Genomics has changed rapidly since the 1980s. Andrew Gane discusses the history and future of genomics from PCR to NGS and more. He also discusses the promise of using AI, the need for better bioinformatics, and the broadening of all omics fields.
Whether it’s Archimedes jumping out of the bathtub yelling, “Eureka!” or Anakin Skywalker turning to the Dark Side, origin stories are as much a part of science as a part of science fiction. The origin story of genomics could be the tale told by Kary Mullis about how he came up with the idea for the polymerase chain reaction (PCR) on a nighttime drive on a California highway in 1983. He was so inspired that he stopped the car on the side of the road to jot down notes on an old receipt he found in his glove compartment (1).
From PCR to NGS
Recently, I sat down with Andrew Gane, Genomics and Diagnostics Solutions Strategy and Technology Manager at Cytiva, to talk about the history and future of genomics. When I asked him to name a few key achievements in genomics, he started with Mullis and PCR.
But Mullis’s idea was just a start. Over the next few years, many scientists and engineers worked to turn the idea into a reliable laboratory technique. The first automated thermocycler was built around 1985 or 1986 and was affectionately called Mr. Cycle (2). Today, Mr. Cycle lives at the Smithsonian Institute (3). “Without PCR, there would be no genomics,” says Gane. “The principles [of PCR] are fundamental to most genomics related applications.”
The next key achievement in genomics named by Gane was, unsurprisingly, the Human Genome Project (HGP). Though the project officially kicked off in 1990, the earliest talks about undertaking such a project happened as early as 1984 (4). Completed in 2003, scientists around the world worked for years on this massive endeavor to sequence an entire human genome. “The sheer determination to get that done was incredible and we can thank that achievement for driving my third choice [of key achievements], the arrival and evolution of next generation sequencing technologies, which opened the door to practical high-throughput genomics,” Gane said.
Next generation sequencing (NGS) refers to massively parallel or deep sequencing methods that allow for simultanous reading of the sequence of nucleotides in a section of a genome or an entire genome (Fig 1). With NGS, scientists can sequence entire genomes at a low cost — a big change from the 12 years and hundreds of millions of dollars spent on the HGP*. NGS is big today, but of course, it’s not the last word in genomics. NGS is leading to new technologies called third-generation sequencing (3GS).
Fig 1. The tools and technolgies used in NGS evolved from those used in the early days of PCR and the HGP.
The impact of NGS on data and the need for better bioinformatics
With NGS, genomics has become an extremely fast-moving field. How much and how quickly it changed can be visualized by looking at a graph of the cost per genome over time (Fig 2). The graph includes a line showing what the cost per genome would have been if it followed Moore’s law. Moore’s law is a prediction that the number of transistors on a silicon chip would double every two years†. Today, Moore’s law is used as a benchmark for measuring the progress of various technologies. As you can see, genomics was following the line closely until the late 2000s when it suddenly outpaced the prediction. What happened? NGS happened.
Fig 2. The cost to sequence an entire human genome dropped slowly at first, but quicky became cheaper with the advent of NGS. Note that the y-axis of this graph is logarithmic. Source: https://www.genome.gov/about-genomics/fact-sheets/DNA-Sequencing-Costs-Data
But with great sequencing comes great responsibility — or rather a great amount of data. “Higher throughput driven by cost reductions will continue to allow data generation at unprecedented levels,” Gane told me. He believes that researchers will need to depend on bioinformatics and artifical intelligence (AI) to analyze all the data being generated. “Data generation is always increasing, and bioinformatics continues to be a bottleneck — mainly in the interpretation field, so more autonomous AI tools that take the heavy lifting will accelerate progress.”
Gane says that AI and machine learning are starting to be used in gemomics, but it’s in the early stages. For example, a team of researchers from the UK and the US used machine learning to analyze sequencing data from hundreds of samples to predict how cancer may evolve in a given patient (7). The method they developed could be used to develop personalized treatments for future patients.
According to Gane, AI could be used to compare across people, genotypes, phenotypes, or even individual cells to look for correlations. It could find mutations in genomes and connect them to dieases. It could help researchers understand how different genes work together to contribute to a disease. AI could even find things that a researcher wasn’t looking for. Gane hopes that AI tools and machine learning will soon be much more accessable and simpler to use. “Right now, there aren’t enough people who can do bioinformatics. They need a special skill set.” Gane lamented.
