ADAMIXTURE Showcase

ADAMIXTURE is a fast CPU/GPU implementation of ADMIXTURE for biobank-scale genetic clustering. .P and .Q outputs remain compatible with ADMIXTURE.

Use Cases

Connectivity and dispersal mode shape the landscape genetics of a carnivorous pitcher plant-arthropod metacommunity

Authors: Nonno Hasegawa, Asa E. Conover, Matin Miryeganeh, David W. Armitage (2026)

Institution: Okinawa Institute of Science and Technology (OIST), University of California, Berkeley

Preprint: bioRxiv (DOI: 10.64898/2026.08.09.743144)

This study investigates how landscape connectivity, dispersal mode, and environmental resistance shape genetic structure in the carnivorous pitcher plant Darlingtonia californica (cobra lily) and two of its obligate arthropod mutualists (the midge Metriocnemus edwardsi and the mite Sarraceniopus darlingtoniae) sampled across sites in Oregon and northern California.

Application: The authors used ADAMIXTURE to estimate individual ancestry proportions across a nuclear genomic dataset of 19,166 SNPs in 199 individuals. They evaluated ancestral components across K = 2 to 15 with 5-fold cross-validation, which supported K = 3–4 as the most plausible regional clustering structure. The results revealed limited admixture across regional boundaries, supporting the central–marginal hypothesis and demonstrating that gene flow in the host plant is restricted by landscape resistance rather than geographic distance alone.

ADAMIXTURE ancestry proportions in Darlingtonia californica Figure: Estimated individual ancestry proportions for Darlingtonia californica from ADAMIXTURE (K = 3, K = 4 supported by 5-fold CV).
Darwin's Cats: Cat Ancestry Methods Overview

Author: Kasia Bryc (2026)

Institution: Darwin’s Ark, Broad Institute of Harvard and MIT

Report: Methods Overview (PDF) • Darwin’s Cats Ancestry Initiative

Darwin’s Cats is a community genomics initiative by the non-profit organization Darwin’s Ark in collaboration with researchers at the Broad Institute of Harvard and MIT. By analyzing low-pass whole-genome sequence data imputed against a high-coverage reference panel of 992 felines across 99.2 million variants, the project discovered that pet cat genetics in the United States is overwhelmingly shaped by geography and isolation-by-distance rather than breeds.

Application: Darwin’s Ark uses ADAMIXTURE to model each cat’s individual ancestry as an optimal mixture across an ancestry reference panel of 591 cats spanning 42 reference populations (13 North American regional populations, 3 world regions, 21 recognized cat breeds, and wild cat species including the European Wildcat Felis silvestris) on a linkage-disequilibrium thinned panel of 155,210 genome-wide SNPs. Ancestry proportions inferred by ADAMIXTURE classify participating felines into four major ancestry cohorts: Regional Roots (multigenerational place-based cats, 69.1%), Rich Tapestry (multi-population ancestry, 20.2%), Breed Legacy (purebred and cross-breed cats, 7.3%), and Uncharted Territory (underrepresented regional populations or breeds, 3.4%).

Darwin's Cats Ancestry Cohorts inferred with ADAMIXTURE Figure: Four broad cat ancestry cohorts identified in Darwin’s Cats using ADAMIXTURE (Regional Roots, Rich Tapestry, Breed Legacy, and Uncharted Territory across 1,958 sequenced cats).

Tools

JanusX: an integrated and high-performance platform for scalable genome-wide association studies and genomic selection

Authors: Jingxian Fu, Anqiang Jia, Haiyang Wang, Hai-Jun Liu (2026)

Institution: Huazhong Agricultural University, Yazhouwan National Laboratory

Preprint: bioRxiv (DOI: 10.64898/2026.01.20.700366)

JanusX is an integrated, high-performance computational platform developed for large-scale Genome-Wide Association Studies (GWAS) and Genomic Selection (GS), delivering up to 19× speedup over traditional tools like GEMMA for Linear Mixed Models (LMM).

Application: JanusX directly integrates ADAMIXTURE through its fastpop module to perform rapid unsupervised population structure inference. This integration enables researchers to calculate ancestry matrices (Q matrices), generate stacked ancestry proportion bar plots, and control for population stratification directly inside automated high-throughput GWAS and genomic prediction pipelines without separate manual pre-processing.

Submit a Use Case

Did you use ADAMIXTURE in your work? Please submit a short entry via GitHub:

Submit Use Case via GitHub Issues or open a Pull Request at AI-sandbox/adamixture-showcase.

Citation

If you use ADAMIXTURE in your research, please cite:

@article{saurina2026adamixture,
  title={ADAMIXTURE: adaptive first-order optimization for biobank-scale genetic clustering},
  author={Saurina-i-Ricos, Joan and Mas Montserrat, Daniel and Ioannidis, Alexander G.},
  journal={Bioinformatics},
  volume={42},
  number={Supplement\_1},
  pages={btag236},
  year={2026},
  doi={10.1093/bioinformatics/btag236},
  url={https://doi.org/10.1093/bioinformatics/btag236}
}