APIPred: An XGBoost-Based Method for Predicting Aptamer-Protein Interactions.

Abstract

Aptamers are single-stranded DNA or RNA oligos that can bind to a variety of targets with high specificity and selectivity and thus are widely used in the field of biosensing and disease therapies. Aptamers are generated by SELEX, which is a time-consuming procedure. In this study, using in silico and computational tools, we attempt to predict whether an aptamer can interact with a specific protein target. We present multiple data representations of protein and aptamer pairs and multiple machine-learning-based models to predict aptamer-protein interactions with a fair degree of selectivity. One of our models showed 96.5% accuracy and 97% precision, which are significantly better than those of the previously reported models. Additionally, we used molecular docking and SPR binding assays for two aptamers and the predicted targets as examples to exhibit the robustness of the APIPred algorithm. This reported model can be used for the high throughput screening of aptamer-protein pairs for targeting cancer and rapidly evolving viral epidemics.