BioMoR

BioMoR (Bioinformatics Modeling with Recursion and Autoencoder-Based Ensemble) is an R package for advanced bioinformatics predictive modeling. It provides tools for constructing recursive transformer-inspired architectures, autoencoders, random forests, XGBoost models, and stacked ensembles. BioMoR includes utilities for cross-validation, calibration, benchmarking, and threshold optimization in predictive modeling workflows.


Installation

Install the development version directly from GitHub:

# install.packages("remotes") # if not installed
# remotes::install_github("Sulkysubject37/BioMoR")
install.packages("BioMoR", repos = c('https://sulkysubject37.r-universe.dev', 'https://cloud.r-project.org'))

Dependencies:
	•	Imports: caret, recipes, themis, xgboost, magrittr, dplyr, pROC
	•	Suggests: randomForest, testthat (>= 3.0.0), PRROC, ggplot2, purrr, tibble, yardstick, knitr, rmarkdown

⸻

Features
	•	Recursive transformer-inspired architectures for complex bioinformatics datasets.
	•	Autoencoders for dimensionality reduction and feature extraction.
	•	Ensemble learning: stacked models combining random forests, XGBoost, and other algorithms.
	•	Predictive modeling utilities:
	•	Cross-validation and calibration
	•	Benchmarking and evaluation
	•	Threshold optimization for classification tasks
	•	Integration with caret, recipes, and themis for preprocessing and balanced modeling workflows.
	•	Visualization and reporting using ggplot2, tibble, yardstick, knitr, and rmarkdown.

⸻

Example Usage

library(BioMoR)
library(dplyr)

# Example dataset
data <- matrix(rnorm(500), nrow = 50, ncol = 10)

# Train an autoencoder-based ensemble
model <- BioMoR::train_autoencoder_ensemble(data)

# Evaluate predictions
preds <- BioMoR::predict_ensemble(model, newdata = data)

# Benchmark performance
results <- BioMoR::benchmark_model(preds, true_labels = sample(0:1, 50, replace = TRUE))


⸻

Documentation

Full documentation is available in R:

help(package = "BioMoR")
?BioMoR::train_autoencoder_ensemble

You can also generate vignettes:

vignette("BioMoR")


⸻

Testing

Unit tests are implemented with testthat (>=3.0.0):

library(testthat)
test_package("BioMoR")


⸻

Contributing

Contributions, bug reports, and feature requests are welcome. Please use GitHub Issues or submit pull requests: GitHub - BioMoR

⸻

License

BioMoR is licensed under MIT License. See LICENSE for details.

⸻

Citation

If you use BioMoR in your research, please cite it as:

MD. Arshad (2025). BioMoR: Bioinformatics Modeling with Recursion and Autoencoder-Based Ensemble. R package version 0.1.2
https://doi.org/10.5281/zenodo.17927462


⸻

Acknowledgements
	•	Inspired by modern bioinformatics and machine learning approaches.
	•	Built with R, leveraging xgboost, randomForest, and deep learning-inspired ensemble methods.