Iterative Supervised Principal Component Analysis-Driven Ligand Design for Regioselective Ti-Catalyzed Pyrrole Synthesis

Abstract

<div> <div> <div> <p>Herein, we describe the use of iterative supervised principal component analysis (ISPCA) in de novo catalyst design. The regioselective synthesis of 2,5-dimethyl-1,3,4-triphenyl-1H- pyrrole (C) via Ti- catalyzed formal [2+2+1] cycloaddition of phenyl propyne and azobenzene was targeted as a proof of principle. The initial reaction conditions led to an unselective mixture of all possible pyrrole regioisomers. ISPCA was conducted on a training set of catalysts, and their performance was regressed against the scores from the top three principal components. Component loadings from this PCA space along with k-means clustering were used to inform the design of new test catalysts. The selectivity of a prospective test set was predicted in silico using the ISPCA model, and only optimal candidates were synthesized and tested experimentally. This data-driven predictive-modeling workflow was iterated, and after only three generations the catalytic selectivity was improved from 0.5 (statistical mixture of products) to over 11 (> 90% C) by incorporating 2,6-dimethyl- 4-(pyrrolidin-1-yl)pyridine as a ligand. The successful development of a highly selective catalyst without resorting to long, stochastic screening processes demonstrates the inherent power of ISPCA in de novo catalyst design and should motivate the general use of ISPCA in reaction development. </p> </div> </div> </div>

Computational Discovery of Transition-metal Complexes: From High-throughput Screening to Machine Learning

“…490 When points are distant in a feature space but similar in property, they also provide evidence that shallow, transparent models (e.g., MLR or KRR) would struggle to use these features to obtain a suitable prediction of properties. Thus, selection of optimal features by PCA has been carried out 491 as a feature selection technique, including in an iterative fashion, 492 for ML model training on data sets, as discussed further in section 4.4.…”

Synthesis of pentasubstituted 2-aryl pyrroles from boryl and stannyl alkynes via one-pot sequential Ti-catalyzed [2 + 2 + 1] pyrrole synthesis/cross coupling reactions

“…The functional groups involved in the initial screen included boronic acid pinacol ester ( Initial reaction conditions were based off from previously successful conditions for TMS-substituted alkyne substrates, 13 using chloride-based Ti catalysts that are typically most robust for [2 + 2 + 1] reactions. 22,23 All new heteroatom-substituted reactions resulted in signicantly lower yields than the corresponding TMS-substituted alkyne reactions, highlighting the challenges of conserving a reactive transmetallating agent through another organometallic transformation.…”

Navigating through the Maze of Homogeneous Catalyst Design with Machine Learning

“…In the past decade, homogeneous catalysis experienced a rise in new optimization strategies, such as data-driven modeling of both reactivity and selectivity with descriptors [17,18], and the systematic application of ML is becoming more common (Box 2) [19]. Specifically, the use of multivariate linear regression combined with the development of descriptors allowed the rapid design of catalysts to improve yields, reaction rates, and (enantio)selectivities, as demonstrated by several experimental applications [20][21][22]. These approaches can be utilized more readily in homogeneous catalysis as computing of parameter sets is more straightforward when good models of the active catalysts exist.…”