Abstract
Iterative syntheses comprise sequences of organic reactions in which the substrate molecules grow with each iteration and the functional groups, which enable the growth step, are regenerated to allow sustained cycling. Typically, iterative sequences can be automated, for example, as in the transformative examples of the robotized syntheses of peptides, oligonucleotides, polysaccharides and even some natural products. However, iterations are not easy to identify—in particular, for sequences with cycles more complex than protection and deprotection steps. Indeed, the number of catalogued examples is in the tens to maybe a hundred. Here, a computer algorithm using a comprehensive knowledge base of individual reactions constructs and evaluates myriads of putative, but chemically plausible, sequences and discovers an unprecedented number of iterative sequences. Some of these iterations are validated by experiment and result in the synthesis of motifs commonly found in natural products. This computer-driven discovery expands the pool of iterative sequences that may be automated in the future.
Highlights
Iterative syntheses comprise sequences of organic reactions in which the substrate molecules grow with each iteration and the functional groups, which enable the growth step, are regenerated to allow sustained cycling
We demonstrate that iterative sequences can be identified without serendipity using computers equipped with the knowledge of individual reaction steps, which the machine concatenates into sequences and recognizes those that are mutually compatible and iterative
These expert-coded rules were chosen owing to their high quality[20], which has been documented by numerous experimental validations of computer-designed syntheses, including those of drugs[16,21], prebiotic compounds[19] and complex natural products18—this quality is essential for the correctness of the reaction sequences we set out to discover
Summary
As the first set of examples of previously undescribed iterations, we consider the syntheses of various poly(heterocycles) shown in Fig. 3 that are potentially useful in molecular electronics[24,25,26] and sensing[27] We chose this family for illustration because the iterations do not conventionally couple the heterocyclic subunits but instead build them from acyclic precursors (allowing more flexible substitution patterns; groups marked as R in the figure) and because the key, ring-forming steps have close literature precedents (Supplementary Fig. 9) that support their feasibility within the proposed sequences. The reader can perform these and other analyses via the web application posted at https://iterator.allchemy.net/ (for a short user manual see Supplementary Section 12)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have