by Giorgia Andrei | Imec is an international reference hub in the field of nanoelectronics. Founded in 1984, it is best known for its work in microchip technology. Among the various applications to which the chips developed by imec are dedicated is sequencing, a process whose peculiarity was highlighted during the pandemic: the sequence of the Sars-CoV-2 virus was published just four weeks after the first samples were processed from patients, enabling the identification and development of molecular tests and the design of Rna vaccines. And it is also thanks to sequencing that mutations and variants of the virus can be monitored.
Semiconductors have already met a challenge
It has to be said, however, that compared with a human genome, a virus is a relatively simple sequence, and that to apply sequencing to individual cells or in biopsies for cancer diagnosis, for example, is still very time-consuming and expensive. However, as imec's experts explain in their white paper 'Semiconductor technologies and system concepts to revolutionize genomics', there is no shortage of hopeful developments: sequencing tools will be able to provide more and more information, and along with them, data storage and visualisation, machine learning and artificial intelligence will become increasingly important. In this crescendo of complexity, semiconductor technology will be crucial. Like no other industry, the semiconductor industry has succeeded, year after year, in building more complex and scalable systems, in compact form factors and at low prices, and in guaranteeing reliability and performance even in the face of ever-smaller solutions: think, for example, of mobile phones, which in just a few years have become veritable pocket supercomputers. The semiconductor industry has met its challenge thanks to 'cross-layer optimisation', i.e. a co-design effort involving the different layers of the electronic system. A similar approach, according to imec, could drive innovation in the field of genomics, which in a few years' time will be able to rely on portable point-of-care sequencing devices: it will be a question of performing operations such as those already practised today for rapid tests, but with the possibility of obtaining much more information on viruses and genetic risk factors of the patient. Below are some of the developments that are affecting the sequencing process, from sample preparation to data analysis.
Sample preparation
Before the DNA code can be read from a sample to be analysed, a number of steps have to be taken, requiring the use of different instruments and a rather long time. Microfluidic structures based on semiconductors can be used to automate these steps and shorten the time required. These structures are very precise and ultra-small, in the order of a few micrometers, and therefore require small, well-defined volumes of reagents, allowing faster chemical reactions. To understand the advantages offered by microfluidics, let us take the example of the analysis of circulating free Dna (Free Cell Dna), which is considered to be the basis for early screening and diagnosis in the near future. Sample preparation steps for this type of application include plasma separation, DNA extraction, amplification, fragmentation, library generation and cleaning, and DNA quality and quantity control. Most of these steps can be performed by semiconductor-based microfluidic structures: precision micropillars can be used for plasma separation by size filtration or Dna extraction via surface interactions, while dielectrophoretic sorting structures isolate relevant Dna molecules for further analysis. In single-cell genomics applications, which focus on sequencing DNA from a single cell, semiconductors are also an ideal option. Sample preparation in this case consists of selecting and manipulating target cells and barcoding the cells prior to sequencing. Specifically, imec has developed a cell-sorter device for cell sorting, integrated with sensing (based on photonics, microscopy or impedance) and synchronisation circuits (to perfectly time the cell passage and cell sorting event).
The evolution of sequencing methods
Until the early 2000s, the dominant method for DNA sequencing was that invented in 1977 by Fred Sanger, also known as the chain-termination method or the dideoxynucleotide method. In 2005, next-generation sequencing techniques, based on sequencing by synthesis, emerged: sequencing platforms became more compact and used multi-channel systems to achieve much higher performance than Sanger-based equipment through parallel reads of short DNA fragments.
Illumina's technology based on highly-parallelised optical detection has been particularly successful, as has the technology developed by ThermoFisher Scientific, which, by combining CMOS with synthesis sequencing, has made it possible to sequence shorter fragments in less than five hours. Third-generation sequencing is characterised by real-time reads, single-molecule resolution and much longer reads.
Examples in this case are Pacific Biosciences, which has launched a long read system based on zero-mode waveguides, and Oxford Nanopore, which has launched long read systems based on biological nano pores. Fourth-generation sequencing will be characterised by integrated electrical and/or photonic sensing and a switch from standard CMOS to advanced nano-dimension technology.
An important element of next-generation sequencing is nanopore sensors, which enable long reads of single molecules and high-throughput sequencing. Nanopores are small holes in thin membranes through which a Dna/RNA molecule or protein can move. The movement generates an electronic signal indicative of the molecule's structure, which can be used to decode its composition. Nanopores can be made from biological or semiconductor materials, in which case they are more robust and can be integrated with on-chip electronics. Using state-of-the-art lithography, imec is able to make nanopores based on semiconductors with a diameter of less than 5 nm. In addition, a new type of nanopore sensor can be produced with lithography patterns on FinFET arrays: in detail, a FinFET transistor and its sensitive gate are wrapped around the nanopore in order to directly detect the electric field in the nanopore and speed up readout times.
Data analysis
Software is a key component of point-of-care sequencing tools. Intelligent software solutions, with machine learning and artificial intelligence algorithms, will be crucial for handling huge amounts of data when millions of sequencing tests are performed in the daily practice of hospitals around the world. imec's ExaScience Life Lab has developed an open source tool, called el- Prep, to speed up the process of reconstruction and variant identification that is 16 times faster than the genome analysis toolkit used as a standard reference. elPrep is validated by Janssen Pharmaceutica and Seven Bridges Genomics. It is written in Go, an open source programming language, and can be run on any standard server, locally or in the cloud.
You might also be interested in:
