通过AlphaGenome推进监管变体效应预测
Žiga Avsec1, Natasha Latysheva2, Jun Cheng2
1Google DeepMind, London, UK. avsec@google.com.
Nature
|January 28, 2026
概括
阿尔法基因组 (AlphaGenome) 是一种新型的深度学习模型,可以从1兆基基因基因组序列预测成千上万个功能性基因组轨迹,以单基因对分辨率. 这种统一的方法克服了现有方法的局限性,增强了跨多种基因组模式的变异效应预测.
科学领域:
- 基因组学和生物信息学
- 计算生物学 计算生物学
- 深度学习应用程序
背景情况:
- 从DNA序列预测功能性基因组测量对于理解遗传调节代码至关重要.
- 当前的深度学习模型面临的局限性是由于输入序列长度和预测分辨率之间的权衡,限制了它们的范围和性能.
- 多种功能性基因组模式,包括基因表达,染色质可访问性和转录因子结合,需要综合分析.
研究的目的:
- 开发一个统一的深度学习模型,AlphaGenome,能够从高分辨率的长DNA序列中预测众多功能基因组轨迹.
- 克服现有方法的局限性,将多种基因组模式集成到一个框架内.
- 在多个功能性基因组层中提供准确的变异效应预测.
主要方法:
- 开发AlphaGenome,这是一个深度学习模型,可以接受1兆基 (Mb) 的DNA序列作为输入.
- 在人类和小鼠基因组上训练AlphaGenome,以预测数千个功能性基因组轨迹,包括基因表达,染色质可访问性,转录因子结合等,在单基因对分辨率下.
- 在26个不同的评估中对AlphaGenome在变异效应预测方面的表现与现有的最先进模型进行了评估.
主要成果:
- 阿尔法基因组成功地预测了成千上万的功能性基因组轨迹,通过单基因对分辨率跨越多种模式.
- 该模型在26个变量效应预测评估中的25个中与最强大的外部模型的性能相匹配或超过.
- 阿尔法基因组准确地回顾了临床相关变异的机制,通过对TAL1瘤基因附近变异的分析来证明这一点.
结论:
- 阿尔法基因组代表了基因组学深度学习的重大进步,提供了一种统一的方法来从长DNA序列中预测功能基因组信息.
- 该模型的高分辨率和广泛的模式范围使得更准确和更全面的变体效应预测成为可能.
- 提供了工具,以促进AlphaGenome用于基因组跟踪和变异效应预测的使用,促进更广泛的研究应用.
相关概念视频
02:02Cis-regulatory Sequences
11.8K
Cis-regulatory sequences are short fragments of non-coding DNA that are present on the same chromosomes as the genes that they regulate. These fragments serve as binding sites for transcriptional regulators, proteins that are responsible for controlling gene transcription and differential gene expression across cell types in eukaryotes. Cis-regulatory sequences can be close to the gene of interest or thousands of bases away in the DNA sequence; however, those sequences that are further away are...
11.8K
02:02Cis-regulatory Sequences
4.1K
4.1K
02:30Histone Variants at the Centromere
5.0K
Histone variants are the histone proteins with structural and sequence variations. These variants may be regarded as “mutant” forms that replace their canonical histone counterparts in the nucleosomes. Specific post-translational modifications on the histone variants enable further chromatin complexity and regulate tissue-specific gene expression. The most common histone variants are from histone H2A, H2B, and linker histone H1 families. However, several variants of histone H3...
5.0K
02:27Predicting Molecular Geometry
45.7K
VSEPR Theory for Determination of Electron Pair Geometries
45.7K
01:28Global Regulatory Systems
666
Global regulatory systems in bacteria enable rapid and coordinated responses to environmental changes by integrating sensory inputs with gene expression, ensuring efficient adaptation to fluctuating conditions. Key global regulatory mechanisms include regulons, two-component systems, sigma factors, and secondary messengers.Regulons and Global RegulatorsA regulon is a collection of genes and operons controlled by a common global regulator. These regulators enable bacteria to prioritize resource...
666
01:03Prediction Intervals
3.4K
The interval estimate of any variable is known as the prediction interval. It helps decide if a point estimate is dependable.
However, the point estimate is most likely not the exact value of the population parameter, but close to it. After calculating point estimates, we construct interval estimates, called confidence intervals or prediction intervals. This prediction interval comprises a range of values unlike the point estimate and is a better predictor of the observed sample value, y.
However, the point estimate is most likely not the exact value of the population parameter, but close to it. After calculating point estimates, we construct interval estimates, called confidence intervals or prediction intervals. This prediction interval comprises a range of values unlike the point estimate and is a better predictor of the observed sample value, y.
3.4K