Global approaches to quantitative analysis of gene-expression patterns observed by use of two-dimensional gel electrophoresis.

A major difficulty in the use of two-dimensional protein maps to identify and classify cell types is the problem of acquiring, selecting, and analyzing quantitative data on hundreds of protein spots. Here we use methods of multivariate statistics to analyze the differences among a panel of human cell lines, in some cases involving quantitative data on more than 250 proteins. Principal-component and cluster-analysis techniques show that the lines can be easily distinguished, even by using the subset of proteins present in all cells. A preliminary analysis of the protein changes brought about by phorbol ester-induced differentiation of the line U937 is included.