WYMM Tour: Ann Arbor

Nanopore sequencing is primarily known for its ability to sequence arbitrarily long DNA molecules in order to more accurately assemble complex genomes, and easily identify large-scale, abnormal structural genomic variation common in many cancers. Another unique (and under-appreciated) aspect of the technology is its rapid time-to-result, form-factor, and economics that allow for truly personalized bed-side sequencing. Our research group investigates potential use-cases where rapid nanopore sequencing could have a transformative impact on the diagnosis and monitoring of brain tumors. We have developed “ultra-rapid” targeted sequencing protocols that can call somatic mutations from tumor tissue within 30 minutes. This protocol could be used to diagnose and grade tumors intra-operatively within the same time-frame as frozen section histology. We have developed a protocol to sequence DNA recovered from surgical waste during brain tumor biopsies. This approach is able to return results from targeted cancer panels within a few hours of a biopsy, well before normal histo-pathologic molecular analysis (e.g. immunohistochemistry) is available and with increased sensitivity/precision. We have also developed same-day, targeted liquid biopsy approaches that are able to accurately measure small fluctuations in cell-free tumor DNA content in blood. While normal nanopore sequencing can have a relatively high error rate, sequencing redundant copies of DNA contained on the same strand (concatemers) enables random error detection and correction. Utilizing this error correction approach we are able to quantify single nucleotide polymorphisms down to ~0.05% allele frequency and measure small, but clinically relevant changes in cell-free tumor DNA content from patient plasma. We show these results match that of droplet digital PCR-based measurement indicating that in certain cases, nanopore sequencing is as accurate as gold standard molecular diagnostic tools. Overall, this platform is poised to transform the way we diagnose, and monitor cancers with clear translational impact immediately on the horizon.

Transgenic animal models have been a critical component of advancing biomedical research for more than 30 years. Methods of generating transgenic animals have evolved from random transgenesis to highly targeted CRISPR/Cas9 mediated genome engineering. The methods are characterized by differing intended and bystander changes to the genome including complex transgene associated structural variation and on target allele mosaicism that are important to understand. Standard characterization efforts have relied on PCR and Sanger sequencing which have limitations for detecting common genetic alterations in transgenic animals. We have developed targeted long read sequencing approaches using Oxford Nanopore Technologies to fully and better characterize transgenic animal models. Our targeted approaches are separated into characterizing CRISPR/Cas9 associated genomic mosaicism and determining the structure and function of randomly integrated transgenes. We utilize PCR based amplicon sequencing or targeted capture with NCATS with high coverage to characterize genomic mosaicism. For identify random transgene integration site, we first use LRS whole genome sequencing (WGS) to achieve 5-25X coverage and blast aligned sequences for transgene sequences to determine chromosome positions for potential random integration locations. ONT Adaptive sampling real time sequencing, which allows for real time enrichment based on alignment to a reference sequence, is then utilized to enhance overall target read coverage to 60-100X for the chromosome allowing for identification and characterization of transgene associated structural variation. By furthering characterizing these genomic changes, we can identify the full picture of mutations in transgenic animals. This will further increase the translational relevance of these animal model systems to help advance their impact on improving human disease.

m6A modifications have been shown to play a fundamental role in ALS pathogenesis, potentially mediated by TDP43-dependent misprocessing of m6A-modified RNA. By using Oxford Nanopore Technology Direct RNA Sequencing (DRS), we characterized the m6A methylome of blood samples from ALS patients at early disease stages. DRS allows whole transcriptome-wide direct sequencing of native RNA molecules and detecting RNA modifications at single-molecule resolution. Our preliminary results support the feasibility of detecting meaningful m6A signals from ALS blood samples, suggesting that blood may be a surrogate of the brain/spinal cord epitranscriptome.

The excision of exons and introns from pre-mRNAs is a key process that cells use to generate proteomic diversity. Over 95% of intron-containing human genes undergo alternative splicing, a mechanism crucial for guiding stem cell fate. However, the transcriptional dynamics and splicing patterns in muscle stem cells (MuSCs) and other adult stem cells are not well understood, partly due to the limitations of short-read sequencing (<300 bp) in capturing full-length transcripts. In this study, we utilized single-cell long-read sequencing (Oxford Nanopore) and short-read sequencing (Illumina) to investigate RNA splicing dynamics in MuSCs during skeletal muscle regeneration. Single cells were isolated from the tibialis anterior muscles of mice before and 7 days post-injury (7 dpi) following barium chloride-induced injury. This combined approach provided a comprehensive view of the splicing landscape across various skeletal muscle cell types (42,647 total cells), including MuSCs (N = 2,776), revealing the diverse RNA isoforms at different muscle repair stages. Long reads from ONT sequencing were deduplicated and annotated using cell barcodes and UMI sequences from 10X Genomics libraries. Isoform detection and gene expression analysis were conducted with IsoQuant, and integration with short-read scRNA-Seq data showed a strong correlation in mean gene expression measured with CellRanger (r = 0.92, p < 0.001). Differential expression analysis of pseudobulked MuSCs from control and 7 dpi conditions identified 1,671 differentially regulated isoforms and 1,241 differentially expressed genes, along with 1,743 significant isoform switches. Many of these isoform switches involved transitions from exons without known functional domains to those with domains critical for muscle differentiation. Our analysis showed an increase in isoforms with functional protein domains at 7 dpi compared to controls (58% vs. 54%). Control MuSCs had more isoforms with retained introns and alternative transcription termination sites, while 7 dpi MuSCs had more alternative 3' and 5' splicing events. This dataset is a valuable resource for future research on alternative splicing in myogenesis and muscle repair.

Tandem repeat sequences comprise approximately 8% of the human genome and have been linked to more than 50 neurodegenerative disorders. While accurate characterization of disease-associated repeat loci using short reads or amplification based techniques remains resource intensive and often lacks high resolution genotype calls, advances in long-read sequencing from Oxford Nanopore Technlogies have allowed for improved investigation into tandem repeat variation. We developed an nCATs based, multiplexed nanopore sequencing panel and HMMSTR, a sequence-based tandem repeat copy number caller to genotype more than 50 loci in parallel. The flexible panel allows us to capture disease associated regions at an average coverage of >150x. Here, we investigate repeat expansions in patient derived samples from individuals with CANVAS and ALS and highlight the utility of the panel in characterizing samples with multiple repeat expansions. This genotyping approach for tandem repeat expansions is scalable, simple, flexible, and accurate, offering significant potential for the investigation of repeat expansions in neurodegenerative diseases.