Jake Leighton
Computational Biologist & Data Scientist · MD Anderson Cancer Center

I build the computation
behind cancer discovery.

I use single-cell genomics and artificial intelligence to study how cancer evolves and spreads — across breast, lung, and hematological disease — and I teach the next generation of scientists to do the same in R.

What I work on
01

Single-cell genomics

scDNA, scRNA, and scATAC to reconstruct mutational and copy-number lineages — resolving how tumors evolve one cell at a time.

02

Spatial transcriptomics

Integrating spatial expression with digital pathology to read the tumor microenvironment in its native tissue context.

03

AI for oncology

Graph transformers, CNNs, and graph neural networks applied to phylogenetic and pathology data to predict metastasis and outcomes.

04

Teaching & open science

A free video course, a division-wide computational training program, and a forthcoming R-based textbook in computational cancer biology.

In brief

A scientist who ships tools, and teaches.

I'm a computational biologist and data scientist at MD Anderson Cancer Center, in the Department of Translational Molecular Pathology. I apply single-cell sequencing, spatial transcriptomics, AI modeling, and digital pathology to understand disease biology and identify precision biomarkers for new cancer therapies — and I lead computational training across the division, embedding bioinformatics support in faculty research and mentoring the next wave of fellows.

I care as much about making these methods learnable as advancing them — which became Leighton Lab, a free course teaching computational cancer biology in R.

The path here
2025 — now

Computational Biologist & Data Scientist · MD Anderson

Department of Translational Molecular Pathology

I wear several hats across the division: DPLM Computational Data Consultant, DPLM Computational Training Lead, and TMP T32 Fellowship Consultant — providing embedded bioinformatics across faculty research, leading computational training, and mentoring fellows.

2023 — 2025

Data Scientist · Institute for Data Science in Oncology

with Yinyin Yuan, PhD · MD Anderson

Built a graph transformer predicting metastasis from AI-annotated phylogenetic trees in TRACERx lung adenocarcinoma; integrated spatial transcriptomics and digital pathology to study the tumor microenvironment; designed AI surveillance for transfusion medicine.

2016 — 2023

PhD · NCI T32 Fellow · MD Anderson

with Nicholas Navin, PhD

Developed single-cell DNA sequencing approaches for triple-negative breast cancer using multi-patient panels and computational phylogenetics; integrated scDNA/scRNA/scATAC to find actionable biomarkers in AML.

2013 — 2016

Research Assistant · UT Health San Antonio

with Pei Wang, PhD

Designed bulk NGS pipelines profiling Hippo-pathway-disrupted pancreatic tumors; combined molecular biology and computation to study tumor and stem-cell biology.

2008 — 2013

B.S. Chemistry · B.A. Physics · B.A. Philosophy

The University of Texas at Austin

Three degrees across the physical, life, and formal sciences — plus early biophysics research with laser optical tweezers and molecular motors.

Selected publications
First author — Delineating mutational lineages with single-cell DNA sequencing using multi-patient specific panels.
Cell Genomics · 2022
Co-first author — Integrated genomic analysis of the ubiquitin pathway across cancer types.
Cell Reports · 2018
Co-first author — Predicting lung-cancer metastasis using evolution-aware graph transformers.
In preparation
Contributing author — Spatial transcriptomics reveals the impact of carbon deposition in lung adenocarcinoma.
In preparation

Additional co-authored work in Diabetologia, Regenerative Medicine, and Biochemistry Insights. Full list on Google Scholar.

Toolkit
R  tidyverse · expert Python  PyTorch · graph transformers Single-cell  DNA · RNA · ATAC Spatial  transcriptomics AI  CNN · GCN · image analysis HPC  Docker · Singularity · Kubernetes Digital  pathology
Off the clock

Before genomics there was a tennis court and a music stand — a competitive junior tennis player and a multi-instrumentalist long before a scientist. I still think the best problem-solving borrows from both: the pattern-recognition of sport and the structure of music. The philosophy degree just gave me the vocabulary to argue about it.

Get in touch

Let's talk.

Open to conversations about computational oncology, collaborations, and roles where rigorous data science meets real biological questions.