arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2409.19482q-bio.QM
收藏

破译细胞系统:单细胞数据分析的机器学习视角与方法

Deciphering Cell Systems: Machine Learning Perspectives And Approaches For The Analysis Of Single-Cell Data

Yongjian Yang

AI总结:

本论文利用现代神经网络和系统生物学方法,针对单细胞RNA测序数据,提出流形学习优化、基因调控网络生成模型和单细胞基因敲除模拟等新方法,并增强黑盒模型可解释性,从而深化对基因型-表型关系及疾病机制的理解。

摘要:

本论文探讨了机器学习在分子生物学中的应用,重点关注单细胞水平上的基因表达调控和细胞行为。利用现代神经网络,本研究解决了细胞间通讯、基因功能推断和蛋白质表达分析中的关键挑战,特别关注单细胞RNA测序(scRNA-seq)数据。研究开发了整合系统生物学和神经网络技术的先进计算方法,以应对单细胞数据的复杂性和高维性,从而加深了对基因型-表型关系的理解。本工作为流形学习中的优化问题提出了新颖的解决方案,探索了基因调控网络的生成模型,并在单细胞分辨率下模拟了基因敲除。此外,本研究增强了黑盒神经模型对多模态数据的可解释性。对细胞生物学的关键贡献包括:分析细胞间相互作用及其在塑造细胞行为中的作用,通过敲除模拟进行基因功能预测,以及研究基因表达模式如何转化为蛋白质表达。这些发现对理解疾病机制和治疗开发具有重要意义。总体而言,本论文通过为单细胞分析提供新的工具和见解,推动了计算生物学的发展,为未来研究细胞行为和多种疾病的潜在治疗方法提供了宝贵资源。

英文摘要:

This dissertation explores the application of machine learning in molecular biology, focusing on gene expression regulation and cellular behavior at the single-cell level. Using modern neural networks, the research addresses key challenges in cell-cell communication, gene function inference, and protein expression analysis, with a special focus on single-cell RNA sequencing (scRNA-seq) data. Advanced computational methodologies integrating systems biology and neural network techniques were developed to handle the complexity and high-dimensionality of single-cell data, leading to a deeper understanding of genotype-phenotype relationships. This work proposes novel solutions to optimization problems in manifold learning, explores generative models for gene regulatory networks, and simulates gene knockout at single-cell resolution. Furthermore, the research enhances the interpretability of black-box neural models for multimodal data. Key contributions to cellular biology include the analysis of cell-cell interactions and their role in shaping cellular behavior, gene function prediction through knockout simulations, and the investigation of how gene expression patterns translate into protein expression. These findings have implications for understanding disease mechanisms and therapeutic development. Overall, this dissertation advances computational biology by providing new tools and insights into single-cell analysis, offering a valuable resource for future studies on cellular behavior and potential treatments for various diseases.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