In this paper several neural network classification algorithms have been applied to a real-world data case of electron microscopy data classification. Using several labeled sets as a reference, the parameters and architecture of the classifiers, LVQ (Learning Vector Quantization) trained codebooks and BP (backpropagation) trained feedforward neural-nets were optimized using a genetic algorithm. The automatic process of training and optimization is implemented using a new version of the g-lvq (genetic learning vector quantization) and G-Prop (genetic back-propagation) algorithms, and compared to a non-optimized version of the algorithms, Kohonen's LVQ and MLP trained with QuickProp. Dividing the all available samples in three sets, for training, testing and validation, the results presented here show a low average error for unknown samples. In this problem, G-Prop outperforms G-LVQ, but G-LVQ obtains codebooks with less parameters than the perceptrons obtained by G-Prop. The implication of this kind of automatic classification algorithms in the determination of three dimensional structure of biological particles is finally discused.