Regarding our original paper, concerning the actual flexibility of the PDLA and consequently the values of the internal distances presented in Figures 11, 13, and 15, it is important to note that the same force field parameters have been used for both l- and d-monomers; i.e., these two structures differ only in the spatial position of the hydrogen in the monomer. More specifically, according to the force field published in ref 45 of the original manuscript the l- and d-forms of poly(lactic acid) should differ in the parameters for CMAP dihedrals and the tabulated potentials for the backbone dihedrals. We did not implement those differences, as it would require the usage of two external tabulated potentials for the copolymer, which is not feasible in Gromacs simulation package. Thus, due to simplicity, consistency, and feasibility, we implemented the same set of force field parameters for our model systems, and we only changed the chemical structure to obtain multiple stereoisomers. Note also that the full implementation of the force field as reported in ref 45 of the original manuscript would be therefore possible only in the case of the pure PDLA and would lead to a reduced flexibility and thus to a higher terminal plateau in Figures 11(a) and 15(a) for PDLA100 and PDLA30, respectively. The full implementation, with the corresponding verification and the modification of the force field for the copolymers, is a topic of our current work. Nevertheless, concerning the methodology and the algorithm for backmapping presented here, we would like to stress again one of our conclusions, namely, that the presented computational tools are versatile and do not rely on the actual force field parameters used for the polymer chains. This article has not yet been cited by other publications.

中文翻译：

---

对“在多种聚乳酸立体异构体的粗粒度配置中重新引入原子细节的物理知情深度学习方法”的修正


关于我们的原始论文，关于 PDLA 的实际灵活性以及图 11、13 和 15 中显示的内部距离值，重要的是要注意l - 和d使用了相同的力场参数-单体；即，这两种结构仅在单体中氢的空间位置上有所不同。更具体地说，根据原始手稿参考文献 45 中发表的力场，聚乳酸的l型和d型在 CMAP 二面体的参数和主链二面体的表列势方面应有所不同。我们没有实现这些差异，因为它需要使用共聚物的两个外部表势，这在 Gromacs 模拟包中是不可行的。因此，出于简单性、一致性和可行性的考虑，我们为我们的模型系统实现了相同的力场参数集，并且我们仅改变了化学结构以获得多种立体异构体。还请注意，原始手稿的参考文献 45 中报告的力场因此只有在纯 PDLA 的情况下才可能实现，并且会导致灵活性降低，从而导致图 11 中更高的终端平台（a PDLA100 和 PDLA30 分别为 ) 和 15(a)。全面实施，以及相应的验证和共聚物力场的修改，是我们当前工作的主题。 然而，关于这里提出的反向映射的方法和算法，我们想再次强调我们的结论之一，即所提出的计算工具是通用的，并且不依赖于用于聚合物链的实际力场参数。这篇文章尚未被其他出版物引用。