin to generating good sequencing data. We discuss our six top tips to help you improve your sequencing library.
Read
Visit USA site
Visit Canada site
