SPORCalc

From Wikipedia, the free encyclopedia

The Substrate Product Occurrence Ratio Calculator ("SPORCalc") is a chemical informatics process for interrogating reaction databases. The SPORCalc process was designed for high-throughput analysis of enzyme transformations of biochemical agents, toxins, xenobiotics and intermediates during endogenous metabolism.

An aim of SPORCalc is to identify labile atom positions in diverse enzyme substrates and products and to derive the conditional probabilities (propensities) that given enzymatic reactions occur at those atom positions. SPORCalc requires two complementary analyses.

1. A comparison of the tails of a Normal distribution of substrate substructures from a Reaction Class subset from a metabolic reaction database and the Normal distribution of similar substructures in the superset of substrates from the entire database.

2. A comparison of the tails of the Normal distribution of product substructures from a Reaction Class subset from a metabolic reaction database and the Normal distribution of similar substructures in the superset of products from the entire database.

Each analysis gives a ratio (an Occurrence Ratio), OR_substrates and OR_products, respectively. The product of the two ratios gives the reaction Occurrence Ratio (OR_RXN) equal to the propensity . Every OR_RXN value gives a likelihood of an enzyme reaction occurring at atom positions throughout the candidate compounds. The orders of magnitude of the OR_RXN values can be calibrated with the observations from experimental studies of enzyme metabolism and therefore used to interpret the half-lives of metabolic reaction intermediates. The reliability of SPORCalc and the OR_RXN values depend on the diversity of metabolites in the enzyme reaction database used and the number of annotated enzyme isoform reactions used.

SPORCalc was designed as an initial filter for metabolism databases containing two-dimensional molecular subgraph information. The process has not been designed for identifying three-dimensional recognition, enzyme isoform-specific or reaction-intermediate, pharmacophores. SPORCalc generates the propensities for given enzyme reactions using separate distributions from both substrate and product substructures. The resulting OR_RXN values can be generated rapidly for every atom position in a candidate compound library for addition or elimination meatbolic reactions^[1].

Keywords: metabolic prediction, computational pharmacology, ligand-based drug design, medicinal chemistry, enzyme biochemistry.