As illustrated in the figure below, ViennaRNA provides the secondary structure results along with a diagram, where '&' denotes hybridization between the two strands. The green represents the designed ogRNA, while the red indicates the mismatched mRNA.

如下图所示，ViennaRNA给出二级结构结果并附图，其中&代表两条链杂交。其中绿色代表提前设计的ogRNA，红色代表与ogRNA错配的mRNA。

(((((((((((((.(((((((((.(....(((((((((((((.((((((((((((((..((((((..((((((((((.(((((((((((((((((((((((((((((((((((.((((((((((((((((((((((((((((((&............................(((((((.((....))))))))).))))))))))))))))))))))(((((...)))))..)))))).)).))))))))))))))))))))))))))))))))))).)))))))))).(((....)))......))))))...)))))))))))))).)))))))))))))(((((((.((....)))))))))).))))))))).))))))))))))).. (-294.50)

(((((((((((((.(((((((((.(....(((((((((((((.((((((((((((((..((((((..((((((((((.(((((((((((((((((((((((((((((((((((.((((((((((((((((((((((((((((((&............................(((((((.((....))))))))).))))))))))))))))))))))(((((...)))))..)))))).)).))))))))))))))))))))))))))))))))))).)))))))))).(((....)))......))))))...)))))))))))))).)))))))))))))(((((((.((....)))))))))).))))))))).))))))))))))).. (-294.50)

Note: no Enter in the sequence

注：序列中间无空格

When examining the secondary structure, we can gradually zoom in to clearly identify the AC base mismatches on the RNA, indicating a relatively successful prediction of the secondary structure. In some predicted results, multiple AC mismatches are present on the double-stranded RNA, suggesting that these mismatches may enhance the recognition efficiency of ADAR enzymes. Additionally, most results also contain other base pair mismatches, further highlighting the complexity and accuracy of the structural predictions.

在观察二级结构时，我们可以逐步放大，从中清晰地识别出 RNA 上的 AC 碱基错配，这表明二级结构预测的结果相对成功。在某些预测结果中，双链 RNA 上存在多个 AC 错配，推测这些错配可能有助于提高 ADAR 酶的识别效率。此外，多数结果还同时包含其他碱基对的错配泡，这进一步表明了结构预测的复杂性和准确性。

The following figure displays the tertiary structure of the double-stranded RNA predicted by trRosettaRNA, compared with the previous secondary structure. The two predictions are closely aligned, indicating that the generation of the tertiary structure maintains a reasonable consistency based on the secondary structure predictions.

下图展示了 trRosettaRNA 预测得到的双链 RNA 三级结构图，并与之前的二级结构进行了对比。两者的预测结果相近，表明在二级结构预测的基础上，三级结构的生成也能保持一定的合理性。

By analyzing these two prediction results, we find that the tertiary structure predicted by trRosettaRNA is consistent with the secondary structure predicted by ViennaRNA in multiple key regions, especially concerning important binding sites and the spatial arrangement of base pairs. This consistency indicates that the secondary structure predictions provide a solid foundation for the generation of tertiary structures, leading to a more accurate spatial configuration of RNA molecules.

通过分析这两种预测结果，我们发现 trRosettaRNA 的三级结构与 ViennaRNA 预测的二级结构在多个关键区域保持一致，特别是在重要的结合位点和碱基对的空间排布上。这种一致性表明，二级结构预测为三级结构的生成提供了良好的基础，使得 RNA 分子的空间构型更为准确。

Despite the consistency observed in many double-stranded RNA samples between trRosettaRNA and ViennaRNA predictions, significant discrepancies also exist in the predictions of certain RNA molecules. Some samples exhibit pronounced differences in structural features between the two methods, which may stem from the inherent diversity of RNA molecules and their stability under different conditions. Specifically, certain predictions of the secondary structure may indicate a specific spatial conformation, while the tertiary structure predicted by trRosettaRNA may reveal an alternative potential conformation, including additional mismatches or atypical base pairings in certain regions.

