Prediction of siRNA Efficacy Using BP Neural Network

Abstract:

In the last decade, RNA interference (RNAi) by small interfering RNAs (siRNAs) has become a hot topic in both molecular biology and bioinformatics. The success of RNAi gene silencing depends on the specificity of siRNAs for particular mRNA sequences. As a targeted gene could have thousands of potential siRNAs, finding the most efficient siRNAs among them constitutes a huge challenge. Previous studies such as rules scoring or machine learning aim to optimize the selection of target siRNAs. However, these methods have low accuracy or poor generalization ability, when they used new datasets to test. In this study, a siRNA efficacy prediction method using BP neural network (BP-GA) was proposed. For more efficient siRNA candidate prediction, twenty rational design rules our defined were used to filter siRNA candidate and they were used in the neural network model as input parameters. Furthermore, the performance optimization of network model has been done by using genetic algorithm and setting optimal training parameters. The BP-GA was trained on 2431 siRNA records and tested using a new public dataset. Compared with existing rules scoring and BP methods, BP-GA has higher prediction accuracy and better generalization ability.

• Periodical: Applied Mechanics and Materials (Volumes 644-650)
• Pages: 5341-5345
• DOI: https://doi.org/10.4028/www.scientific.net/AMM.644-650.5341
• Online since: September 2014
• Authors: Xuan Wang*, Fa Zhang
• Keywords: BP Neural Network (BPNN), Genetic Algorithm (GA), Sirna Design, siRNA Efficiency Prediction

References

[1] A. Fire, S. Xu, M.K. Montgomery, S.A. Kostas, S.E. Driver and C.C. Mello: Nature Vol. 391 (1998), pp.806-811.

[2] S.M. Hammond, E. Bernstein, D. Beach and G.J. Hannon: Nature Vol. 404 (2000), pp.293-296.

[3] G.J. Hannon: Nature Vol. 418 (2002), pp.244-251.

[4] M. Amarzguioui, H. Prydz: Biochem Biophys Res. Commun. Vol. 316 (2004), pp.1050-1058.

[5] D. Quan, T. Hakan, W. Jue, W. Claes, L. Zicai: Nucleic Acids Res. Vol. 33 (2005), pp.1671-1677.

[6] O. Matveeva1, Y. Nechipurenko, L. Rossi: Nucleic Acids Res. Vol. 35 (2007), p. e63.

[7] Y. Ren, W. Gong, Q. Xu, X. Zheng, D. Lin, Y. Wang and T. Li: Bioinform. Vol. 22 (2006), pp.1027-1028.

[8] P. Saetrom: Bioinform. Vol. 20 (2004), pp.3055-3063.

[9] D. Huesken, J. Lange, C. Mickanin: Nat Biotechnol. Vol. 23 (2005), pp.995-1001.

[10] SA. Shabalina, AN. Spiridonov, AY. Ogurtsov: BMC Bioinform. Vol. 7 (2006), p.65.

[11] M. Mysara, M. Elhefnawi, J. M. Garibaldi: J. Biomed Inform. Vol. 45 (2012), pp.528-534.

[12] M. Ichihara, Y. Murakumo, A. Masuda, T. Matsuura, N. Asai and M. Jijiwa: Nucleic Acids Res. Vol. 35 (2007), p. e123.

DOI: 10.1093/nar/gkm699

[13] J. Moody, in: From Statistics to neural Networks, edited by V. Cherkassky, J.H. Friedman and H. Wechsler, Part II , Springer Publishers(1994).

[14] D.F. Wang, X. Chen, F. Zhang: International Symposium on Information Science and Engineering (Shanghai, China, December 20-22, 2008). Vol. 2, pp.84-87.

[15] T. Katoh , T. Suzuki: Nucleic Acids Res. Vol. 35 (2007), p. e27.