深度学习揭示了元基因组中的环境适应性
Johanna C Winder1, Simon Poulton2,3, Taoyang Wu4
1School of Environmental Sciences, University of East Anglia, Norwich Research Park, Norwich, NR4 7TJ, UK. j.winder@uea.ac.uk.
BMC biology
|August 10, 2025
概括
深度学习模型根据源环境对蛋白质进行分类,从而达到高精度. 结合方法可以提高生物洞察力和可解释性,用于多种序列分类.
科学领域:
- 生物信息学是一种生物信息学.
- 计算生物学 计算生物学
- 基因组学就是基因组学.
背景情况:
- 深度学习 (DL) 提供了强大的生物数据分析,但在成本,复杂性和可解释性方面面临挑战.
- 人工神经网络 (ANN) 是复杂的,限制了生物洞察力提取.
- 未知功能的域3494 (DUF3494) 在微生物中广泛存在,这表明其生态作用多样化.
研究的目的:
- 通过转移学习方法,根据它们的源环境对DUF3494蛋白进行分类.
- 在序列分类中探索预测准确性和生物解释性之间的平衡.
- 开发一个框架,将DL与其他方法结合起来,以增强生态洞察力.
主要方法:
- 应用转移学习与ESM-2蛋白质结构预测和较小的ANN.
- 从公共元基因组中分析了50,669个DUF3494序列.
- 将ANN性能与基因算法 (GA) 进行比较,以获得可解释性.
主要成果:
- 通过源环境 (例如,极地海洋,冰川冰,岩石) 成功分类了DUF3494序列.
- 获得了75.9%至97.8%之间的分类准确度,最佳ANN.
- 确定特定环境特征和关键氨基酸残留物驱动分类.
- GA提供了透明的规则,补充了DL的预测能力.
结论:
- 深度学习对于分类各种生物序列是有效的.
- 将DL与GA等方法结合起来,可以提高模型的解释性和生态理解.
- 该研究为生物信息学中整合预测和可解释模型提供了一个框架.
相关概念视频
01:30Environmental Applications of Microorganisms
242
Microorganisms play a pivotal role in maintaining ecosystem balance by recycling essential elements such as carbon, nitrogen, and phosphorus, as well as supporting processes like bioremediation, wastewater treatment, and biofuel production.Microbes in Elemental CyclesIn the carbon cycle, microorganisms decompose organic matter, releasing carbon dioxide via aerobic respiration. This carbon dioxide is subsequently used by photosynthetic organisms to synthesize organic compounds, closing the...
242
01:27Diversity of Archaea III
73
Crenarchaeota, a prominent phylum of Archaea, is remarkable for its ability to thrive in extreme environments characterized by high temperatures and acidity. These microorganisms inhabit sulfuric hot springs, volcanic systems, and submarine hydrothermal vents, where temperatures often exceed 100°C. The unique adaptations of Crenarchaeota not only allow survival under such extreme conditions but also provide insights into the mechanisms of life in primordial Earth-like...
73
01:28Microbial Nutrition
292
Organisms exhibit remarkable metabolic diversity, categorized based on how they acquire energy and carbon. These strategies enable survival in various ecological niches and are essential for maintaining energy flow and nutrient cycling within ecosystems.Energy and Carbon SourcesOrganisms are classified as phototrophs or chemotrophs based on energy acquisition. Phototrophs use light as their energy source, while chemotrophs rely on oxidizing chemical compounds. Further differentiation arises...
292
01:29Overview of Archaea
137
Archaea, named after the Archaean eon, represent a unique domain of life, distinct from bacteria and eukaryotes, with remarkable traits. Their cellular and molecular features, ecological adaptability, and industrial relevance highlight their importance in understanding life processes and leveraging biotechnology.Cellular and Molecular CharacteristicsA defining feature of archaea is their unique membrane composition. Archaeal membranes contain ether-linked isoprenoid lipids, which confer...
137
01:24Diversity of Archaea II
89
Archaea, one of the three domains of life, exhibit remarkable diversity and adaptability, thriving in both extreme and moderate environments. Historically, most identified archaea have been classified into two major phyla: Euryarchaeota and Crenarchaeota. However, recent molecular studies have expanded this classification to include three additional phyla: Thaumarchaeota, Nanoarchaeota, and Korarchaeota, each exhibiting unique characteristics and ecological roles.Thaumarchaeota: Mesophiles...
89
01:29Modern Molecular Taxonomy
141
Advancements in molecular biology have revolutionized the identification and characterization of bacteria, with multiple methods leveraging DNA sequencing for enhanced precision. As sequencing technologies improve and costs decline, these approaches are increasingly used in clinical, environmental, and evolutionary studies.Multilocus Sequence Typing (MLST) examines several housekeeping genes, essential chromosomal genes encoding cellular functions, to distinguish strains. Approximately...
141