尽管 trRosettaRNA 和 ViennaRNA 的预测结果在许多双链 RNA 样本中显示出一致性，但在某些 RNA 分子的预测中也存在显著的不一致。有些样本的结构特征在两种方法中表现出明显的差异，这种不一致可能源于 RNA 分子自身的多样性及其在不同条件下的稳定性。具体来说，某些双链 RNA 的二级结构预测可能指向一种特定的空间构型，而 trRosettaRNA 的三级结构预测则可能揭示出另一种潜在的构型，甚至在某些区域出现了额外的错配或非典型的碱基配对。

We conducted a comprehensive molecular docking analysis of all 40 hybrid double-stranded RNA pairs with ADAR1. Utilizing PyMOL software, we obtained 80 detailed molecular docking images that fully demonstrate the binding characteristics of each hybrid double-stranded RNA with ADAR1. These docking results reveal not only the spatial conformation relationship between RNA and ADAR1 but also provide interaction information regarding key amino acid residues and nucleotides.

我们对所有40组杂交双链RNA与ADAR1进行了全面的分子对接分析。通过使用Pymol软件，我们获得了80张详细的分子对接图，充分展现了每组杂交双链RNA与ADAR1的结合特征。这些对接结果不仅揭示了RNA与ADAR1之间的空间构象关系，还提供了关键氨基酸残基、关键核苷酸的相互作用信息。

Firstly, in the docking analysis, we focused on amino acid residues that are functionally relevant to ADAR1. The selective participation of these amino acids in binding with RNA is a crucial factor influencing RNA editing efficiency. Through statistical analysis, we identified the main ADAR1 amino acid residues that interact with hybrid RNA, revealing their interaction preferences in the RNA editing process, as shown in the figure below.

首先，在对接分析中，我们重点关注了与 ADAR1 功能相关的氨基酸残基。这些氨基酸的选择性参与与 RNA 的结合，是影响 RNA 编辑效率的重要因素。通过统计和分析，我们识别出与杂交 RNA 结合的主要 ADAR1 氨基酸残基，揭示了其在 RNA 编辑过程中的相互作用偏好性，如下图所示。

ADAR1 exhibits significant selectivity and preference when binding to different double-stranded RNAs, particularly with certain key amino acid residues playing a central role in RNA interactions. In analyzing the docking results of 40 hybrid double-stranded RNA pairs with ADAR1, we observed that several amino acid residues frequently appeared in multiple docking models, indicating their high conservation and functionality in RNA recognition and editing processes. For instance, the amino acids R790, R714, and R736 displayed frequent RNA-binding capabilities across most docking models, suggesting they may be core residues involved in RNA recognition and binding. These residues may stabilize the RNA-ADAR1 complex through charge interactions with RNA bases or backbones, thereby promoting RNA editing.

ADAR1 在与不同双链 RNA 的结合过程中，表现出了显著的选择性和偏好性，特别是某些关键氨基酸残基在与 RNA 相互作用时起到了核心作用。我们在分析 40 组杂交双链 RNA 与 ADAR1 的对接结果时，观察到多个氨基酸残基频繁出现在多个对接模型中，显示出它们在 RNA 识别和编辑过程中具有高度的保守性和功能性。例如，R790、R714 和 R736 这三个氨基酸在大多数对接模型中表现出频繁的 RNA 结合能力，表明它们可能是参与 RNA 识别和结合的核心残基。这些残基可能通过与 RNA 碱基或主链的电荷相互作用，稳定了 RNA-ADAR1 复合体，进而促进了 RNA 编辑的发生。

Next, we examined the binding characteristics of each hybrid double-stranded RNA with ADAR1, noting that several results indicated that ADAR1 can form tight interactions with bases located on the AC mismatch bubbles, as depicted in the following figure.

