Scholarlink

<p>Single-cell analyses have transitioned from descriptive atlasing towards inferring causal effects and mechanistic relationships that capture cellular logic. Technological advances and the growing scale of observational and interventional datasets have fuelled the development of machine learning methods aimed at identifying such dependencies and extrapolating perturbation effects. Here, we review and connect these approaches according to their modelling concepts (including representation learning, causal inference, mechanistic discovery, disentanglement and population tracing), underlying assumptions and downstream tasks. We propose a unifying ontology to&#xa0;guide practitioners in selecting the most suitable methods for a given biological question, with detailed technical descriptions provided in an <a href="https://interp-extrap-perturb.readthedocs.io/">online resource</a>. Finally, we identify promising computational directions and underexplored data properties that could pave the way for future developments.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Interpretation, extrapolation and perturbation of single cells

  • 期刊论文

  • 发表日期：

    2026年1月2日

查看全文
  • Daniel Dimitrov,
  • Stefan Schrod,
  • Martin Rohbeck,
  • Oliver Stegle

摘要

Single-cell analyses have transitioned from descriptive atlasing towards inferring causal effects and mechanistic relationships that capture cellular logic. Technological advances and the growing scale of observational and interventional datasets have fuelled the development of machine learning methods aimed at identifying such dependencies and extrapolating perturbation effects. Here, we review and connect these approaches according to their modelling concepts (including representation learning, causal inference, mechanistic discovery, disentanglement and population tracing), underlying assumptions and downstream tasks. We propose a unifying ontology to guide practitioners in selecting the most suitable methods for a given biological question, with detailed technical descriptions provided in an online resource. Finally, we identify promising computational directions and underexplored data properties that could pave the way for future developments.