Computer automated lanes detection and profiles evaluation of one‐dimensional gel electrophoretic autoradiograms

A new fast, precise and reproducible method for lanes detection and profiles evaluation of one‐dimensional gel electrophoretic autoradiograms is presented. It is fully automated, and its principles are based on image and signal processing techniques. The operator usually has only to specify the number of lanes that are present in the preparation, but still may retain overall control. In the first phase, the program automatically locates the lateral position of the lanes, and approximates their left and right boundaries by polynomials (which allow curved as well as straight lanes). The operator can then intervene, if necessary, to adjust these positions. In the second phase, the background variations are corrected; two types of density profiles are computed, representing the average density as well as the total amount of material in the lanes. This calculation is carried out with correction for tilt or curvature for each lane. Finally, prinouts are produced that contain graphs of the profiles together with quantitative measurements. Use of this method is illustrated by analysis of electrophoretic gels containing proteins and lipopolysaccharides of varying mobilities. Experiments made on these gels as well as others yield accurate quantitation of the relative amount of each component.