WYMM Tour: San Diego

Next-generation sequencing (NGS) has transformed cancer research, yet its effectiveness is often constrained by sample quality, tumor fraction, and limitations inherent to short-read sequencing—particularly in detecting structural variants, complex rearrangements, and alterations in repetitive or polymorphic regions. Long-read sequencing offers a solution, but cost and throughput challenges persist. To address these limitations, we developed a flexible, automation-compatible library prep and target enrichment workflow that supports DNA inputs as low as 200 ng, accommodates both enzymatic and mechanical shearing, and enables fast hybridization (90 minutes). Coupled with Oxford Nanopore’s long-read sequencing and the SureSelect Cancer Pan Heme assay, this platform enables detection of diverse genomic alterations—including SNVs, indels, CNVs, and gene fusions—within a single assay. We demonstrate high enrichment efficiency using a novel fast hybridization buffer, achieving on-target rates of ~80% for libraries with insert sizes up to 4–5 kb . Comparative analysis reveals superior coverage in challenging genomic regions using enriched long-read sequencing versus short-read approaches. This solution offers a scalable, automation-compatible, and streamlined workflow for molecular laboratories, supporting comprehensive genomic profiling applications relevant to hematologic malignancies.

Pediatric central nervous system (CNS) tumors remain the leading cause of cancer-related mortality in children. While molecular profiling has transformed tumor classification and risk stratification, integrating genomic information into real-time clinical decision-making remains a significant challenge. Together with Rady Children's Health, we have developed a Functional Precision Medicine program that combines comprehensive molecular characterization with functional drug-response testing to identify molecular markers associated with differential therapy response for children with high-risk and recurrent brain tumors.

In this presentation, I will describe our implementation of long read sequencing for rapid CNS tumor classification using DNA methylation profiling and discuss its integration into multidisciplinary molecular tumor board workflows. I will present our experience leveraging long-read sequencing technologies to accelerate diagnosis, improve molecular characterization, and support treatment decision-making.

Quantitative trait locus (QTL) mapping connects genotypes to molecular phenotypes, especially expression (eQTLs), chromatin accessibility (caQTLs) and histone mark (hQTL) readouts on enhancer-associated chromatin, but these layers sample only part of cis-regulatory control. Prior work indicates that active DNA demethylation can occur before chromatin opening and depositioning of active histone marks, pointing to 5-hydroxymethylcytosine (5hmC), a stable oxidized form of 5-methylcytosine (5mC) enriched at active and poised enhancers, as an early marker of enhancer activity and in general of epigenetically dynamic loci with regulatory potential. Standard bisulfite conversion and array-based methylation assays cannot separate 5hmC from 5mC, so population-scale cis-5hmC QTL landscapes remain largely unknown. Emerging short and long-read DNA sequencing techniques now enable genome-wide 5hmC measurements. Utilizing Nanopore (ONT) long-read sequencing, in this work, we profiled naive CD4+ T cells from 39 healthy donors. In order to concurrently measure chromatin accessibility alongside 5mC and 5hmC, we employed nano-NOMe-seq (enzymatic GpC methylation-based labeling of accessible DNA) to measure these epigenetic features on single chromatin fibers at allele-specific resolution (n=20). We observe widespread cis-associations and uncover 5hmC-QTL loci that colocalize with eQTL but not with hQTLs, and 5hmC-QTL loci without eQTL or hQTL overlap, patterns supported by allele-resolved long-read evidence. Using matched ATAC-seq and RNA-seq from the same cohort and chromatin loops we previously mapped, we prioritize putative targets of 5hmC-QTLs including THEMIS2 gene where 5hmC-QTLs colocalize with eQTLs and show genotype-dependent chromatin accessibility and gene expression within our cohort. Our work establishes a cis-5hmC QTL mapping framework and defines a demethylation-centered genotype–phenotype layer that is complementary to molecular QTL studies probing other epigenetic ...

Many tissue samples of interest for microbiome research are dominated by host cells and host DNA, and it remains unclear what functional role, if any, the microbes present in these tissues play in human health. Progress toward answering that question depends on first reliably detecting these low-abundance microbial populations. Existing comparisons of host-DNA depletion methods have focused almost exclusively on how much host DNA is removed, not on whether the same microbial taxa are recovered before and after depletion. This gap limits confidence that current methods preserve an accurate picture of the microbial community, a prerequisite for any downstream functional study. Here, we compare molecular depletion and in situ spatial transcriptomics approaches, and outline our future direction using long-read sequencing as a novel strategy for separating microbial from host genomic compositions. Together, these efforts are a step toward the more reliable microbial detection needed to eventually investigate the functional role these organisms play in human health.