F. N. David

The analysis of chromosome patterns in the abnormal cell

1. In a previous paper (Barton & David, 1 9 6 1 ~ ) we discussed briefly the test of the hypothesis that homologous pairs of chromosomes in the normal cell are situated randomly with respect to one another. The application of this analysis to a number of observed normal patterns was made in a further paper. We now turn to study the analysis of chromosome patterns when there are either an abnormal number of chromosomes in toto in the cell, or where the number is the normal 46 but these are not 23 pairs. It will be clear that another application of the analysis could be for the normal cell of 23 pairs where there is difficulty in distinguishing between certain chromosomes with adjacent numbering. The mathematical derivation of the results which we use here has been given in Barton & David (1961b). 2. It is assumed as before that the coordinates of each of the N centromeres are known referred to some convenient origin and a convenient pair of axes of co-ordinates. Write the means of the x and y co-ordinates, respectively, as E and fj and define

The relative positions of the chromosomes in the human cell in mitosis

In a recent paper on the positions of the sex chromosomes in the human cell in mitosis ( I

A TEST FOR BIRTH ORDER EFFECT

164) we gave a method of analysis depending on the counts of the chromosome centromeres lying outside a given circle. If there are N chromosomes with (si,yi) the co-ordinates of the centromere of the ith chromosome (i = 1,2, . , , N ) referred to any arbitrary pair of axes we may plot these points as a centromere diagram.

Research papers in statistics : festschrift for J. Neyman

In 1948 Haldane & Smith put forward a test to determine whether offspring suffering from a given defect occurred randomly as regards their place in the family. Their test consisted of assigning all sibs in the family a weight which was the number of their place in the birth order. These weights were added up for affected sibs. For small numbers this sum can be referred to known tables, for large numbers it is permissible to assume that it is normally distributed with a mean and variance which are known. Thus, if in a family there are r2 affected sibs and rl normal (rl + r2 = r ) and if the weights of the r2 affected sibs are Z,, Z, .... Z,,, then