Data-Driven Soft Labeling Scales DNA Read Classification to Whole-Body Cell-Type Deconvolution
Abstract
Cell-type deconvolution, the task of estimating the proportions of constituent cell types in a heterogeneous biological sample, is a core problem in computational biology. Methods that rely on epigenetic marks such as DNA methylation typically operate on aggregated methylation estimates, discarding the pattern-level information carried by individual DNA reads. Existing read-level approaches that exploit this information are scarce, and all remain restricted to few-class settings; scaling them further is an open problem because, at scale, non-discriminative reads dominate and hard labels conflict with the many-to-many mapping between methylation patterns and cell types, preventing classifier convergence. To overcome this, we propose data-driven soft labels that estimate the conditional cell-type distribution for each read, and integrate this scheme into Syto, a new modular framework for read-level classification-based deconvolution. On a whole-body atlas of 39 human cell types, Syto reduces MSE by 3.7× over the best examined baseline, with gains transferring to an out-of-distribution dataset spanning 16 tissues. Syto lays the foundation for modeling increasingly large cell-type panels, with improved applications in biology and healthcare. The proposed soft-labeling scheme is further translatable to any setting with a many-to-many signal-to-label mapping.
Related
- Whole genome similarity provides a rapid, robust framework for classification of fungal taxa from the genus rank to intraspecies variants
- Graph-Based Classification of Penetrating Arteries and Veins in Whole-Mouse Vasculature Datasets
- Interferon‐α–Driven Stratification of B Cells Reveals Metabolic Reprogramming of Double‐Negative, Naïve, and Transitional Cell Subsets and Refines Molecular Classification in Sjögren's Disease
- Benchmarking of cell type deconvolution techniques for microarray data
- Altered hemispheric lateralization of white matter pathways in developmental dyslexia: Evidence from spherical deconvolution tractography
BibTeX
@misc{rizdvanetskyi2026datadriven,
title = {{Data-Driven Soft Labeling Scales DNA Read Classification to Whole-Body Cell-Type Deconvolution}},
author = {Dmytro Rizdvanetskyi and Nathan Roos and Pavlo Lutsik},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2607.04987},
eprint = {2607.04987},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2607.04987},
note = {CC BY-SA 4.0},
}
RIS
TY - UNPB
TI - Data-Driven Soft Labeling Scales DNA Read Classification to Whole-Body Cell-Type Deconvolution
AU - Dmytro Rizdvanetskyi
AU - Nathan Roos
AU - Pavlo Lutsik
PY - 2026
DA - 2026/10/20
PB - arXiv
DO - 10.48550/arxiv.2607.04987
UR - https://arxiv.org/abs/2607.04987
AB - Cell-type deconvolution, the task of estimating the proportions of constituent cell types in a heterogeneous biological sample, is a core problem in computational biology. Methods that rely on epigenetic marks such as DNA methylation typically operate on aggregated methylation estimates, discarding the pattern-level information carried by individual DNA reads. Existing read-level approaches that exploit this information are scarce, and all remain restricted to few-class settings; scaling them further is an open problem because, at scale, non-discriminative reads dominate and hard labels conflict with the many-to-many mapping between methylation patterns and cell types, preventing classifier convergence. To overcome this, we propose data-driven soft labels that estimate the conditional cell-type distribution for each read, and integrate this scheme into Syto, a new modular framework for read-level classification-based deconvolution. On a whole-body atlas of 39 human cell types, Syto reduces MSE by 3.7× over the best examined baseline, with gains transferring to an out-of-distribution dataset spanning 16 tissues. Syto lays the foundation for modeling increasingly large cell-type panels, with improved applications in biology and healthcare. The proposed soft-labeling scheme is further translatable to any setting with a many-to-many signal-to-label mapping.
ER -