Production of Stimulated Emission-Encoded Beads for Large Molecular Library Generation

Project Summary/Abstract Combinatorial synthesis is a simple and effective strategy to generate large libraries of ligands for diagnostic tool and therapeutic lead development – all without access to a high-throughput screening (HTS) facility. Efforts to miniaturize combinatorial library synthesis and screening can dramatically accelerate molecular discovery, but require an accompanying encoding strategy to push below the detection limits for direct chemical determination techniques (e.g., NMR, mass spectrometry). Chemical encoding has emerged as the most popular strategy, but requires that “winning” ligands are selected for deconvolution. Only a small fraction of hits can be resynthesized and validated (<0.01% of the library), finally establishing the informatic link between molecular identity and ac- tivity. Compared to HTS, chemically-encoded library screening is a deceptively data-poor endeavor. An encoding medium that allows for simultaneous assay readout and decoding would eliminate data scarcity, but a scalable solution for large libraries (>105 members) does not exist. This proposal will explore optical-encoded library tech- nology, which is capable of encoding diverse molecular libraries and decoding them instantaneously during assay readout. Libraries will be synthesized on polymeric beads encoded with optical particles that emit narrow-band- width light upon excitation. An alphanumerical bead identifier is derived from the spectrum for a random combi- nation of particles within a single bead. Bead identifiers are tracked through each round of split-and-pool synthe- sis to record the molecular identity of each bead in a library database. During screening, the assay readout associ- ated with each bead identifier is recorded in a screening database. Database integration via bead identifier match- ing links each library member to its cognate activity. Validation of the optical encoding platform will proceed by the microfluidic droplet-templated production of encoded beads, which will be used to examine on-bead binding assays using positive control authentic ligands and their known targets in model libraries. The platform will be assessed for library encoding depth, bead identifier mismatch frequency, and false discovery rate during model library screens. Successful development of the optical encoding platform will allow for the efficient, miniaturized screening of large combinatorial libraries without additional bead processing (e.g., hit sorting). Such a compre- hensive encoding/decoding approach can rapidly generate millions of quantitative primary screening data points to train machine learning models, producing a platform for next-generation therapeutic and diagnostic discovery efforts.