digitalSELEX

Abstract

Molecules that have high affinity and specificity for their target are critical for functioning biosensors and effective therapeutics. Aptamers, or single-stranded oligonucleotides, are one type of molecule capable of both high affinity and specificity. Systematic Evolution of Ligands by EXponential enrichment (SELEX) is the iterative in vitro process for identifying aptamers with high affinity and specificity from an initial pool of approximately 1015 randomized nucleotide molecules. There have been a multitude of SELEX variations developed over the years to include incorporation of machine learning algorithms to address the limited success (~30%), cost, and time required to identify high affinity and specific aptamers. While some SELEX variations have been more successful than others in addressing some of the challenges, issues remain. To confront these challenges, the digitalSELEX platform introduces a novel de novo design approach. The platform has two main components. The first component analyzes the target molecule identifying clusters of amino acids along the molecule’s surface based on their accessibility and proximity of atoms relevant to target-aptamer binding. The platform then proposes aptamers built from sequences of nucleotides that paired to the amino acids in the clusters. The second component improves these aptamers sequentially. This is done via simulation-based optimization procedure which uses molecular docking and stochastic optimization techniques. It explores small adjustments made on the starting aptamer that increase the affinity and specificity that is calculated extracting binding related features from the output of the docker. Once in silico counter-selection is complete, the best possible sequences are extracted for in vitro validation. To validate digitalSELEX, aptamers were designed for four different target molecules of varying size ranging from 18 – 140 kDa. Some of the aptamers were designed with specific counter-targets while others did not have counter-target molecules. In total, 19 oligonucleotides were chemically synthesized, and their affinity and specificity tested for five explicit validation problems. All 19 aptamers demonstrated high affinity for their respective target molecules. Sixteen of the 19 oligonucleotides were tested for specificity with nine meeting the 4-times Kd-value difference specificity criteria. Depending on the computational capacity being employed for each problem, the approximate time required from initiating the de novo design to the point of validation was 170 hours. The cost of in silico oligonucleotide design is negligible while validation of a few aptamers is few hundred dollars. The digitalSELEX platform was comprehensively tested examining the initial de novo design through affinity and specificity determination. The digtalSELEX platform is a prototype that has the opportunity for further development such as employing different molecular simulators.