Transform-MinER (Transforming Molecules in Enzyme Reactions) is a web application facilitating the exploration of chemical biosynthetic space, guiding the user toward promising start points for enzyme design projects or directed evolution experiments. Two types of search are possible: Molecule Search allows a user to submit a source substrate enabling Transform-MinER to search for enzyme reactions acting on similar substrates, whereas Path Search additionally allows a user to submit a target molecule enabling Transform-MinER to search for a path of enzyme reactions acting on similar substrates to link source and target. Transform-MinER searches for potential reaction centers in the source substrate and uses chemoinformatic fingerprints to identify those that are situated in molecular environments similar to native counterparts, prioritizing steps that move closer to the target using reactions most similar to native in its exploration of search space. The ligand-based methodology behind Transform-MinER is presented, and its performance is validated yielding 90% success rates: first, on a data set of native pathways from the KEGG database, and second, on a data set of de novo enzyme reactions.

中文翻译：

---

使用Transform-MinER探索化学生物合成设计空间。

Transform-MinER（酶反应中的转化分子）是一个网络应用程序，可促进化学生物合成空间的探索，指导用户朝着酶设计项目或定向进化实验的有希望的起点迈进。两种类型的搜索是可能的：分子搜索允许用户提交源底物，从而使Transform-MinER可以搜索作用在相似底物上的酶反应，而路径搜索还允许用户提交目标分子，从而使Transform-MinER可以搜索酶反应作用在相似底物上以连接源和靶的路径。Transform-MinER在源底物中搜索潜在的反应中心，并使用化学信息学指纹来识别与天然对应物相似的分子环境中的那些指纹，使用最接近搜索目标的本机反应，对接近目标的步骤进行优先级排序。介绍了Transform-MinER背后基于配体的方法，并对其性能进行了验证，得出90％的成功率：第一，来自KEGG数据库的天然途径数据集，第二，来自KEGG数据库的天然途径数据集。从头酶反应。