Interpretable DNA Sequence Classification via Dynamic Feature Generation in Decision Trees

Nicolas Huynh, Krzysztof Kacprzyk, Ryan M Sheridan, David L. Bentley, Mihaela van der Schaar
Proceedings of The 29th International Conference on Artificial Intelligence and Statistics, PMLR 300:2890-2898, 2026.

Abstract

The analysis of DNA sequences has become critical in numerous fields, from evolutionary biology to understanding gene regulation and disease mechanisms. While deep neural networks can achieve remarkable predictive performance, they typically operate as black boxes. Contrasting these black boxes, axis-aligned decision trees offer a promising direction for interpretable DNA sequence analysis, yet they suffer from a fundamental limitation: considering individual raw features in isolation at each split limits their expressivity, which results in prohibitive tree depths that hinder both interpretability and generalization performance. We address this challenge by introducing DEFT, a novel framework that adaptively generates high-level sequence features during tree construction. DEFT leverages large language models to propose biologically-informed features tailored to the local sequence distributions at each node and to iteratively refine them with a reflection mechanism. Empirically, we demonstrate that DEFT discovers human-interpretable and highly predictive sequence features across a diverse range of genomic tasks.

Cite this Paper

BibTeX
@InProceedings{pmlr-v300-huynh26a, title = { Interpretable DNA Sequence Classification via Dynamic Feature Generation in Decision Trees }, author = {Huynh, Nicolas and Kacprzyk, Krzysztof and Sheridan, Ryan M and Bentley, David L. and van der Schaar, Mihaela}, booktitle = {Proceedings of The 29th International Conference on Artificial Intelligence and Statistics}, pages = {2890--2898}, year = {2026}, editor = {Khan, Emtiyaz and Li, Yingzhen and Solin, Arno and Ramdas, Aaditya}, volume = {300}, series = {Proceedings of Machine Learning Research}, month = {02--05 May}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v300/main/assets/huynh26a/huynh26a.pdf}, url = {https://proceedings.mlr.press/v300/huynh26a.html}, abstract = { The analysis of DNA sequences has become critical in numerous fields, from evolutionary biology to understanding gene regulation and disease mechanisms. While deep neural networks can achieve remarkable predictive performance, they typically operate as black boxes. Contrasting these black boxes, axis-aligned decision trees offer a promising direction for interpretable DNA sequence analysis, yet they suffer from a fundamental limitation: considering individual raw features in isolation at each split limits their expressivity, which results in prohibitive tree depths that hinder both interpretability and generalization performance. We address this challenge by introducing DEFT, a novel framework that adaptively generates high-level sequence features during tree construction. DEFT leverages large language models to propose biologically-informed features tailored to the local sequence distributions at each node and to iteratively refine them with a reflection mechanism. Empirically, we demonstrate that DEFT discovers human-interpretable and highly predictive sequence features across a diverse range of genomic tasks. } }
Endnote
%0 Conference Paper %T Interpretable DNA Sequence Classification via Dynamic Feature Generation in Decision Trees %A Nicolas Huynh %A Krzysztof Kacprzyk %A Ryan M Sheridan %A David L. Bentley %A Mihaela van der Schaar %B Proceedings of The 29th International Conference on Artificial Intelligence and Statistics %C Proceedings of Machine Learning Research %D 2026 %E Emtiyaz Khan %E Yingzhen Li %E Arno Solin %E Aaditya Ramdas %F pmlr-v300-huynh26a %I PMLR %P 2890--2898 %U https://proceedings.mlr.press/v300/huynh26a.html %V 300 %X The analysis of DNA sequences has become critical in numerous fields, from evolutionary biology to understanding gene regulation and disease mechanisms. While deep neural networks can achieve remarkable predictive performance, they typically operate as black boxes. Contrasting these black boxes, axis-aligned decision trees offer a promising direction for interpretable DNA sequence analysis, yet they suffer from a fundamental limitation: considering individual raw features in isolation at each split limits their expressivity, which results in prohibitive tree depths that hinder both interpretability and generalization performance. We address this challenge by introducing DEFT, a novel framework that adaptively generates high-level sequence features during tree construction. DEFT leverages large language models to propose biologically-informed features tailored to the local sequence distributions at each node and to iteratively refine them with a reflection mechanism. Empirically, we demonstrate that DEFT discovers human-interpretable and highly predictive sequence features across a diverse range of genomic tasks.
APA
Huynh, N., Kacprzyk, K., Sheridan, R.M., Bentley, D.L. & van der Schaar, M.. (2026). Interpretable DNA Sequence Classification via Dynamic Feature Generation in Decision Trees . Proceedings of The 29th International Conference on Artificial Intelligence and Statistics, in Proceedings of Machine Learning Research 300:2890-2898 Available from https://proceedings.mlr.press/v300/huynh26a.html.

Related Material