CryoConvNeXt enables robust identification of low-abundance molecular species in experimental cryo-EM data
- 1University of Basel
Abstract
Rare but biologically important molecular species are easily lost when reconstruction-based cryo-EM classification is applied to heterogeneous samples dominated by more abundant particles. We developed CryoConvNeXt, a deep-learning classifier that combines cyclic equivariance with adaptive frequency filtering. It is trained on simulated projections and adapted to experimental data by self-training, an unsupervised domain adaptation technique where the model acts as its own annotator. We tested CryoConvNeXt on cryo-EM data collected for this study from controlled binary and ternary mixtures of Catalase, Apoferritin, and HSP60. Manual particle curation provided reference labels for benchmarks with class ratios from 1:1 to 1:16. CryoConvNeXt retained minority-species recall across this range. In contrast, cryoSPARC's recall collapsed in four of six pairwise conditions, as early as 1:4 when Catalase was the minority species. By recovering low-abundance species that conventional classification fails to recover reliably, CryoConvNeXt takes an important step towards solving a central problem in quantitative visual proteomics: measuring the molecular composition of complex biological samples.
Related
- An Integrated Experimental and Informatics Strategy Enables Large Proteoform Identification in Top-Down Proteomics
- SpaMOAL is a deep learning method that enables accurate spatial domain identification from multi-omics data
- Spatial proteomic analysis in human Alzheimer’s disease brains enables identification of microenvironment-dependent microglial cell states
- From in silico prediction to experimental validation: Identification of drugs and novel synergistic combinations that inhibit growth of inflammatory breast cancer cells
- Experimental workflows for the accurate identification of mitochondrial redox events
BibTeX
@misc{glass2026cryoconvnext,
title = {{CryoConvNeXt enables robust identification of low-abundance molecular species in experimental cryo-EM data}},
author = {L. Glass and J. P. Abrahams and T. Braun},
howpublished = {bioRxiv},
year = {2026},
month = sep,
doi = {10.64898/2026.09.08.750038},
url = {https://www.biorxiv.org/content/10.64898/2026.09.08.750038v1},
note = {CC BY-NC-ND},
}
RIS
TY - UNPB
TI - CryoConvNeXt enables robust identification of low-abundance molecular species in experimental cryo-EM data
AU - L. Glass
AU - J. P. Abrahams
AU - T. Braun
PY - 2026
DA - 2026/09/08
PB - bioRxiv
DO - 10.64898/2026.09.08.750038
UR - https://www.biorxiv.org/content/10.64898/2026.09.08.750038v1
AB - Rare but biologically important molecular species are easily lost when reconstruction-based cryo-EM classification is applied to heterogeneous samples dominated by more abundant particles. We developed CryoConvNeXt, a deep-learning classifier that combines cyclic equivariance with adaptive frequency filtering. It is trained on simulated projections and adapted to experimental data by self-training, an unsupervised domain adaptation technique where the model acts as its own annotator. We tested CryoConvNeXt on cryo-EM data collected for this study from controlled binary and ternary mixtures of Catalase, Apoferritin, and HSP60. Manual particle curation provided reference labels for benchmarks with class ratios from 1:1 to 1:16. CryoConvNeXt retained minority-species recall across this range. In contrast, cryoSPARC's recall collapsed in four of six pairwise conditions, as early as 1:4 when Catalase was the minority species. By recovering low-abundance species that conventional classification fails to recover reliably, CryoConvNeXt takes an important step towards solving a central problem in quantitative visual proteomics: measuring the molecular composition of complex biological samples.
ER -