Finding the K best synthesis plans

Abstract

In synthesis planning, the goal is to synthesize a target molecule from available starting materials, possibly optimizing costs such as price or environmental impact of the process. Current algorithmic approaches to synthesis planning are usually based on selecting a bond set and finding a single good plan among those induced by it. We demonstrate that synthesis planning can be phrased as a combinatorial optimization problem on hypergraphs by modeling individual synthesis plans as directed hyperpaths embedded in a hypergraph of reactions (HoR) representing the chemistry of interest. As a consequence, a polynomial time algorithm to find the K shortest hyperpaths can be used to compute the K best synthesis plans for a given target molecule. Having K good plans to choose from has many benefits: it makes the synthesis planning process much more robust when in later stages adding further chemical detail, it allows one to combine several notions of cost, and it provides a way to deal with imprecise yield estimates. A bond set gives rise to a HoR in a natural way. However, our modeling is not restricted to bond set based approaches—any set of known reactions and starting materials can be used to define a HoR. We also discuss classical quality measures for synthesis plans, such as overall yield and convergency, and demonstrate that convergency has a built-in inconsistency which could render its use in synthesis planning questionable. Decalin is used as an illustrative example of the use and implications of our results.

Compositionality of Rewriting Rules with Conditions

“…sequences of rule applications to a given input graph), we posit that a viable alternative approach may consist in focusing instead on sequential rule compositions, which in particular when combined with application conditions is anticipated to yield a powerful framework to study the causality of rewriting systems. For certain specialized applications of DPO-type graph rewriting, namely those in the well-established field of chemical graph rewriting [3,7,17], such types of analyses have already proven very fruitful [4][5][6]43]. Therefore, we believe that our compositional refinements described in the present paper can provide a significant contribution to future algorithm developments in this field.…”

“…e.g. [1, 23, 31-33, 44, 56, 61] and [4,6,7,17,43]) is beyond the scope of the present paper, suffice it here to comment on a salient technical point: notably, many of the developments in the aforementioned frameworks were not explicitly rooted in categorical rewriting theory itself, despite the foundations of the two fields upon this type of theory. We recently demonstrated in [10] that based upon the results of the present paper, one may not only reformulate equivalently the current implementations of chemical rewriting theories, but one may also take advantage of our novel compositionality and associativity results in order to formulate rule algebra and tracelet theories (see below) to develop new approaches to the static analysis of complex reaction systems and their dynamics.…”

Selection of cost-effective yet chemically diverse pathways from the networks of computer-generated retrosynthetic plans

“…In the subsequent step, the costs of all nodes in the network are calculated bottom-up ( i.e. , from the starting materials to the target) using a Dijkstra-like algorithm similar to the one for finding minimum-weight B-paths in weighted hypergraphs24 (see also ref. 26).…”

“…We note that although the problems of (i) finding a desired number of the best/lowest-cost solutions within the so-called directed graphs with weighted nodes ( e.g. , in random time-dependent networks,21–23 transit networks,21,24,25 or reaction networks19,20,26) and also (ii) identifying qualitatively different pathways ( e.g. , within transportation networks27) have been individually studied in graph theory, the specific approaches are not easily extendable to realistic synthetic–organic planning ( cf.…”

Predicting Retrosynthetic Pathways Using a Combined Linguistic Model and Hyper-Graph Exploration Strategy

“…The dawn of AI-driven chemistry. While human chemical knowledge will keep fueling the organic chemistry research in the years to come, a careful analysis of current trends [5,[7][8][9][10][11][12][13][14][15][16][17][18][19][20] and the application of basic extrapolation principles undeniably shows that there are growing expectations on the use of Artificial Intelligence (AI) architectures to mimic human chemical intuition and to provide research assistant services to all bench chemists worldwide.…”