Share Email Print
cover

Proceedings Paper

Utilizing molecular data descriptors as basis vectors for clustering
Author(s): Sydney Sukuta
Format Member Price Non-Member Price
PDF $14.40 $18.00
cover GOOD NEWS! Your organization subscribes to the SPIE Digital Library. You may be able to download this paper for free. Check Access

Paper Abstract

The goal of this study is to test the feasiblity of directly using molecular descriptor data to generate clusters of similar molecules. We have developed an approach that utilizes the "most orthogonal" molecular descriptor variables as a basis set for clustering. In this study we have specifically utilized normal skin tissue and melanoma cancer data derived via Fourier transform infrared (FTIR) spectroscopy to generate these clusters, but the approach presented should be applicable to any other molecular descriptor or response data. Using the three most orthogonal FTIR frequencies as a basis set for cluster analysis, normal skin and melanoma tumors' clusters were resolved and localized in the three-dimensional variable/frequency space. Such clusters can be used to rapidly identify molecules with similar structures, and biological activity given their physico-chemical descriptors or molecular response data. This study also points out possible fallacies when inspecting clusters and how they can be avoided.

Paper Details

Date Published: 18 July 2003
PDF: 9 pages
Proc. SPIE 4966, Microarrays and Combinatorial Technologies for Biomedical Applications: Design, Fabrication, and Analysis, (18 July 2003); doi: 10.1117/12.477783
Show Author Affiliations
Sydney Sukuta, San Jose City College (United States)


Published in SPIE Proceedings Vol. 4966:
Microarrays and Combinatorial Technologies for Biomedical Applications: Design, Fabrication, and Analysis
Dan V. Nicolau; Ramesh Raghavachari, Editor(s)

© SPIE. Terms of Use
Back to Top