Establishing Multivariate Analysis of Materials through Intelligent Data Analysis System (IDAS)

14 Dec 2020 2:16 PM

The DOST- Advanced Science and Technology Institute (DOST-ASTI), in partnership with DOST- Industrial Technology Development Institute Advanced Device and Materials Testing Laboratory (DOST-ITDI-ADMATEL), developed a project that will be useful in conducting Cluster Analysis and Kinetic Stability Modelling on input spectra dataset. Through the project titled Application of Multivariate Analysis on Methamphetamine-HCL Chemical Fingerprints and Kinetic Stability Modelling, both agencies aim to create an Intelligent Data Analysis System (IDAS) for drug trafficking investigation in the Philippines.

This one-year partnership of the DOST-ASTI with Dr. Araceli Monsada and her team of chemists and chemical engineers from DOST-ITDI-ADMATEL involved identifying their requirements for the system through User Experience (UX) workshops. Through this, the implemented software tool design will allow users an easy access to its functionalities.

The current version of the IDAS software tool includes functionalities for multivariate analysis of many-feature datasets generated by ADMATEL using their Time-of-Flight Secondary Ion Mass Spectrometer (TOF-SIMS) for material fingerprinting, Gas Chromatography-Mass Spectrometer (GC-MS) for quantification, and Fourier Transform Infrared (FTIR) Spectrometer for material composition. The IDAS software tool is supported by a graphics processing unit that can be used for the acceleration of computation when accessing the multivariate analysis functionality.

IDAS Architecture Diagram

The DOST-ASTI Project Development Team used an agile project management and development methodology, called Scrumin implementing the software requirements of DOST-ITDI-ADMATEL. This involved a series of cycles of requirements elicitation, requirements implementation, and feedback elicitation between DOST-ASTI and DOST-ITDI-ADMATEL. Frequent discussions between the two groups provided a clearer understanding of what needs to be implemented into the software tool and what needs to be prioritized during the 1-year project implementation. 

The IDAS software tool provides the user with a list of analysis transactions that was performed using its analysis functionalities, see Figure 1. In the current version of the IDAS software tool, each input dataset is associated with a single user transaction and the multivariate analysis is performed only with respect to the samples in the input dataset. Comparison of samples in an input dataset with other datasets previously inputted into the software tool may be implemented in the succeeding versions of the software. 

Figure 1. List view of analysis transactions performed in the software tool.

Using the multivariate analysis functionality of the IDAS software tool, sample data points in the input dataset can be clustered based on their distances in a Euclidean space. The K-Means clustering algorithm is the implemented algorithm in the current version of the IDAS software tool. Other clustering algorithms that would compensate for the disadvantages of the K-Means algorithm may be implemented in the succeeding versions of the software. Aexample of clustering of samples in an input dataset is shown in Figure 2.  

Figure 2. Visualization of clustering of samples in the input dataset. Clustering is performed using the K-Means algorithm.

Users may download the results of the multivariate analysis performed on an input dataset through a Download PageThey may use the files included in the downloadable result for further analyses that are currently not supported in the IDAS software toolUsers may also use these files in their analysis report. 

Figure 3. Download page wherein users can download the results of analysis performed on their input dataset.

Additional functionalities will be implemented into the IDAS software tool in 2020. These include the kinetic modeling of the stability of materials under different temperatures. 

The Research Team

Joanna G. Syjuco
Chief Science Research Specialist

Elmer C. Peramo
Senior Science Research Specialist

Jeffrey A. Aborot
Senior Science Research Specialist

Roxanne S. Aviñante
Senior Science Research Specialist

Vanesa O. Osiana
Science Research Specialist II

Elson M. Crisologo, Jr.
Science Research Specialist I

Michelle P. Neverida
Science Research Specialist I

Rother Jay B. Copino
Science Research Specialist I

Eva Mae B. Sitoy
Science Research Specialist I

John Kevin C. Abonita
Science Research Specialist I

Lorenzo Miguel S. Gabarda
Science Research Specialist I

Trixie Shane G. Maningding
Science Research Specialist I

Carl Christian D. Raquid
Science Research Specialist I

Abegail F. Bactol
Project Development Officer I

Elisa C. Espares
Administrative Aide IV