Ke Li; Xiaohong Liu; Sixiu Liu; Yulong An; Yanfang Shen; Qingxia Sun; Xiaodong Shi; Wenji Su; Weiren Cui; Zhiqiang Duan; Letian Kuai; Hongfang Yang; Alexander L. Satz; Kaixian Chen; Hualiang Jiang; Mingyue Zheng; Xuanjia Peng; Xiaojie Lu
iScience, 2020, 23(6), 101142
https://doi.org/10.1016/j.isci.2020.101142

Abstract

The application of machine learning towards DNA encoded library (DEL) technology is lacking despite obvious synergy between these two advancing technologies. Herein, a machine learning algorithm has been developed that predicts the conversion rate for the DNA compatible reaction of a building block with a model DNA-conjugate. We exemplify the value of this technique with a challenging reaction, the Pictet-Spengler, where acidic conditions are normally required to achieve the desired cyclization between tryptophan and aldehydes to provide tryptolines. To avoid damaging the DNA our reaction conditions must be exceptionally mild, and therefore most building blocks fail to provide acceptable yields of desired product (<20% pass rate) in a test reaction employing our optimized protocol. In contrast, building blocks selected by our trained machine learning algorithm have a >78% pass rate. This is the first demonstration of using a machine learning algorithm to cull potential building blocks prior to their purchase and testing for DNA encoded library synthesis. Importantly, this allows for a challenging reaction, with an otherwise very low building block pass rate in the test reaction, to still be used in DEL synthesis. Furthermore, we discuss herein the rational design of DNA conjugated tryptophan substrates for our Pictet-Spengler reaction, and optimization of the reaction protocols. Lastly, because our protocol is solution-phase it is directly applicable to standard plate-based DEL synthesis.