Cancer treatment is a race against evolution

Cancer is a disease of the genome, but a tumor is not static. It is an evolving population of genetically diverse cells, continually shaped by mutation, cellular stress, and selection. These forces influence which cells survive and expand, how tumors progress, and how they respond to treatment.

By identifying the genetic alterations and cellular pathways that shape tumor fitness, we can better understand how cancers evolve and uncover vulnerabilities that may be exploited therapeutically.

I study these evolutionary dynamics by investigating how tumors tolerate the cellular stress associated with high tumor mutational burden (TMB), and which genetic alterations and dependencies emerge under this burden. More broadly, I am interested in using the concepts of cancer evolution to identify therapeutic opportunities, understand tumor adaptation, and develop more effective and durable treatments.

My research

Do different cancer types share a common response to high tumor mutational burden?

Tumors with high TMB carry large numbers of somatic mutations. Although some mutations benefit cancer cells, most are passenger mutations with neutral or mildly deleterious effects. Individually, these mutations may have little impact, but their cumulative burden can destabilize proteins, disrupt protein folding, and create protein-misfolding stress that reduces tumor fitness.

I use human cancer genomic and transcriptomic data to investigate how tumors respond to increasing TMB and to identify the pathways and protein complexes that may help highly mutated tumors tolerate this burden. My recent work reveals a consistent transcriptional response to TMB across diverse cancer types, suggesting that the cellular consequences of high TMB extend beyond the tissue in which a tumor originates.

By revealing how tumors compensate for the cellular costs of mutation accumulation, this work identifies dependencies associated with high TMB that may represent tissue-agnostic therapeutic vulnerabilities.

Building experimental models of a high-TMB tumor state

High TMB cannot be understood solely by comparing naturally occurring human tumors, in which TMB is entangled with tumor type, driver mutations, environmental exposures, and treatment history.

I develop experimental models that reproduce key features of high-TMB human tumors while allowing these factors to be controlled. These systems make it possible to test how the protein misfolding stress observed in high-TMB tumors affects tumor initiation, development, and response to treatment.

The goal is not simply to reproduce a molecular signature. It is to create a system in which the causal effects of this stress on tumor fitness can be measured directly.

Mapping the tumor fitness landscape

Once high-TMB-associated stress can be modeled experimentally, I use multiplexed perturbation approaches to determine which genetic changes help or hinder tumor growth under protein misfolding stress.

Using CRISPR editing, tumor barcoding, and deep sequencing in vivo, I systematically measure how tumor suppressor loss and the perturbation of context-dependent essential genes alter tumor fitness. This approach maps how genetic alterations affect tumor growth under stress and can uncover promising therapeutic targets.

More broadly, this work provides a framework for understanding how genetic variation and selection shape tumor fitness, and how evolutionary adaptations can create vulnerabilities that may be exploited therapeutically.

Turning abstract biological observations into testable systems

Many important biological concepts are easy to describe but difficult to model experimentally. High TMB is one example: it is not a single mutation or pathway, but an accumulated state with consequences for protein stability, cellular stress, and tumor fitness.

My PhD research translates this abstract evolutionary state into experimentally tractable systems. I combine human cancer genomics and transcriptomics, genetically engineered mouse models, tumor barcoding, CRISPR editing, and deep sequencing to investigate how tumors tolerate the fitness costs associated with high TMB.

Where I want to go next

I want to extend my work on cancer evolution by developing computational and AI-based approaches for cancer therapeutics, treatment resistance, and tumor adaptation.

My current research asks how tumors tolerate the fitness costs associated with high TMB and which genetic changes allow them to survive under stress. Going forward, I want to investigate these questions across a wider range of biological and therapeutic contexts: How do tumors respond to treatment pressure? Which evolutionary paths lead to resistance? Can those paths be predicted, and ultimately constrained, before resistant populations emerge?

I am especially interested in approaches that integrate genomic, functional, and experimental data to model tumor behavior across biological contexts. I want to move beyond pattern recognition and candidate generation toward systems that support reliable, mechanistically grounded predictions:

  • Can we reliably distinguish driver mutations from passenger mutations?
  • How likely are tumors to adapt under a particular treatment?
  • Which vulnerabilities will persist as a tumor evolves?
  • Which interventions could prevent, delay, or redirect resistance?

Answering these questions requires more than producing lists of genes, pathways, or therapeutic targets. It requires determining whether a model has sufficient biological and contextual information to support its predictions, whether those predictions can be distinguished from plausible alternatives, and which experiments would provide decisive evidence.

I am interested in developing computational–experimental workflows in which biological representation and validation are part of the scientific problem from the beginning. My goal is to connect AI-based prediction with rigorous experimentation to generate actionable insights and develop more durable strategies for cancer treatment.

Previous work

Before my PhD, I studied how cells detect and respond to reductive stress. This work uncovered the structural and molecular basis of a conserved stress-response pathway involving a ubiquitin ligase and its substrate. Components of this pathway are recurrently altered in cancer.

Work with me

I am open to scientific collaborations, interdisciplinary AI-for-biology projects, and consulting opportunities.

I can contribute expertise in cancer evolution, cancer genomics/transcriptomics, computational biology, functional genomics, and the biological evaluation of AI-based research tools. I am particularly interested in working with people who value mechanistic depth, rigorous validation, and biologically meaningful questions.

If you are working on a related problem or think my perspective could help your team, I would be glad to talk.

Publications

  1. Shih KY, Brandman O, Winslow MM, Petrov DA. Transcriptional response to tumor mutational burden is consistent across cancer types. bioRxiv. 2026 Aug 27. doi:10.64898/2026.08.27.744035

  2. Diaz-Jimenez A, Shuldiner EG, Somogyi K, Shih KY, González-Velasco Ó, Najajreh M, Kim S, Akkas F, Murray CW, Andrejka L, Tsai MK, Brors B, Hofmann I, Sivakumar S, Sisoudiya SD, Sokol ES, Cai H, Petrov DA, Winslow MM, Sotillo R. EML4-ALK variant-specific genetic interactions shape lung tumorigenesis. Cancer Discovery. 2026 Jan 12;16(1):46-65. doi:10.1158/2159-8290.CD-24-1417 · PMID: 40986428 · *Equal contribution.

  3. McMinimy R, Manford AG, Gee CL, Chandrasekhar S, Mousa GA, Chuang J, Phu L, Shih KY, Rose CM, Kuriyan J, Bingol B, Rape M. Reactive oxygen species control protein degradation at the mitochondrial import gate. Molecular Cell. 2024 Dec 5;84(23):4612-4628.e13. doi:10.1016/j.molcel.2024.11.004 · PMID: 39642856

  4. Manford AG, Rodríguez-Pérez F, Shih KY, Shi Z, Berdan CA, Choe M, Titov DV, Nomura DK, Rape M. A Cellular Mechanism to Detect and Alleviate Reductive Stress. Cell. 2020 Oct 1;183(1):46-61.e21. doi:10.1016/j.cell.2020.08.034 · PMID: 32941802

  5. Manford AG, Mena EL, Shih KY, Gee CL, McMinimy R, Martínez-González B, Sherriff R, Lew B, Zoltek M, Rodríguez-Pérez F, Woldesenbet M, Kuriyan J, Rape M. Structural basis and regulation of the reductive stress response. Cell. 2021 Oct 14;184(21):5375-5390.e16. doi:10.1016/j.cell.2021.09.002 · PMID: 34562363