Protein-nanoparticle conjugation provides unique interactions be-tween biological systems and synthetic materials used for analytical, diagnostic and therapeutic applications. This paper presents the development of Artificial Neural Network (ANN) for quantification of proteins concentration adsorbed on gold nanoparticles (Au NPs). Single hidden layer feedforward ANN is based on concentration of free proteins as input parameters, while concentration of con-jugated proteins is desired output. Totally, 210 samples were used, 200 of them derived from experiment and 10 additionally added as blank samples. Training data sets contain 120 samples, out of which 108 samples are used for estimation and 12 for validation. The ANN system is subsequently validated with 90 samples, 80 samples from experiment and 10 additionally added. From 80 of samples with known protein concentration, 74 are successfully quantified as proteins adsorbed on nanoparticles, which gives sensitivity of 92.5%. Out of experiment data, 10 blank samples are correctly classified as free of proteins giving the specificity of 100%. Developed system can be used in laboratory conditions and further validated on new experimental samples.
CITATION STYLE
Fojnica, A., Osmanović, A., Tarakčija, D., & Demirović, S. (2017). Quantification of protein concentration adsorbed on gold nanoparticles using artificial neural network. In IFMBE Proceedings (Vol. 62, pp. 142–146). Springer Verlag. https://doi.org/10.1007/978-981-10-4166-2_22
Mendeley helps you to discover research relevant for your work.