scLong:一个十亿参数基础模型,用于捕捉单细胞转录组学中的远程基因上下文
Ding Bai1, Shentong Mo1, Ruiyi Zhang2
1Mohamed bin Zayed University of Artificial Intelligence, Masdar City, Abu Dhabi, UAE.
Nature communications
|February 4, 2026
概括
新的基础模型scLong分析了单细胞RNA测序数据中的所有基因,包括低表达的基因. 它整合了基因知识,以改善基因调节和药物反应的预测.
科学领域:
- 基因组学就是基因组学.
- 计算生物学 计算生物学
- 生物信息学是一种生物信息学.
背景情况:
- 单细胞RNA测序 (scRNA-seq) 提供高分辨率的基因表达数据,揭示细胞异质性.
- 对于scRNA-seq数据的现有基础模型经常忽视低表达基因和外部生物知识.
- 分析复杂的基因相互作用对于理解细胞功能和疾病机制至关重要.
研究的目的:
- 介绍scLong,一个用于scRNA-seq数据分析的大规模基础模型.
- 为了实现全面的基因表达建模,包括低表达和未表达的基因.
- 整合外部基因知识,以增强生物背景和预测能力.
主要方法:
- 在4800万个细胞上预训练一个十亿参数基础模型 (scLong).
- 在所有28,000个人类基因中实施自我注意力,以捕捉长期依赖.
- 整合基因本体学知识使用图形卷积网络.
主要成果:
- scLong在各种任务中表现出与最先进的模型相比的卓越性能.
- 该模型有效地捕捉了涉及低表达和未表达基因的依赖关系.
- scLong在预测转录反应,癌症药物疗效和基因调控网络方面表现出强大的能力.
结论:
- scLong代表了scRNA-seq数据分析的基础模型的重大进步.
- 该模型处理所有基因和整合外部知识的能力增强了生物洞察力.
- scLong在理解基因调节,疾病机制和治疗开发方面具有广泛的应用.
相关概念视频
Range
14.3K
The range is one of the measures of variation. It can be defined as the difference between a dataset's highest and lowest values. For example, in the study of seven 16-ounce soda cans, the filled volume of soda was measured, thus producing the following amount (in ounces) of soda:
15.9; 16.1; 15.2; 14.8; 15.8; 15.9; 16.0; 15.5
Measurements of the amount of soda in a 16-ounce can vary since different subjects record these measurements or since the exact amount - 16 ounces of liquid, was not...
15.9; 16.1; 15.2; 14.8; 15.8; 15.9; 16.0; 15.5
Measurements of the amount of soda in a 16-ounce can vary since different subjects record these measurements or since the exact amount - 16 ounces of liquid, was not...
14.3K
Self Within Cultural Contexts
239
Cultural frameworks for understanding the self are often categorized into two broad orientations: individualism and collectivism. These paradigms influence how people define themselves, relate to others, and interpret their social worlds. Each orientation offers distinct perspectives on autonomy, responsibility, and the role of the individual within a community.Individualistic CulturesIn individualistic cultures like North America and Western Europe, identity is understood as autonomous and...
239
Social Foundations of Self II: The Generalized Other
266
According to George Herbert Mead, as children progress beyond the game stage, they develop a more comprehensive understanding of societal rules and norms. This cognitive and social development enables them to internalize the expectations of the broader community, refining their ability to regulate behavior.Consistent participation in organized activities is crucial in helping children recognize that their actions are not isolated but contribute to a more significant, interconnected group...
266
Physiological Foundation of Stress
681
Stress triggers a coordinated physiological response involving the sympathetic nervous system (SNS) and the hypothalamic-pituitary-adrenal (HPA) axis. This dual activation ensures that the body is prepared for both immediate and prolonged stress management. The process begins with the perception of a stressor. This initial phase activates the SNS, leading to the rapid release of adrenaline (epinephrine) from the adrenal glands.
Role of the Sympathetic Nervous System
Adrenaline triggers the...
Role of the Sympathetic Nervous System
Adrenaline triggers the...
681
Theoretical Foundations of Nursing Practice
17.5K
Theories play an essential role in organizing patient care. Theories refer to a proposed or followed belief, policy, or procedure that is the basis for action. Nursing theories are knowledge-based concepts that guide nurses' actions, influence nursing education and practice, and allow nurses to care for their patients.
Theories provide a perspective to assess patients' conditions and organize data and methods. They also assist in analyzing and interpreting information. They represent a...
Theories provide a perspective to assess patients' conditions and organize data and methods. They also assist in analyzing and interpreting information. They represent a...
17.5K
Model Approaches for Pharmacokinetic Data: Distributed Parameter Models
249
Pharmacokinetic models are mathematical constructs that represent and predict the time course of drug concentrations in the body, providing meaningful pharmacokinetic parameters. These models are categorized into compartment, physiological, and distributed parameter models.
The distributed parameter models are specifically designed to account for variations and differences in some drug classes. This model is particularly useful for assessing regional concentrations of anticancer or...
The distributed parameter models are specifically designed to account for variations and differences in some drug classes. This model is particularly useful for assessing regional concentrations of anticancer or...
249


