对属性矩阵和LLM嵌入的两阶段概率增强回归使得通过修改的siRNAs能够对基因敲除进行最先进的预测
Ivan Golovkin1, Denis Shatkovskii1, Nikita Serov1
1Center for Artificial Intelligence in Chemistry, ITMO University, 191002 Saint-Petersburg, Russia.
International journal of molecular sciences
|December 30, 2025
概括
这项研究引入了一种新的机器学习管道,用于预测小干扰RNA (siRNA) 基因淘汰活性. 该模型增强了化学修饰siRNAs的设计,提高了治疗疗效.
科学领域:
- 生物技术和生物信息学
- 计算化学计算化学
- 基因组医学是基因组医学.
背景情况:
- 自2018年以来批准的六种小型干扰RNA (siRNA) 突出显示了它们通过选择性基因淘汰的治疗潜力.
- siRNA设计复杂,化学修饰对稳定性和治疗半衰期至关重要.
- 机器学习 (ML) 为预测siRNA有效性和非目标效应提供了先进的分析.
研究的目的:
- 开发一种用于预测化学修饰siRNAs基因淘汰活性的新管道.
- 为了利用构成意识的属性矩阵和大型语言模型 (LLM) 嵌入来增强siRNA设计.
- 为了对预测siRNA活性中的目标基因编码进行各种LLM的基准测试.
主要方法:
- 一个新的管道,集成siRNA化学成分感知性质矩阵和LLM嵌入用于目标基因编码.
- 对包括Mistral 7B在内的通用和特定领域的LLM进行基准测试,以预测siRNA活动.
- 一个两阶段的概率增强模型,以解决数据不平衡并提高预测准确性.
主要成果:
- 与在基因组数据上预先训练的模型相比,Mistral 7B LLM表现出卓越的性能.
- 提出的模型在未见的数据上实现了R2 = 0.84和RMSE = 12.27%的最先进的质量.
- 留下一个基因的实验证实了该模型对未见的基因进行概括的能力,表明了强大的特征和嵌入式表示.
结论:
- 开发的管道有效地预测了化学修饰siRNAs的基因淘汰活性.
- 该模型通过整合化学特性和基因嵌入来增强siRNA设计,提高治疗疗效.
- 这项工作推进了下一代核酸疗法的组成意识siRNA设计领域.
相关概念视频
Experimental RNAi
7.2K
RNA interference (RNAi) is a cellular mechanism that inhibits gene expression by suppressing its transcription or activating the RNA degradation process. The mechanism was discovered by Andrew Fire and Craig Mello in 1998 in plants. Today, it is observed in almost all eukaryotes, including protozoa, flies, nematodes, insects, parasites, and mammals. This precise cellular mechanism of gene silencing has been developed into a technique that provides an efficient way to identify and determine the...
7.2K
In-vitro Mutagenesis
16.0K
To learn more about the function of a gene, researchers can observe what happens when the gene is inactivated or “knocked out,” by creating genetically engineered knockout animals. Knockout mice have been particularly useful as models for human diseases such as cancer, Parkinson’s disease, and diabetes.
16.0K
MicroRNAs
3.7K
MicroRNA (miRNA) are short, regulatory RNA transcribed from introns (non-coding regions of a gene) or intergenic regions (stretches of DNA present between genes). Several processing steps are required to form biologically active, mature miRNA. The initial transcript, called primary miRNA (pri-mRNA), base-pairs with itself, forming a stem-loop structure. Within the nucleus, an endonuclease enzyme, called Drosha, shortens the stem-loop structure into hairpin-shaped pre-miRNA. After the pre-miRNA...
3.7K


