基於ReaDDy2的粒線體裂變-融合動力學反應擴散模型

Abstract

粒線體形成動態的絲狀網絡，其結構由融合和裂變之間的持續相互作用所塑造。 了解這些結構轉變如何隨時間和在不同細胞條件下演變仍然是粒線體生物學領域的關鍵挑戰。 在本研究中，我們提出了一個計算框架，該框架透過將基於粒子的擴散與結構和空間反應機制相結合來模擬粒線體網絡重塑。 該模型基於節點連通性和空間鄰近性，編碼了受生物學啟發的融合和裂變規則，使分支、伸長和碎片化等拓撲事件能夠隨著時間的推移自然發生。 此模擬系統由實驗影像產生的骨架圖初始化，並透過雙層反應方案演化：結構反應基於局部圖規則重建內部拓撲結構，而空間反應在滿足鄰近性標準時合併各個組件。 一系列詳細的輸出包括:粒子軌跡、拓撲檔案、反應日誌和度分佈:支援可視化和定量分析。 單次運行模擬揭示了網路複雜性的動態變化，例如端點頻率的增加和平均聚合物長度的減少。 對100次重複實驗進行多重運行統計平均，證明了度機率的穩健收斂性，並允許與實驗數據直接比較。 透過時序性顯微鏡影像的定量驗證，在對照組條件下表現出高度一致性，但在其他藥物如 FCCP 和 Mdivi-1 等條件下，會有誤差產生。這些結果表明，該模型雖然能捕捉粒線體重塑，但也指出了需要納入其他可能潛在的生物學機制，例如局部降解或生化回饋，才能獲得完全的準確性。 總體而言，此模擬平台為探索粒線體動力學提供了一個分析的工具，在實驗假設檢定、藥物反應建模和細胞能量學的系統級研究中具有應用價值。

Mitochondria form dynamic, filamentous networks whose architecture is shaped by a continuous interplay between fusion and fission. Understanding how these structural transformations evolve over time and under different cellular conditions remains a key challenge in mitochondrial biology. In this study, we present a computational framework that simulates mitochondrial network remodeling by integrating particle-based diffusion with both structural and spatial reaction mechanisms. The model encodes biologically inspired rules for fusion and fission based on node connectivity and spatial proximity, enabling topological events such as branching, elongation, and fragmentation to emerge naturally over time. The simulation system is initialized from experimental image-derived skeleton graphs and evolves through a dual-layer reaction scheme: Structural reactions restructure internal topology based on local graph rules, while spatial reactions merge separate components when proximity criteria are met. A series of detailed outputs—including particle trajectories, topology files, reaction logs, and degree distributions, both visualization and quantitative analysis. Single-run simulations reveal dynamic transitions in network complexity, such as increases in endpoint frequency and reductions in average polymer length. Multi-run statistical averaging across 100 replicates demonstrates robust convergence of degree probabilities and allows for direct comparison with experimental data. Quantitative verification of timing microscopy images showed high consistency under the control group conditions, but errors occurred under other drugs such as FCCP and Mdivi-1. These results suggest that while capturing metaphysical remodeling, the model also points to the need to incorporate other possible potential biological mechanisms, such as local degradation or biochemical feedback, in order to achieve complete accuracy. Overall, this simulation platform provides an analytical tool for exploring metasoma dynamics, with application value in systematic research on experimental hypothesis assays, drug response modeling, and cell energy.