The need for longer sequencing reads
Other developments that Gane sees coming include longer sequencing reads, greater resolution, and single-cell analysis. He hopes that sequencing reads will become longer and eventually cover entire chromosomes. Long reads are important to understanding the relationships between genes on a single chromosome.
”Increased resolution and sensitivity will drive deeper understanding,” explained Gane. ”Previously we analyzed cells in batches, now we analyze them individually in many ways (Fig 3). This is the increased resolution we talked about, but the next goal is to make sure we can understand how each cell behaves not on its own, but in conjunction with all the other cells around it. This is the next big challenge and it is already starting to be addressed by spacial multi-omics, for example, but better approaches are still required.”
Fig 3. Single-cell sequencing enabled the characterization of multiple sub-populations of cells in specific regions of the brain.
It’s an omics world after all
The mention of multi-omics made me wonder if the other omics disciplines would replace genomics in importance, but Gane quickly corrected me. “I see ’omics’ as more how we compartmentalize our knowledge in terms of how DNA is arranged, how RNA is transcribed, or how protein interacts. But as disciplines, it’s unusual that each is studied in isolation. Biology does not work in this way.”
Gane explained that we need to start looking at a biological system as a whole and not just stare at one cell at a time and look at how it functions. We need to look at a cell and the cells around it to see how they are arranged, how they interact, and how they are different from one another. To do this, however, the cells have to be taken out of the body. And when you take the cells out of their home, they react. Gane compared this problem to the observer effect in quantum mechanics: if you observe the cells, they change. As a result, you don’t measure what you were hoping to measure.
In the future, Gane hopes there will be a method to study cells in-situ. Scientists will have to layer tools and techniques to fully understand what’s going on in each cell and how that affects an organism as a whole. “All these omics applications will develop as there is knowledge to be gained. With so much unknown about health and disease, all avenues need to be explored with open minds. Science is about discovery, and our aim in Cytiva is to put that to best use in understanding, preventing or treating disease. Making use of the knowledge translating or applying it to health and disease situations has to be the target.” concluded Gane.
Learn more about supporting your NGS workflows here.
* The HGP explains that the final cost of the first human genome sequence is difficult to pin down because what “counts” and “doesn’t count” wasn’t always clear. However, they estimate that the cost was somewhere between USD 500 million and USD 1 billion (5).
† The original prediction of Moore’s law was that the number of transistors on a chip would double every year, but that time estimate has gotten longer over time. Hence the “law” is not scientific law, but an observation that changes as limits constrain technology (6).
Andrew Gane is the Product Strategy and Technology Manager within the Genomics and Diagnostic Solutions business responsible for building the innovation pipeline in collaboration with the R&D and commercial teams. His knowledge and understanding of emerging trends and new applications have been fundamental to developing the product portfolio into workflow-based solutions, with a particular focus on next-generation sequencing (NGS). Andrew has more than 30 years’ experience in immunodiagnostics and molecular diagnostics in both lab-based and product development roles. Andrew lives in Cardiff, Wales.
About the author: Laura Prescott is a science writer at Cytiva. She majored in chemistry at Williams College and has a master’s degree in chemistry from the University of Chicago. During and after graduate school, she worked in several molecular biology labs and knows how to pour a polyacrylamide gel without bubbles. Laura lives in Indianapolis, Indiana.
References
- The nobel prize in chemistry 1993. NobelPrize.org. Accessed August 29, 2022.
- Henry A. Erlich PD. Development and evolution of PCR. GEN. Published October 30, 2018. Accessed August 29, 2022.
- Mr. Cycle, thermal cycler. National Museum of American History. Accessed August 29, 2022.
- Human genome project timeline. Genome.gov. Accessed August 29, 2022.
- The cost of sequencing a human genome. Genome.gov. Accessed August 29, 2022.
- Moore's law. Intel. Accessed August 29, 2022.
- Caravagna, G., Giarratano, Y., Ramazzotti, D. et al. Detecting repeated cancer evolution from multi-region tumor sequencing data. Nat Methods 15, 707–714 (2018).