A Transformer-Based Deep Diffusion Model for Bulk RNA-Seq Deconvolution

Yunqing Liu

doi:10.3390/biology14091150

2025-09-01

王国强 (Guoqiang Wang) *

Jianwei Wu *

DOI：10.3390/biology14091150

原文链接

原文求助

摘要

En 中文

背景：Bulk RNA-seq是一种经济高效的方法，用于测量组织样本中基因表达的均值，但其缺乏单细胞分辨率限制了人们对细胞异质性的理解。计算去卷积旨在从Bulk RNA-seq数据中推断细胞类型比例；然而，现有方法的准确性有待提高，尤其是在复杂组织中。方法：本研究中，我们介绍了DiffFormer，一种新颖的去卷积模型，首次将条件扩散模型与Transformer架构相结合。我们在四个伪Bulk数据集上对DiffFormer进行了系统评估，并在基于FACS的金标准真实数据集上进行了验证。结果：DiffFormer在所有测试数据集上均表现出一致且优异的性能，优于现有方法和基于MLP的基准扩散模型（DiffMLP）。例如，在pbmc3k数据集上，DiffFormer将均方根误差（RMSE）从0.1060降低至0.0120，相较于DiffMLP。这一优势在真实数据集上得到进一步确认，DiffFormer实现了最高的皮尔逊相关系数（PCC）。结论：本研究提供了一种高精度、可复现的细胞去卷积工具。关键在于，与基于MLP的扩散模型的直接比较提供了确凿证据，证明Transformer架构是其成功的关键，突显了此类模型解决复杂生物信息学问题的潜力。

Keyword:

cellular deconvolution

diffusion model

Transformer architecture

bulk RNA-seq

single-cell resolution

期刊

Biology

IF:

3.5

论文数:

7.1K

被引数:

2.3W

机构

暂无机构信息

引用论文

Bulk tissue cell type deconvolution with multi-subject single-cell expression reference 具有多受试者单细胞表达参考的大块组织细胞类型反卷积

NATURE COMMUNICATIONS

IF15.7

2019-01-22

866

OAAI

Wang, Xuran; Park, Jihwan; Susztak, Katalin; Zhang, Nancy R.; Li, Mingyao

Orchestrating single-cell analysis with Bioconductor 使用Bioconductor协调单细胞分析

NATURE METHODS

IF32.1

2019-12-02

438

OAAI

Amezquita, Robert A.; Lun, Aaron T. L.; Becht, Etienne; Carey, Vince J.; Carpp, Lindsay N.; Geistlinger, Ludwig; Marini, Federico; Rue-Albrecht, Kevin; Risso, Davide; Soneson, Charlotte; Waldron, Levi; Pages, Herve; Smith, Mike L.; Huber, Wolfgang; Morgan, Martin; Gottardo, Raphael; Hicks, Stephanie C.

Single-cell RNA-seq reveals new types of human blood dendritic cells, monocytes, and progenitors 单细胞rna-seq揭示了新型的人血树突状细胞，单核细胞和祖细胞

Science

IF0

2017-04-21

OAAI

Alexandra-Chloé Villani; Rahul Satija; Gary Reynolds; Siranush Sarkizova; Karthik Shekhar; James Fletcher; Morgane Griesbeck; Andrew Butler; Shiwei Zheng; Suzan Lazo; Laura Jardine; David Dixon; Emily Stephenson; Emil Nilsson; Ida Grundberg; David McDonald; Andrew Filby; Weibo Li; Philip L. De Jager; Orit Rozenblatt-Rosen; Andrew A. Lane; Muzlifah Haniffa; Aviv Regev; Nir Hacohen

A comparison of marker gene selection methods for single-cell RNA sequencing data 单细胞RNA测序数据标记基因选择方法的比较

GENOME BIOLOGY

IF9.4

2024-02-26

OAAI

Pullin, Jeffrey M.; McCarthy, Davis J.

A Single-Cell Transcriptome Atlas of the Human Pancreas 人类胰腺的单细胞转录组图谱

CELL SYSTEMS

IF7.7

2016-10-01

875

OAAI

Muraro, Mauro J.; Dharmadhikari, Gitanjali; Gruen, Dominic; Groen, Nathalie; Dielen, Tim; Jansen, Erik; van Gurp, Leon; Engelse, Marten A.; Carlotti, Francoise; de Koning, Eelco J. P.; van Oudenaarden, Alexander

Cell2location maps fine-grained cell types in spatial transcriptomics Cell2location在空间转录组学中映射细粒度的细胞类型

NATURE BIOTECHNOLOGY

IF41.7

2022-01-13

378

PREAI

Kleshchevnikov, Vitalii; Shmatko, Artem; Dann, Emma; Aivazidis, Alexander; King, Hamish W.; Li, Tong; Elmentaite, Rasa; Lomakin, Artem; Kedlian, Veronika; Gayoso, Adam; Jain, Mika Sarkin; Park, Jun Sung; Ramona, Lauma; Tuck, Elizabeth; Arutyunyan, Anna; Vento-Tormo, Roser; Gerstung, Moritz; James, Louisa; Stegle, Oliver; Bayraktar, Omer Ali

SimBu: bias-aware simulation of bulk RNA-seq data with variable cell-type composition SimBu: 具有可变细胞类型组成的批量rna-seq数据的偏差感知模拟

Bioinformatics

IF0

2022-09-18

OAAI

Alexander Dietrich; Gregor Sturm; Lorenzo Merotto; Federico Marini; Francesca Finotello; Markus List

Comparison of Values of Pearson's and Spearman's Correlation Coefficients on the Same Sets of Data

QUAGEO

IF0

2011-06-24

OAAI

Jan Hauke; Tomasz Kossowski