PubMed1988
During DNA synthesis, each pair of homologous chromosomes replicates its bands in a precise order and at a specific time. When using asynchronous cell populations, this replication programme has to be reconstructed from a series of "stills"--serial samples taken at intervals through S-transit. Obviously, the result obtained is dependent upon the kinetic progression of cells through the cycle, and any perturbation of the cycle. A difficulty arises when we wish to compare the replication programme of a chromosome in two different cultures (e.g. cells from different origins, or after different treatments). Kinetic differences between cultures make it almost impossible to obtain two samples for analysis containing the same "mixture" of cells. Thus, a false programme difference could be introduced, or a real one masked. In this paper, we present a method of comparison that overcomes this problem. It is based upon the observation that with serial sampling of steady-state cell populations through S-phase, band appearance curves are sigmoidal and are very well approximated by cumulative Normal distributions with very similar standard deviations. If a family of such curves, closely spaced in time, is sampled twice, the two observed frequencies for each curve are related, their probits all lying on a single straight line. This line has a slope of 45 degrees and its displacement from the origin is a function of the time interval between the two samples. Given two identical families of such curves, exactly the same relationship will hold if one sample is drawn from each. If, however, the two families differ (in order, spacing, standard deviation etc.), the probit/probit plot will deviate in various ways from a straight line with 45 degrees slope. Any two subsets of chromosome replication-band frequencies can be regarded as derived from a family of cumulative Normal curves and probit/probit comparisons used to test the similarity of their replication programmes.