Matthew Montierth
Computational biologist working on cancer genomics, tumor evolution, and statistical methods for messy biological data.
I build statistical methods for cancer genomics and apply them at scale. My research has focused on quantifying tumor heterogeneity through the lenses of large scale -omics.
I hold a PhD in Quantitative and Computational Biosciences from Baylor College of Medicine, completed in the lab of Dr. Wenyi Wang at MD Anderson Cancer Center, and a B.S. in Genetics and Biotechnology with a statistics minor from Brigham Young University. I am currently a data scientist at MD Anderson.
What I work on
Tumor evolution
Tumors are not uniform. They carry populations of cells with distinct mutations, and that internal structure shapes how a tumor grows and whether it responds to treatment. I reconstruct subclonal architecture from bulk sequencing data and connect it to patient outcomes. Most recently that meant 7,827 tumors spanning 32 cancer types, where subclonal mutational load turned out to predict survival and immunotherapy response in cancers with low to moderate mutation burden.
Transcriptomic deconvolution
A bulk RNA-seq sample is a mixture. Deconvolution recovers the tumor-specific signal from that mixture, but the standard methods assume counts are dense and well-behaved, which fails for microRNA-seq and spatial transcriptomics. I developed DeMixNB, a semi-reference-based model built on a sum of negative binomial distributions, and applied it to miRNA-seq from 856 breast cancer patients and 3,755 spatial spots from lung tumors.
Cancer genomics at scale
Methods matter in analyzing real cohorts. I have optimized somatic mutation calling across 5,000+ whole exomes, integrated ATAC-seq with RNA-seq to identify tumor signature genes, and managed multi-terabyte genomic datasets on HPC infrastructure. A good share of my time goes to the unglamorous work of making pipelines reproducible and fast enough to iterate on.
Biomarker discovery
I work with clinicians to turn molecular signatures into usable markers that predict progression or therapy response. That has covered anaplastic thyroid carcinoma survival, immunotherapy response in metastatic prostate cancer, and total mRNA expression as a pan-cancer prognostic signal.
Selected work
| Subclonal mutation load predicts survival and immunotherapy response | Pan-cancer analysis of 7,827 tumors across 32 cancer types. Details |
| DeMixNB: deconvolution for sparse-count RNA sequencing | Genome Biology, 2026. Details |
| A guide to transcriptomic deconvolution in cancer | Nature Reviews Cancer, 2025. Details |
| Tumor cell total mRNA expression predicts disease progression | Nature Biotechnology, 2022. Details |
A full publication list is kept current, along with talks and posters and the software I have contributed to.
Currently
I am open to roles in computational biology, bioinformatics, and data science, either industry or academic, where statistical rigor and large-scale genomic data both matter. I am based in Newport, Kentucky, and work well remotely. If that sounds like your team, get in touch.
I also write occasionally on the blog about the practical side of this work: benchmarking, storage formats, and the tooling that makes large datasets tractable.