AUTOMATIC SELECTION OF TWO-DIMENSIONAL ELECTROPHORESIS GEL WITH LEAST GEOMETRIC DISTORTIONS

Abstract

The paper presents an algorithm for automated selection of the highest quality two-dimensional gel electrophoresis image. The quality criterion is the amount of vertical geometric distortions of the gel. The aim is to select the least distorted gel from the group received during the same biochemical experiment. Vertical geometric distortions displace proteins of the same molecular mass from the horizontal line and have a greater impact on the determination of protein characteristics than horizontal distortions. After presenting algorithm for evaluation of distortions and selection of base gel results are compared to expert's made selections. If necessary, algorithm may be adapted for horizontal distortion evaluation.