# BaGGLS: A Bayesian Shrinkage Framework for Interpretable Modeling of Interactions in High-Dimensional Biological Data

arXiv: https://arxiv.org/abs/2511.15330 · HTML: https://mlsanity.com/paper/2511.15330/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 16 of its 55 references that are in the map.
- First posted on arXiv 19 Nov 2025.

## Builds on (most cited first, 16 of 16)

- [Axiomatic Attribution for Deep Networks](https://mlsanity.com/paper/1703.01365/) (2017)
- [Variational Inference: A Review for Statisticians](https://mlsanity.com/paper/1601.00670/) (2016)
- [Variational Bayes for high-dimensional linear regression with sparse priors](https://mlsanity.com/paper/1904.07150/) (2019)
- [Spike and slab variational Bayes for high dimensional logistic regression](https://mlsanity.com/paper/2010.11665/) (2020)
- [Conjugate Bayes for probit regression via unified skew-normal distributions](https://mlsanity.com/paper/1802.09565/) (2018)
- [Scalable and Accurate Variational Bayes for High-Dimensional Binary Regression Models](https://mlsanity.com/paper/1911.06743/) (2019)
- [Conditionally conjugate mean-field variational Bayes for logistic models](https://mlsanity.com/paper/1711.06999/) (2017)
- [Bayesian conjugacy in probit, tobit, multinomial probit and extensions: A review and new results](https://mlsanity.com/paper/2206.08118/) (2022)
- [Technical Note on Transcription Factor Motif Discovery from Importance Scores (TF-MoDISco) version 0.5.6.5](https://mlsanity.com/paper/1811.00416/) (2018)
- [Feature selection revisited in the single-cell era](https://mlsanity.com/paper/2110.14329/) (2021)
- [Variational inference based on a subclass of closed skew normals](https://mlsanity.com/paper/2306.02813/) (2023)
- [Skew-symmetric approximations of posterior distributions](https://mlsanity.com/paper/2409.14167/) (2024)
- [Bayesian Sparse Global-Local Shrinkage Regression for Selection of Grouped Variables](https://mlsanity.com/paper/1709.04333/) (2017)
- [The R2D2 Prior for Generalized Linear Mixed Models](https://mlsanity.com/paper/2111.10718/) (2021)
- [Skewed Bernstein-von Mises theorem and skew-modal approximations](https://mlsanity.com/paper/2301.03038/) (2023)
- [Variational inference for hierarchical models with conditional scale and skewness corrections](https://mlsanity.com/paper/2503.18075/) (2025)

## Abstract (arXiv)

Biological data sets are often high-dimensional, noisy, and governed by complex interactions among sparse signals. This poses major challenges for interpretability and reliable feature selection. Tasks such as identifying motif interactions in genomics exemplify these difficulties, as only a small subset of biologically relevant features (e.g., motifs) are typically active, and their effects are often non-linear and context-dependent. While statistical approaches often result in more interpretable models, deep learning models have proven effective in modeling complex interactions and prediction accuracy, yet their black-box nature limits interpretability. We introduce BaGGLS, a flexible and interpretable probabilistic binary regression model designed for high-dimensional biological inference involving feature interactions. BaGGLS incorporates a Bayesian group global-local shrinkage prior, aligned with the group structure introduced by interaction terms. This prior encourages sparsity while retaining interpretability, helping to isolate meaningful signals and suppress noise. To enable scalable inference, we employ a partially factorized variational approximation that captures posterior skewness and supports efficient learning even in large feature spaces. In extensive simulations, we can show that BaGGLS outperforms the other methods with regard to interaction detection and is many times faster than MCMC sampling under the horseshoe prior. We also demonstrate the usefulness of BaGGLS in the context of interaction discovery from motif scanner outputs and noisy attribution scores from deep learning models. This shows that BaGGLS is a promising approach for uncovering biologically relevant interaction patterns, with potential applicability across a range of high-dimensional tasks in computational biology.