The idea of the inheritance of biological processes, such as the
developmental process or the life cycle of an organism, has been discussed in
the biology literature, but formal mathematical descriptions and plausible data
analysis frameworks are lacking. We introduce an extension of the nested
Dirichlet Process (nDP) to a multiscale model to aid in understanding the
mechanisms by which biological processes are inherited, remain stable, and are
modified across generations. To address these issues, we introduce Nested
Inheritance Dynamics Algorithm (NIDA). At its primary level, NIDA encompasses
all processes unfolding within an individual organism's lifespan. The secondary
level delineates the dynamics through which these processes evolve or remain
stable over time. This framework allows for the specification of a physical
system model at either scale, thus promoting seamless integration with
established models of development and heredity.

通过 Nested Inheritance Dynamics Algorithm (NIDA) 模型，研究了生物过程的继承、稳定性和变异性机制。

嵌套的继承动力学

Nested Inheritance Dynamics

We present the Thought Graph as a novel framework to support complex
reasoning and use gene set analysis as an example to uncover semantic
relationships between biological processes. Our framework stands out for its
ability to provide a deeper understanding of gene sets, significantly
surpassing GSEA by 40.28% and LLM baselines by 5.38% based on cosine similarity
to human annotations. Our analysis further provides insights into future
directions of biological processes naming, and implications for bioinformatics
and precision medicine.

我们提出了 Thought Graph 作为一种新颖的框架，用于支持复杂推理，并以基因集分析为例，揭示生物过程之间的语义关系。我们的框架在提供对基因集的更深入理解方面表现出色，与基于余弦相似性到人类注释的 GSEA 相比，超过了 40.28％，并且超过了 LLM 基线 5.38％。我们的分析进一步为生物过程命名的未来方向和生物信息学以及精准医学的意义提供了见解。

思维图：生物推理的思维过程生成

Thought Graph: Generating Thought Process for Biological Reasoning

Fitting generative models to sequential data typically involves two recursive
computations through time, one forward and one backward. The latter could be a
computation of the loss gradient (as in backpropagation through time), or an
inference algorithm (as in the RTS/Kalman smoother). The backward pass in
particular is computationally expensive (since it is inherently serial and
cannot exploit GPUs), and difficult to map onto biological processes.
Work-arounds have been proposed; here we explore a very different one:
requiring the generative model to learn the joint distribution over current and
previous states, rather than merely the transition probabilities. We show on
toy datasets that different architectures employing this principle can learn
aspects of the data typically requiring the backward pass.

生成模型对顺序数据的拟合通常涉及两个循环计算，一个向前，一个向后。本文探讨了一种不同的解决方法：要求生成模型学习当前状态和前一个状态之间的联合分布，而不仅仅是转移概率。在玩具数据集上展示了不同架构使用这一原则能够学习通常需要向后传递的数据方面的内容。