RaptScore: a large language model-based algorithm for versatile aptamer evaluation

Abstract RNA aptamers are a high-potency tool in the life sciences, offering promising applications in drug discovery and beyond. They are typically obtained through systematic evolution of ligands by exponential enrichment (SELEX), which imposes constraints on sequence length and diversity. Several metrics, such as frequency and enrichment, have been developed to identify high-activity aptamers from SELEX. However, existing evaluation metrics are limited to sequences that appear within SELEX and cannot assess sequences of varying lengths, limiting their utility in optimizing aptamer design. To overcome these limitations, we developed RaptScore, a novel binding activity evaluation metric leveraging large language models. RaptScore enables the assessment of arbitrary sequences, including those absent from SELEX, and accommodates variations in sequence length. RaptScore exhibited a strong correlation with binding activity, allowing the identification of shorter aptamers with enhanced binding properties. By integrating RaptScore with in silico maturation, we achieved a 10-nucleotide truncation while maintaining binding efficiency. Furthermore, we demonstrated improved aptamer discovery efficiency by combining RaptScore with RaptGen, a variational autoencoder-based aptamer discovery tool. By enabling efficient sequence evaluation and optimization, RaptScore provides a powerful tool for aptamer research, facilitating the discovery of high-activity candidates while reducing experimental effort.