其次，我们关注了每组杂交双链 RNA 与 ADAR1 的结合特征，其中有几组结果显示，ADAR1可以与AC错配泡上碱基比较贴合的结合，如下图所示。

However, while certain results demonstrate a strong binding of ADAR1 to mismatch bubbles, the majority of double-stranded RNA binding sites are somewhat distanced from these mismatch bubbles. In most models, ADAR1 shows a tendency to bind at the junctions between the branched structures and the main chain of the RNA. This observation suggests that RNA molecules may exhibit considerable conformational flexibility in three-dimensional space, leading ADAR1 to more readily find binding sites in these relatively open structural regions. We hypothesize that this is due to the reduced spatial hindrance at the junctions, allowing for a more stable binding environment for ADAR1, thus enhancing editing efficiency. As illustrated in the figure below, ADAR1 frequently localizes to these junction regions, likely due to their local structural openness.

然而，尽管某些结果显示 ADAR1 能与错配泡处形成紧密结合，大多数双链 RNA 的结合位点却与这些错配泡存在一定的距离。在大多数模型中，ADAR1 更倾向于结合 RNA 的分支结构与主干结构交接处。这一观察表明，RNA 分子在三维空间中可能存在较大的构象自由度，导致 ADAR1 在这些相对开放的结构区域更易找到结合位点。我们推测，这是由于交接处的空间位阻较小，能够更好地容纳 ADAR1 的结合。RNA 的分支结构通常伴随较大的自由能变化，而这些交接处可能为 ADAR1 提供了一个较为稳定的结合环境，从而提升其编辑效率。如下图所示，ADAR1 经常定位于这些交接区域，可能是由于局部的空间结构开放性所导致。

Additionally, we conducted a statistical analysis of the energy changes observed in the molecular docking. By calculating the binding free energy between ADAR1 and various hybrid double-stranded RNAs, we found that the Gibbs free energy (ΔG) for most docking models hovered around -200 kcal/mol. This value indicates a relatively stable binding between ADAR1 and RNA molecules. A lower binding free energy implies a more stable complex, reflecting the strong binding affinity of ADAR1 for RNA. Furthermore, we recorded the LGscore values, which represent the scoring of the docking models, assessing the geometric structure's alignment with experimental data. Most models achieved an LGscore of 2.083, suggesting an ideal binding pattern between ADAR1 and the hybrid RNA. These binding free energies and scoring values provide robust evidence for further experimental validation, as well as a quantifiable basis for the binding preferences of ADAR1 in recognizing RNA.

此外，我们对分子对接中的能量变化进行了统计分析。通过计算 ADAR1 与不同杂交双链 RNA 之间的结合自由能，我们发现大部分对接模型的 Gibbs 自由能（ΔG）在 -200 kcal/mol 左右。这一数值表明，ADAR1 与 RNA 分子之间形成了相对稳定的结合。结合自由能越低，意味着复合体越稳定，反映了 ADAR1 与 RNA 的强结合力。除此之外，我们还记录了 LGscore 值，该值代表对接模型的评分，衡量了对接的几何结构与实验数据的吻合程度。大多数模型的 LGscore 达到了 2.083，表明 ADAR1 与杂交 RNA 之间的结合模式较为理想。这些结合自由能和评分值为进一步实验验证提供了有力的依据，同时也为 ADAR1 识别 RNA 的偏好性提供了量化依据。

The table below presents the results of all Gibbs free energy calculations, with units in kcal/mol.

下表将展示所有的吉布斯自由能结果，单位：kcal/mol.

The code(in python) for model fitting and the corresponding fitting results are presented below.

模型拟合的代码（语言：python）和拟合效果图如下所示。

The suboptimal fitting results may stem from multiple biological factors that are crucial in the design and application of RNA sensors.

