Classification of fentanyl analogues through Principal Component Analysis (PCA) and Hierarchical Clustering of GC-MS data

Abstract The emergence of a wide variety of fentanyl analogues has become a problem for the identification of seized drug samples. While chemical databases are largely reactive to the emergence of new analogues, efforts should focus on the development of predictive models which can discern how new analogues differ from the parent drug. Principal component analysis (PCA) was performed on mass spectral data from 54 fentanyl analogues. Hierarchical clustering was used to group these analogues into meaningful classes. The model was able to classify 67 analogues not previously included in the model with high accuracy, based on the nature and position of the chemical modification.