Long-Read DNA Sequencing Using High-Q Resonators Enhanced Raman Spectroscopy

Despite their transformative impact, next-generation sequencing (NGS) technologies based on short reads struggle to resolve complex genomic regions containing repetitive sequences, structural variations, and tandem repeats, among others. Certain long-read technologies circumvent these hurdles as they can generate reads spanning tens to hundreds of kilobases, but they suffer from relatively low raw accuracy and high-costs that hinder widespread adoption. Furthermore, existing short and long-read technologies struggle to directly capture crucial epigenetic modifications like DNA methylation and histone marks that regulate gene expression. Profiling the epigenome alongside the genome is vital for comprehensively understanding gene regulation, disease mechanisms, and phenotypic variations. As we move towards routinely sequencing entire genomes, transcriptomes, and epigenomes at high throughput and accuracy, cost-effective, integrated platforms are needed for sequencing and subsequent remedial strategies.<br/><br/>Here, we propose a long-read DNA sequencing platform leveraging high-Q resonator-enhanced Raman scattering for low-cost, accurate nucleotide sequencing, epigenetic variation detection, and single-point mutation repair. Unlike fluorescence-based approaches susceptible to photobleaching, Raman spectroscopy, relying on inherent molecular vibrational modes, has the potential for higher raw accuracy and longevity, which is crucial for longer reads. It also enables direct epigenetic profiling, such as DNA methylation or RNA glycosylation detection. Our approach combines: 1) Fab-friendly high-density resonator arrays for on-chip chemistry and surface-enhanced Raman spectroscopy (SERS), providing &gt;100 million sites/cm² density; 2) Raman-active nucleotide chemistry with high scattering efficiency; and 3) Cost-effective, high-throughput Raman spectroscopy techniques.<br/><br/>We have converted our recently demonstrated VLSI-inspired resonator platform^[1] into a high-density SERS platform, with ≥ 10⁸ Raman signal enhancement by incorporating metallic nanoparticles into the Silicon-resonator slots. We demonstrate selective functionalization strategies to localize a single polymerase to each metal dot within each resonator, developing microfluidic strategies to enable on-chip chemistry, and screening for effective Raman-labeled nucleotide reporters. We assess the performance of different tagged nucleotides in the presence of polymerase, across different concentrations, and detect epigenetic marks. With machine-learning models for data analysis, we demonstrate we can differentiate each nucleotide with &gt; 80% accuracy. Future steps will integrate these components to demonstrate single-molecule sequencing for hundreds to thousands of base pairs using a 3'-blocking approach, transitioning to long-read sequencing, all using a portable Raman microscope.<br/><br/>Reference:<br/>1. Dolia, V., Balch, H., Dagli, S., Abdollahramezani, S., Delgado, H. C., Moradifar, P., ... & Dionne, J. A. (2024). Very-Large-Scale Integrated High-Q Nanoantenna Pixels (VINPix). <i>In Press, Nature Nanotechnology.</i>