我们分析拟合结果不理想的原因可能源于多个生物学因素，这些因素在RNA传感器的设计与应用中至关重要。

First, the functionality of RNA sensors involves complex biological processes, including the secondary structure of RNA, protein interactions, and enzyme catalysis. The dynamic characteristics of these processes are often nonlinear or complex, making it difficult for simple kinetic models to accurately capture the true behavior of biological systems. For instance, the interaction force fields, binding affinities, and corresponding kinetic properties between RNA molecules and their binding proteins may exhibit significant variability under different physiological conditions. If these variations are not adequately represented in the model, they may lead to discrepancies in predicted outcomes. Additionally, the selected model parameters may not fully represent the actual biological system. For example, kinetic parameters such as Vmax and Km might be estimated based on insufficient or inaccurate experimental data, resulting in substantial differences between model predictions and actual experimental results.

首先，RNA传感器的功能涉及复杂的生物过程，包括RNA的二级结构、蛋白质相互作用及酶催化等。这些过程的动态特征通常并非线性或简单，因此简单的动力学模型难以准确捕捉生物系统的真实行为。例如，RNA分子与其结合蛋白之间的相互作用力场、结合亲和力以及相应的动力学特性在不同的生理条件下可能表现出明显的变化，这些变化在模型中如果未得到充分体现，就可能导致预测结果的偏差。其次，所选模型参数可能未能充分代表实际生物系统。例如，动力学参数如Vmax和Km的估计可能基于不充分或不准确的实验数据，导致模型预测与实际实验结果之间的显著差异。

Moreover, the secondary and tertiary structures of RNA play a critical role in its functionality. There are significant differences in the stability and reactivity of various structures. If the model fails to consider the diversity of RNA structures adequately, or if overly simplistic structural descriptions are used in the input, the fitting performance may be further compromised. RNA structure prediction inherently carries a high degree of uncertainty, leading to low confidence in predicted results, which may misrepresent actual biological interactions. Furthermore, in docking experiments, we observed that many RNA-protein bindings did not occur at the expected sites, adding to the complexity of model predictions. Therefore, these challenges of structure prediction and docking must be emphasized in the design of RNA sensors, as they may significantly influence functionality and interactions.

此外，RNA的二级和三级结构对其功能起着至关重要的作用。不同结构的稳定性和反应性存在显著差异。如果模型未能充分考虑RNA结构的多样性，或者在输入中使用了过于简单的结构描述，可能会进一步削弱拟合效果。RNA的结构预测本身具有较大的不确定性，预测结果的可信度不高，这可能导致实际生物相互作用的误判。此外，在对接实验中，我们观察到许多RNA与蛋白质的结合并未发生在预期的位点上，这进一步增加了模型预测的复杂性。因此，在进行RNA传感器设计时，必须重视这些结构预测和对接的挑战，并考虑它们对功能及相互作用的潜在影响。

The limitations of model assumptions are also a critical factor. In establishing differential equation models, certain reactions or processes may be simplified, failing to accurately reflect true reaction rates. For instance, the influence of some low-abundance molecules may be overlooked, or certain reactions may be assumed to follow quasi-steady-state approximations. These simplifications can significantly impact the final fitting performance. Thus, selecting an appropriate model form and considering more complex interactions and feedback mechanisms when necessary is crucial for enhancing model accuracy.

同时，模型假设的局限性也是一个重要因素。在建立微分方程模型时，某些反应或过程可能被简化假设，未能准确反映真实反应速率。例如，可能忽略某些低丰度分子的影响，或者假设某些反应为准稳态近似，这些简化可能会显著影响最终的拟合效果。因此，选择合适的模型形式，并在需要时考虑更复杂的相互作用和反馈机制，是提高模型准确性的重要步骤。

Finally, external environmental factors such as changes in temperature, pH, and ion concentration may also significantly influence the activity of RNA sensors. RNA stability and functionality are highly sensitive to these factors; if the model does not account for these variations, it may fail to effectively describe actual conditions. Additionally, the adaptability of RNA sensors in different cell types or physiological conditions may influence their design and application. Therefore, the process of combining computational simulations with experimental validation requires continuous adjustments and optimizations of the model. Conducting multiple experiments and parameter adjustments is essential for calibrating the model and improving its accuracy.

