Overview

AmiRNAs offer advantages over other RNAi tools due to their high gene silencing specificity and reduced off-target effects. However, efficient and safe siRNA production relies on Drosha and Dicer's precise processing of amiRNAs. Since the sequence, structure, and variability of the pri-miRNA strongly influence this processing, predicting the outcome of replacing a miRNA with an exogenous siRNA sequence is challenging. Poor amiRNA design can lead to inefficient processing, the generation of siRNA variants with altered seed sequences, or arm switching - a process that results in the unintended release of the passenger strand. Nevertheless, studies have shown that modifying the 21-nt sequence of a mature miRNA within a natural miRNA precursor can preserve proper biogenesis and maturation, offering a pathway for designing amiRNAs with novel targets.

miRarchitect was developed to facilitate the rational design of amiRNA therapeutics. It integrates state-of-the-art computational techniques, including Feedforward Neural Networks (FNNs), to analyze target sequences, model siRNA structure and thermodynamics, select optimal native pri-miRNA scaffolds, and generate functional amiRNAs.

Team

miRarchitect tool is managed by researchers at the Department of Genome Engineering and at the Department of Structural Bioinformatic, Institute of Bioorganic Chemistry, Polish Academy of Sciences, Poznan, Poland and Institute of Computing Science, Poznan University of Technology, Poznan, Poland.