de Koning Lab
Data•Models•Computation
We are a multidisciplinary group of computational biologists interested in genomics, evolution, and algorithms for understanding the fundamental processes of life in health and disease. We use probabilistic modelling, clever computation, and artificial intelligence to enable new insights and approaches for high-throughput biological discovery.
Hover or tap a portrait for the full profile.
Recent talks, posters and software from the lab. Select an entry to read it in full.
The “inside-out likelihood sandwich” removes the dependence of phylogenetic EM and Quasi-Newton methods on the number of free parameters, producing the exact gradient at only the cost of readout.
Exact quantities under Wright-Fisher models — fixation and absorption probabilities, sojourn and absorption times with their full distributions, allele age, establishment, and allele frequency spectra — by direct solution of the complete transition matrix, with neither simulation nor diffusion approximation. A C++ core, eleven command-line programs, and a new graphical interface, now with native Apple Silicon support.
A multiple sequence alignment viewer and editor that runs entirely in the browser — a GPU-accelerated renderer over a WebAssembly engine, fast enough to scroll hundreds of sequences smoothly. Try it in the page.
Standard branch-site tests have almost no power when selection is as sparse as it really is. Site-specific covariates from ESM-2 recover much of it.
Multi-nucleotide substitutions without multi-nucleotide mutations: rescue mutations hitchhiking out of a fitness valley, and what ignoring them does to inference.
Watterson's θ and θ_π hold up at θ = 0.001 and fall apart past ~0.05. Once recurrent mutation is common, the standard estimators systematically underestimate.
The complete Bayesian pipeline behind the infant-attachment paper — imputation, hierarchical ordinal regression, model comparison and every table and figure in the manuscript.
Forward-time population genetics over phylogenetic timescales: generates and runs SLiM from a Newick tree, keeping polymorphism and finite sites that standard phylogenetic models exclude by assumption.
Posterior expected convergent and divergent substitutions for every pair of independent branches in a phylogeny, with an interactive explorer for the results. The live demo reproduces the squamate analysis from Castoe, de Koning et al. (2009).
Phylogenetic simulation that returns the complete substitution history of every run, event by event, under mutation-selection models with site-specific fitness, epistasis between sites, and population sizes that change across the tree. Histories plot as interactive figures, with every substitution on its branch.
Open-source tools from the lab, on GitHub.
Rapid, interactive, probabilistic analysis of convergent molecular evolution
An educational project that began as a classroom exercise for MDSC 308 at the University of Calgary in fall 2025. Sequences Lined-up with Overwhelming Precision: a GPU-accelerated multiple sequence alignment viewer and editor that runs entirely in the browser.
WFES3 - Wright-Fisher Exact Solver: C++ library, CLI, and Electron GUI with native Apple Silicon support
SLiM-Tree a flexible simulation tool which automates pure population genetics simulations over phylogenetic timescales under realistic models of sequence-fitness relationships.
Simulate and visualize sequence evolution with full substitution histories under site- and time-heterogeneous mutation-selection models (and related approaches)
🦘 Clangaroo: Fast C++ code intelligence for LLMs via MCP
Wright-Fisher Exact Solver
NGS pipeline for viral metagenomics analysis
Wright Fisher Exact Solver
Handy quick reference for bioinformatics-related learning resources
Simulator to approximate the allele age probability distribution for a Wright-Fisher model
High performing variant-impact prediction scores for all missense mutations in the human exome
GUI implementation of wfes2.
No description
Ideal human genome reference sequences estimated from population and evolutionary data
A modified version of tabix query which allows to filter by variants
Repository data refreshed 2026-08-21
A.P. Jason de Koning
Department of Biochemistry & Molecular Biology
Cumming School of Medicine, University of Calgary
HSC 1151 (Genomics and Bioinformatics Annex)
Foothills Medical Campus
Calgary, Alberta, Canada
[email protected]
github.com/dekoning-lab
Administrative assistant to Dr. de Koning:
Monica Flohr Mauchline
[email protected]
To prospective recruits
16 August 2026
Thank you for your interest in our research group. Unfortunately, we are not generally recruiting new graduate students or research staff at this time. However, highly motivated and well-prepared candidates should feel free to reach out to Jason to inquire if a funded position could be coming available in the near future. Students who have secured their own external funding are also warmly encouraged to inquire.
