This package brings together HyperTraPS(-CT), HyperHMM, HyperLAU, HyperDAGs and HyperMk, different algorithms for accumulation modelling. All are based around the idea of inferring hypercubic transition graphs that describe the accumulation of binary features.
Install with
remotes::install_github("StochasticBiology/hyperinf")
This includes content from:
- HyperTraPS [1] https://github.com/StochasticBiology/hypertrapsct
- HyperMk [2] https://github.com/StochasticBiology/hypermk
- HyperHMM [3] https://github.com/StochasticBiology/hyperhmm
- HyperDAGs [4] https://github.com/StochasticBiology/hyperdags
- HyperLAU [5] https://github.com/StochasticBiology/hyperlau
- ideas (but not code) from Cluster-HyperHMM [6]
Wrapper function hyperinf produces fitted hypercubic inference models from data; plot_hyperinf produces summary plots of the inferred transition graph. Arguments to hyperinf specify which fitting approach to use (default is to choose based on data structure). full_to_squared_fit converts a fully-parameterised hypercube (including those output from HyperHMM, HyperMK, HyperLAU) to a best estimate of an L2-parameterised HyperTraPS model (encoding pairwise interactions between features). plot_hyperinf_comparative and plot_hyperinf_bubbles produce comparative plots summarising dynamics across a list of model fits; plot_hyperinf_bootstrap compares bubble plots from two bootstrapped model fits. Functionality in ordering_matrix, compare_orderings, and plot_hyperinf_compare_orderings supports comparisons between matrices describing probabilities that feature i is acquired before j, where j is either another feature or an ordering.
This will check and demo some of the functionality, based on HyperHMM. A more illustrative test bed is in inst/test-bed.R.
library(hyperinf)
# construct a simple dataset
data = matrix(rep(c(0,0,1, 0,1,1, 1,1,1), 10), byrow = TRUE, ncol=3, nrow=30)
# do model fits with bootstrapping to this dataset and its "inverse"
fit.1 = hyperinf(data, boot.parallel = 50)
fit.2 = hyperinf(1-data, boot.parallel = 50)
# plot the data
plot_hyperinf_data(data)
# plot a single model fit
plot_hyperinf(fit.1)
# compare transition networks
plot_hyperinf_comparative(list(fit.1, fit.2))
# compare summary "bubble" plots
plot_hyperinf_bubbles(list(fit.1, fit.2), p.scale = 0.2)
# compare bootstrapped bubble plots
plot_hyperinf_bootstrap(fit.1, fit.2)
[1] Aga, O.N., Brun, M., Dauda, K.A., Diaz-Uriarte, R., Giannakis, K. and Johnston, I.G., 2024. HyperTraPS-CT: Inference and prediction for accumulation pathways with flexible data and model structures. PLOS Computational Biology, 20(9), p.e1012393.
[2] Johnston, I.G. and Diaz-Uriarte, R., 2025. A hypercubic Mk model framework for capturing reversibility in disease, cancer, and evolutionary accumulation modelling. Bioinformatics, 41(1), p.btae737.
[3] Moen, M.T. and Johnston, I.G., 2023. HyperHMM: efficient inference of evolutionary and progressive dynamics on hypercubic transition graphs. Bioinformatics, 39(1), p.btac803.
[4] Giannakis, K., Aga, O.N., Moen, M.T., Drange, P.G. and Johnston, I.G., 2024. Identifying parsimonious pathways of accumulation and convergent evolution from binary data. bioRxiv, pp.2024-11.
[5] Renz, J., Brun, M. and Johnston, I.G., 2025. Flexible inference of evolutionary accumulation dynamics using uncertain observational data. arXiv preprint arXiv:2502.05872.
[6] Dauda, K.A., Aga, O.N. and Johnston, I.G., 2025. Clustering large-scale biomedical data to model dynamic accumulation processes in disease progression and anti-microbial resistance evolution. IEEE Access, 13, pp.13816-13831.