Josh P. Roberts has an M.A. in the history and philosophy of science, and he also went through the Ph.D. program in molecular, cellular, developmental biology, and genetics at the University of Minnesota, with dissertation research in ocular immunology.
Clinical diagnostic applications continue to push the limits of different technologies by demanding greater sensitivity and quantitation of often precious or limited-quantity samples. The search for the rare mutation—whether uncovering a translocation or an extra copy of a gene, for minimal residual disease (MRD) monitoring or to search for a variant pathogen from a liquid biopsy, or to decide on potential therapies—has driven innovative development of PCR instruments and assays alike. Although quantitative PCR (qPCR) using hydrolysis probes will likely continue to dominate the field for some time to come, digital PCR (dPCR) has shown its worth, especially when it’s important to identify if there’s a rare mutation in the sample and to quantify how much of it is there.
What is a rare mutation?
Ask different people what the term “rare mutation” means, and you’re likely to get different answers. To some, it’s the frequency in the population—a person may carry only that mutated allele (or is heterozygous or a mosaic for it), but it’s considered rare because very few people have that same mutation.
Another meaning of “rare” is the relatively low-level presence of a mutated sequence in a given organism. It could be rare in the first sense, but it’s also possible that many individuals might have that self-same mutation. For example, “there are certain very, very hot-spot mutations that give rise to cancer,” explains Lance Wakida, senior product manager for digital PCR systems at Thermo Fisher Scientific. “So in the world of breast cancer, there are a small set of mutations—like a particular exon in EGFR—that occur with some regularity among that cancer type.”
The term has different quantitative interpretations, as well. To a sequencing person, given that error rates are in the 1% to 1.5% range, rare is anything occurring at less than 5%. If you’re talking about qPCR, that number might be less than one in one thousand, says S. Roopom Banerjee, president and CEO of RainDance Technologies. But “the reality is that we see sequences that are scientifically relevant even out at one in one million—latency in HIV is a great example of that.”
There are two principal considerations when using PCR to look for a rare mutation, notes Tania Nolan, founder and CEO of the molecular biology technical consultancy The Gene Team. The first is “sensitivity, so that you are detecting your rare mutation as soon as you possibly can—you might be detecting only one copy of it.” The second is “specificity, so that you can distinguish between your mutant and your wild type.” These are often interconnected, in that sensitivity and specificity are (at least in part) dependent on both the instrumentation technology and the chemistry of the assay itself.
Quantitative real-time PCR is so called because (unlike the predecessor end-point instruments) these isystems measure the amount of product that accumulates as the reaction progresses and double-stranded DNA is formed. By running a standard curve of known concentration alongside the experimental sample, the amount of sample can itself be quantitated. Compared with dPCR, the instrumentation is nearly ubiquitous, is quick and easy to use, has a very wide dynamic range and boasts a low cost per reaction.
Yet, the reliable lower limit of detection (LoD) of a qPCR instrument is typically considered to be between 10 to 100 copies. This has perhaps begun to change. Some newer systems, such as the PCRMax Eco 48 with advanced thermal control and optical systems “will reliably pick up a single copy of something within the reaction mixture,” claims Andrew Birnie, business development manager for PCRMax.
Whether that single copy is truly the mutant allele, though, may depend on the background sequence you’re trying to distinguish. If you’re looking for a single base change in the middle of a run of five guanine (Gs), for example, “it’s highly unlikely that a normal qPCR probe would be able to tell the difference at all,” cautions Nolan. In such difficult instances, Nolan recommends playing with the primer and probe sequences. You can use modified bases, such as locked nucleic acids (LNAs), which are very effective at altering oligo sequences, and you can design LNAs to be highly specific. LNAs and design tools for LNAs are available from Euorgentec, Exiqon, Integrated DNA Technologies and Sigma-Aldrich, as well as other tool providers.
Not every PCR reaction requires both primer and probe, and about 50% use intercalating dye to detect duplex DNA, says Banerjee. “I would expect that some of the non-probe-based approaches will continue to get traction.”
How many copies?
Because qPCR optimally doubles the number of DNA duplexes during each cycle, it can in theory discriminate a two-fold difference in the number of target copies initially present in a sample.
With current dPCR instrumentation, a sample is partitioned into thousands or millions of individual reactions, which are allowed to go to completion. Each partition is scored either negative (meaning no target template was present) or positive (a template was present). This conveys several benefits to dPCR, relative to qPCR, to find rare mutations. With dPCR, as samples are diluted out, there is less competitive inhibition from wild-type sequences, and quantitation is more absolute (no standard curve is necessary). Discrimination of fold-difference is based on the percentage of positive partitions rather than at which cycle a well was deemed positive relative to background.
dPCR can be used to detect copy number disorders like Fragile X syndrome, or to follow the accumulation or loss of resistance and driver mutations in cancer from a blood sample. For example, in a recent Nature paper, researchers used Bio-Rad’s Droplet Digital PCR to show a link between schizophrenia risk and variation in the copy number of diverse alleles of a complement gene, implicating the complement cascade in the development of the disease, points out Carolyn Reifsnyder, marketing manager at Bio-Rad’s Digital Biology Center [1].
What about the assays?
Currently, millions of pre-dsigned qPCR assays exist, notes Wakida.
But “if it’s only primers and probes, then it will work in digital. If the assay is already in reaction buffer as well, then that’s when you could have problems putting it in a digital instrument,” says Nolan. And “you might have the same problems going from a Roche to a Qiagen” qPCR instrument, Nolan shares.
Independent of the specific platform used, some optimization of the protocol is required. qPCR and dPCR both have pros and cons, and many labs will go back and forth depending on what kind of work they’re doing. For routine screening for rare mutations, attributes like cost per sample, preparation time or throughput might make qPCR more attractive. Or maybe the need to finely discriminate copy number necessitates a digital instrument. If you’re trying to decide which one to purchase, though, make sure you do your homework not only as to which category, but which instrument within that category, best suits your needs.
Reference
[1] Sekar, A, et al., “Schizophrenia risk from complex variation of complement component 4,” Nature, 530:177-83, 2016. [PubMed ID: 26814963]
Image: Shutter Stock images