As interest in biobased chemicals grows, and their application space expands, computational tools to navigate molecule space as a complement to experimental approaches are imperative. This work expands upon previous work that identified candidate bioprivileged molecules from the C₆H_xO_y (C6) subspace. It refines the framework that was developed previously to better refine the molecules according to their biological origin and applies it to three new subspaces of chemical structure: C₄H_xO_y (C4), C₅H_xO_y (C5), and C₇H_xO_y (C7). For C5 and C7, roughly the top 100 bioprivileged candidates were identified, and the enhanced framework was applied to recast slightly the previous list of the top 100 C6 molecules. In addition, all top candidates were analyzed for their key functional moieties using a random forest model, and this algorithm was applied to compare the functional group space occupied by bioprivileged molecules of various databases of molecules with a focus on evaluating how closely the molecules were aligned with those known to biology. Furthermore, with the present work's focus on automation and data science principles, the framework can be easily expanded to include other chemical formulae to screen for bioprivileged candidates. This in turn facilitates the retrosynthesis process inherent in the framework to identify those bioprivileged intermediates in other subspaces that lead to target molecules.

中文翻译：

---

鉴定生物特权分子：将计算方法扩展到更广泛的分子空间

随着对基于生物的化学物质的兴趣不断增长，其应用空间不断扩大，用于导航分子空间以作为实验方法的补充的计算工具势在必行。这项工作是在以前的工作的基础上进行的，该工作从C ₆ H _x O _y（C6）子空间中识别出了候选的生物特权分子。它改进了先前开发的框架，以根据分子的生物学起源更好地改进分子，并将其应用于化学结构的三个新子空间：C ₄ H _x O _y（C4），C ₅ H _x O _y（C5）和C ₇高_x O _y（C7）。对于C5和C7，大致确定了前100个生物特权候选者，并使用增强的框架稍微重铸了前100个C6分子的列表。此外，使用随机森林模型分析了所有顶级候选者的关键功能部分，并将该算法用于比较各种分子数据库的生物特权分子所占据的官能团空间，重点是评估分子排列的紧密程度与那些生物学知道的。此外，由于本工作着重于自动化和数据科学原理，因此可以轻松扩展该框架，使其包含其他化学式来筛选生物特权候选人。