Charles Vaske, PhD

Charles Vaske, PhD is Founder and Principal of Vaske Genomics, a computational biology consultancy.

I love to study biology through the lens of computation, mostly in the field of genomics. I’m passionate about translating discovery all the way to practice, such as developing and getting regulatory approval for comprehensive genomic/transcriptomic profiling of tumors.

ORCID · Google Scholar · GitHub · LinkedIn · Bluesky · Publications

For consulting or collaboration, contact Vaske Genomics at info@vaskegenomics.com.

Experience

Vaske Genomics, Founder and Principal, Feb 2024 – present. Founded as a sole proprietorship; incorporated as Vaske Genomics, Inc., a Delaware corporation, in September 2026.
Clear Labs, Strategic Consultant, Mar 2024 – present; VP of Bioinformatics, Mar 2022 – Feb 2024.
Claret Bioscience, VP Bioinformatics, Dec 2019 – Mar 2022. Led the cell-free DNA fragmentomics research program and the bioinformatic analysis of rootless hair for forensic genealogy.
NantOmics (which acquired Five3 Genomics), Executive Vice President, Genomic Research, Aug 2014 – Dec 2019. Took tumor/normal whole-genome sequencing and whole-transcriptome RNA-seq from FFPE tumor samples through CLIA approval, and led a research team applying deep learning to DNA sequence and histopathology images.
Five3 Genomics, Chief Scientific Officer, Apr 2011 – Aug 2014. Directed the cancer research program and designed the computational infrastructure for whole-genome and whole-transcriptome sequencing analysis.
University of California, Santa Cruz, Postdoctoral Fellow, Jul 2010 – Mar 2011. Integration of multiple cancer genomics data types with pathway information.
Princeton University, Postdoctoral Research Associate, Sep 2009 – Jul 2010.

Education

University of California, Santa Cruz, PhD, Bioinformatics, 2009.
Washington University in St. Louis, BS, Computer Science and Mathematics, 2003.

Favorite work

Paradigm

Genomics provides multi-modal views of biological samples, and cancer has some of the most extreme changes compared to normal body tissues. Paradigm is a method for integrating gene expression levels, DNA methylation, protein abundance measurements, and gene copy numbers all on top of known interactions between proteins and the genome. It’s an attempt to combine all the available data with all the available knowledge of regulation in a cell.

By modeling individual genes, complexes of molecules, and gene families, we construct a factor graph that models a whole cell:

visual display of the factor graph model for genes, complexes, and gene families, and an illustration of the interactions used to construct a whole cell network

These models make predictions about activity profiles in individual tumors, which can be more informative than the source data types. For example in a Norwegian cohort of breast cancer micro-metastases, the subtyping was more informative for progression than the other data types:

Heatmap of per-patient per-gene pathway activities, with cluster annotation. Additionally, the Kaplan-meier survival curve, and the significance of the progression of the subtypings using various data types

Learning subgroup-specific regulatory interactions and regulator independence with PARADIGM. Bioinformatics, 2013. doi:10.1093/bioinformatics/btt229
Inference of patient-specific pathway activities from multi-dimensional cancer genomics data using PARADIGM. Bioinformatics, 2010. doi:10.1093/bioinformatics/btq182

See all publications.