A method for the refinement of data arising from electrophoresis of biological fluids.
Gingival crevicular fluid (GCF) is being analyzed by sodium dodecyl sulfate/polyacrylamide gel electrophoresis (SDS/PAGE) in our laboratory. We wish to characterize the major protein species in GCF and to investigate their association with the progression of disease. Laser densitometry of SDS/PAGE provides estimates of log molecular weight (LMW) and relative abundance for each band in each sample. Our aim was to develop a method for the refinement of these data into clusters which could be treated as distinct protein species, whose relation to disease progression could be assessed. A reproducibility study showed that the estimated LMW would fall within 3% of the true LMW 95% of the time. The method clusters the estimated LMWs of each band in each sample so that clusters form at LMWs which occurred frequently. The data from each band in each sample are thus refined into cluster number and relative abundance. Application of the technique to data from the reproducibility study showed that clusters formed around the individual components in the mixture, with little misclassification. The technique was then applied to two data sets: from SDS/PAGE of 104 GCF samples from 74 adolescents without progressive periodontal disease, and from 2 patients suffering from advanced progressive disease. The clusters appeared accurately to reflect the appearance of the gels, and clear differences were observed between the two data sets. The method will enable changes in the composition of biological fluids to be associated with external factors such as disease status and should be widely applicable.