Lipidomics produces enormous data and its analysis plays a key role, especially in untargeted studies. As such, robust bioinformatics is critical. Prior to statistical analysis, data preprocessing including signal processing, data normalization and transformation are required, such that raw data are transformed into a format compatible with statistical data analysis [91,92]. Given the large degree of lipid variation, the first step of unsupervised and supervised statistical analysis is data reduction. This may be accomplished by a number of methods including orthogonal partial least squares-discriminate analysis, principal components analysis (PCA), and partial least squares-discriminate analysis (PLS-DA). Both unsupervised and supervised methods can be used, depending on the goal of the specific analysis. In unsupervised data analysis, unknown information about different groups is used by PCA and hierarchical cluster analysis. In the supervised approach, each sample or metabolite is associated with known compounds and this prior information is then used for analysis via principal component regression and neural networks [91,92]. Other regression methods including Elastic Net and Least Absolute Shrinkage and Selection Operator are also available for analysis of lipidomic data sets to ascertain the relationship between variables