最后，外部环境因素如温度、pH值和离子浓度的变化也可能对RNA传感器的活性产生重要影响。RNA的稳定性和功能对这些因素非常敏感，如果模型未考虑这些环境因素的变化，可能无法有效描述实际情况。此外，RNA传感器在不同细胞类型或生理条件下的适应性也可能影响其设计和应用。因此，在计算模拟与实验验证相结合的过程中，模型需要不断进行调整和优化。进行多次实验和参数调节是必要的，以校准模型并提高其准确性。

To address the current suboptimal fitting results, we can consider several improvement measures to enhance the accuracy and reliability of the RNA sensor model.

为了解决当前拟合结果不理想的问题，我们可以考虑多项改进措施，以提升RNA传感器模型的准确性和可靠性。

First, improving the methods for acquiring model parameters is vital. We can collect more kinetic data through larger-scale experimental designs to ensure more accurate parameter estimates. Additionally, employing advanced statistical analysis methods, such as Bayesian inference, can more effectively handle data uncertainties, resulting in more reliable parameter estimates. Second, in terms of RNA structure prediction, we can utilize more sophisticated computational tools and algorithms, combined with experimental validation, to improve the accuracy of structural predictions. Simultaneously, we should consider integrating results from various prediction tools to mitigate biases introduced by any single tool. For docking experiments, we can conduct more replicative experiments and use different docking software for cross-validation to ensure the reliability and accuracy of docking results. Furthermore, to address the limitations of model assumptions, we should explore more complex model forms, including dynamic and multiscale models. This will allow us to more comprehensively consider various interactions and feedback mechanisms within biological systems. Moreover, incorporating the principles of systems biology and integrating multi-level data (such as genomics, transcriptomics, and proteomics) will aid in the comprehensive understanding of RNA sensor functionality. Finally, to better account for the impact of external environmental factors, we need to conduct systematic studies analyzing the effects of varying conditions such as temperature, pH, and ion concentration on RNA sensor activity. Establishing dynamic models of environmental changes will help enhance the model's adaptability and predictive capabilities.

首先，改进模型参数的获取方法至关重要。我们可以通过更大规模的实验设计来收集更多的动力学数据，以确保参数的估计更加精确。此外，采用高级的统计分析方法，如贝叶斯推断，可以更有效地处理数据不确定性，从而获得更可靠的参数估计。其次，在RNA结构预测方面，我们可以利用更为先进的计算工具和算法，结合实验验证，来提高结构预测的准确性。同时，我们应考虑使用多种预测工具的结果进行集成，以降低单一工具带来的偏差。对于对接实验，我们可以进行更多的重复实验，并使用不同的对接软件进行交叉验证，以确保对接结果的可靠性和准确性。进一步地，针对模型假设的局限性，我们应当探索更为复杂的模型形式，包括动态模型和多尺度模型。这将使我们能够更全面地考虑生物系统中的各类相互作用和反馈机制。此外，引入系统生物学的理念，通过整合多层次的数据（如基因组学、转录组学和蛋白质组学）将有助于全面理解RNA传感器的功能。最后，为了更好地考虑外部环境因素的影响，我们需要开展系统性研究，分析不同温度、pH值和离子浓度等条件对RNA传感器活性的影响。建立环境变化的动态模型，将有助于提升模型的适应性和预测能力。

Trend: As the number of bases increases, the success rate shows a steady upward trend.

趋势：随着碱基范围的增加，成功率呈稳步上升的趋势。

Initial Growth: At lower base numbers, the success rate is relatively low. However, it rises rapidly as the base number increases.

初始增长：碱基范围较小时，成功率相对较低。然而，随着碱基数的增加，它会迅速上升。

Gradual Increase: From 30 to 60 bases, the success rate continues to rise but at a slower, more gradual pace.

逐步增加：碱基范围从30到60个，成功率继续上升，但速度更慢，更渐进。