Dan Hovelson

Dan Hovelson, Ph.D.

Ph.D. Program
Sr. Bioinformatician
Strata Oncology, Inc.


Dissertation Title

Precision oncology opportunities and disease insights from next-generation-sequencing profiling of routine clinical biospecimens

Research Interest

Rapid technological developments in next-generation sequencing (NGS) and inter-institutional collaborations including The Cancer Genome Atlas (TCGA) have enabled comprehensive characterization of the genomic, transcriptomic, and epigenetic landscapes from bulk tissue specimens in a wide range of cancers. Emerging work has focused on scaling NGS-based profiling strategies to guide precision medicine approaches in clinical oncology using routine clinical biospecimens such as formalin-fixed, paraffin-embedded (FFPE) tissue or less-invasive liquid (e.g., blood or urine) samples. Technical challenges associated with limited tumor lesion size, low nucleic acid quantities, disease-specificity applications, and disease and histological heterogeneity present hurdles to widespread adoption and utility of extant NGS-based precision oncology approaches. Here, several analytical advances are described supporting democratization of precision oncology approaches from clinical tissue and liquid biospecimens, while revealing disease insights and important clinical considerations in the context of both localized and advanced (including multifocal and/or heterogeneous) disease. First, development and validation of a targeted DNA and RNA NGS assay compatible with small quantities of DNA and RNA isolated from routine, archived FFPE tissue specimens is described. This assay, targeting recurrently mutated oncogenic hotspots, tumor suppressors, copy-number-altered genes, and recurrent gene fusions is applied to a cohort of >300 FFPE tissue samples, revealing high sensitivity with orthogonal molecular diagnostic assays for BRAF, KRAS, and EGFR oncogenic alterations. Second, I describe a rapid, inexpensive, low-pass cell-free DNA (cfDNA) whole-genome sequencing (WGS) copy-number profiling approach, including a novel heuristic tumor content approximation method, capable of establishing genome-wide copy-number profiles from 0.01-0.1x sequencing coverage. Application of our approach in plasma samples from patients with advanced cancer with matched comprehensive tissue NGS revealed high concordance with tissue-based molecular profiles, while highlighting important areas of potential utility from noninvasive profiling of overall disease burden. Third, I describe the systematic assessment of expression-based molecular subtypes in histologically heterogeneous bladder cancers, revealing robust identification of basal/luminal molecular subtypes in a cohort of >100 bladder cancer cell lines and tumor tissue specimens, and recapitulation of basal/luminal subtypes in >400 samples profiled by TCGA using selected marker subsets. Importantly, I describe divergent expression profiles in the context of shared genomic alterations for individual histologically divergent tumor components from the same tumor, confounding proposed clinical utility of expression-based subtypes for disease prediction and prognosis. Fourth, I describe the development of a targeted RNAseq panel capable of assessing major transcriptional programs and disease biomarkers across the full spectrum of prostate cancer disease, while deriving commercially available prognostic scores that show limited robustness to disease multifocality. Lastly, I describe extensions of our cfDNA WGS approach to urine cfDNA samples from patients with advanced cancer, while exploring the potential utility of pairing described analytic tools with existing and emerging molecular profiling strategies to improve our understanding of disease biology and maximize clinical utility. 

Current Placement

Strata Oncology, Inc.