On numerical characterization of proteomics maps based on partitioning of 2-D maps into Voronoi regions
- First Online:
- Cite this article as:
- Randić, M. & Orel, R. J Math Chem (2011) 49: 1759. doi:10.1007/s10910-011-9856-x
- 51 Views
We consider the problem of partitioning of the 2-D proteomics maps into regions associated with set of selected protein spots so that one can construct the adjacency matrix for proteomics maps, which will be a source of 2-D map descriptors. We selected a set of N most abundant spots for which we construct the Voronoi regions. Once Voronoi regions have been obtained one can construct the accompanying adjacency matrix A, and adjacency-distance matrix AD. The binary elements A(i, j) of the adjacency matrix are determined by the adjacency of corresponding Voronoi regions, and the elements AD(i, j) of the adjacency-distance matrix are determined by the distances of corresponding centers of Voronoi regions. The approach is illustrated on a smaller map for 20 most abundant proteins from a selection of available data in the literature.