Antibody thermostability is challenging to predict from sequence and/or structure. This difficulty is likely due to the absence of direct entropic information. Herein, we present AbMelt where we model the inherent flexibility of homologous antibody structures using molecular dynamics simulations at three temperatures and learn the relevant descriptors to predict the temperatures of aggregation (T), melt onset (T), and melt (T). We observed that the radius of gyration deviation of the complementarity determining regions at 400 K is the highest Pearson correlated descriptor with aggregation temperature (r = −0.68 ± 0.23) and the deviation of internal molecular contacts at 350 K is the highest correlated descriptor with both T (r = −0.74 ± 0.04) as well as T (r = −0.69 ± 0.03). Moreover, after descriptor selection and machine learning regression, we predict on a held-out test set containing both internal and public data and achieve robust performance for all endpoints compared with baseline models (T R = 0.57 ± 0.11, T R = 0.56 ± 0.01, and T R = 0.60 ± 0.06). In addition, the robustness of the AbMelt molecular dynamics methodology is demonstrated by only training on <5% of the data and outperforming more traditional machine learning models trained on the entire data set of more than 500 internal antibodies. Users can predict thermostability measurements for antibody variable fragments by collecting descriptors and using AbMelt, which has been made available.

中文翻译：

---

AbMelt：从分子动力学学习抗体热稳定性


从序列和/或结构预测抗体热稳定性具有挑战性。这种困难可能是由于缺乏直接的熵信息。在此，我们提出了 AbMelt，其中我们在三个温度下使用分子动力学模拟对同源抗体结构的固有灵活性进行建模，并学习相关描述符来预测聚集温度 (T)、熔化起始温度 (T) 和熔化温度 (T)。我们观察到，400 K 时互补性决定区的回转半径偏差是与聚集温度 (r = -0.68 ± 0.23) 最高的 Pearson 相关描述符，而 350 K 时内部分子接触的偏差是与两者的最高相关描述符。 T (r = -0.74 ± 0.04) 以及 T (r = -0.69 ± 0.03)。此外，在描述符选择和机器学习回归之后，我们对包含内部和公共数据的保留测试集进行预测，并与基线模型相比，在所有端点上实现稳健的性能（TR = 0.57 ± 0.11，TR = 0.56 ± 0.01，并且TR = 0.60 ± 0.06)。此外，仅对 <5% 的数据进行训练，证明了 AbMelt 分子动力学方法的稳健性，并且优于在 500 多个内部抗体的整个数据集上训练的更传统的机器学习模型。用户可以通过收集描述符并使用已推出的 AbMelt 来预测抗体可变片段的热稳定性测量值。