Intelligent Microarray Data Analysis through Non-negative Matrix Factorization to Study Human Multiple Myeloma Cell Lines

• Published in: Applied sciences Vol. 9; no. 24; p. 5552
• Main Authors: Casalino, Gabriella, Coluccia, Mauro, Pati, Maria L., Pannunzio, Alessandra, Vacca, Angelo, Scilimati, Antonio, Perrone, Maria G.
• Format: Journal Article
• Language: English
• Published: Basel MDPI AG 01.12.2019
• Subjects: Acids; Artificial intelligence; Aspirin; Cell culture; Chemotherapy; Clustering; Data analysis; Drugs; Gene expression; Laboratories; Metabolism; Multiple myeloma; Principal components analysis; Proteins; Subject specialists; Italy
• ISSN: 2076-3417; 2076-3417
• DOI: 10.3390/app9245552

Microarray data are a kind of numerical non-negative data used to collect gene expression profiles. Since the number of genes in DNA is huge, they are usually high dimensional, therefore they require dimensionality reduction and clustering techniques to extract useful information. In this paper we use NMF, non-negative matrix factorization, to analyze microarray data, and also develop “intelligent” results visualization with the aim to facilitate the analysis of the domain experts. For this purpose, a case study based on the analysis of the gene expression profiles (GEPs), representative of the human multiple myeloma diseases, was investigated in 40 human myeloma cell lines (HMCLs). The aim of the experiments was to study the genes involved in arachidonic acid metabolism in order to detect gene patterns that possibly could be connected to the different gene expression profiles of multiple myeloma. NMF results have been verified by western blotting analysis in six HMCLs of proteins expressed by some of the most abundantly expressed genes. The experiments showed the effectiveness of NMF in intelligently analyzing microarray data.