ACCURATE NONLINEAR REGISTRATION FOR TWO-DIMENSIONAL GEL ELECTROPHORESIS IMAGES

Abstract

Two-dimensional gel electrophoresis (2DGE) images are an important support for the analysis of proteins in proteomics. The registration of 2DGE images is considered as one of key elements in protein identification while it is a difficult problem. This paper proposes a new accurate nonlinear registration approach for 2DGE images, based on the exploitation of both spot distance measure and spot intensity. The method consists of three steps: multi-resolution affine registration, spot pairing and thin-plate spline interpolation. The results on both simulated and real gel images show that the proposed method significantly improves registration accuracy in comparison with thin-plate spline registration techniques.