A series of simple and robust operations for handling large chromatographic time-of-flight mass spectrometry fingerprinting data has been established and applied to data from extracts of Fusarium graminearum genotypes modified in a non-ribosomal peptide synthase gene by over-expression and deletion. It includes importing data into a computing environment by binning the m/z axis; baseline removal; data transformation and compression across retention times. Each sample represented by a total mass spectrum was normalized to unit sum and variables selected by partial least squares discriminant analysis. Finally, principal component analysis was used for identification of high discriminatory power mass-to-charge ratios (m/z's) separating over-expression, wildtype and deletion genotypes. Two compounds exhibiting a positive correlation to the expected levels in different genotypes were identified. The two compounds were represented by m/z 683.5 with retention time of 8.9 min, and m/z's 774.5 and 775.5 with retention time of 14.1 min. This methodology enables extraction of chemical information from large data sets (>10⁸ entries), and provides a starting point for individual optimization in targeting small molecules from metabolomics data.