Utility-Aware Screening with Clique-Oriented Prioritization

Abstract

Most methods of deciding which hits from a screen to send for confirmatory testing assume that all confirmed actives are equally valuable and aim only to maximize the the number of confirmed hits. In contrast, “utility-aware” methods are informed by models of screeners’ preferences and can increase the rate at which the useful information is discovered. Clique-oriented prioritization (COP) extends a recently proposed economic framework and aims—by changing which hits are sent for confirmatory testing—to maximize the number of scaffolds with at least two confirmed active examples. In both retrospective and prospective experiments, COP enables accurate predictions of the number of clique discoveries in a batch of confirmatory experiments and improves the rate of clique discovery by more than three-fold. In contrast, other similarity-based methods like ontology-based pattern identification (OPI) and local hit-rate analysis (LHR) reduce the rate of scaffold discovery by about half. The utility-aware algorithm used to implement COP is general enough to implement several other important models of screener preferences.

Sharing Chemical Relationships Does Not Reveal Structures

“…The matrix is encoded in a skyline format to improve compression. 45,46 (5) The nearest neighbors were encoded by compressing an adjacency matrix of the whole data set. The molecules in this data set are reordered using the Reverse Cuthill−McKee algorithm to put molecules in the matrix with the same neighbors near one another, and the matrix is encoded in skyline format.…”

“…Nonetheless, sharing molecule IDs with assay data is more useful than might be expected. Molecule IDs provide enough information to find patterns in screening data − and to enable intelligible communication about private chemical entities. For example, the NIH repurposing program depends on sharing molecule IDs with assay and clinical data …”

“…Finding ways to share relevant chemical information about screening data that leaves structures blinded could open the door for this type of collaborative work. These methods include better strategies for identifying active molecules from primary screens, which leverage information from fingerprints, scaffold groupings, , economic modeling, − and improved processing of raw data. − They also include automatic methods of organize screening data into workflows, a series of powerful strategies for visualizing how biological activity maps to chemical space, − and understanding the contribution of specific substitutions to the properties of the molecule …”

An informatic pipeline for managing high-throughput screening experiments and analyzing data from stereochemically diverse libraries

“…In the absence of cheminformatics tools, data analysis from an HTS screen is an ad hoc process and may be biased by scientists’ pre-conceived notions about compounds or data, including biases that may not reflect their actual preferences [13–15]. Moreover, the use of cheminformatics tools allows the capture and documentation of the decision-making process at each stage, facilitating subsequent reanalysis of the data if desired.…”

Combined Analysis of Phenotypic and Target-Based Screening in Assay Networks

“…4 Likewise, public data enable the direct comparison of competing methods of selecting hits from primary screens. 5 Similarly, it enables academics to propose and test improvements to screening experiment design. 6…”