SMEpred workbench: A web server for predicting efficacy of chemicallymodified siRNAs

Area of Science:

• Bioinformatics
• Computational Biology
• RNA Therapeutics

Background:

• Chemical modifications enhance small interfering RNA (siRNA) therapeutic applications.
• Experimental design and testing of siRNAs and chemically modified siRNAs (cm-siRNAs) are resource-intensive.

Purpose of the Study:

• To develop an in-silico tool for predicting the efficacy of both siRNAs and cm-siRNAs.
• To facilitate the rational design of novel cm-siRNAs for therapeutic development.

Main Methods:

• Developed the SMEpred workbench using 3031 cm-siRNA sequences from the siRNAmod database.
• Employed Support Vector Machine (SVM) with sequence features like nucleotide composition and binary patterns.
• Integrated algorithms for predicting normal siRNA efficacy from gene/mRNA sequences and designing multiple modifications.

Main Results:

• Achieved a high Pearson Correlation Coefficient (PCC) of 0.80 in both 10-fold cross-validation and independent validation.
• Demonstrated the predictive accuracy of SMEpred for siRNA and cm-siRNA efficacy.
• Provided a pipeline for designing and evaluating multiple chemical modifications on siRNAs.

Conclusions:

• SMEpred offers a valuable in-silico approach for designing effective siRNAs and cm-siRNAs.
• The webserver accelerates RNA interference-based technology research and therapeutic development.
• This tool will aid the scientific community in advancing RNAi-based therapies.