Improving cell type identification with Gaussian noise-augmented single-cell RNA-seq contrastive learning.

IF 2.5 3区生物学 Q3 BIOTECHNOLOGY & APPLIED MICROBIOLOGY

Briefings in Functional Genomics Pub Date : 2024-07-19 DOI:10.1093/bfgp/elad059

Ibrahim Alsaggaf, Daniel Buchan, Cen Wan

引用次数: 0

Abstract

Cell type identification is an important task for single-cell RNA-sequencing (scRNA-seq) data analysis. Many prediction methods have recently been proposed, but the predictive accuracy of difficult cell type identification tasks is still low. In this work, we proposed a novel Gaussian noise augmentation-based scRNA-seq contrastive learning method (GsRCL) to learn a type of discriminative feature representations for cell type identification tasks. A large-scale computational evaluation suggests that GsRCL successfully outperformed other state-of-the-art predictive methods on difficult cell type identification tasks, while the conventional random genes masking augmentation-based contrastive learning method also improved the accuracy of easy cell type identification tasks in general.

查看原文本刊更多论文

利用高斯噪声增强单细胞 RNA-seq 对比学习改进细胞类型识别。

细胞类型鉴定是单细胞 RNA 序列（scRNA-seq）数据分析的一项重要任务。最近提出了许多预测方法，但对困难的细胞类型鉴定任务的预测准确率仍然很低。在这项工作中，我们提出了一种新颖的基于高斯噪声增强的 scRNA-seq 对比学习方法（GsRCL），以学习一种用于细胞类型鉴定任务的判别特征表征。大规模计算评估表明，GsRCL 在高难度细胞类型鉴定任务中的表现成功地超越了其他最先进的预测方法，而传统的基于随机基因掩蔽增强的对比学习方法也普遍提高了简单细胞类型鉴定任务的准确率。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文求助全文

来源期刊

Briefings in Functional Genomics BIOTECHNOLOGY & APPLIED MICROBIOLOGY-GENETICS & HEREDITY

CiteScore

6.30

自引率

2.50%

发文量

审稿时长

6-12 weeks

期刊介绍： Briefings in Functional Genomics publishes high quality peer reviewed articles that focus on the use, development or exploitation of genomic approaches, and their application to all areas of biological research. As well as exploring thematic areas where these techniques and protocols are being used, articles review the impact that these approaches have had, or are likely to have, on their field. Subjects covered by the Journal include but are not restricted to: the identification and functional characterisation of coding and non-coding features in genomes, microarray technologies, gene expression profiling, next generation sequencing, pharmacogenomics, phenomics, SNP technologies, transgenic systems, mutation screens and genotyping. Articles range in scope and depth from the introductory level to specific details of protocols and analyses, encompassing bacterial, fungal, plant, animal and human data. The editorial board welcome the submission of review articles for publication. Essential criteria for the publication of papers is that they do not contain primary data, and that they are high quality, clearly written review articles which provide a balanced, highly informative and up to date perspective to researchers in the field of functional genomics.