vir2vec:一个全基因组的病毒嵌入
bioRxiv : the preprint server for biology
|December 22, 2025
概括
我们开发了vir2vec,这是一个大型基因组语言模型,用于训练各种病毒基因组,以及vGUE,用于评估病毒基因组理解的基准. vir2vec显著改善了病毒分类任务,推进了基因组监测和发现.
科学领域:
- 基因组学就是基因组学.
- 生物信息学是一种生物信息学.
- 机器学习 机器学习
背景情况:
- 基因组语言模型 (gLMs) 显示了DNA分析的前景,但缺乏针对病毒的特定架构和基准.
- 现有的模型往往侧重于人类DNA或有限的病毒数据集.
- 需要一个全面的评估框架来学习病毒基因组表征.
研究的目的:
- 介绍vir2vec,这是一个大规模的基因组语言模型,在多样化的泛病毒群体上进行预训练.
- 介绍vGUE,这是评估病毒基因组理解和表示学习的统一基准.
- 评估vir2vec在各种病毒基因组预测任务上的表现.
主要方法:
- 在来自295种物种的565,747个完整病毒基因组上持续预训练Mistral-DNA,以创建vir2vec (422M参数).
- 在嵌套交叉验证下,开发了vGUE,这是一个使用vir2vec嵌入在分类器 (逻辑回归,SVM,随机森林) 中的基准.
- 预测任务的评估,包括生物的歧视,进化指纹,物种的分离,变异类型和表型上下文检测.
主要成果:
- 在八个不同的病毒分类任务中,vir2vec在七个任务中取得了卓越的平衡精度.
- 该模型的表现始终优于人类DNA训练和现有的病毒特异性基因组基础模型.
- 来自vir2vec的嵌入物有效地捕获了跨多个任务的生物相关病毒变异.
结论:
- vir2vec和vGUE为病毒基因组建模,监测和发现奠定了坚实的基础.
- 开发的工具为了解病毒多样性和进化提供了更好的能力.
- 由于其潜在的双重用途影响,对vir2vec的负责任部署需要道德考虑和治理监督.
相关概念视频
01:26Size and Structure of Viral Genomes
616
Viral genomes exhibit remarkable diversity in size, structure, and composition, influencing their replication strategies and interactions with host cells. These genomes consist of either DNA or RNA and may be linear or circular. Additionally, they can be single-stranded or double-stranded, with each configuration affecting how the virus propagates within a host. RNA viruses, for instance, generally have smaller genomes than DNA viruses, a factor that contributes to their high mutation rates and...
616
01:29Viruses with RNA Genomes
714
RNA viruses are categorized into positive-strand, negative-strand, or double-stranded groups based on their genomic structure and replication mechanisms. This classification dictates how they exploit host cellular machinery for protein synthesis and replication. Some RNA viruses also utilize reverse transcription as part of their life cycle, further diversifying their replication strategies.Positive-Strand RNA VirusesPositive-strand RNA viruses have genomes that function directly as messenger...
714
00:57Viral Recombination
24.8K
Cells are sometimes infected by more than one virus at once. When two viruses disassemble to expose their genomes for replication in the same cell, similar regions of their genomes can pair together and exchange sequences in a process called recombination. Alternatively, viruses with segmented genomes can swap segments in a process called reassortment.
24.8K
00:36Viral Mutations
39.5K
A mutation is a change in the sequence of bases of DNA or RNA in a genome. Some mutations occur during replication of the genome due to errors made by the polymerase enzymes that replicate DNA or RNA. Unlike DNA polymerase, RNA polymerase is prone to errors because it is not capable of “proofreading” its work. Viruses with RNA-based genomes, like HIV, therefore accrue mutations faster than viruses with DNA-based genomes. Because mutation and recombination provide the raw material...
39.5K
00:56Viral Structure
73.6K
Viruses are extraordinarily diverse in shape and size, but they all have several structural features in common. All viruses have a core that contains a DNA- or RNA-based genome. The core is surrounded by a protective coat of proteins called the capsid. The capsid is composed of subunits called capsomeres. The capsid and genome-containing core are together known as the nucleocapsid.
73.6K
02:33Retroviruses
14.5K
Retroviruses and retrotransposons both insert copies of their genetic elements into the genome of the host cell. Thus, the viral genes are passed on when the host genome is replicated or translated. A typical retroviral DNA sequence contains 3-4 genes that encode the different proteins required for its structural assembly and function as a molecular parasite. This DNA is transcribed into a single mRNA, which is very similar in structure to conventional mRNAs, i.e., it is capped at the 5’...
14.5K