Q: What are some of the differences between single-cell sequencing and standard next generation sequencing [NGS]?
A: In single-cell sequencing, the first step is isolating the cells. There are a number of different technologies, like flow sorting, laser capture microdissection, micromanipulation, and droplet technology, and choosing which one to use depends on whether the cell you are trying to isolate is very rare or whether you are trying to randomly sample a large number of cells from a population. Some techniques, like flow sorting, are well established. Others, like laser capture microdissection, are in active development, as the laser can fragment the DNA or RNA in the cell. The second step is amplifying the DNA or RNA using either whole genome amplification or whole transcriptome amplification. The third step is the sequencing reaction, which involves standard NGS platforms. However, the analysis of the data is much more challenging in single-cell sequencing when compared with standard NGS because there are lots of technical errors—variations in the coverage performance, in the variant calling, and in post-processing.
Q: What are some of the challenges with using single-cell sequencing?
A: For single-cell RNA sequencing, we can generate thousands and thousands of profiles on single cells in a very rapid time frame on various systems, but the challenge lies in analyzing the data [that] tend to be noisy due to the amplification artifacts and the transcriptomes. For single-cell DNA sequencing, the challenge lies in trying to scale it up, because there aren’t good techniques or protocols to sequence thousands of single-cell DNA profiles at a low cost [and] in a short time frame.
Q: When do you use single-cell versus standard sequencing?
A: I think it completely depends on the [situation], and in many cases single-cell sequencing isn’t really needed. With single-cell sequencing, you can profile only a limited number of cells, but when sequencing bulk tissue, you are profiling millions of cells. If you want to know whether a mutation exists broadly across lots of cells, then you might prefer using a deep sequencing approach. However, if you want to know what combination of mutations are present in every cell, then you are not going to get that information from bulk sequencing. The main applications for single-cell sequencing are for resolving complex sub-populations, such as intra-tumor heterogeneity. The other major application is to study rare sub-populations that are responsible for disease progression. For instance, with cancer stem cells or circulating tumor cells, where you can isolate only a few cells, single-cell sequencing becomes very important for genomic profiling.
Q: Can you elaborate on some of the emerging applications of single-cell sequencing?
A: We are really interested in connecting genotypes and phenotypes. By isolating the DNA and RNA from the same cell, you can see how a mutation affects the overall transcriptional profile or cell state. Another area of interest is connecting imaging data from single cells, such as morphology measurements, to genomic data. With imaging, you can see how a cell migrates or interacts with other cells, and genomic analysis helps you understand the mutations or expression changes that are causing such behaviors. Multimodal measurements, which involve profiling the DNA, RNA, or epigenome from the same cell in a way that is accurate and reliable, is also gaining interest. There are a lot of translational applications for single-cell sequencing in developmental biology, neurobiology, and microbiology as well. One important application is for prenatal genetic diagnosis, especially for in vitro fertilization. If parents are disease carriers, you can screen one or two cells from the blastocyst and use the ones for implantation that do not carry the risk alleles.
Q: Do you see single-cell sequencing being used as a routine clinical or diagnostic tool in the near future?
A: We are working on getting single-cell sequencing tools to be accurate, reproducible, cost effective and have high throughput so we can get them into the clinic. We are interested in applying it for noninvasive monitoring of circulating tumor cells in blood samples, trying to replace tumor biopsies that are invasive. You can take lots of blood samples over the course of a therapy and the genomic information from the circulating tumor cells can be used to monitor how patients are responding and determine which mutations are sensitive and which are resistant. We are also using single-cell sequencing to diagnose clinical samples and measure tumor heterogeneity to help predict certain properties of the tumor, such as tumor resistance to therapy and metastasis. A lot of these applications will be used in the clinic very soon, hopefully within the next one or two years. In terms of costs, single-cell sequencing is still a bit expensive. However, the costs are coming down very rapidly, especially with single-cell RNA sequencing, where you can now sequence a cell for about $1. With DNA sequencing, it is still a bit more expensive, about $30–$40 a cell.
Shrikant Mane, PhD, director of the Yale Center for Genome Analysis and Keck Proteomics Laboratory at Yale University School of Medicine, discusses the opportunities and challenges in next-generation sequencing. With the development of single-cell sequencing techniques, improvement in data informatics, and falling costs of analysis, he shares much optimism about the continued demand and growing applications for sequencing in the coming years.
Q: For what types of applications are you using nextgeneration sequencing (NGS)?
A: The majority of our projects involve exome sequencing. We also do RNA transcriptome analysis and other genomic DNA analyses. We do very little whole genome sequencing. We were the pioneers in developing exome sequencing technologies and, given the cost advantage, exome sequencing is definitely better than whole genome sequencing, at least until the costs come down. We work in Mendelian genomics, and there the DNA variants are mostly in the coding regions and hence, exome sequencing makes more sense.
Q: What are the limitations of next-generation sequencing?
A: The biggest challenge in sequencing, especially with short-read sequencing, is that it doesn’t work well in regions with high GC content. Detecting structural and copy number variations is also a challenge with some of the existing sequencing technologies. There are new sequencing systems that are in development [that] will hopefully address these issues. Although costs of sequencing have significantly reduced in the [p]ast five years, large-scale whole genome projects are still a challenge. If one has to identify genes for common disorders, we still have to sequence nearly a million samples. At $1,500–$1,600, whole genome sequencing is still not affordable and feasible for such applications.
Q: In terms of its applications, where do you see sequencing making the biggest impact?
A: The applications of sequencing are mostly around identifying the disease-causing genes, and that’s where its biggest impact is likely to be. The other application is transcriptome analysis, where people are trying to find out how alternative splicing takes place and how RNA and proteins interact. That being said, all applications of sequencing that help us understand the biology of the disease are important.
Q: What do you think of single-cell sequencing and the impact it is likely to have?
A: In terms of its applications, single-cell sequencing is definitely going to be a big player. However, there are limitations. Sometimes it’s very difficult to access the quality of the original RNA that is isolated from the cell. Also, with a single cell, many rounds of amplification are needed to get the right amount of DNA, and that may cause the data to be skewed. These are my two main concerns. However, from a biological point of view, single-cell sequencing makes perfect sense, especially if you are looking to capture the heterogeneity in a system.
Q: Do you see any other technology replacing sequencing in the near future?
A: I don’t see any other technology replacing NGS. From a precision medicine standpoint, we are still discovering biomarkers and looking at disease-causing DNA variants, and for that, sequencing at a whole genome level is very essential. Once that is done, we may use microarrays to identify panels of genes. Microarrays are definitely cheap, but NGS is also getting cheaper. With new sequencing platforms and developments in single-cell technology, NGS is going to be less than $500 in the near future. Once that happens, sequencing will become a more routine tool in the clinical laboratory. At the end of the day, sequencing is the gold standard for genomics and the demand for NGS will continue.
Q: What would you advise people who are new to the field or looking to make new investments in sequencing?
A: For those who are already in the field, I would ask them to constantly keep up to date with the new products and protocols that are rapidly becoming available. They should watch out for updates to the existing sequencing technologies and for new technologies that are in development. For those starting a new lab, the field is a lot easier now, as sequencing is fairly standardized. Sequencing works beautifully and seamlessly, but now the challenge has shifted to analyzing the sequencing data. People have to figure out what works best for them in terms of IT support and data storage.