A framework for selecting analytical techniques in profiling authentic and counterfeit Viagra and Cialis

Area of Science:

• Analytical Chemistry
• Chemometrics
• Forensic Science

Background:

• Drug profiling methods are often expensive, time-consuming, and not readily available.
• Accurate classification of drug samples into authentic and unauthentic categories is crucial.

Purpose of the Study:

• To develop a method for identifying the most relevant analytical techniques for drug sample classification.
• To optimize drug profiling by prioritizing cost-effective and efficient analytical methods.

Main Methods:

• Integration of Principal Component Analysis (PCA) with k-Nearest Neighbor (KNN) and Support Vector Machine (SVM) classification.
• Application of a 'leave one subset out at a time' approach to evaluate PCA score subsets from five analytical techniques (physical profile, XRF, ESI-MS, UPLC-MS, ATR-FTIR).

Main Results:

• The proposed method successfully identified key analytical techniques for improved drug classification accuracy.
• For Viagra and Cialis samples, UPLC-MS, physical profile, and ATR-FTIR were recommended as the most informative techniques.
• Support Vector Machine (SVM) demonstrated higher classification accuracy compared to k-Nearest Neighbor (KNN).

Conclusions:

• The developed method effectively prioritizes analytical techniques for efficient and accurate drug profiling.
• Combining PCA with classification algorithms like SVM can enhance the selection of optimal analytical strategies.
• This approach offers a valuable tool for streamlining drug analysis and quality control.