预测不同物种之间的染色质可访问性差异的挑战
Amy Stephen1,2,3, Arian Raje2,4,5, Heather H Sestili2
1Mathematical Sciences Department, Carnegie Mellon University, Pittsburgh, PA, USA.
bioRxiv : the preprint server for biology
|November 26, 2025
概括
机器学习模型可以预测增强器的活动,但与物种之间的定量差异作斗争. 对多种物种的训练可以提高概括性,但不能提高对染色质可访问性的跨物种预测准确性.
科学领域:
- 基因组学就是基因组学.
- 计算生物学 计算生物学
- 进化生物学 进化生物学
背景情况:
- 增强剂是推动表型多样性的关键转录性调节元素.
- 增强剂的快速序列进化,尽管有功能性保护,但使跨物种功能预测复杂化.
- 用于增强剂活动预测的机器学习模型尚未严格测试跨物种定量差异.
研究的目的:
- 评估机器学习模型能够预测不同物种的正统区域中增强器活动的定量差异的能力.
- 开发和应用一个框架来评估增强剂活动预测模型的跨物种性能.
- 调查多物种训练数据对模型概括和跨物种预测的影响.
主要方法:
- 卷积神经网络 (CNN) 接受了回归任务的训练,以预测五种哺乳动物物种的肝脏中的染色质可访问性 (增强剂活性的代理).
- 开发了一个新的框架来评估这些CNN模型的跨物种预测性能.
- 评估了模型的性能,以预测染色质可访问性差异的物种内和物种间预测.
主要成果:
- 在多种哺乳动物物种上训练CNN,改善了对训练和保持物种的模型概括.
- 模型在预测不同物种之间正统区域之间的染色质可访问性的定量差异方面始终表现不佳.
- 尽管整体模型概括的改进,对染色质可访问性差异的跨物种预测准确性仍然是一个重大挑战.
结论:
- 多种培训可以提高增强剂活动预测模型的概括性,但不能完全解决预测物种之间的定量差异的挑战.
- 使用当前的机器学习回归模型,预测物种之间增强剂活性和染色质可访问性的进化变化仍然很困难.
- 需要进一步开发计算框架,以准确地建模跨进化时间尺度的监管元素的功能分歧.
相关概念视频
Chromatin Immunoprecipitation- ChIP
12.1K
Chromatin immunoprecipitation, or ChIP, is an antibody-based technique used to identify sites on DNA that bind to transcription factors of interest or histone proteins. It also helps determine the type of histone modifications such as acetylation, phosphorylation, or methylation.
Types of ChIP
ChIP can be divided into two types - X-ChIP and N-ChIP. X-ChIP involves in vivo cross-linking of histones and regulatory proteins to DNA, fragmenting the DNA by sonication, and isolating the protein-DNA...
Types of ChIP
ChIP can be divided into two types - X-ChIP and N-ChIP. X-ChIP involves in vivo cross-linking of histones and regulatory proteins to DNA, fragmenting the DNA by sonication, and isolating the protein-DNA...
12.1K
Heterochromatin
17.8K
The extent of chromatin compaction can be studied by staining chromatin using specific DNA binding dyes. Under the microscope, the dense-compacted regions that take up more dye are called heterochromatin. Heterochromatin is further classified into two forms – constitutive heterochromatin and facultative heterochromatin.
Constitutive heterochromatin: It is a highly compact region of chromatin that is mostly concentrated in the centromere and telomere. Unlike euchromatin, the amino acid at...
Constitutive heterochromatin: It is a highly compact region of chromatin that is mostly concentrated in the centromere and telomere. Unlike euchromatin, the amino acid at...
17.8K
Euchromatin
8.8K
The extent of chromatin compaction can be studied by staining chromatin using specific DNA binding dyes. Under the microscope, the dense-compacted regions take up more dye, appearing darker, while the less-compact areas take up less dye and appear lighter. Based on the compaction level, chromatins are classified into two primary forms – euchromatin and heterochromatin.
Euchromatin is the less dense region of the chromatin and stains lighter. Euchromatin contains histone H3 extensively...
Euchromatin is the less dense region of the chromatin and stains lighter. Euchromatin contains histone H3 extensively...
8.8K
Spreading of Chromatin Modifications
9.3K
The histone proteins in the nucleosomes are post-translationally modified (PTM) to increase or decrease access to DNA. The commonly observed PTMs are methylation, acetylation, phosphorylation, and ubiquitination of lysine amino acids in the histone H3 tail region. These histone modifications have specific meaning for the cell. Hence, they are called "histone code". The protein complex involved in histone modification is termed as "reader-writer" complex.
Writers
The writer...
Writers
The writer...
9.3K
Chromatin Position Affects Gene Expression
24.6K
Chromatin is the massive complex of DNA and proteins packaged inside the nucleus. The complexity of chromatin folding and how it is packaged inside the nucleus greatly influences access to genetic information. Generally, the nucleus' periphery is considered transcriptionally repressive, while the cell's interior is considered a transcriptionally active area.
Topologically Associated Domains (TADs)
The 3-dimensional positioning of chromatin in the nucleus influences the...
Topologically Associated Domains (TADs)
The 3-dimensional positioning of chromatin in the nucleus influences the...
24.6K
Position-effect Variegation
7.0K
In 1928, a German botanist Emil Heitz observed the moss nuclei with a DNA binding dye. He observed that while some chromatin regions decondense and spread out in the interphase nucleus, others do not. He termed them euchromatin and heterochromatin, respectively. He proposed that the heterochromatin regions reflect a functionally inactive state of the genome. It was later confirmed that heterochromatin is transcriptionally repressed, and euchromatin is transcriptionally active chromatin.
7.0K


