交叉模式嵌入集成器用于使用多头注意力机制进行疾病基因/蛋白质关联预测
Munyoung Chang1, Junyong Ahn2,3, Bong Gyun Kang3
1Education and Research Program for Future ICT Pioneers, Department of Electrical and Computer Engineering, Seoul National University, Seoul, South Korea.
一个新的计算模型,交叉模式嵌入集成器 (CMEI),准确地预测疾病基因/蛋白质关联. 这种工具有助于发现疾病机制和潜在的治疗点.
科学领域:
- 生物医学信息学是生物医学信息学.
- 计算生物学是一种计算生物学.
- 机器学习在医疗保健中的应用
背景情况:
- 生物医学知识图对于推断新知识和识别疾病-基因/蛋白质关系至关重要.
- 对生物医学实体的准确表示对于预测这些关联至关重要.
- 发现新的疾病-基因/蛋白质联系可以揭示疾病机制和治疗点.
研究的目的:
- 开发一种用于预测疾病-基因/蛋白质关联的计算模型.
- 整合多样化的数据模式,以提高预测准确度.
- 利用生物医学知识图表来增强生物洞察力.
主要方法:
- 使用了精准医学知识图.
- 使用大语言模型 (LLM) 和知识图嵌入 (KGE) 算法生成生物医学实体嵌入.
- 开发了跨模式嵌入集成器 (CMEI) 模型,通过多头关注集成嵌入.
主要成果:
- 该CMEI模型实现了高预测性能,接收器操作特征曲线下的面积为0.9662 (±0.0002).
- 证明了整合LLM和KGE嵌入用于协会预测的有效性.
- 验证了模型在预测疾病-基因/蛋白质关系方面的能力.
结论:
- 开发的计算模型CMEI有效地预测疾病基因/蛋白质关联.
- CMEI显示了加速识别疾病发展机制的潜力.
- 这种方法可能有助于发现疾病的新治疗点.
更多相关视频
09:47Author Spotlight: Advancing Alzheimer's Research – Exploring Early Detection and Multi-Omics Approaches
Published on: December 15, 2023
08:51Author Spotlight: Integrated Multi-Omics Analysis for Unveiling Multicellular Immune Signatures in Clinical Heart Attack Cohorts
Published on: September 20, 2024
相关概念视频
Genome-wide Association Studies-GWAS
GWAS does not require the identification of the target gene involved in...
Genomics
Protein Networks
These interactions can be represented through maps depicting protein-protein interaction networks, represented as nodes and edges. Nodes are circles that are representative of a protein,...
Insertion of Multi-pass Transmembrane Proteins in the RER
The multipass transmembrane proteins are the type IV integral membrane proteins with multiple topogenic sequences determining their spatial arrangement in the ER membrane. Nearly all multipass proteins lack a cleavable signal sequence and use...
Pleiotropy
Multi-pass Transmembrane Proteins and β-barrels
α-Helix containing multi-pass transmembrane proteins
Multi-pass transmembrane proteins such as...
