Generative aptamer discovery using RaptGen

Abstract

Nucleic acid aptamers are generated by an in vitro molecular evolution method known as systematic evolution of ligands by exponential enrichment (SELEX). Various candidates are limited by actual sequencing data from an experiment. Here we developed RaptGen, which is a variational autoencoder for in silico aptamer generation. RaptGen exploits a profile hidden Markov model decoder to represent motif sequences effectively. We showed that RaptGen embedded simulation sequence data into low-dimensional latent space on the basis of motif information. We also performed sequence embedding using two independent SELEX datasets. RaptGen successfully generated aptamers from the latent space even though they were not included in high-throughput sequencing. RaptGen could also generate a truncated aptamer with a short learning model. We demonstrated that RaptGen could be applied to activity-guided aptamer generation according to Bayesian optimization. We concluded that a generative method by RaptGen and latent representation are useful for aptamer discovery.

Machine Learning Directed Aptamer Search from Conserved Primary Sequences and Secondary Structures

“…In the example in Figure , the hairpin contributed 70% of the structure score in this particular sequence, and the structure score carried the most weight in all the three score components. This extra hairpin score sets our algorithm apart from the previous work focused on primary sequence alignment. ,,, Different hairpin sequences might be predicted in different clusters (e.g., the purine-rich sequences in Figure A and A), although they may belong to the same family as illustrated in Figure . The presence of multiple sequences with different hairpins is an indication of successful selection and prediction.…”

“…However, aptamer enrichment happens mainly in the later rounds of selections, and other factors such as PCR bias can also lead to preferentially amplified sequences that are not necessarily aptamers . In addition, motif finding algorithms have been developed, ,, which rely on predicted secondary structures and dividing the structures into multiple subunits. A motif is defined based on the sequence of single-stranded regions, and it still looks for identical sequences.…”

Fluorescent RNA Tags for In Situ RNA Imaging in Living Cells

“…Notably, the development of computer science paves the way for de novo aptamer design without wet-lab SELEX experiments in the future. [104] Besides, for most FLAPs, fluorophore emission wavelengths locate in the blue-tored region. Very few FLAPs with near-infrared emission are reported.…”

In Silico discovery of aptamers with an enhanced library design strategy

“…To cover this gap, several sequence analysis algorithms were developed to identify valid sequences from high-throughput sequencing (HTS) data. These molecular identification algorithms efficiently and accurately identify ligand recognition aptamers by multidimensional evaluation of sequence motif enrichment, aptamer family abundance, and structural stability [12] , [13] . Moreover, molecular simulation based approaches [1] have also been reported for the optimal modification design of aptamers, where the binding free energy calculation in combination with the implicit model significantly helped the experimentalists to optimize the minimum useful length of the aptamer [14] .…”