Towards Foundation Models for Cryo-ET Subtomogram Analysis

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Jiang, Runmin, Feng, Wanyue, Yang, Yuntian, Pingulkar, Shriya, Wang, Hong, Xiao, Xi, Cao, Xiaoyu, Zhang, Genpei, Wang, Xiao, Wu, Xiaolong, Wang, Tianyang, Liu, Yang, Li, Xingjian, Xu, Min
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908576885243904
author Jiang, Runmin
Feng, Wanyue
Yang, Yuntian
Pingulkar, Shriya
Wang, Hong
Xiao, Xi
Cao, Xiaoyu
Zhang, Genpei
Wang, Xiao
Wu, Xiaolong
Wang, Tianyang
Liu, Yang
Li, Xingjian
Xu, Min
author_facet Jiang, Runmin
Feng, Wanyue
Yang, Yuntian
Pingulkar, Shriya
Wang, Hong
Xiao, Xi
Cao, Xiaoyu
Zhang, Genpei
Wang, Xiao
Wu, Xiaolong
Wang, Tianyang
Liu, Yang
Li, Xingjian
Xu, Min
contents Cryo-electron tomography (cryo-ET) enables in situ visualization of macromolecular structures, where subtomogram analysis tasks such as classification, alignment, and averaging are critical for structural determination. However, effective analysis is hindered by scarce annotations, severe noise, and poor generalization. To address these challenges, we take the first step towards foundation models for cryo-ET subtomograms. First, we introduce CryoEngine, a large-scale synthetic data generator that produces over 904k subtomograms from 452 particle classes for pretraining. Second, we design an Adaptive Phase Tokenization-enhanced Vision Transformer (APT-ViT), which incorporates adaptive phase tokenization as an equivariance-enhancing module that improves robustness to both geometric and semantic variations. Third, we introduce a Noise-Resilient Contrastive Learning (NRCL) strategy to stabilize representation learning under severe noise conditions. Evaluations across 24 synthetic and real datasets demonstrate state-of-the-art (SOTA) performance on all three major subtomogram tasks and strong generalization to unseen datasets, advancing scalable and robust subtomogram analysis in cryo-ET.
format Preprint
id arxiv_https___arxiv_org_abs_2509_24311
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Towards Foundation Models for Cryo-ET Subtomogram Analysis
Jiang, Runmin
Feng, Wanyue
Yang, Yuntian
Pingulkar, Shriya
Wang, Hong
Xiao, Xi
Cao, Xiaoyu
Zhang, Genpei
Wang, Xiao
Wu, Xiaolong
Wang, Tianyang
Liu, Yang
Li, Xingjian
Xu, Min
Computer Vision and Pattern Recognition
Cryo-electron tomography (cryo-ET) enables in situ visualization of macromolecular structures, where subtomogram analysis tasks such as classification, alignment, and averaging are critical for structural determination. However, effective analysis is hindered by scarce annotations, severe noise, and poor generalization. To address these challenges, we take the first step towards foundation models for cryo-ET subtomograms. First, we introduce CryoEngine, a large-scale synthetic data generator that produces over 904k subtomograms from 452 particle classes for pretraining. Second, we design an Adaptive Phase Tokenization-enhanced Vision Transformer (APT-ViT), which incorporates adaptive phase tokenization as an equivariance-enhancing module that improves robustness to both geometric and semantic variations. Third, we introduce a Noise-Resilient Contrastive Learning (NRCL) strategy to stabilize representation learning under severe noise conditions. Evaluations across 24 synthetic and real datasets demonstrate state-of-the-art (SOTA) performance on all three major subtomogram tasks and strong generalization to unseen datasets, advancing scalable and robust subtomogram analysis in cryo-ET.
title Towards Foundation Models for Cryo-ET Subtomogram Analysis
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2509.24311