Iterative Stochastic Elimination for Solving Complex Combinatorial Problems in Drug Discovery

Abstract

Iterative Stochastic Elimination (ISE) is a novel algorithm that was originally developed to solve extremely complex problems in protein structure and interactions, and has since been applied to diverse topics that share a few general “ingredients”: they are extremely complex, of combinatorial nature, may be presented as large sets of variables that can each have many alternative values, there is some interdependence of the variables on each other, and there is a scoring function that can evaluate each choice of the problems “configuration”; this is the set of single values of each of the variables that constitute its full presentation. Those are picked randomly in a large sample, the analysis of which allows decisions to be made for rejecting some values for each of the variables; thus resulting in a smaller set of potential combinations. This continues in iterations until the number of combinations allows all the remaining options to be computed exhaustively and to order them by their scores. ISE has been mainly applied to problems that are relevant to drug design and discovery. We demonstrate, among others, the use of ISE to determine the properties of molecular ensembles and to pick the best molecules (“focused libraries”) for hitting a specific target. Future ideas for using ISE are discussed, as well as mentioning its contributions to the construction of two start‐up companies.

In silico Docking Analysis for Blocking JUNO‐IZUMO1 Interaction Identifies Two Small Molecules that Block in vitro Fertilization

“…In order to expand the pool of candidate molecules, a classification model was constructed using the Iterative Stochastic Elimination (ISE) algorithm ( Stern and Goldblum, 2014 ) and was based on the docking results to the 5JKC JUNO structure: Docked molecules that met the geometric criteria were divided into 68 “Best” molecules (top quartile) and 69 “Worst” molecules (bottom quartile) according to the criteria listed in Table 3 . The two sets were combined into a learning set, for which 206 molecular descriptors were calculated by MOE2018.0101 [ Molecular Operating Environment (MOE) and Chemical Computing Group ULC, 2018 ].…”

Computational design of substrate selective inhibition

“…Using ISE [26][27][28][29][30], a five-fold model (see Materials and Methods) of the ncP52 set (31 fragments) vs. the random set (7104) was constructed. The AUC was 0.97 (see ROC curve in S4A Fig) [31].…”

“…The fourth method was Visual Inspection (VI), in which we visually examined each of the 50 conformations of each molecule for presence or absence in the S1-S1' area of POP (in this option we demand that at least 30 poses will not interact with the S1-S1' area), a highly tedious and time consuming examination. ISE classification modeling [26]. In the learning set we have actives and inactives (actives are diluted by at least 100-fold inactives).…”

Candidate Therapeutics by Screening for Multitargeting Ligands: Combining the CB2 Receptor With CB1, PPARγ and 5-HT4 Receptors

“…Our generic ISE algorithm has been applied to many problems related to drug discovery and has been presented in reviews, with details of the mathematical and statistical criteria to distinguish between two activities based on physicochemical properties (descriptors) of known active vs. inactive compounds ( Stern and Goldblum, 2014 ; El-Atawneh and Goldblum, 2017 ). For each model, five cross-validations were performed ( James et al, 2013 ), with 4 out of the five-folds producing the model, and the fifth fold was used as a test set.…”

“…Our research combines ligand and structure-based methods. Our algorithm for solving complex combinatorial problems, the 'Iterative stochastic elimination algorithm’ (ISE) ( Stern and Goldblum, 2014 ; El-Atawneh and Goldblum, 2017 ), has been applied in recent years to molecular discovery ( Zatsepin et al, 2016 ; Da’adoosh et al, 2019 ; El-Atawneh et al, 2019 ), including one example of multitargeting modeling: modeling the properties of molecules that may be remotely loaded to nanoliposomes and the properties that enable them to be stable inside the nanoliposomes, in a biological fluid ( Cern et al, 2017 ). Molecules that had high scores in both loading and stability models were chosen.…”