Use of artificial neural network techniques to model proteomics of cellular stress

Commercial scale bio processing is a major challenge when developing a whole cell biological vaccine. You must ensure that the end product has a sufficient level of potency to invoke an immune response within the host. Making sure that these key antigens remain on the cell surface during the manufacturing process is essential to make a viable product. Performing cellular stress experiments on an industrial scale would prove far too costly; instead an ultra-scale down model is used here to mimic the effects of a large scale industrial plant in a laboratory, for example to study the effect of hydrodynamic shear on cell membrane and cell surface makers (mimicking shear forces created in the mechanical pumps of the large scale system), by using a rotational shear device. Here, bioinformatic modelling techniques, such as Artificial Neural Networks have been used to predict the levels of surface markers CD9 CD147 and HLA A-C based on experimental parameters of data obtained from ultra-scaled down experiments. These models where used to predict how changing the parameters would affect the density and abundance of the surface markers on the cell lines outer membrane. The surface markers CD9, CD147 and HLA A-C were chosen because of their relevance to the immune system and because of their strong expression on the surface of the P4E6 cells