Balanced DNA interpolation improves learning of genetic distance-informed embeddings in plants
13 affiliations
- 1Biodiversity, Ecosystems and Society Group, Max Planck Institute for Biogeochemistry, Jena, Germany
- 2Data-intensive Systems and Visualization (dAI.SY), Technische Universität Ilmenau, Ilmenau, Germany
- 3Laboratory of Molecular Systematics and Evolutionary Genetics, Florida Museum of Natural History, University of Florida, Gainesville, Florida, United States of America
- 4Professorship of Integrative Plant Taxonomy, Friedrich Schiller University Jena, Jena, Germany
- 5Unit of Evolutionary Biology and Ecology CP160/12, Université Libre de Bruxelles, Brussels, Belgium
- 6Evolutionary Biology Research Group (GIBE), Departamento de Bioloxía, Facultade de Ciencias, Universidade da Coruña, A Coruña, Spain
- 7Department of Biological Sciences, Boise State University, Boise, Idaho, United States of America
- 8Evolutionary Biology Research Group (GIBE), Centre of Information and Communication Technologies (CITIC), Universidade da Coruña, A Coruña, Spain
- 9Singapore Botanic Gardens, National Parks Board, Singapore, Singapore
- 10Taxonomy and Macroecology Group, Royal Botanic Garden Edinburgh, Edinburgh, United Kingdom
- 11Evolutionary Biology Research Group (GIBE), Interdisciplinary Centre of Chemistry and Biology (CICA), Universidade da Coruña, A Coruña, Spain
- 12Institute of Biodiversity, Ecology and Evolution, Friedrich Schiller University Jena, Jena, Germany
- 13German Centre for Integrative Biodiversity Research (iDiv) Halle-Jena-Leipzig, Leipzig, Germany
Abstract
In taxonomic research, traditional phylogenetic tree and structure analyses of genetic data are increasingly complemented by machine-learning-based identification and representation learning. Although the amount of DNA data needed to train state-of-the-art machine learning models often exceeds what can realistically be collected and sequenced in biological studies, the number of samples can be extended artificially through data augmentation. Genetic data augmentation usually refers to the introduction of random base variations, translocations, and reverse complementing. These augmentations do not take into account the inherent structures of populations and species, potentially blurring the lines between entities within genetic datasets. Here, we propose DNAInterpolator, an approach based on interpolation of DNA sequences within a given dataset that presents a neighbor-guided alternative to random mutations. We tested interpolation as an augmentation technique using four flowering plant datasets and an artificial neural network trained to predict genetic distances between paired samples. To address unequally distributed distances within our training datasets, we examined the effect of balancing the distance distribution by curating interpolated sequences. We found that balancing helps models capture genetic distances across the full distance range by strengthening performance in underrepresented regions of the distribution. Our new approach leverages the potential of taxonomic DNA datasets for modern machine learning applications. Author summary: In recent years, interest in the study of genetic data using machine-learning-based approaches is increasing. However, for research focusing on fine-grained species classification or delimitation, limited genetic data exist. The finite amount of readily available fine-grained data and the challenges researchers face in material collection, e.g., narrow endemics or remote areas, constitute a significant issue when aiming to finetune state-of-the-art genome foundation models or train large models from scratch. Up-sampling genetic data through DNA augmentation methods has been shown to improve model performance. Typically, these augmentation methods introduce random transmutations, potentially blurring species boundaries and population structures inherent to a given dataset. We examined the potential of an offline augmentation technique that entails interpolating between samples of the same species to add semi-artificial DNA sequences to the dataset. Our approach is transparent, interpretable, and independent from encoding methods or subsequent analyses.
Related
- NeuroLens: Learning Latent Embeddings of Neural Semantics from Chronic Recordings
- Chemical Descriptors and Deep Learning Embeddings for Scoring de novo Peptide Designs
- MOSAIC: Learning Graph Node Embeddings from Spectrally Isolated Dominant Subspaces of Accumulated Diffusion Operators
- Interrogating contrastive learning embeddings for structure-based virtual screening: a case study on DrugCLIP
- BCFL-DAI-PUB: Blockchain-integrated federated learning with dynamic adaptive incentives for privacy-utility balanced credit risk modelling
BibTeX
@article{kosters2026balanced,
title = {{Balanced DNA interpolation improves learning of genetic distance-informed embeddings in plants}},
author = {Lara M. Kösters and Kevin Karbstein and Ladislav Hodač and Laura Albreht and Elvira Sahuquillo Balbuena and Daniel Botello and Olivier Hardy and Phebian Odufuwa and Eva Pardo Otero and Aireen Phang and Manuel Pimentel and Rosalía Piñeiro and James Smith and Peter Wilkie and Patrick Mäder and Jana Wäldchen},
journal = {PLOS Computational Biology},
year = {2026},
month = sep,
doi = {10.1371/journal.pcbi.1014722},
url = {https://journals.plos.org/ploscompbiol/article?id=10.1371/journal.pcbi.1014722},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - Balanced DNA interpolation improves learning of genetic distance-informed embeddings in plants
AU - Lara M. Kösters
AU - Kevin Karbstein
AU - Ladislav Hodač
AU - Laura Albreht
AU - Elvira Sahuquillo Balbuena
AU - Daniel Botello
AU - Olivier Hardy
AU - Phebian Odufuwa
AU - Eva Pardo Otero
AU - Aireen Phang
AU - Manuel Pimentel
AU - Rosalía Piñeiro
AU - James Smith
AU - Peter Wilkie
AU - Patrick Mäder
AU - Jana Wäldchen
PY - 2026
DA - 2026/09/03
JO - PLOS Computational Biology
DO - 10.1371/journal.pcbi.1014722
UR - https://journals.plos.org/ploscompbiol/article?id=10.1371/journal.pcbi.1014722
AB - In taxonomic research, traditional phylogenetic tree and structure analyses of genetic data are increasingly complemented by machine-learning-based identification and representation learning. Although the amount of DNA data needed to train state-of-the-art machine learning models often exceeds what can realistically be collected and sequenced in biological studies, the number of samples can be extended artificially through data augmentation. Genetic data augmentation usually refers to the introduction of random base variations, translocations, and reverse complementing. These augmentations do not take into account the inherent structures of populations and species, potentially blurring the lines between entities within genetic datasets. Here, we propose DNAInterpolator, an approach based on interpolation of DNA sequences within a given dataset that presents a neighbor-guided alternative to random mutations. We tested interpolation as an augmentation technique using four flowering plant datasets and an artificial neural network trained to predict genetic distances between paired samples. To address unequally distributed distances within our training datasets, we examined the effect of balancing the distance distribution by curating interpolated sequences. We found that balancing helps models capture genetic distances across the full distance range by strengthening performance in underrepresented regions of the distribution. Our new approach leverages the potential of taxonomic DNA datasets for modern machine learning applications. Author summary: In recent years, interest in the study of genetic data using machine-learning-based approaches is increasing. However, for research focusing on fine-grained species classification or delimitation, limited genetic data exist. The finite amount of readily available fine-grained data and the challenges researchers face in material collection, e.g., narrow endemics or remote areas, constitute a significant issue when aiming to finetune state-of-the-art genome foundation models or train large models from scratch. Up-sampling genetic data through DNA augmentation methods has been shown to improve model performance. Typically, these augmentation methods introduce random transmutations, potentially blurring species boundaries and population structures inherent to a given dataset. We examined the potential of an offline augmentation technique that entails interpolating between samples of the same species to add semi-artificial DNA sequences to the dataset. Our approach is transparent, interpretable, and independent from encoding methods or subsequent analyses.
SN - 1553-7358
ER -